The polychaetes Australonereis ehlersi (Augener) and Simplisetia aequisetis (Augener) within the eutrophic Swan River Estuary, Western Australia: Life history, population structure and effects on sedimentary microbial nitrogen cycling Robert John De Roach, B.Sc. (Hons.) This thesis is presented for the degree of Doctor of Philosophy, The University of Western Australia Zoology School of Animal Biology The University of Western Australia December, 2006
255
Embed
The polychaetes Australonereis ehlersi (Augener) and ......The polychaetes Australonereis ehlersi (Augener) and Simplisetia aequisetis (Augener) within the eutrophic Swan River Estuary,
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
The polychaetes Australonereis ehlersi (Augener) and Simplisetia
aequisetis (Augener) within the eutrophic Swan River Estuary,
Western Australia: Life history, population structure and
effects on sedimentary microbial nitrogen cycling
Robert John De Roach, B.Sc. (Hons.)
This thesis is presented for the degree of Doctor of Philosophy,
and Kristensen, 1991; Aller and Aller, 1998; Welsh, 2003).
Nitrogen is the chief element limiting primary production in most unpolluted estuaries and coastal
marine systems (Ryther and Dunstan, 1971; Carpenter and Capone, 1983; Seitzinger, 1990), and
also in some freshwater lakes, streams and rivers (Keeney, 1973; Gerhart and Likens, 1975;
Seitzinger, 1990). However, as aquatic systems continue to receive an increasing load of nitrogen
1
and other nutrients, primarily sourced from fertilisers, eutrophication has become a major problem,
both throughout the world and in south-western Australia (Heathwaite et al., 1996; Colney, 2000;
Meyer-Reil and Köster, 2000; Peters and Donohue, 2001). Eutrophication is the process by which
increasing nutrients (typically nitrogen or phosphorous but also carbon) cause a change in the
nutritional status of a given body of water (Meyer-Reil and Köster, 2000). Increasingly, water
bodies with resident algal populations previously limited in productivity by low nitrogen
availability are becoming ‘unlimited’ in nitrogen supply, often resulting in algal growth of
excessive intensity and frequency (Pedersen and Borum, 1996). Within the Swan River Estuary,
nitrogen is more limiting to algal biomass production than phosphorous, by ≤20 times (Horner
Rosser and Thompson, 2001).
The environmental, health and economic implications of algal blooms can be far-reaching. Excess
accumulations and subsequent biodegradation of phytoplankton, epiphytic and/or ephemeral algae
may be displeasing aesthetically but, more importantly, can cause highly detrimental secondary
effects (Pedersen and Borum, 1996). Blooming phytoplankton species may contain or produce
toxins that directly impact on the health and mortality of trophically-related or surrounding
organisms (including humans) (Atkins et al., 2001; de Jonge et al., 2002). Turbidity and shading
caused by fast-growing algal species may reduce the growth or abundance of benthic microphytes
and macrophytes such as seagrasses and kelp (Pedersen and Borum, 1996). Biodegradation of
excess algae consumes oxygen that can lead to anoxia of both bottom and overlying waters (Meyer-
Reil and Köster, 2000) and to significant mortality of oxygen-dependent organisms including fish
and benthic fauna (Elliott and de Jonge, 2002). Further, loss of benthic fauna and flora may
destabilise sediments, thereby reinforcing the effects of increases in turbidity (Meyer-Reil and
Köster, 2000). Changes in species distribution, abundance, diversity and physiological state have
the potential to transfer cascading effects to higher trophic levels, ranging from slight to major
disruptions to fish, bird and mammal populations, to complete ecosystem collapse in extreme cases
(Meyer-Reil and Köster, 2000; Elliott and de Jonge, 2002). Consequential deleterious impacts on
human values and economies associated with problem waterways include reduced fisheries, direct
threats to public health and lowered potential for recreational activities and tourism.
Consequently, any natural means by which nitrogen can be unloaded from eutrophic waters
deserves attention. The process of sedimentary microbial denitrification, whereby soluble inorganic
forms of nitrogen (nitrate and nitrite) are converted to gaseous forms (nitrous oxide and dinitrogen
gas) that are free to percolate from aquatic systems, is particularly important since it represents an
ultimate sink of nitrogen with the capacity to regulate the eutrophic status of water bodies
(Seitzinger, 1990). For example, in Port Phillip Bay, Victoria, Australia, denitrification removes (as
2
dinitrogen gas) an amount of nitrogen that is approximately equivalent to the combined total
loading of nitrogen to that system each year from catchment drainage, atmospheric inputs and
outfall from the major wastewater treatment plant servicing the city of Melbourne (Harris et al.,
1996). Whilst the large embayment remains highly productive due to tight internal cycling of
nitrogen, sedimentary denitrification primarily keeps Port Phillip Bay in an oligotrophic to
mesotrophic (low nutrient) state (Harris et al., 1996). Although a critical terminal process,
denitrification needs to be considered in the context of the entire nitrogen cycle and not regarded as
a separate event (see Chapter 3) (Kuenen and Robertson, 1988). Further, a holistic understanding of
the many biological and physico-chemical controls on the flows and transformations of nitrogenous
species within the affected system is crucial to successful eutrophication management strategies.
In this regard, there has been increasing awareness recently of the role benthic fauna plays in
regulating the size, diversity, distribution and activity of populations of denitrifiers and other
nitrogen-transforming bacteria (e.g. see reviews by Petr, 1977; Aller, 1982, 1988; Krantzberg, 1985;
Henriksen and Kemp, 1988; Kristensen, 1988, 2000; Andersen and Kristensen, 1991; Svensson et
al., 2001; Welsh, 2003). For example, the presence of burrows considerably extends the sediment-
water interface area where microbial populations can reside (Hylleberg and Henriksen, 1980;
Kristensen, 1984a; Kristensen et al., 1991), compared to sediment devoid of fauna. Burrow walls
are often bound by organic-rich exudates of the inhabitant fauna which further supports elevated
microbial activity (Aller, 1983, 1988; Aller et al., 1983; Reichardt, 1988). Irrigation of burrows
transports substrates of denitrification (nitrate and nitrite) from the overlying water to depths greater
than expected by diffusion, potentially enhancing the rate of nitrogen removal (Aller, 1988;
Kristensen, 1988; Kristensen et al., 1991). However, the supply of oxygenated water can suppress
anaerobic denitrification but favours other microbial nitrogen transformations such as nitrification
(the conversion of ammonium to nitrate) (Svensson et al., 2001). Intermittent ventilation is typical
of many infaunal species, and burrow waters may quickly turn anoxic during periods of ventilatory
rest (Kristensen, 2000), often imparting alternating redox conditions within the burrow. Such
circumstances may lead to significantly increased rates of inorganic nitrogen removal when
ammonium is first aerobically nitrified to nitrate during active ventilation periods, then to be
anaerobically denitrified during rest periods (coupled nitrification-denitrification) (Jenkins and
Kemp, 1984; Henriksen and Kemp, 1988; Capone, 2000). De Roach et al. (2002) previously
indicated that the burrowing behaviour of adult S. aequisetis may enhance significantly the
denitrification process of sediments in the Swan River Estuary.
The many ways in which benthic fauna, particularly polychaetes, can affect the flow and microbial
transformations of nitrogen within aquatic systems is a major theme of my thesis. A particular aim
3
of my study was to determine the influence of A. ehlersi and S. aequisetis upon denitrification and
broader microbial nitrogen cycling within their surrounding sediment environment. Highlighting
the importance of faunal impacts on sedimentary nitrogen cycling underscores the potential for
development of benthic fauna management strategies aimed at preventing the deleterious effects of
eutrophication and problem algal blooms.
1.2 THESIS QUESTION, AIMS AND STRUCTURE
In my thesis I address two, over-arching questions:
1. What are the life histories and population structures of A. ehlersi and S. aequisetis
within the Swan River Estuary, Western Australia?
2. What effects do members of both species of polychaete have on microbial nitrogen
cycling within sediments of the estuary?
Three specific aims are addressed, as follows:
I. I present information concerning the life history, geographical population structure and
production of the polychaetes A. ehlersi and S. aequisetis from the Swan River Estuary, in
Chapter 4.
II. Results of my measurements of the effect of A. ehlersi and S. aequisetis individuals on
microbial denitrification and nitrogen cycling within the sediments of the estuary are
presented in Chapter 5.
III. In Chapters 6 and 7 I integrate knowledge on fine-scale effects of benthic fauna on
microbial nitrogen cycling, with a broad-scale understanding of faunal population dynamics
to highlight the functional role of benthic fauna in estuarine nitrogen cycling, with
important implications for eutrophication management.
In Chapter 1, I present a general overview of my thesis content and intent. In Chapter 2, I present a
detailed description of the structure and function of the Swan River Estuary, leading into a review
of the causes of the nitrogenous-based eutrophication and related environmental and socio-
economic problems, and a review of the current state of knowledge regarding A. ehlersi and
S. aequisetis taxonomy, biology, and of the composition of the Swan River Estuary’s macrobenthic
community.
Such an assessment of the species’ basic biology and ecological niche is considered requisite to a
comprehension of their biogeochemical role in the environment. Prior to providing a historical
4
perspective on the realisation of the effects of benthic fauna on sedimentary microbial nitrogen
cycling, Chapter 3 defines the component processes of the nitrogen cycle, placing emphasis on the
importance of denitrification as an ultimate sink of nitrogen in aquatic environments. Chapter 3
also provides an ideological framework for the methodological consideration of nitrogen cycling,
notably defining denitrification as an enzymatic process. The Appendices provide supporting
information relevant to these contextual issues.
5
6
CHAPTER 2
THE SWAN RIVER ESTUARY: HYDRODYNAMICS,
EUTROPHICATION AND MACROBENTHIC ECOLOGY
2.0 INTRODUCTION
The growth of the world’s population from 1.6 billion in 1900 to over 6 billion in 2006 has been
sustained, in large part, by increased crop productivity attributable to development of ammonia-
based fertilisers (Smil, 1999). Unfortunately, the prolonged and intensive use of fertilisers on a
global scale, compounded by the consequences of human population expansion (changing land use
patterns, deforestation, increased discharge of municipal and industrial wastes), has lead to
significantly increased export of nitrogen and other nutrients to near-shore aquatic ecosystems
(Rabouille et al., 2001). Where these increases in nutrients (typically nitrogen or phosphorous but
also carbon) cause a change in the nutritional status of a given body of water, the process can be
defined as eutrophication (Meyer-Reil and Köster, 2000). The world’s rivers, estuaries and coastal
ocean are becoming increasingly eutrophic and subject to the effects of shifts, often deleterious, in
primary productivity. In non-impacted systems, estuarine and near-shore primary productivity is
typically limited by nitrogen availability, whilst phosphorous availability typically limits freshwater
primary production (Nixon, 1995; Peters and Donohue, 2001). Within eutrophic systems, nutrient
limitation of algal productivity is dampened, often resulting in algal growth of excessive intensity
and frequency (Pedersen and Borum, 1996). The environmental and socio-economic implications
of nuisance algal blooms are extensive: problems range from slight aesthetic displeasure to
complete ecosystem collapse (Elliott and de Jonge, 2002; see Section 2.1.3 below). In this section I
outline the pertinent causes and implications of eutrophication and associated algal blooms using
the eutrophication of the Swan River Estuary, Western Australia, as a case study. Such
consideration is intended to highlight not only the broad-scale issues relevant to nitrogenous
eutrophication but also to provide the local context of this thesis.
2.1 THE SWAN RIVER ESTUARY AND EUTROPHICATION
2.1.1 Catchment Geography
The Swan River Estuary, 60 km in length from its mouth to the Indian Ocean at Fremantle to its
confluence with Ellen Brook (Hodgkin, 1987) within a catchment of 121 000 km2 in south-west
Western Australia, covers an area of 53 km2 (Figure 2.1). The riverine source is 375 m above sea
level on the Darling Scarp (Stephens and Imberger, 1996). The estuary has been a major node for
7
the State’s population growth since European settlement in 1829. The State capital city, Perth, sits
astride the estuary with a present population of approximately 1.4 million (Atkins et al., 2001),
expected to double over the next 30 years (Stephens and Imberger, 1996). Water quality of the
Swan River Estuary, post-settlement, has been related to changes in land-use patterns within the
catchment (Atkins and Klemm, 1987; Kurup et al., 1998).
Drainage into the estuary originates from two main areas:
• The Avon catchment comprising a high escarpment (the Darling Scarp) and dissected
plateau, primarily draining into the Avon River and its tributaries (Figure 2.1; Atkins and
Klemm, 1987; Kurup et al., 1998; Viney and Sivapalan, 2001).
• The lower Swan Coastal Plain sub-catchments extend approximately 30 km inland and are
drained via streams and stormwater drains.
The Avon catchment contributes 85% of total flow and is the larger of the two source areas,
extending 500 km inland and covering 120 000 km2 in area. It is partially forested with remnant
native dry sclerophyll woodland vegetation but has been cleared extensively for agriculture, mainly
between 1940 and 1970 (Atkins and Klemm, 1987; Kurup et al., 1998; Peters and Donohue, 2001;
Viney and Sivapalan, 2001). The Darling Scarp elicits a strong west-east rain-shadow effect,
ranging from mean annual rainfall >800 mm on the scarp, to ~250 mm in the eastern parts of the
catchment (Viney and Sivapalan, 2001). Mean annual rainfall over the entire catchment is
~400 mm.
The lower sub-catchments of the Swan Coastal Plain collectively contain much of the metropolitan
population within 1 000 km2. The area is highly urbanised but also retains much agricultural land
use (Atkins and Klemm, 1987; Kurup et al., 1998). Groundwater discharge exerts primary control
on streamflow on the coastal plain (Peters and Donohue, 2001). The systems responsible for the
discharge are the Gnangara Mound to the north of the estuary, the Cloverdale Groundwater Flow
System in the region between the Swan and Canning Rivers, Jandakot Mound to the south of the
estuary, and the Cottesloe Mound is a small freshwater lens between coast and estuary (Linderfelt
and Turner, 2001). The groundwater systems are recharged solely by rainfall (Donohue et al.,
2001). Soils of the coastal plain are generally coarse and sandy with poor nutrient retention, but
interdigitate with a silt and clay unit at the base of the Darling Scarp (Peters and Donohue, 2001;
Viney and Sivapalan, 2001).
8
PERTH
5 km
Avon River
Ellen Brook
N
WESTERN AUSTRALIA
FREMANTLE
Helena River
Darling Scarp
Swan
Estuary
River
Canning River
150 km
[A]
[B]
Indian Ocean
300 mm
400 mm
500 mm
600 mm
800 mm
1200 mm
Inset [A]
AVON CATCHMENT BOUNDARY
Figure 2.1: The Swan River Estuary and its sub-catchments: [A] Swan Coastal Plain sub-
catchments; [B] the greater Avon Catchment, illustrated with rainfall contours. (Adapted from Peters and Donohue, 2001).
9
2.1.2 Hydrology
The hydrodynamics exert primary control over both the ecology, particularly of the spatio-temporal
distribution and abundance of algal and benthic populations, and also of nutrient transport and sinks
within the Swan River Estuary: both aspects constitute a primary focus of this thesis. It is therefore
relevant to review here the hydrology of the estuary, particularly of the main factors governing the
dynamics of this seasonal system. Spencer (1956), Hodgkin (1987), Stephens and Imberger (1996)
and Kurup et al. (1998) have reviewed thoroughly the Swan River Estuary hydrodynamics.
Seasonality and Tides
The hydrological dynamics of the Swan River Estuary are a consequence of its extreme climatic
seasonality, small tidal range and characteristic bathymetry (Hodgkin, 1987; Stephens and
Imberger, 1996; Kurup et al., 1998). The climate is Mediterranean with typically hot, dry summers
and mild, wet winters and greater than 90% of annual rainfall is usually received between May and
October (Peters and Donohue, 2001). As an oversimplification, received freshwater inflows during
winter flush the upper reaches of the estuary thereby reverting the system to a freshwater to
brackish state (Stephens and Imberger, 1996; Linderfelt and Turner, 2001). Decreasing river
discharge in spring allows the upstream intrusion of marine-derived saline water as a salt-wedge,
resulting in marine conditions within the estuary during summer (Stephens and Imberger, 1996;
Linderfelt and Turner, 2001). The Swan River Estuary is therefore classified as a seasonal estuary
(Hodgkin, 1987).
Mean tidal range at Fremantle is 0.6 m, such that the Swan River Estuary is also classified as
microtidal (i.e. estuaries with a tidal range <2 m; Monbet, 1992; Kurup et al., 1998). The low tidal
amplitude allows the persistence of a seasonal cycle of salinity variation, since tidal forcing in
winter is not sufficient to overcome the seaward flow of freshwater (Kurup et al., 1998). This
pattern contrasts with those of meso- and macrotidal estuaries (i.e. tidal range >2 m) where tidal
forcing permits a semi-diurnal, diurnal or fortnightly periodicity in salinity variation, with less
dominance of freshwater discharge (Kurup et al., 1998). Whilst tidal amplitude in the Swan River
Estuary is small, it is nevertheless sufficient to exert an influence up to 60 km upstream during
summer, when freshwater discharge is restricted (Linderfelt and Turner, 2001).
Physiognomy
Underlying the general characterisation of the Swan River Estuary as a seasonal microtidal estuary,
is much variation in estuarine dynamics both on a local geographical scale (i.e. across the spatial
extent of the estuary) and on a temporal scale (i.e. inter-annual variation). The estuary has a free
10
connection with the open sea and as such approximates Pritchard’s (1967: p. 3) definition of an
estuary as “a semi-enclosed body of water within which sea water is measurably diluted with fresh
water derived from land drainage”. However, Hodgkin (1987) also points out that such a
superficial characterisation of the system as a typical estuary is inadequate for a useful
consideration of its hydrology. Estuarine classification systems may be important for delineating
broad-scale differences between estuaries, but an appreciation of locally-specific features is
necessary to a full understanding of the hydrology of any given estuary. There are several physical
features of the Swan River Estuary that impart implicit effects on its hydrodynamics.
Hodgkin (1987) and Stephens and Imberger (1996) subdivide the Swan River Estuary into five
major regions of hydrological importance: the main basin, two adjacent sills, Fremantle Harbour
and the upper estuary (Figure 2.2). From Preston Point (~5 km upstream of the mouth), a deep (up
to 17 m) and narrow 5 km long trough opens into an approximately 10 km long by 2 km wide
shallower basin (>5 m deep in the centre). The upstream limit of the basin is the constriction
known as the Narrows ~20 km from the mouth. The trough and basin areas are referred to
collectively as the deep or main basin. The Canning River also discharges into the Swan River
Estuary, and for 6 km from its confluence point to the Kent Street Weir is basically an extension of
the main basin. Average depth decreases from 4-5 m at the confluence to less than 1 m upstream of
Salter Point (~4 km upstream of the confluence). The weir acts as a barrier to penetration of marine
water, such that the upstream Canning River remains fresh year round and is not considered a part
of the estuary.
Sills on either side of the main basin regulate the flow of water within the estuary. Downstream of
the deep trough, a shallow (5 m deep) and narrow sill extends for 2 km from Preston Point to the
Fremantle Traffic Bridge (~3 km upstream of the mouth). From the bridge, a 3 km long, 11 m deep,
channel is dredged to the mouth of the estuary to comprise Fremantle Harbour. The downstream
Fremantle Sill controls water exchange between the ocean and the estuary. Upstream of the main
basin, Perth Water comprises a secondary sill, a wide, expanse of shallow water (<2 m deep)
between the Narrows and Heirisson Island. The second sill regulates water flow between the main
basin and the 35 km stretch of upper estuary, i.e. the narrow riverine reaches between Heirisson
Island and the confluence with Ellen Brook (~60 km upstream of the mouth). Average water depth
of the upper estuary is 2 to 3 m, punctuated by pockets of 5 to 6 m, but becomes much shallower
with many sand banks upstream of the confluence with Jane Brook (~50 km from the mouth).
11
Bathymetry
Dep
th (m
)
0
5
10
15
20
0 10 20 30 40Distance from mouth (km)
Upper Estuary
Main
Basin
Fremantle
Harbour
Fremantle Sill
Perth Water
Kent Street
Weir
The Narrows
Heirisson Island
Fremantle Traffic Bridge
Salter Point
Preston Point
Deep Trough
5 km
N
Figure 2.2: Physiognomy and bathymetry of the Swan River Estuary. Adapted from Stephens and
Imberger (1996).
12
Temporal Variation and Stratification
The described physical characters of the Swan River Estuary markedly influence the
hydrodynamics of this seasonal microtidal system. Stephens and Imberger (1996) and Kurup et al.
(1998) consider the hydrodynamics of the system in relation to the annual upstream migration of
tidally forced marine water as a salt wedge, when freshwater discharge decreases following the
cessation of winter rains. Kurup et al. (1998) describe the hydrology of the estuary in relation to
three phases of upstream salt wedge positioning controlled primarily by river discharge:
(1) Salt wedge (marine) dominated in summer and autumn (January to May), when freshwater
discharge is very low. High salinity water extends from the mouth to 40 to 60 km
upstream, i.e. between the confluence with the Helena River and Ellen Brook (Figure 2.3).
(2) (a) Salt wedge waning during early winter (June to July); or, (b) salt wedge emplacement
(propagation) during late spring and early summer (November to December); when
freshwater discharges are low. Fresh to brackish waters extend from the upper estuary to
surficial (<2 m deep) Perth Water (20 to 25 km upstream of the mouth), whilst high salinity
bottom waters (>2 m deep) may extend from Perth Water to 10 – 15 km into the upper
estuary. Thus, the resulting halocline from upstream bottom sediments to downstream
surface waters, i.e. the salt wedge, may extend for >10 km. (Figure 2.3).
(3) Salt wedge absent (freshwater dominated) in late winter and early spring (August to
October), when freshwater discharge is high. Fresh to brackish waters extend from the
upper estuary across the surficial (<5m deep) waters of the main basin to the Fremantle Sill
(5 km upstream of the mouth; Figure 2.3).
These descriptions are valid for the shallow waters (<5 m deep) of the estuary wherein the annual
hydrological cycle depends upon the seasonal variation in freshwater discharge regulating the
upstream migration of a tidally-forced salt wedge. However, the deep waters (>5 m deep) of the
main basin remain at high salinity year-round, with freshwater overlay causing stratification during
phase 3. Detailed consideration of important variations in hydrology occurring within the estuary
due to bathymetry, and also due to inter-annual variation in rainfall and resultant freshwater
discharge, is given in Appendix A. A similar stratification of the deeper pockets of the upper
estuary may also occur during phases 2 and 3, depending upon whether the magnitude of freshwater
discharge has been sufficient (>23-50 m3 s-1) to flush all high salinity water from the upper reaches
(Appendix A). Thus bathymetry, in addition to strong seasonality of river inflow and weak tidal
forcing, is also important in determining the hydrodynamics of the Swan River Estuary. In this
13
respect the presence of the Fremantle Sill is particularly significant, since its shallow depth
ultimately defines the limits of water exchange between ocean and estuary. Differences in estuarine
hydrology between years are driven by inter-annual variation in rainfall and resultant freshwater
discharge, which governs both the timing and period of the salt-wedge phases and related
stratification events.
Phase 1 – Salt wedge dominated ~ January to May
Phase 2a – Salt wedge waning ~ June to July
Phase 3 – Salt wedge absent ~ August to October
Phase 3 – Salt wedge propagation ~ November to December
0 5 10 15 20 25 30 35 40Distance from mouth (km)
Dep
th (m
)
Figure 2.3: Phases of salinity-depth contours within the Swan River Estuary throughout the
course of a typical year of freshwater discharge. (Adapted from Kurup et al., 1998; Swan River Trust, 2004a).
14
2.1.3 Eutrophication and Algal Blooms
The Swan River Estuary is a system highly suitable for a case study consideration of the causes and
implications of nitrogenous eutrophication. Nutrients have entered and accumulated within the
estuary over past years, and many authors suggest that levels are now sufficient for it to be
classified as eutrophic (e.g. John, 1994; Stephens and Imberger, 1996; Twomey and John, 2001).
Nitrogen has been identified as the chief nutrient limiting algal productivity within the estuary
(Thompson and Hosja, 1996; Douglas et al., 1997; Thompson et al., 1997), but increased
anthropogenic nitrogen loading has caused a dampening of this limitation on algal growth. The
estuary has a long-standing history of periodic algal blooms which, to some extent, are natural
phenomena (Hodgkin and Vicker, 1987); however, the heightened nutrient load within its more
recent history has increased both the frequency and intensity of problem phytoplankton blooms
(Thompson and Hosja, 1996). Marked anoxia upon algal biodegradation, fish kills (amongst other
fauna), loss of seagrasses and general changes in species’ composition have been associated with an
intensification of both toxic and non-toxic phytoplankton blooms within the Swan River Estuary
(Atkins and Klemm, 1987; Horner Rosser and Thompson, 2001). Further, these environmental
problems have major socio-economic implications including related public health costs, reduced
fisheries, decreased aesthetic and recreational value and loss of tourism (Twomey and John, 2001).
My aim in this section is to detail the main causes that have lead to the increased nutrient loading
and eutrophication of the Swan River Estuary, and to examine the consequences of such
eutrophication upon algal productivity and associated environmental and societal problems.
Nutrient Loading, Transport and Regeneration
Change in catchment land-use patterns has been the driving force of increased nutrient loading to
the Swan River Estuary. As the primary mechanism for nutrient transport, drainage to the estuary is
dominated by winter discharge from the Avon sub-catchment, contributing about 85% of total flow
(Atkins and Klemm, 1987; Kurup et al., 1998). However, during the critical summer period, inflow
is dominated by urbanised coastal plain drainage which typically has nutrient concentrations greater
than that of Avon drainage (Atkins and Klemm, 1987). The combined load is such that over the
period 1986 to 1996 the Swan River Estuary received an input of 550 tonnes of nitrogen per year.
Some of this nitrogen load may be transported to the ocean during freshwater discharge events in
winter and early spring, or during unseasonal rain events; however, ocean exports of nitrogen
during late spring, summer and autumn will typically be limited due to upstream tidal forcing of the
salt wedge (see Figure 2.3). The net effect of land-use changes upon nutrient contributions to the
estuary are separately considered for the two major sub-catchments below. The importance of
sediment nutrient regeneration as an internal supply of nitrogen is then discussed.
15
Agriculture is the main, present day land-use within the Avon sub-catchment. Only 25 to 35% of
the catchment retains its native vegetation of dry sclerophyllous woodland, which was mostly
cleared between 1940 and 1970 to enable sheep grazing and cereal cropping (Peters and Donohue,
2001; Viney and Sivapalan, 2001). Pastures typically are fertilised in autumn (April/May) to
promote good establishment of crops and growth in spring (Donohue et al., 2001). The timing of
fertiliser application heightens the availability of nutrients for surface and leachate transport at the
onset of winter rains (Donohue et al., 2001). Whilst the widespread application of fertilisers within
the Avon catchment has provided the source of nutrients primarily responsible for eutrophication of
the Swan River Estuary (Peters and Donohue, 2001), the potential capacity for nutrient transport
from this area has also increased. During the period 1986 to 1996, it was estimated that the
cumulative effect of land conversion within the Avon catchment facilitated a contribution of 356
tonnes or 65% of the total nitrogen entering the Swan River Estuary each year (Peters and Donohue,
2001).
The remaining 35% of nitrogen loading is generated within the much smaller sub-catchments of the
coastal plain. Over a similar period of monitoring (1987 to 1994), it was estimated that combined
load of nitrogen from the Avon and Ellen Brook sub-catchments was 484 tonnes per year (Viney
and Sivapalan, 2001). The coastal plain sub-catchment of Ellen Brook therefore contributes
approximately 128 tonnes or 23% of total nitrogen entering the Swan River Estuary each year.
Ellen Brook is the largest sub-catchment (664 km2) of the coastal plain and primarily retains an
agricultural land-use, mainly pasture for stock (Donohue et al., 2001). The impact of agriculture
within the Ellen Brook sub-catchment upon nutrient load and transport to the estuary is similar to
that detailed for the Avon catchment; however, the disproportionately higher nitrogen load for its
comparatively small area is a consequence of several other factors. The catchment comprises large
areas of highly permeable siliceous sand and groundwater is a major component of the stream’s
annual ephemeral flow, which is generally maintained from June to November following winter
rains (Donohue et al., 2001). A relatively small amount of superphosphate fertiliser is applied
directly to the discharge zone of the Gnangara groundwater system (Donohue et al., 2001; Peters
and Donohue, 2001). Thus, nutrient transport is significantly elevated via leachate flow and
overland flow when the water table is seasonally high in riparian and near-channel areas (Donohue
et al., 2001). Leguminous pasture has also been suggested to contribute to the large nitrogen pool
of the Ellen Brook sub-catchment (Donohue et al., 2001).
Within the urbanised portion of the coastal plain, nutrient inputs to the Swan River Estuary derive
mainly from industrial discharges, fertilisers (from gardens, parks and golf-courses) and seepage
from septic tanks and landfill sites (Atkins and Klemm, 1987). Soils of the residential and
16
industrial areas of the catchment are characteristically sandy and highly permeable (Donohue et al.,
2001). Reduced evapotranspiration via loss of native vegetation, combined with importation of
water for irrigation, have increased recharge and nutrient leaching to the unconfined groundwater
aquifers which govern stream flow on the coastal plain (Donohue et al., 2001). As a result, urban
nutrient sources are estimated to contribute 66 tonnes or 12% of total nitrogen to the estuary each
year (Peters and Donohue, 2001; Viney and Sivapalan, 2001). The quantity is small compared with
the agricultural inputs. It is thought that considerable amounts of nitrogen are lost via
denitrification within the groundwater aquifers which are comprised of naturally high levels of
dissolved organic carbon and a pH between 5 and 7 (i.e. ideal conditions for denitrification, see
Appendix D, Section D.7) (Donohue et al., 2001). However, the timing of urban discharge is
highly significant to algal productivity within the estuary (Peters and Donohue, 2001). As opposed
to the mainly ephemeral discharge of the agricultural catchment following winter rains, many of the
urban drains are perennial, with groundwater discharge sustaining flows throughout the dry season
(Donohue et al., 2001). Whilst groundwater discharge on the coastal plain contributes only 6% of
total annual flow to the estuary, the percentage increases to 55% between the summer period of
November to April (Linderfelt and Turner, 2001). Nitrogen-laden groundwater drives urban
drainage and is forced through sediments underlying the estuary (Linderfelt and Turner, 2001), thus
providing the major allochthonous nitrogen source utilised by summer algal blooms. Thompson
and Hosja (1996) suggest that unseasonal summer rainfall may also be important in providing an
external nitrogen supply to surface waters.
More importantly, autochthonous benthic regeneration of dissolved inorganic nitrogen may be
equivalent to or greater than annual allochthonous loads (Pennifold and Davis, 2001). The
sedimentary decomposition of organic nitrogen (including that contained within collapsed and
deposited phytoplankton blooms) by benthic fauna and bacteria, and resupply of ammonium via
excretion and microbial mineralisation may be the most important source of inorganic nitrogen
during the summer period. Transport of the regenerated ammonium from the bottom and pore-
waters of the benthos to the overlying pelagic zone of phytoplankton productivity is driven by
physical and biological processes (Douglas et al., 1997). Physical transport mechanisms include
diffusion, groundwater forcing, wind-induced hydrodynamic forcing and density-driven
displacement of pore-waters by advancement of the salt-wedge (Douglas et al., 1997; Oldham and
Lavery, 1999); bioturbation, resulting from sediment mixing and burrow irrigation of a ubiquitous
benthic fauna, likely is a sediment-water nutrient transfer process of pervasive importance
(Pennifold and Davis, 2001).
17
Algal Blooms
The combined effect of catchment modification and consequential increases in loading and
transport rate of nitrogen and other nutrients, both urban- and agriculturally-derived, has been an
apparent increase in the intensity and frequency of algal blooms within the Swan River Estuary
(Horner Rosser and Thompson, 2001). While this eutrophication has been well documented, the
occurrence of nuisance blooms of algae appears to have been a regular phenomenon of the estuary,
even before the onset of eutrophication. Further, the annual succession of population growth of
various algal species within the estuary has been an on-going topic of relevance to the residents of
Perth since settlement. Hodgkin and Vicker (1987) describe the history of algal pollution in the
Swan River Estuary, which is briefly summarised in Appendix B together with a consideration of
the present day, annual succession of bloom-forming phytoplankton species and the conditions
necessary for bloom-forming species.
Presently, it is the increasingly eutrophic status of the estuary and shift from productive macroalgal
species towards both toxic and non-toxic phytoplankton species, together with the consequences of
an intensification of frequency and severity of blooms, which is driving environmental and public
concern. The spatio-temporal distribution pattern of resident phytoplankton within the Swan River
Estuary is determined primarily by the salinity preference and tolerance of individual species, and
therefore by the factors (e.g. bathymetry, seasonality, rainfall events and freshwater discharge)
governing the hydrology of the system (John, 1987). However, excessive growth (blooms) of
phytoplankton is ultimately regulated by nutrient, mainly nitrogen, availability (John, 1987;
Thompson and Hosja, 1996).
A Recent Example
During February 2000, within the study period of this thesis (see Chapter 3), a massive, toxic
phytoplankton bloom occurred throughout the entire 60 km length of the Swan River Estuary
(Atkins et al., 2001). The rainfall events and subsequent hydrological conditions that facilitated the
bloom within the estuary are detailed in Appendix A. Briefly, heavy and cyclonically-derived rain
fell over the whole catchment during January and the subsequent discharge had reverted the entire
estuary to a winter-like freshwater state by February. The typical catchment of the estuary was
somewhat extended, and areas of the Avon sub-catchment that do not usually receive enough
rainfall to exert considerable flow (even during winter), produced waterways with their first
significant flows in forty years (Atkins et al., 2001). As a result, the discharge deposited over 800
tonnes of nitrogen (i.e. more than the average total annual load) and 35 tonnes of phosphorous into
the subsequently freshwater estuary under summer-time influence (i.e. with long day lengths and
relatively warm, calm and clear water) (Atkins et al., 2001). In the upper and mid-estuary,
18
concentrations of dissolved inorganic nitrogen were five to twelve times higher than normal
summer conditions (Atkins et al., 2001).
Whilst the high nutrient availability was highly conducive to phytoplankton growth, the low salinity
precluded exploitation by typical summer-time marine diatoms and dinoflagellates. Rather, several
species of freshwater algae proliferated, seeded either from small populations extant in the upper
reaches of the system (Atkins et al., 2001), or possibly from overflow of nearby eutrophic lakes (as
has been suggested by John, 1994, for previous algal blooms). Of these species, the non-nitrogen
fixing cyanobacteria Microcystis aeruginosa out-competed other phytoplankton in utilising the
abundant nutrient source to grow and reproduce. Microcystis aeruginosa has an effective gas
vacuole system which, combined with other carbohydrate metabolic features, enables it to float at
the water’s surface and receive the greatest share of sunlight for photosynthesis (Atkins et al.,
2001). Such a capability during the atypical hydrological circumstance, combined with high
nutrient availability, resulted in an extensive and spectacular bloom of M. aeruginosa. The bloom’s
appearance varied from a slight green tinge in open water to a dense bright green scum in many
sheltered bays and beaches (personal observation). The average cell density reached 15,000 ml-1 in
the general water column and 130 million ml-1 in the accumulated bright green scums at the peak of
the bloom (Atkins et al., 2001).
Figure 2.4: The Microcystis aeruginosa bloom in the Swan River, February, 2000. (Swan River Trust, 2000).
19
During one period, water concentrations of 8 μg l-1 microcystin toxin produced by this species
exceeded the level recommended by the World Health Organisation by a factor of sixteen (Atkins et
al., 2001). Indeed, 20-25 μg of M. aeruginosa biomass per kilogram of mammal biomass is fatal to
mice (Atkins et al., 2001). The high toxicity posed a tangible hazard to public health and
authorities closed the entire estuary for a period of two weeks, disallowing any contact with the
water until the bloom collapsed with the re-entrance of marine water (Atkins et al., 2001). The
upper salinity tolerance of M. aeruginosa is 10 g kg-1 (Atkins et al., 2001). As freshwater flow
subsided by mid-February, re-establishment of the salt wedge caused the M. aeruginosa bloom to
die-off first in the lower estuary, with subsequent mortality occurring upstream concurrent with
progression of the salt wedge. An estuary-wide collapse of the bloom occurred by the end of
February (Atkins et al., 2001).
Consequences of Eutrophication and Algal Blooms
The environmental, social and economic implications of the February 2000 M. aeruginosa bloom,
and moreover of the recently heightened frequency and intensity of both toxic and non-toxic blooms
within the Swan River Estuary, are far-reaching. Aside from posing a direct threat to human health,
the toxicity of several species of blooming cyanobacteria (e.g. Microcystis spp., Anabaena spp.,
Nodularia spp.) and dinoflagellates (Karlodinium spp., Gymnodinium spp.) may severely impact on
proximate fauna (Blackburn et al., 1997; Kanandjembo, 1998). For example, blooms of the
dinoflagellate Karlodinium micrum, whilst not toxic to humans, clog the gills of fish and cause
fatalities through asphyxiation (Swan River Trust, 2004b). Initial nutrient supply to the Swan River
Estuary provided by the onset of winter rains caused significant blooms of K. micrum and over
100,000 and 30,000 respective fish deaths in 2003 and 2004 (Swan River Trust, 2004b).
Furthermore, microbial oxygen consumption during the biodegradation of blooms and dead fish
may cause or exacerbate water column anoxia, potentially resulting in further fatalities of fish
and/or sessile fauna and flora (e.g. Jørgensen, 1980; Llansó, 1992). For example, the collapse of an
intense dinoflagellate (Gymnodinium spp.) bloom within the upper Swan River Estuary during
March 1994, produced water column anoxia that resulted in numerous fish kills (Kanandjembo,
1998). While it may be hypothesised that the comparatively sessile benthic fauna similarly suffers
mortality during periods of anoxia related to bloom collapse, the severity of stress to both benthic
and fish fauna populations will likely depend on the intensity of the algal bloom. Kanandjembo
(1998) found that an upper Swan River Estuary bloom of dinoflagellates during the summer of 1996
did not effect fish species composition but may have had a detrimental effect on their abundance.
However, there was no evidence that the bloom led to a significant change in either composition or
abundance of macrobenthic fauna. Finally, seagrasses may also be susceptible to mortality
20
resulting from prolonged periods of anoxia and/or excessive growth of epiphytic algae (Scanes et
al., 1998). Aside from seagrasses providing important habitat, their importance to nutrient cycling
within the Swan River Estuary has been described by Connel and Walker (2001) (also see
Appendix D, Section D.3).
The fish community of the estuary has an inherent ecological and preservation value due to its high
diversity (137 species from 70 families), while the high productivity of commercial species
provides a valuable fishery (John, 1994). More than 300 tonnes of fish are commercially caught on
an annual basis within the estuary, with an equivalent catch rate estimated for the recreational
fishery (John, 1994). Problem algal blooms can have marked impacts on these fisheries. The loss
of fish may also infer cascading effects to higher (and lower) trophic levels. Water bird diversity
(87 species) and abundance (≤10 000 birds at any one time) of the Swan River Estuary is amongst
the highest of any wetland in Western Australia (John, 1994). There is great potential for negative
impacts of algal blooms on bird populations (via either direct toxic effects or indirect effects
through loss of food source), since bird abundance is typically greatest at the time of year
coinciding with most problem blooms, i.e. from mid-spring to autumn. However, it must be
recognised that ecological relationships are complex. A reduction in zooplanktivorous fish may
allow the proliferation of zooplankton that graze on phytoplankton, thereby regulating bloom sizes
(Kanandjembo, 1998; Griffin et al., 2001). Increasing zooplankton abundance may help to
facilitate the re-establishment of planktivorous fish larvae and adults (Kanandjembo, 1998). Within
the Swan River Estuary, examination of these trophic relationships has only recently been instigated
(e.g. see Kanandjembo, 1998; Griffin et al., 2001).
Whilst the administrative and public health costs associated with problem algal blooms may be
high, other social and economic costs become apparent when the Swan River Estuary is considered
as a highly important resource to recreation and tourism (John, 1994). More than 25,000 vessels are
registered for use within the estuary whilst activities of water skiing, rowing, fishing and swimming
are highly popular to both local residents and tourists (John, 1994). The substantial commercial
pleasure cruise industry uses the estuary as a basis for tourism (John, 1994). The economic impact
of closure, or disruptions in use, of the estuary therefore results in loss of income to the many
dependant operators of commercial fisheries and recreational activities. Perhaps more important in
both an economic and cultural sense is the somewhat intangible aesthetic value of the estuary. The
visual landscape of Perth surrounded by the picturesque waterways of the Swan River Estuary is
important to the identity of the city. Moreover, promotion of that image is undefinably valuable for
the attraction of tourists. A sustained decrease in the aesthetic quality of the Swan River Estuary
due to algal blooms could lead to substantial losses in tourism revenue.
21
Consequently, the eutrophication of the Swan River Estuary is an environmental, social and
economic problem in need of scientific investigation, management and remediation. Our
understanding of the many complex interactions between algal production, nutrients, estuarine
ecology, sediment biogeochemistry, hydrology and physical control factors is still in its infancy. It
is not my intent to summarise here the plethora of worthwhile avenues of research regarding
eutrophication management within the Swan River Estuary. However, it is important to emphasise
the central role of nitrogen, rather than phosphorous, as the nutrient with the greatest capacity to
limit phytoplankton blooms. Thompson (1998) notes that phosphate concentrations in the estuary
would have to be reduced by a factor of 15 to 30 times to become more limiting to algal biomass
than nitrogen; and that since such a reduction is unfeasible in the short term, control of summer-
time algal blooms is most likely to be achieved through reduction in nitrogen availability. Horner
Rosser and Thompson (2001: p. 2593) further state that: “If workable strategies for reducing bloom
intensity and frequency, through the manipulation of nitrogen availability, are to be successful for
the Swan River Estuary, then the various flux processes influencing available nitrogen need to be
considered. These processes appear to play an important role throughout the water column in
maintaining nutrient levels and to promoting and sustaining phytoplankton biomass, and thereby
influencing the pattern of phytoplankton bloom succession.”
In this respect, a consideration of the flux of nitrogen from the sediment surface is critical,
especially a recognition of the availability of ammonium near the sediment surface during the
period of highest algal productivity (November to May). While the heightened load and transport
of land-derived nutrients is ultimately responsible for the increased eutrophic status of the estuary
(and catchment nutrient reduction strategies are incredibly important!), it appears that management
of sediment nitrogen fluxes may play a vital role in the short to long term health of the estuary.
Benthic fauna are ubiquitous within the estuary and must be considered an integral constituent of
the sediment composition. An understanding of how benthic fauna influence sedimentary microbial
nitrogen cycling may ultimately allow management of benthic faunal communities to help regulate
the supply of nitrogen to the water column and prevent the deleterious consequences of algal
blooms.
22
2.2 AUSTRALONEREIS EHLERSI AND SIMPLISETIA AEQUISETIS: POLYCHAETE
CONSTITUENTS OF THE SWAN RIVER ESTUARY’S BENTHIC
MACROFAUNAL COMMUNITY
2.2.1 Benthic Macrofaunal Community Composition of the Swan River Estuary
The overall theme of this thesis is a functional consideration of the two subject species,
Australonereis ehlersi and Simplisetia aequisetis, in the ecology and nitrogen-transforming
processes of the Swan River Estuary. Interpretation of the functional-importance of these two
polychaetes first necessitates an understanding of their positioning within the estuarine benthic
community. In terms of species richness, abundance and productivity, estuarine benthic
macrofaunal communities typically are dominated by polychaetes (nereidids, capitellids and
spionids), molluscs (bivalves and gastropods) and crustaceans (amphipods and decapods) (Day et
al., 1989; Rose, 1994; Kanandjembo et al., 2001). The overall dominance or ecological importance
of any one taxon (e.g. for the Polychaeta see Knox, 1977 and Wilson et al., 2003a) varies both
within and between estuaries, and over time (Day et al., 1989; Rose, 1994; Kanandjembo et al.,
2001; Pennifold and Davis, 2001). Spatio-temporal dynamism in community composition is a
manifestation of the response of constituent species to complex variations and interrelations of
governing environmental factors, both physicochemical (e.g. salinity, temperature and other
chemistry determined by hydrological regime and estuarine physiognomy) and/or biological (e.g.
competition, predation, trophic relations, dispersal mechanisms and evolutionary biogeographical
constraints etc.). As such, every estuary possesses a unique benthic faunal community, and scrutiny
of the distribution and ecological role, dominance or importance of constituent taxa must
necessarily be conducted on an estuary-specific basis. In this Section I consider briefly the broad
expression, and major determinants, of benthic species distribution, population structure and
cumulative community dynamics within the Swan River Estuary; in Appendix C, the position of
A. ehlersi and S. aequisetis within the benthic macrofaunal community of the estuary is discussed.
The Swan River Estuary may be defined by five distinct hydrological regions or zones: (i) the upper
estuary, (ii) the wide sill of Perth Water, (iii) the main basin (i.e. the deep trough and wide basin),
(iv) the shallow Fremantle sill and (v) Fremantle Harbour (see Section 2.1.2; Figure 2.2). To a
large extent, the characteristic hydrological, particularly salinity, regime of these zones governs the
species composition and distribution of the estuarine community (Chalmer et al., 1976; Hodgkin,
1987; Loneragan et al., 1989; Loneragan and Potter, 1990; Rose, 1994). Utilising similar species
composition (mainly of molluscs), Chalmer et al. (1976) separated the estuary into three main
regions or biotopes: the upper, middle and lower estuary (Figure 2.5). The biological and
hydrological definitions of the upper estuary are synonymous. The middle estuary is defined by the
23
stretch of Perth Water to the wide main basin, excluding the deep trough. The deep trough,
Fremantle Sill and Fremantle Harbour together comprise the lower estuary. A similar classification
based on the community composition and distribution of teleost fish was presented by Loneragan
and Potter (1990).
Hodgkin (1987) argued that salinity regime and the tolerance of constituent taxa primarily
determine this tri-part zonation, which is likely applicable to the broad patterns of distribution and
composition of most biotic groups within the Swan River Estuary, including the Polychaeta. Based
on salinity tolerance, estuarine organisms are classified into the following groups (Chalmer et al.,
1976; Hodgkin, 1987; Rose, 1994):
(i) oligohaline species – freshwater affinity, limited estuarine representation and no marine
occurrence (salinity tolerance <0.5 g L-1)
(ii) ‘true’ euryhaline species – estuarine endemics with neither marine nor freshwater
representation (2 - 60 g L-1);
(iii) marine euryhaline species – marine affinity but with continuous estuarine
representation (5 - 36 g L-1)
(iv) stenohaline species - marine affinity with only temporary or sporadic estuarine
representation (25 - 40 g L-1).
UPPER ESTUARY
Main
Basin
Fremantle
Harbour
Fremantle Sill
Perth Water
Kent Street
Weir
The Narrows
Heirisson Island
Fremantle Traffic Bridge
Salter Point
Preston Point
Deep Trough
MIDDLE ESTUARY
LOWER ESTUARY
N
5 km
Figure 2.5: Biotopic regions of the Swan River Estuary (after Chalmer et al., 1976).
24
The concept of a biological zonation of the Swan River Estuary into its lower, middle and upper
components implies the definition of boundaries to these zones, and implies a certain level of
stability in community composition within these bounds. While it is conceptually important to
define regions of ecological similarity, the interpretation that regional boundaries are fixed or that
the compositions of constituent taxa are immutable should be rejected. In large part, the valid
broad-scale resolution and recognition of the three biotic zones is determined by the fixed
physiognomic / bathymetric features of the estuary, primarily the sills at the junction between each
zone (Fremantle Sill and Perth Water). These features largely govern the hydrological processes
(including salinity regime) of each region which in turn regulate the broad-scale expression of
community composition. However, while important, physiognomy is only one regulator of
hydrological processes and other factors also impart control and variance on faunal distributions.
Many physicochemical and biological regulatory factors, and their complex relationships, will
ultimately define the expression of species’ population structure and community constituency
(further see Appendix C).
In the following sections I describe current knowledge about the distribution, habit and life history
of the study species, S. aequisetis and A. ehlersi, in the Swan River Estuary and elsewhere. The
environmental factors that may influence polychaete ecology, and particularly the distribution of
each species, will be discussed. However, prior to a consideration of the population biology of
A. ehlersi and S. aequisetis it is pertinent to first briefly discuss recent changes and confusion in
their taxonomy, since each species has previously been referenced under various synonyms and/or
misidentifications.
2.2.2 Taxonomy of A. ehlersi and S. aequisetis
Australonereis ehlersi and S. aequisetis belong to the Nereididae family. Within Australia, about
93 species of nereidids (representing 22 genera) are presently known, with the principal estuarine
genera represented by Australonereis, Olganereis and Simplisetia (formerly Ceratonereis sensu
lato, see below) (Wilson et al., 2003). Nereidids generally are distinguished by having the
following characters: an eversible pharynx with one pair of lateral jaws, a first segment with
tentacular cirri and a prostomium which is bluntly conical to trapezoidal (Wilson, 2000; Wilson et
al., 2003).
Detailed morphological description of A. ehlersi is provided by Hartman (1954), Hutchings and
Turvey (1982), Hutchings and Reid (1990), Wilson et al. (2003b) and Bakken and Wilson (2005).
The species was first described by Augener (1913) as Nereis (Leonnates) ehlersi and by Monro
(1938) as Leptonereis ehlersi. Hartman (1954) first defined the monospecific genus of
25
Australonereis, which differs most notably from other nereidid genera for having, on the ventrum,
paired fleshy transverse ridges/papillae lateral to the ventral cirri of anterior segments (these ridges
are absent in S. aequisetis) (Figures 2.6 and 2.7). Further, compared to S. aequisetis, A. ehlersi has a
flattened body and no simple falcigers (P. Hutchings, pers. comm.). For individuals of similar
length, A. ehlersi is broader and has a greater body volume than S. aequisetis (Hartman, 1954; pers.
obs.; Figure 2.6). Maximum body length may be dependant upon environmental setting. Hutchings
(1982) documents that A. ehlersi specimens may be up to 6 cm long and 0.5 cm wide; however,
Hartman (1954) notes that individuals from Lakes Entrance in Victoria may measure up to 14 cm
long and 1.2 cm wide. In Port Phillip Bay, A. ehlersi attains a length of 10-12 cm (Dorsey, 1981),
while the maximum body length in the Swan River Estuary is ~8 cm (Chapter 4).
Simplisetia aequisetis is described morphologically by Hutchings and Glasby (1985), Wilson et al.
(2003b) and Bakken and Wilson (2005). In comparison to A. ehlersi, the differentiating features of
S. aequisetis are a pharynx with scleroprotein paragnaths and posterior parapodia with simple
falcigers; in contrast, A. ehlersi has a pharynx with fleshy papillae in place of paragnaths
(P. Hutchings and R. Wilson, pers. comm.). Adults from both Port Phillip Bay and the Swan River
Estuary attain a body length of 4-6 cm (Dorsey, 1981; Chapter 4). Augener (1913) first described
the polychaete as Nereis (Ceratonereis) aequisetis1. The species descriptions by Hartmann-
Schröder (1982) and Hutchings and Glasby (1985) elevated the subgenus to full generic status, i.e.
Ceratonereis aequisetis. Hartmann-Schröder (1985) further reviewed species assigned to
Ceratonereis and recognised that the genus (comprising a large number of species ranging from
muddy estuaries to the deep sea), contained several unrelated groups of species. She therefore
formally defined three subgenera, two of which were new (R. Wilson, pers. comm.). Ceratonereis
(Simplisetia) included species with fused heterogomph falcigers in the dorsal neuropodial position,
including aequisetis (Bakken and Wilson, 2005). Finally, Khlebovich (1996) elevated the
Simplisetia subgenera to full generic status. Other synonym attributions have included C. cf.
anchylochaeta2 (Hartmann-Schröder, 1979) and C. pseudoerythraeensis (Hutchings and Turvey,
1982; Hutchings and Murray, 1984). The latter synonym was instigated as a new species to
delineate Australian nereidids which had previously been incorrectly identified as C. erythraeensis 1 The type material of both A. ehlersi and S. aequisetis collected by Augener (1913) originated from Melville
Water in the middle section of the Swan River Estuary. This is within the immediate vicinity of the long-term
sampling site for this study (see Chapter 4). 2 The synonymising of Hartmann-Schröder’s (1979) record of C. cf. anchylochaeta from Port Hedland likely
is incorrect; the species is more similar to the Indonesian C. anchylochaeta. Thus, S. aequisetis probably does
not occur this far north on the Western Australian coastline (Glasby, pers. comm.; Glasby, Wilson and
Bakken, in prep.).
26
(Hutching and Turvey, 1982; also see below). Arguments for synonymy of S. aequisetis are detailed
in Hutchings and Glasby (1985), Wilson et al. (2003b) and Bakken and Wilson (2005).
(i) A. ehlersi (ii) S. aequisetis
(a) Dorsal
(b) Ventral
3 mm
3 mm
2 mm
2 mm
Figure 2.6: Dorsal (a) and ventral (b) photographs of A. ehlersi (i) and S. aequisetis (ii).
Figure 2.7: Ventral view of A. ehlersi setigers 10-17, illustrating the ventral papillae (‘pap’) that
characterise the genus (Hartman, 1954).
27
Given the previous taxonomic revisions and inaccuracies, when referencing Australian nereidids in
this thesis, I have made every effort to denote species under the present classification system of
Wilson et al. (2003b). Where known, the currently accepted taxonomy of other species groups has
also been utilised; otherwise, species are generally referenced as provided in source documents. It is
highly likely that previous identifications of C. erythraeensis in the Swan River estuary are
erroneous (e.g. Monro, 1938; Chalmer et al., 1976; Pennifold and Davis, 2001) and the species
should be referred to as S. aequisetis. While C. erythraeensis (now Simplisetia erythraeensis)
remains a valid species (described from Madagascar), all published reports of the species in
Australian waters are incorrect (R. Wilson, pers. comm.). Depending on the area, Australian
references to C. erythraeensis more likely indicates the presence of one or more of the following
species: S. aequisetis (southern distribution), S. amphidonta (South Australia, Victoria, Tasmania
and southern New South Wales), S. transversa (South Australia and Western Australia, including
near Bunbury), S. limnetica (Hawkesbury River, New South Wales) or S. turveyi (Merimbula, New
South Wales) (Hutchings and Turvey, 1982; Hutchings and Glasby, 1982, 1985). The only other
Simplisetia species (formerly Ceratonereis sensu lato) known from Australia is S. lizardensis from
a creek on Lizard Island, north-eastern Australia (Wilson et al., 2003b). Australian species of
Ceratonereis sensu stricto are not closely related to the Simplisetia and typically are marine
(Wilson, 2003a). Australian Ceratonereis spp. include C. australis (Hartmann-Schröder, 1985), C.
lapinigensis (Grube, 1878), C. longiceratophora (Hartmann-Schröder, 1985), C. mirabilis (Kinberg,
1866), C. perkinsi (Hartmann-Schröder, 1985), Ceratonereis singularis (Treadwell, 1929) and C. cf.
singularis (Treadwell, 1929).
Regarding populations in the Swan River Estuary, delineation of A. ehlersi and S. aequisetis is
important since they are apparently the only nereidids that inhabit the system. It is likely that past
taxonomic confusion has led not only to erroneous identifications of the two species, but perhaps
also wrongful combinations of the two species into one species group. For example, in a recent
study on the effect of benthic macrofaunal communities on nutrient cycling in the middle Swan
River Estuary, Pennifold and Davis (2001) list C. erythraeensis (i.e. S. aequisetis) amongst the ten
most abundant species in both winter and summer, but do not include the presence of A. ehlersi.
Their study site was adjacent to the long-term study site of this thesis, where both A. ehlersi and
S. aequisetis were found in considerable abundance, at least in one season (see Chapter 4). While it
is possible that A. ehlersi did not rank amongst the ten most abundant species in the study of
Pennifold and Davis (2001), a more likely suggestion is that they failed to recognise that two
species were present in the study and referred to all material as a single species (C. erythraeensis).
28
2.2.3 Distribution and Feeding Habit of A. ehlersi
Swan River Estuary Population
Noting a general paucity of macrobenthic studies of estuaries, Dürr and Semeniuk (2000) comment
that little is known about estuarine polychaete population biology in Western Australia. As such,
limited information is available on the autecology of either A. ehlersi or S. aequisetis within the
Swan River Estuary. The review of Chalmer et al. (1976) noted that A. ehlersi has been recorded
only from the lower and middle estuary, with no records from the upper estuary or Perth Water
portion of the middle estuary (Figure 2.2). Studies by McShane (1977) and Kanandjembo (1998)
further detail an absence of A. ehlersi in the latter regions. Within the shallows of the Melville
Water portion of the middle estuary (Figure 2.2), Rose (1994) found A. ehlersi to be the sixth most
abundant polychaete but comprised only 0.5% of the total polychaete abundance (27 specimens out
of 5516 polychaetes sampled in total) (Appendix C, Table C1[B]). Individuals collected from this
region by McShane (1977) weighed between 50 and 500 mg (presumably wet weight). McShane
(1977) documented that A. ehlersi reside in densely aggregated burrows or tubes to a sediment
depth no greater than 15 cm. Further, he found that A. ehlersi is intolerant of depressed oxygen
conditions, but broadly tolerant of salinities ranging from 2 to 36 g kg-1 (and capable of osmotic
regulation of its coelomic fluid). During the winter study period, individuals in a reproductive state
were observed swimming in overlying waters of lowered salinity (~7 g kg-1), which was interpreted
as indicative of a pelagic spawning event, with larval stages tolerant of brackish salinities
(McShane, 1977). To my knowledge, no further information exists regarding the distribution,
biomass, productivity or life history of the Swan River Estuary population of A. ehlersi.
Australian Distribution
The known Australian distribution of A. ehlersi stretches around the southern half of the continent
from the Swan River Estuary in the southwest to the Calliope River Estuary in Gladstone,
Queensland (Hartman, 1954; Day and Hutchings, 1979; Hutchings, 1982; Hutchings and Turvey,
1982; Hutchings and Murray, 1984; Hutchings and Reid, 1990; Wilson et al., 2003b; Bakken and
Wilson, 2005). The species has also been documented from estuaries in most regions of Tasmania
(Edgar, 1999). This Great Southern biogeographic distribution largely equates to the temperate (i.e.
non-tropical) waters of Australia and is broadly shared with numerous marine fauna, including other
nereidids such as S. aequisetis and Olganereis edmondsi (Figure 2.8; Glasby et al., 2000).
Factors Determining Distribution
Similar to its abundance in the Swan River Estuary, A. ehlersi is not a dominant species in terms of
abundance within the neighbouring Peel-Harvey Estuary, Mandurah (Rose, 1994); nor in the nearby
29
Figure 2.8: The Great Southern Australian biogeographic distribution (dashed line) of
Australonereis ehlersi and Simplisetia aequisetis. Adapted from Glasby et al. (2000), after Wilson and Allen (1987).
Leschenault Estuary, Bunbury (Dürr and Semeniuk, 2000). A comparatively low density of
A. ehlersi may be a general characteristic of the species. In the estuary of the Hawkesbury River
(New South Wales), densities of A. ehlersi are about 6-fold lower than S. aequisetis (MacFarlane
and Booth, 2001), a pattern repeated in Tasmanian estuaries (Edgar, 1999). Nevertheless, dense
beds of A. ehlersi have been reported from some locations, e.g. at Lakes Entrance and in Port
Phillip Bay, Victoria (Hartman, 1954; Dorsey, 1981). Similar to the Swan River Estuary, A. ehlersi
was recorded only within the lower to middle region of the Leschenault Estuary (Dürr and
Semeniuk, 2000), which is suggestive of a marine euryhaline habit. The Lakes Entrance and Port
Phillip Bay populations of A. ehlersi in Victoria are located in waters of marine salinity (Hartman,
1954; Dorsey, 1981). Within the estuaries of New South Wales, A. ehlersi has been found to
underlie waters with salinities ranging from 5 to 37 g kg-1 (Hutchings and Murray, 1984). Given
McShane’s (1977) determination of A. ehlersi’s great capacity for osmoregulation, the general
absence of the species in the upper regions of Australian estuaries is confounding and perhaps
relates less to salinity than to other determining factors.
30
Rose’s (1994) macrobenthic surveys of the Swan River and Peel-Harvey Estuaries sampled only
sediments of shallow waters (depth <1.5 m). Dürr and Semeniuk (2000) found A. ehlersi to be
present in the deeper muddy sediments of Leschenault Estuary; however, they did not observe a
depth-preference as the species was also recorded in the shallows. Australonereis ehlersi is found
within both subtidal and intertidal sediments of Port Phillip Bay (Dorsey, 1981). In seeking to
identify the causal reasons for A. ehlersi distribution patterns in the bay, Dorsey (1981) implicated
factors other than water depth as being important, including a positive association with well-sorted
fine sand. In contrast, Dürr and Semeniuk (2000) did not find A. ehlersi to demonstrate a substrate-
preference, since the species was found within sand, mud and muddy sand. For estuaries of New
South Wales, Hutchings and Murray (1984) similarly note that the species may be found in
sediment substrates ranging from fine to coarse-grained sand to muddy sand. Australia-wide,
Hutchings and Turvey (1982) report that A. ehlersi may be found within sandy mud of estuarine or
sheltered bays, often associated with seagrass beds. An apparent association with stands of
Posidonia, Halophila or Zostera seagrasses, or Rhizophora mangroves, in muddy sand or sand-flats
has been further documented by Hutchings and Reid (1990).
Citing McShane’s (1977) findings of an intolerance towards low dissolved oxygen conditions,
Dorsey (1981) suggested that oxygen depression, associated with organic matter loading, was
implicated in excluding A. ehlersi from near sewage outfalls in Port Phillip Bay. A similar
argument was invoked to help explain the lack of A. ehlersi within the high intertidal zone, wherein
burrow waters are susceptible to hypoxia at low tide (Dorsey, 1981). Additionally, dense shell beds
in some parts of Port Phillip Bay may impart a physical barrier that preclude the construction of
A. ehlersi burrows. Within the bay, A. ehlersi is a prey item of wading birds and a species of
stingarie; predation is probably a significant influence on the size and structure of populations
(Dorsey, 1981; Wilson, 2000). Furthermore, Dorsey (1981) suggested that dense assemblages of
opportunistic species in areas of high organic matter loading may out-compete A. ehlersi for
infaunal space. Thus, many biotic and abiotic factors may potentially control the distribution,
abundance and population structure of A. ehlersi, likely in a synergistic manner (Dorsey, 1981).
Feeding Habit
Australonereis ehlersi is a strictly tubiculous species, inhabiting angular, U-shaped burrows to a
depth of 15 to 20 cm, the ends of which protrude slightly above the sediment surface (Hartman,
1954; McShane, 1977; Dorsey, 1981; Hutchings, 1982; Hutchings and Turvey, 1982; Hutchings and
Reid, 1990). However, documentation regarding the feeding habit of the species is largely
speculative. Rose (1994) considered A. ehlersi within the Swan River Estuary to be a non-specific
deposit feeder. Dorsey (1981) hypothesised that individuals may construct a mucous suspension
31
feeding net to capture food; however, he too primarily considered A. ehlersi to be a deposit feeder.
In contrast, Rouse (2000) considered A. ehlersi to be raptorial and probably omnivorous, using its
powerful jaws attached to the eversible pharynx to grasp food that comes near the burrow
entrance/s. The common presence of large, curved jaws in the Nereididae is not necessarily
indicative of a predatory or raptorial habit; the jaws may be used for gathering sediments, as well as
for scavenging large food items (Dorsey, 1981; Wilson, 2000). Surface deposit feeding and
herbivory are the most common nereidid feeding habits, but the Family also contains omnivores,
carnivores and suspension-feeders (Wilson, 2000; Wilson et al., 2003a). Feeding behaviour can
vary even within and between populations of a single species (Wilson et al., 2003a).
2.2.4 Distribution and Feeding Habit of Simplisetia aequisetis
Swan River Estuary Population
Simplisetia aequisetis has been recorded from the lower, middle and upper regions of the Swan
River Estuary, typically comprising one of the most abundant macrobenthic species (Chalmer et al.,
1976; Rose, 1994; Kanandjembo, 1998; Kanandjembo et al., 2001; Pennifold and Davis, 2001).
Within shallow waters of the middle estuary (<1.5-2.0 m deep), Rose (1994) found that
S. aequisetis was more numerous than any other benthic macroinvertebrate collected (Appendix C,
Table C1[B]). At Pelican Rocks, a shallow water site in the middle estuary (very close to the long-
term study site of this thesis, see Chapter 4), Pennifold and Davis (2001) found S. aequisetis to be
the most abundant species in winter and the second most abundant species in summer (after the
bivalve mollusc Xenostrobus securis). The density of S. aequisetis retained by Rose (1994) on a 1
mm sieve ranged between 173 individuals m-2 at Dalkeith in the autumn of 1987 and 4074
individuals m-2 at Applecross during the spring of 1986. For the middle estuary, he found that S.
aequisetis density was significantly greater in summer than winter, and at Applecross compared to
other sites. Sampling both shallow and deep waters of the upper estuary using a 500 µm mesh
sieve, Kanandjembo (1998) found S. aequisetis to be the third most abundant macrobenthic species,
with the bivalve mollusc (Arthritica semen) and spionid polychaete (Pseudopolydora sp.) being
more numerous (Appendix C, Table C1[A]). The average density of S. aequisetis over her
sampling period from winter 1995 to autumn 1997 was approximately 250,000 individuals m-2 in
shallow waters and 13,000 individuals m-2 in deep waters. Rose (1994) documented that a sieve size
of 1 mm captured about the same abundance (>95%) as a mesh of 500 µm; thus, it appears that the
density of S. aequisetis may be greater in the upper than middle estuary. No information is
available on the typical abundance or other population statistics of S. aequisetis in the lower Swan
River Estuary.
32
A wet weight of about 80 mg previously has been determined for large adult S. aequisetis
individuals in the middle estuary (De Roach et al., 2002)3. The only estimates of S. aequisetis
biomass within the Swan River Estuary are provided by Rose (1994), for the shallows of the middle
region. In terms of biomass, S. aequisetis is less dominant than three bivalve mollusc species
(Sanguinolaria biradiata, Xenostrobus securis and Musculista senhousia) but is the most dominant
polychaete species in this region (Appendix C, Table C1[C]; Rose, 1994). At his study sites, wet
weight biomass of S. aequisetis ranged between 5.2 g m-2 at Matilda Bay in winter 1986 and 100.5 g
m-2 at Applecross in spring 1987 (approximately corresponding to a minimum dry weight4 biomass
of 0.68 g m-2 and a maximum of 19.4 g m-2). Similar to patterns of density, S. aequisetis biomass
was significantly greater during spring and summer than autumn and winter, and at Applecross and
Deepwater Bay compared to other sites of the middle estuary. However, there was also a
significant season by site interaction, with the greatest biomass of S. aequisetis at Deepwater Bay
occurring in winter, indicating some site-specificity in the determinants of biomass (Rose, 1994).
Suggested determinants of the spatio-temporal distribution of S. aequisetis in the Swan River
Estuary and elsewhere are discussed below.
Australian Distribution
As noted above, the records of occurrence for S. aequisetis represent a Great Southern, Australian
distribution largely congruent with that of A. ehlersi, stretching around the southern half of the
continent from the Swan River Estuary in the southwest to the Calliope River Estuary in Gladstone,
Queensland (Figure 2.8; Hartman, 1954; Day and Hutchings, 1979; Hutchings, 1982; Hutchings and
Turvey, 1982; Hutchings and Murray, 1984; Hutchings and Glasby, 1985; Glasby et al., 2000;
Wilson et al., 2003b; Bakken and Wilson, 2005). Like A. ehlersi, S. aequisetis has also been
documented from Tasmania, but not from the estuaries of the southwest region of the island state
(Edgar, 1999). Where present, S. aequisetis typically is one of the most abundant constituents of
the polychaete assemblage, if not entire benthic macrofaunal community; including not only within
the Swan River Estuary, but also the Peel-Harvey Estuary, Leschenault Estuary, Hardy Inlet and
Wilson Inlet in Western Australia (Chalmer and Scott, 1984; Rose, 1994; Platell and Potter, 1996;
Dürr and Semeniuk, 2000); Port Phillip Bay in Victoria (Dorsey, 1981, 1982); and the estuary of
Hawkesbury River and many other estuaries of New South Wales (Hutchings and Murray, 1984;
MacFarlane and Booth, 2001) and Tasmania (Edgar, 1999).
3 N.B. – A typographical error exists in the Material and methods section of De Roach et al. (2002): ~ 0.8 g
wet weight should read ~ 0.08 g wet weight. 4 Dry weight calculation based on the assumption that: dry weight = [0.130 to 0.193] x wet weight (Cammen,
1980; Kristensen 1984; Sardá et al., 2000). For ash-free dry weight and/or carbon weight conversion factors, see the methodology section of Chapter 4.
33
Factors Determining Distribution
The inverse relation between S. aequisetis density and overlying water depth reported by
Kanandjembo (1998) within the upper Swan River Estuary has also been documented in other
Western Australian estuaries. Platell and Potter (1996) likewise found S. aequisetis to be far less
abundant in deeper than shallow waters of Wilson Inlet. Noting a reduction in abundance with
increasing sampling depth along transects, Dürr and Semeniuk (2000) described a depth-controlled
concentric distribution pattern of S. aequisetis in the Leschenault Estuary. By contrasting
incidences of S. aequisetis with tidal depth in Tasmanian estuaries, Edgar (1999) denoted the
greatest occurrence of specimens at ~30 cm below the low-water mark, with a trend of decreasing
records for both deeper waters and shallower (including intertidal) depths. In most estuaries, this
concentric distribution pattern is generally concomitant with the distribution of seagrasses and
benthic algae, a purported component of S. aequisetis’ diet (Platell and Potter, 1996). Australia-
wide, it is true that the distribution of S. aequisetis is often associated with estuarine algae,
seagrasses and mangroves (Hutchings and Turvey, 1982; Hutchings and Murray, 1984; Hutchings
and Glasby, 1985). Within the Swan River Estuary, Rose (1994) found S. aequisetis density to be
significantly positively correlated with the biomass of Ulva spp. (green algae), Gracilaria spp. (red
algae) and Halophila ovalis (seagrass). In Wilson Inlet, the relationship was true for S. aequisetis
and the resident seagrass Ruppia spp. (Plattel and Potter, 1996).
Whether or not the plant materials provide a food source that is a direct determinant of S. aequisetis
distribution and abundance, or the relationship is rather more indirect or coincident, requires further
investigation. For example, Platell and Potter (1996) also offered that seagrass habitat may provide
protection from predators. Indeed, Dorsey (1981) suggested that fish and invertebrate predators
consume S. aequisetis at subtidal depths in Port Phillip Bay, thereby helping to explain the species’
absence in this zone. An alternate suggestion is that the shallower photosynthetic zone of estuaries
is largely coincident with other determinant factors such as increased dissolved oxygen levels and
higher temperatures (Platell and Potter, 1996; Dürr and Semeniuk, 2000). Other physicochemical
(e.g. salinity, sediment particle size and organic matter content) and biological parameters (intra-
and inter-specific effects), and their interactions, may also correlate with small and large scale
distribution patterns of S. aequisetis.
In addition to the positive relationships with plant material described above, Rose (1994) also found
Swan River Estuary densities of S. aequisetis to be negatively correlated with the biomass of
content. Furthermore, his analytical stepwise regression methodology could not define a single
environmental variable to explain spatio-temporal variations in S. aequisetis density. Within their
long-term study of patterns of S. aequisetis (and other polychaetes) abundance within the
Leschenault Estuary, Dürr and Semeniuk (2000) similarly noted a lack of coupling to seasonal
variation in abiotic factors. Furthermore, much variability in abundance occurred between sites,
within a single year and from year to year; with the population dynamics at different sites being
35
largely asynchronous. They argued that such variability is a common and definitive trait of many
estuarine polychaete populations. Rather than a single dominating causal factor, it is likely that a
synergy of determinant variables governs the distribution and population dynamics of S. aequisetis
(and many other polychaete species) within any given estuary; and the combination of determinant
variables may be quite different in another estuary. Experimental manipulative population studies
are largely required, on an estuary-specific basis, to tease out the most important combinations of
physicochemical and biological determinants of polychaete distributions.
It is important to here reiterate that investigations of biological determinants of polychaete
distributions, e.g. competition, predation, other trophic relations, dispersal/reproductive
mechanisms, intraspecific interactions and biogeographical constraints, are often neglected,
particularly in multivariate studies which usually relate only easily measurable abiotic factors to
population parameters of species. To some extent, insufficient study of potential biological
determinants may contribute to the uncertainties concerning the factors controlling the spatio-
temporal distributions of S. aequisetis and other polychaete species. In a relatively unique study
arising out of the Ph.D. thesis of Dorsey (1981), Kent and Day (1983) found that wading shore birds
(red-necked stints, sharp-tailed sandpipers and curlew sandpipers) and fish (flounders) within Port
Phillip Bay feed selectively on larger individuals of S. aequisetis, which has a significant effect on
the structure of the nereidid population. Further, they document that a decreased density of
S. aequisetis adults increases juvenile recruitment, which apparently counter-balances losses from
predation (in terms of abundance). The Swan River Estuary is also an important habitat for
migratory waders, thus avian predatory effects may similarly be expected within intertidal areas.
Further, Simplisetia aequisetis is a known prey item of four species of fish (Sillaginodes punctatus,
Silago schomburgkii, Pomatomus saltatrix and Argyrosomus hololepidotus) from estuaries of
south-western Australia (Chalmer and Scott, 1984); with at least another two fish species
(Pseudogobius olorum and Amniataba caudavittata) found to consume unidentified nereidids –
S. aequisetis and/or A. ehlersi, within the Swan River Estuary (Kanandjembo, 1998). Largely
unknown are the influences of predation, interspecific competition and other biological interactions
as determinants of nereidid distribution and population structure in Australian estuaries.
Feeding Habit
A degree of herbivory was above insinuated in considering the determinants of S. aequisetis
distribution patterns. Whilst Rose (1994) notes that serrations on the jaws of S. aequisetis are well
adapted to manipulating large pieces of plant material, nonetheless caution should be exercised
when extrapolating jaw morphology to feeding habit. In addition to plant debris, Dorsey (1981)
found the diet of S. aequisetis to include amphipods, arthropod limbs, faecal pellets, foraminiferans,
36
diatoms, the remains of other polychaetes and also freshwater cladocerans (i.e. Daphnia spp.
discharged with sewage wastewater). Gut contents included a large range of sediment particle
sizes, broadly suggestive of a non-selective deposit feeding habit (although very fine sediments may
not be ingested). The formation of faecal pellets was not observed in the gut. Whether these food
items are consumed with the sediment, or are raptorially seized, remains to be tested. Wilson (2000)
classified S. aequisetis as a scavenging deposit feeder.
Importantly, S. aequisetis may be non-tubiculous throughout most of its life but reproductive adults
have been found to construct burrows or tubes in which eggs are laid for brood protection (Dorsey,
1981, 1982; Hutchings and Turvey, 1982; De Roach et al., 2002). In relation to reproductive traits
(e.g. egg positioning), Dorsey (1981) describes the morphology of these tubes in great detail (also
see Chapter 4). Mucous is used to bind and line the burrows to make a resilient sheath (often even
withstanding the sieving collection process) and that these burrows may be irrigated (Dorsey,
1981). De Roach et al. (2002) also documented burrow construction to a depth of ~6cm in an
experimental situation and vigorous intermittent irrigation by large mature individuals of
S. aequisetis. The possibility that burrow ventilation may equate to some form of suspension
feeding requires further investigation.
37
CHAPTER 3
THE NITROGEN CYCLE: PROCESSES, METHODOLOGY AND
BENTHIC MACROFAUNAL EFFECTS
3.1 PROCESSES DEFINING THE NITROGEN CYCLE
Prior to outlining the impacts of the polychaete and benthic fauna on sedimentary nitrogen
regulation, it is relevant to define the components of the estuarine nitrogen cycle. While an
increased total nitrogen supply is fundamental to the eutrophication of an estuary, and the
dynamics of physical nutrient transport processes are highly important (Section 2.1), it is the
rate of the various microbial nitrogen transforming processes that determine the availability of
nitrogen for algal growth. A summary of the estuarine nitrogen cycle is presented in Figure 3.1,
and the biochemistry, biology and relevance of each nitrogen-transforming process, including:
(i) description of bacterial and enzymatic processes; (ii) consideration of the main factors
regulating ambient rates; (iii) a review of ambient rates in aquatic environments; and (iv) a
review of previous studies examining the enzyme kinetics of denitrification, is provided in
Appendix D.
Denitrification is the ultimate process that liberates gaseous nitrogen from an aquatic
environment (Figure 3.1). The process therefore has potential to act as a nitrogen sink and
alleviate the symptoms of eutrophication. Consequently, denitrification is often the focus of
aquatic research in situations where an abundance of nitrogen has caused deleterious effects.
However, denitrification is only one component of the nitrogen cycle and must be put in context
amongst other components. Each step of the nitrogen cycle is related integrally to all other
processes occurring within it (Figure 3.1).
The potential rate of denitrification within any ecosystem depends upon the reaction rates of all
the other nitrogen transforming processes occurring within that system. Within estuarine
environments, large inorganic nitrogen inputs may be assimilated rapidly into organic material
(definitive of eutrophic estuaries with algal blooms), with a subsequent tight recycling of the
stored nitrogen pool (Capone, 2000). Alternatively, physical processes or typically high
denitrification rates may transport rapidly inorganic nitrogen inputs from the system.
Assimilatory and dissimilatory nitrate reduction to ammonium (DNRA) pathways of nitrate
reduction will be important as competing pathways of nitrate utilisation and will conserve
nitrogen within the system rather than release it. Nitrification may be an important process in
providing the substrate (nitrate) required for denitrification. Both of the latter processes are
typically of quantitative importance in estuarine systems (Capone, 2000). The rate of
nitrification may depend, in turn, on the rates of ammonification and nitrogen fixation, both of
39
40
which provide ammonia to the system (though the latter process typically is not prevalent).
Alternately, ammonia may be taken up by assimilatory processes and away from the
denitrification pathway. Therefore, if ammonium rather than nitrate is the main source of
nitrogen assimilated into algal blooms, the process of denitrification may be relatively
unimportant in the management of those blooms, unless there is a high rate of coupled
nitrification. It is also becoming increasingly apparent that the novel processes of nitrifier-
denitrification, anammox and heterotrophic nitrification / aerobic denitrification (Appendix D),
add complexity of quantitative importance to ecosystem nitrogen budgets, with the processes
perhaps rivalling conventional denitrification in the production of nitrogenous gases. The
potential for the presence of these processes in estuarine nitrogen cycling systems should, at the
very least, be considered (see Appendix D).
Consequently, an understanding of each nitrogen cycling component is crucial if effective
management of aquatic ecosystems with nitrogen related environmental problems is to be
achieved. Furthermore, the quantification of the rate of turnover of the various nitrogen
compounds, together with the quantitative estimation of the extent to which different biological
and abiotic processes contribute to the cycle in different ecosystems, may allow the successful
management of those ecosystems. For this reason, I have attempted in this study to elucidate
the biological impact of benthic fauna on both denitrification and also on the other important
components of microbial nitrogen cycling within estuarine sediments (Chapter 5).
3.2 NITROGEN CYCLING METHODOLOGY
Quantification of denitrification, in addition to being a focal process of importance, is often a
useful first step in a holistic investigation of the nitrogen cycle. Denitrification activity may be
measured directly by dinitrogen gas production or via a range of indirect approaches based on
(i) measurements of nitrate flux, (ii) application of metabolic inhibitors (e.g. acetylene),
(iii) utilisation of nitrogen isotope tracers (including the isotope pairing technique), (iv) mass
balance and stoichiometric modelling and/or (v) examining kinetic parameters of the
denitrifying community (Seitzinger, 1988, 1993; Nielsen, 1992; Joye et al., 1996). The isotope
pairing technique (Nielsen, 1992) has recently become a common and efficient method to adopt
when access to nitrogen isotope supplies is viable; however, each methodology has its own
assumptions and limitations (Tiedje, 1988), of which Seitzinger (1988, 1993) provides a
comparative review. No one approach has been the obvious method of choice for studies
measuring denitrification, but since 1980 the acetylene inhibition method has often been the
method of first consideration because of its low cost and convenience (Tiedje, 1988; Groffman
et al., 1999). The combination of this method with a kinetic approach, i.e. the kinetic-fix
adaptation of the acetylene block (Joye et al., 1996), provides a powerful technique to measure
not only denitrification but also other nitrogen cycling processes in estuarine sediments. I used
41
NO3-
NH4+
Organic N
NO2-
N2O N2Organic N
NH4+ NO3
- NO2-
N2O N2
NH4+
NO3-
NO2-
Organic N
Export
NH4+
NO3-
NO2-
Organic N
Import
NH4+, NO3
-, NO2-
Groundwater Import
Ammonification / Mineralisation
Nitrification
Denitrification
Sediment / Water Column / Atmosphere Solute Flux
ATMOSPHERE
WATER COLUMN
SEDIMENT
Nitrogen Fixation
Ammonium Assimilation
Assimilatory Nitrate Reduction
Dissilmilatory Nitrate
Reduction to Ammonium
LEGEND
Figure 3.1: The estuarine nitrogen cycle, showing the main transformations occurring in the sediment and water column. For simplicity many intermediates have been omitted. See also Appendix C.
a hybridised kinetic-fix acetylene block method in my study. The background theory, advantages
and limitations of both the acetylene block method and general kinetic approach to denitrification
are given below. To give context to results using combined methodology, a comprehensive review
of outcomes is provided from studies based on a kinetic approach to denitrification, from a variety
of environments, in Appendix D.
3.2.1 The Acetylene Block
Nitrous oxide accumulation in soil during exposure to acetylene (C2H2) was first reported in 1973
(Oremland and Capone, 1987). The gas inhibits the nitrous oxide reductase of denitrifying bacteria
(Figure 3.2; Balderston et al., 1976; Yoshinari and Knowles, 1976). The rate of nitrous oxide
evolution is thus a measure of denitrification. The finding was significant because the direct
approach of measuring dinitrogen gas previously had been difficult to achieve (Tiedje, 1988). In
sedimentary environments, most applications of the acetylene inhibition technique for the study of
denitrification have been based on incubation of sediment cores with acetylene and subsequent
extraction of the evolved nitrous oxide after a relatively short (a few hours) incubation period
(Revsbech and Sørensen, 1990).
A potential major problem when using the acetylene inhibition method alone to investigate
denitrification is concurrent blockage of nitrification activity (Figure 3.2; Oremland and Capone,
1987; Kuenen and Robertson, 1988; Tiedje, 1988; Revsbech and Sørensen, 1990; Groffman et al.,
1999). Acetylene blocks the nitrification process by inhibiting the oxidation of ammonium to nitrite
(Hynes and Knowles, 1978, 1982; Walter et al., 1979; Mosier, 1980). The effect of nitrification
inhibition may be small in sediments where there is a high concentration and supply of nitrate from
the overlying water (i.e. where this source of nitrate dominates substrate supply for denitrification);
however, there have been many cases where the acetylene inhibition technique did not satisfactorily
quantify denitrification in situ because of the absence of a natural nitrate supply from nitrification
(i.e. where there is tight coupling of denitrification with nitrification) (Jenkins and Kemp, 1984;
Oremland et al., 1984; Slater and Capone, 1989; Revsbech and Sørensen, 1990). Therefore, nitrate
is often supplied artificially as a substrate at natural or saturating concentrations, and the resultant
nitrous oxide flux is considered the ‘potential’ denitrification rate.
The lack of nitrification in the presence of acetylene is beneficial to studies where the nitrate
substrate supply for denitrification is required to be somewhat precisely controlled or manipulated
(e.g. in approaches based on enzyme kinetics, see below), and/or where it is useful to discount the
nitrification component. In such experimental circumstances (i.e. with C2H2), the absence of
biological nitrate production allows the initial nitrate concentration to be fixed (such that ambient
42
N2O
N2 NO2-
NH4+ NO3
-
Organic N
Ammonification / Mineralisation
Nitrification
Denitrification
Process blocked by acetylene
Nitrogen Fixation
Ammonium Assimilation
Assimilatory Nitrate Reduction
Dissilmilatory Nitrate
Reduction to Ammonium
LEGEND
C2H2
\
C2H2
\
C2H2
\
Figure 3.2: Acetylene inhibition of nitrous oxide reduction and ammonium oxidation and effects on nitrogen cycling.
concentration cannot increase due to nitrification) and the resultant rate of nitrous oxide production
is a measure of denitrification, expressed relative to the initial nitrate concentration. When nitrate
consumption is concurrently determined, the resultant rate of nitrate loss is due to denitrification,
dissimilatory nitrate reduction to ammonium (DNRA) and assimilatory nitrate reduction (Figure
3.2). Hence, subtraction of the known rate of denitrification (determined via nitrous oxide
production) yields the DNRA plus assimilation component of nitrate loss (Blackburn, 1986;
Revsbech and Sørensen, 1990). Furthermore, dark incubation (and high background ammonium
concentrations) inhibits photosynthetic nitrate assimilation (Kristensen et al., 1985) such that the
component of nitrate consumption by DNRA alone can be determined. This is denoted overleaf.
43
Where ‘Δ [NO3--N]’ and ‘Δ [N2O-N]’ are the measured fluxes of nitrogen (N) respectively
incorporated in nitrate and nitrous oxide; and ‘Nitrification’, ‘Denitrification’, ‘Nitrate
Assimilation’, ‘DNRA’, ‘Ammonification’, ‘Ammonium Assimilation’ and ‘Nitrogen Fixation’
refer to the rates of N turnover via these processes (all set positive), then:
Finally, if both DNRA and nitrogen fixation can be determined independently to be negligible
(which often is the case in estuarine sediment; sensu Postma et al. (1984) and Capone (1988, 2000);
but see Appendix D, Sections D.2 and D.6), then ammonium production provides a good measure
of ammonification (although animal excretion and microbial decomposition components cannot be
separated, unless independently determined). Furthermore, acetylene (an analogue of N2) is
reduced preferentially to ethene (C2H4) by nitrogenase activity of nitrogen fixers, such that
ammonium production by nitrogen fixation is masked (Capone, 2002). That is:
(8) In the presence of acetylene, in darkness and when both DNRA ≅ 0 and nitrogen fixation ≅
0 (Nfix):
Δ [NH4+-N] +C2H2,dark,DNRA,Nfix = Ammonification
Manipulation of the presence/absence of acetylene and darkness in an experimental system,
combined with quantification of nitrate, nitrous oxide and ammonium flux rates therefore can yield
substantial information regarding denitrification and other nitrogen transformations in that system,
particularly when it is possible to verify independently the contribution of DNRA and/or nitrogen
fixation.
Limitations to the acetylene block method need to be considered. One, acetylene inhibition of
nitrous oxide reduction may be incomplete at low (<10 μM) nitrate concentrations, leading to an
underestimation of denitrification where nitrous oxide production is measured (Kaspar, 1982;
Oremland et al., 1984; Slater and Capone, 1989; Rudolf et al., 1991; Seitzinger, 1993; Groffman et
al., 1999). Two, the presence of sulphide hinders acetylene inhibition (Sørensen et al., 1987), and
methods other than the acetylene block (e.g. nitrogen isotope tracing) generally should be applied
where high sulphide concentrations are naturally present (Revsbech and Sørensen, 1990). Revsbech
and Sørensen (1990) have also highlighted the problem of the production of a more oxidised end
45
product of denitrification after acetylene addition: per nitrate molecule, the reduction to nitrous
oxide requires four electrons whilst reduction to dinitrogen gas requires five electrons. Where the
amount of available organic carbon (as an electron donor) is a limiting factor, the fewer electrons
required in the presence of acetylene may lead to an overestimation of the natural denitrification
rate. It is difficult to quantify this overestimation in the realistic situation where nitrate is supplied
by diffusion to a stratified microbial community (Revsbech and Sørensen, 1990). In the presence of
acetylene, a decreased penetration depth of nitrate may be expected due to the increased demand for
nitrate in surficial and successively deeper layers (Revsbech and Sørensen, 1990). As Revsbech
and Sørensen (1990: p. 263) have stated, “the exact elevation of the total rate of nitrate reduction by
addition of acetylene then depends on the microdistribution of available electron donors”.
The use of acetylene is complicated further when using the inhibitor with sediments inhabited by
fauna. In vitro injection of acetylene saturated water into uninhabited sediment cores through
silicon-rubber injecting ports at appropriate depths has been used successfully by, for example,
Sørensen (1978a), Andersen et al. (1984), Jensen et al. (1988) and Jørgensen and Sørensen (1988)
to quantify large-scale regional and seasonal variation in denitrification activity in sediments.
However, injections into small cores inhabited by infauna are not possible since there is a danger of
damaging or killing the burrow residents (Kristensen et al., 1991; Svensson, 1997). Kristensen et
al. (1991) overcame the problem by introducing inhabited cores to an acetylene-sparged reservoir
for a long period of time (20 h) to ensure an even distribution of acetylene into the denitrifying
sediment volume prior to incubation. Furthermore, Revsbech and Sørensen (1990: p. 262) have
noted that “the denitrification zone within shallow sediments is generally found so close to the
surface or to ventilated infaunal burrows that diffusional supply of acetylene from the overlying
water should suffice to stop nitrous oxide reduction, and it may thus not be necessary to perform the
injection of acetylene in such sediments”. Additionally, “the injection of relatively large volumes
of water does disturb the natural chemical gradients in the sediment core to some extent, and it
would be an advantage to avoid this step” Revsbech and Sørensen (1990: p. 262). An additional
consideration is that acetylene may directly affect the physiology or metabolism of benthic fauna.
Whilst Caffrey and Miller (1995) observed acetylene to have negligible effect on faunal excretion
(of a polychaete and mollusc), elevated rates of ammonium excretion (Kristensen et al., 1991) and
depressed irrigation activity (Chapter 5) by polychaetes are known to occur from exposure to
acetylene. It is important to be aware of such limitations in the application of the acetylene block
method so that appropriate control measures can be taken and/or recognition of potential
implications can be given.
46
3.2.2 Kinetics of Denitrification
Kinetic analysis is a useful and informative approach to the study of enzyme-driven denitrification.
Kinetic description defines the rate of denitrification as a function of another parameter such as
substrate (nitrate) concentration, enzyme concentration, temperature or pH (Saier Jr., 1987). The
most common method of kinetic analysis is to plot the steady state rate of denitrification (V0)
against substrate concentration (S) (Figure 3.3). The resultant rectangular hyperbola (Figure 3.3) is
defined by a Michaelis-Menten equation:
S KSV V
m
max 0
+×
= Equation 1
where Vmax is the maximal denitrification rate and Km is the nitrate concentration at which
denitrification rate is half maximal, i.e. the Michaelis or half-saturation constant (Firestone, 1982;
Knowles, 1982; Copeland, 1996; Joye et al., 1996). This model can be used to describe systems
more complex than a single-enzyme substrate reaction (Nedwell, 1975). The definition of Km as it
relates to denitrification activity by heterogeneous bacterial populations is not the same as in single
enzyme kinetics, rather it is a useful ecological parameter that describes a population’s hyperbolic
response of denitrification activity to increasing ambient nitrate concentration (Williams, 1973;
Nedwell, 1975). Where denitrification is the net result of enzyme activity from a cohort of
denitrifying organisms in sediment, for example, the kinetic constants should be classed as
“apparent” (Joye et al., 1996). Joye et al. (1996) refer to the apparent Michaelis constant as Kapp
and to the maximum potential denitrification rate as Vmp. Kapp represents a measure of the
physiological adaptation of a denitrifying population to reduce nitrate to nitrogenous gases
(Nedwell, 1975). However, the value of Kapp may be influenced by the physical transport rate of
nitrate to the denitrifiers (i.e. by diffusion or active transport processes), and nitrate availability
must be considered in addition to the general ambient concentration (Nedwell, 1975). Vmp may be
related positively to denitrifier population size but increases in denitrification rate will remain
proportional to nitrate concentration for a steady-sized population (Nedwell, 1975). Therefore, Kapp
should also remain constant for an increased denitrifier population size assuming that no size-
related physiological adaptation occurs (Nedwell, 1975). These kinetic constants are inherently
informative of the denitrification system under study; furthermore, their approximation and
substitution into the Michaelis-Menten equation also allows for an estimate of denitrification rate at
any given ‘natural’ background nitrate concentration (Joye et al., 1996). Natural in situ rates of
denitrification are generally much lower (e.g. 0.1-2.5%; Kaspar, 1982) than maximum potential
rates (Kristensen et al., 1985).
47
Nitrate concentration (S)
Den
itrifi
catio
n ra
te (V
0)
Km
Vmax
1/2 Vmax
Figure 3.3: Michaelis-Menten curve of enzyme-catalysed denitrification. The maximal denitrification rate is defined as Vmax. The Michaelis or half-saturation constant, Km, is the nitrate concentration at ½ Vmax. (Adapted from Saier Jr., 1987, p. 40).
Typically, the Michaelis-Menten (S vs V0) curve is derived empirically and the kinetic constants are
estimated via a linear transformation. For example, by rearrangement of Equation 1 we get:
⎟⎠⎞
⎜⎝⎛+⎟
⎠⎞
⎜⎝⎛=
max
m
max0 VKSx
V1
VS
Equation 2
The resultant linear Hanes-Woolf plot of S/V0 as a function of S (Figure 3.4), allows the kinetic
constants to be estimated from the gradient (1/Vmax) and y intercept (Km/Vmax) (Copeland, 1996;
Joye et al., 1996). Other linear transformations include the Lineweaver-Burk and Eadie-Hofstee
plots (Copeland, 1996). Whilst these rearrangements are each mathematically equivalent, the
measurement of V0 and, to a lesser extent, S is subject to experimental error which affects the
aptitude of the different plots (Dowd and Riggs, 1965). For instance, plotting the reciprocal of a
variable (V0 or S) places undue emphasis on the error associated with the smallest values of that
48
variable (Dowd and Riggs, 1965). The Lineweaver-Burk method involves plotting against each
other the reciprocal of both the independent variable (S) and the dependent variable (V0) (Copeland,
1996). As such, Dowd and Riggs (1965) strongly argue that the method should be abandoned in
favour of one of the other two aforementioned linear transformations.
The Michaelis-Menten equation describes steady state kinetics. The term steady state applies to
experimental conditions in which the enzyme-substrate complex is at an appreciable steady state
level throughout the course of the denitrification reaction (Copeland, 1996). Copeland (1996: p. 95)
notes that “Experimentally one finds that the time course of product appearance and substrate
depletion is well modelled by a linear function up to the time when about 10% of the initial
substrate concentration has been converted to product.” It is the denitrification rate measured
during the initial linear period of substrate loss that comprises V0 (Copeland, 1996). To satisfy the
assumption of steady state kinetics during the empirical determination of V0 for any given initial
nitrate concentration (S), it is important, therefore, to ensure that background nitrate levels do not
depreciate by more than 10%.
Nitrate concentration (S)
S /
V0
Gradient = 1 / Vmax
Intercept = Km / Vmax
Figure 3.4: Hanes-Woolf linear transformation of a Michaelis-Menten curve. Nitrate
concentration (S) divided by denitrification rate (V0) is plotted against S. The reciprocal of the gradient can be used to estimate the maximal denitrification rate (Vmax). The y intercept multiplied by Vmax then yields an approximation of the Michaelis constant (Km) (Adapted from Copeland, 1996, p. 93).
49
3.3 THE IMPORTANCE OF BENTHIC MACROFAUNA IN SEDIMENTARY
NITROGEN CYCLING
Darwin championed the idea that infauna are capable of profoundly influencing their habitat, both
biochemically and biomechanically (Johnson, 2002). In his book The formation of mould through
the action of worms, Darwin (1881) demonstrated how the subsurface feeding activity of
earthworms and deposition of faecal castings at the surface actually created the topsoil and stratified
environment in which they live. Johnson (2002: p. 8) notes how Darwin’s ideas “spawned a spate
of biochemical and biomechanical observations in the late 19th and early 20th centuries” across
many fields of study, including marine sedimentology (e.g. Davison, 1891). However, it has been
only in the last 30 years that similar biogeochemical observations have been applied to concepts of
explanatory importance in the paradigms of related disciplines (e.g. archaeology, geomorphology,
pedology). This gap in recognition of faunal influence upon substrate processes results from many
complex factors, chiefly among them: (i) the early practice of limiting world-views and an adoption
of conceptual frameworks in which biomechanical/biochemical processes were largely ignored; and
(ii) lack of a genetic language and supporting theory, including terms such as bioturbation, to
describe and showcase the importance of such processes (Johnson, 2002). The interpretations hold
particularly true when considering the historical development of sedimentary biogeochemistry.
Prior to the 1980s, Aller and Yingst (1978: p. 201) observed that “most studies of near interface
sediments assume that chemically and biologically important properties are stratified vertically in a
deposit”. Their study was one of the first to highlight the importance of the benthic fauna that
bestow localised effects and impart a three-dimensional heterogeneity to the typical sediment
environment: animal burrows, for example, induce a radial plane of diffusion (also see: Gray, 1974;
Rhoads, 1974; Goldhaber et al., 1977; Grundmanis and Murray, 1977; Kikuchi and Kurihara, 1977;
Petr, 1977; Schink and Guinasso, 1977; Aller, 1978, 1980a-c; Billen, 1978; Rhoads et al., 1978; Lee
and Swartz, 1980; McCaffrey et al., 1980; Zeitzschel, 1980). The consolidation of ideas pertaining
to the role of benthic fauna in sediment biogeochemistry (particularly nitrogen cycling) is a major
theme of my thesis and is discussed further in Chapters 4 to 7. Historically, the catalyst driving
early research and insights into benthic faunal effects was eutrophication of aquatic ecosystems and
the realisation that nutrient regeneration in the benthos is an important regulator of primary
productivity in shallow aquatic ecosystems.
In the early 1970s it was realised that nutrient availability generally tended to determine the extent
of primary production in aquatic ecosystems (Rhyther and Dunstan, 1971). Whilst phosphorous
supply was typically implicated in limiting productivity of freshwater environments (Nixon, 1995),
50
it became understood that nitrogen availability governed the magnitude of primary production in
most estuarine and coastal marine systems (Rhyther and Dunstan, 1971; Goldman et al., 1973;
Nixon et al., 1976; Carpenter and Capone, 1983; Seitzinger, 1990) and in some freshwater lakes,
streams and rivers (Keeney, 1973; Gerhart and Likens, 1975; Seitzinger, 1990). As detailed in
Section 2.1, the discovery and extensive global use of fertilisers during the past century,
compounded by the consequences of human population expansion, led to significantly increased
exports of nitrogen and other nutrients to near-shore aquatic environments (Rabouille et al., 2001).
Many freshwater lakes had become eutrophic in the earlier part of the century, but it was not until
the 1970s that eutrophication of the world’s rivers, estuaries and coastal oceans began to occur on a
broad scale (Cloern, 2001). These water bodies, in which primary productivity had previously been
limited by nutrient availability, were now increasingly subject to algal growth of excessive intensity
and frequency (Pedersen and Borum, 1996). The environmental and socio-economic implications
of eutrophication and problem algal blooms were, and continue to be, far-reaching; ranging from
unsightliness to acute toxic effects and/or complete ecosystem collapse (Section 2.1.3; for an
excellent review see de Jonge et al., 2002). The widespread manifestation of these eutrophic side-
effects in the coastal aquatic environments of the developed world during the 1970s (particularly in
western Europe and in the north-east and south-west of the U.S.A.), is likely the main reason that
spurred researchers to investigate thoroughly the factors regulating primary productivity in aquatic
ecosystems.
Increased nutrient loading arising from accelerated movement of nutrients from the lithosphere to
the hydrosphere had been implicated as the primary causative agent of excessive primary
productivity (Cloern, 2001). However, allochthonous nutrient supplies alone often could not
explain the high rates of algal growth in eutrophic waters. For example, nutrient-laden freshwater
runoff is typically diminished during the dry summer of temperate regions and cannot account for
the high rates of near-shore productivity frequently observed during this period (Rowe et al., 1975).
For many water bodies, it became obvious that tight internal mechanisms of organic algal matter
breakdown and nutrient re-supply (i.e. autochthonous nutrient regeneration), within both the pelagic
and benthic zones, were highly important factors regulating primary productivity (at least during
certain times of the year). Indeed, Nixon et al. (1976: p. 271) predicted that benthic nutrient
regeneration likely played only a minor role in the surface productivity of deep oceans (wherein
pelagic regenerative processes were expected to dominate5); however, they argued that: “the
5 Eppley and Peterson (1979) later highlighted the fact that the majority (82-94%) of nitrogen necessary to support primary production in deep oceanic waters was regenerated from organic matter in the pelagic zone; and Smith et al. (1978) and Zeitschel (1980) documented that nutrient effluxes of deep ocean sediments are two to four orders of magnitude less than those of shallow coastal sediments.
51
importance of benthic communities may be expected to increase dramatically in coastal waters, and
it is surprising that so little is known about their role in nutrient cycling or in maintaining the
characteristically high levels of productivity found in shallower marine and estuarine regions”.
The significance of benthic nitrogen regeneration in supporting primary production of aquatic
ecosystems had previously been highlighted by the classic limnological study of Mortimer (1941,
1942). A similar role of the marine benthos was conceptualised by Rittenberg et al. (1955);
however, their findings that the deep-sea basin sediments adjacent to southern California provided
only 0.4% of the overlying phytoplankton’s nitrogen requirements perhaps retarded recognition of
the process within the marine environment. Nevertheless, Barnes (1957) hypothesised that near-
shore nutrient regeneration occurred in bottom sediments and Pomeroy et al. (1965) showed benthic
phosphate regeneration to be an important nutrient source for estuarine plants and phytoplankton.
However, it was not until the widespread coastal eutrophication crisis of thirty years ago that
nitrogen recycling by the benthos was fully recognised to be a major determinant of primary
Callender and Hammond, 1982; Hammond et al., 1985). The flux of nutrients across the sediment-
water interface was considered a net result of their consumption or production within the benthos,
as governed by pertinent diagenetic reactions and transport processes.
Factors controlling sedimentary reaction rates and transport processes were acknowledged to play a
role in bentho-pelagic nutrient cycling and were therefore duly investigated. However, it was the
process of bioturbation (Rhoads, 1974) that began to receive a large proportion of research attention
and soon challenged conventional thought regarding benthic biogeochemistry. The classical
sediment geochemistry paradigm characterised the average sedimentary deposit as dominated by
one-dimensional, vertical transport processes with stratified zones of diagenetic reactions (Berner,
53
1976, 1980; Aller and Yingst, 1978). Given the vertical pore-water concentration profile and the
diffusion coefficient of a constituent solute (e.g. nutrients), Fick’s First Law could be used to
determine the diffusive flux of that solute across the sediment-water interface (Berner, 1976).
Discrepancies between diffusive nutrient fluxes predicted via pore-water profiling and directly
measured fluxes were demonstrated and the enhanced flux rates explained by advective processes
(Rhoads, 1974; Goldhaber et al., 1977; Aller, 1978; McCaffrey et al., 1980; Callender and
Hammond, 1982). Whilst the importance of bioturbation was acknowledged, the one-dimensional
formalism was retained by integrating the process (and other advective processes) into an ‘apparent’
or ‘effective’ diffusion coefficient that was ascertained empirically (Goldberg and Koide, 1962;
Berner, 1976, 1980; Aller, 1982). However, the diffusion analogy “is based on the assumption that
the activity of sediment-dwelling and irrigating organisms causes random movement of particles or
solutes, resulting in a transport of particles or pore-water along a concentration gradient in a similar
manner as molecular diffusion” (Timmermann et al., 2002: p. 152). The treatment of a sediment
deposit as an averaged or spatially-integrated entity only works well “where the length scale for
biological transport of material is less than the scale of the solute profile or where the mixing is
approximately stochastic” (Timmermann et al., 2002: p. 152).
It was the insightful research of Robert Aller and his co-workers during the late 1970s and early
1980s, from which understanding of the integral controls that benthic animals impart on diagenetic
processes and nutrient transport at the sediment-water interface, progressed beyond the view of the
sedimentary environment as one dominated by one-dimensional processes. Large- and fine-scale
patchiness in the magnitude and direction of sediment-water nutrient fluxes had previously been
attributed to differences in benthic faunal composition, abundance and biomass (e.g. Smith et al.,
1978), but the underlying mechanisms were still poorly known. In many instances, the assumptions
of utilising the diffusion analogy in considering bioturbative transport were breached. For example,
in the presence of irrigated burrows, exchange of solutes and solids between nonadjacent points in
the sediment cannot be modelled as a diffusive transport because transport distances are too great
and directional (i.e. non-random) (Timmerman et al., 2002). The radial diffusion model was
developed (Aller and Yingst, 1978; Aller, 1978, 1980a-c, 1982) to simulate these situations
whereby water and constituent solutes are transported over large vertical distances into burrows,
during and after which solutes are locally and radially exchanged with the surrounding pore-water
via diffusion6 (Timmerman et al., 2002). The benefit of this model was not only an improved
practical consideration of the presence of irrigated burrows on sediment-water nutrient fluxes; but
6 The related and practically simpler nonlocal irrigation model (Emerson et al. 1984; Bordreau, 1984) was developed later for practical use in similar situations (accounting for the presence of irrigated burrows), but reverted to a one-dimensional plane of reference.
54
much more importantly by invoking a radial plane of diffusion and an allowance for lateral
heterogeneity, conceptualisation of the typical sediment environment had been transformed from a
one-dimensional diagenetic stratification into a dynamic, three-dimensional mosaic of diagentic
stratifications and biogenic microenvironments (Aller, 1982, 1988; Kristensen, 2000).
The consideration of irrigated burrows introduced a major advance in elucidating the faunal effects
on sedimentary biogeochemistry. Subsequently, other faunal affects on diagenetic processes (e.g.
the influence of epibenthic filter-feeders upon algal biodeposition and nutrient regeneration; Dame
et al., 1980; Cloern, 1982; Officer et al., 1982) were considered in a more holistic framework.
Aller’s radial diffusion model marked the beginning of a realisation that the pervasive presence,
diverse habits and complex interactions of benthic fauna were integral components of aquatic
ecosystems with much explanatory power in the consideration of sediment-water nutrient fluxes and
regulation of primary productivity. The increasing frequency of global eutrophication has ensured
the continuing relevance and scientific interest in benthic faunal effects on bentho-pelagic nutrient
cycling. Indeed, in the early 1980s, Aller (1982: p. 94) stated, prophetically: “The interactions
between different groups of sediment-dwelling organisms as determined by, or reflected in,
sediment chemistry remains one of the most intriguing and understudied aspects of marine
biogeochemistry”. Consequently, my study of the role of two polychaete species (as a case study)
in the Swan River Estuary, Western Australia contributes to understanding of the effects and
interactions of sediment dwelling organisms on sedimentary nitrogen biogeochemistry.
55
PART II– THE STUDY
57
CHAPTER 4
LIFE HISTORY, POPULATION STRUCTURE AND PRODUCTION
OF AUSTRALONEREIS EHLERSI AND SIMPLISETIA AEQUISETIS IN
THE SWAN RIVER ESTUARY, WESTERN AUSTRALIA
4.1 INTRODUCTION
Polychaete worms typically constitute an important component of the benthic fauna of marine
and estuarine ecosystems. In reviewing their functional significance in sedimentary
communities, Knox (1977) noted that polychaetes are one of the dominant elements of the
macrobenthic fauna, with species contributing, on average, >40% to both community richness
and abundance. In southern Australia, polychaetes often comprise about half of the species
richness of all macrobenthic invertebrates (Wilson et al., 2003). Further, polychaetes typically
contribute substantially to overall community biomass, productivity and respiration (Knox,
1977). Secondary production of polychaetes often is of pivotal ecological importance in the
food-web dynamics and trophic inter-relationships including other macroinvertebrates, fish,
birds and humans of estuaries and nearshore marine environments (Dorsey, 1981; Kent and
Day, 1983; Petti et al., 1996; Omena and Amaral, 2000; Morris and Keough, 2003). Species
such as Nereis virens and Perinereis aibuhitensis are cultured or harvested in many places
around the world and represent considerable commercial value as fish bait (Dorsey, 1981; Choi
and Lee, 1997; Wilson, 2000; Kristensen, pers. comm.). Knox (1977) highlighted the fact that
benthic polychaetes may exert substantial influence on their biophysical surroundings, including
having synergistic effects on sediment reworking, particle size distribution, stability and
community composition. The role of polychaetes and other benthic fauna in sedimentary
biogeochemical processes (including nutrient cycling) has gained increasing recognition over
the past thirty years (Section 3.3). To ascertain fully the role of benthic polychaetes in these
ecosystem functions/processes, a detailed knowledge of the basic biology of constituent species
is required – yet is often lacking.
Polychaetes represent a significant faunal component of the macrobenthos of the Swan River
Estuary, Western Australia. Within those habitats studied most intensively, namely in shallow
water <2 m deep, members of the families Nereididae (with two species recorded,
Australonereis ehlersi and Simplisetia aequisetis), Spionidiae, Capitellidae, Sabellidae and
Orbiniidae together comprise about half of the total macrobenthic abundance of the main basin
and upper estuary (Rose, 1994; Kanandjembo et al., 2001). The distribution of A. ehlersi is
restricted to the middle and lower estuary (Chalmer et al., 1976; Rose, 1994). Simplisetia
aequisetis occurring in the lower, middle and upper regions of the Estuary, typically is one of
59
the most abundant elements of the macrobenthic fauna (Chalmer et al., 1976; Rose, 1994;
Kanandjembo, 1998; Kanandjembo et al., 2001; Pennifold and Davis, 2001). Burrowing adult
specimens may enhance significantly denitrification within sediments of the Estuary (De Roach
et al., 2002). I included both species of nereidid in this study; their distribution and abundance
in the Estuary and elsewhere are described in Chapter 2 (Section 2.2).
The most detailed account of the life history and reproductive biology of A. ehlersi and
S. aequisetis relate to populations near a major treated-sewage outfall in Port Phillip Bay,
Victoria (Dorsey, 1981). The only other life history study of an Australian nereidid, noted by
Wilson (2000), is by Glasby (1984, 1986) of the freshwater species S. limnetica from the
Hawkesbury River (New South Wales). Both A. ehlersi and S. aequisetis display rapid growth
and complete their life cycles in about 1-1.5 years, reproducing from spring to autumn (Dorsey,
1981). Australonereis ehlersi is tubiculous but likely produces a free-swimming larvae;
however, details of post-larval development are unknown (Dorsey, 1981). Simplisetia
aequisetis is dioecious, and the adult male broods embryos in specially constructed burrows or
tubes (Chapter 2, Section 2.2).
In this Chapter I describe the life history, geographical population structure and production of
the polychaetes A. ehlersi and S. aequisetis in the Swan River Estuary. The dynamics and
reproductive biology of both species from the estuary are compared in the Discussion with
Dorsey’s (1981) results. My investigation into the reproductive biology of both nereidid species
was intended to resolve not only their inherent ecological importance, but also to facilitate
insight and quantification of their respective roles in the biogeochemical cycling of nitrogen
within the estuary that constitute the substance of Chapters 5, 6 and 7.
4.2 MATERIALS AND METHODS
4.2.1 Study Area
The spatial distribution of A. ehlersi and S. aequisetis during summer (February) and winter
(August) conditions within the estuary was investigated by sampling at 12 shallow water
locations during 2000 (Figure 4.1; Table 4.1). Recruitment and progression of nereidid cohorts
were ascertained following monthly sampling from December 1999 to March 2001 at the Como
sampling site in the middle estuary. Temperature and salinity were determined for bottom water
(~2 m depth) on all sampling occasions via a conventional laboratory mercury thermometer and
AO temperature-compensated salinity refractometer.
60
Applecross
East Fremantle
South Perth
Heirisson Island
Bassendean
Ascot
Attadale
The Spit
Point Walter
Pelican Point
Matilda Bay
Como
N
5 km
Figure 4.1: Study sites within the Swan River Estuary.
61
Table 4.1: Swan River Estuary sampling sites and bottom water salinities and temperatures, 2000. Salinity throughout the estuary in February, 2000, was atypical, depressed due to freshwater runoff from a large, unseasonal rain event (further see Section 2.1.3). Salinity concentrations in February 2001 (Water and Rivers Commission, 2001), representing typical summer-time conditions, are given in parentheses.
Salinity
(g kg-1)
Temperature
(°C) Sampling Site
Location Substratum
Distance
from mouth
of estuary
(km) Feb
2000*
Aug
2000
Feb
2000
Aug
2000
East Fremantle Clayey/sand bank 4 32
(36)
36 26 17
Point Walter Sand/shell bed with lime-stone
rocks, macroalgal cover
8 11
(36)
28 28.5 19
‘The Spit’ Sand bank
9 10
(36)
28 28 16.5
Attadale Sand with sparse seagrass
(Halophila sp.) cover
12 9.5
(36)
16 28.5 19
Applecross Sand with sparse Halophila
cover
14 8.5
(35)
17 28 16
Como Sand/silt/shell bed
15 6
(35)
8 28 15
Pelican Point Sand/shell with patchy
Halophila cover
15 8
(35)
5 28 14.5
Matilda Bay Sand/shell bed
16 6
(35)
6 29 17
South Perth Mud/sand/shell bed
20 5
(34)
8 28.5 17
Heirisson
Island
Muddy sand 22 5
(31)
4 29 15
Ascot Sand bank
32 6
(27)
5 28.5 14.5
Bassendean Muddy sand
37 5.5
(23)
2 28 14
4.2.2 Sampling Procedure
Sediment was sampled by snorkellers using PVC pipe to extract a 25 cm long core, cross-
sectional area of 60 cm2. On every sampling occasion at each site, cores were taken within a
20 m wide transect extending from the shoreline (at low tide) to the 2 m deep water mark.
Within this transect, 12 cores were taken in total; with 4 cores taken randomly in water of three
depths: (i) <0.7 m deep, (ii) between 0.7 and 1.3 m deep, and (iii) between 1.3 m and 2 m deep
62
(Figure 4.2). Each core was considered the primary sampling unit, such that between site
comparisons of parameters were based on the mean of 12 cores. The contents of cores were
transferred carefully to plastic jars and returned to the laboratory. Sediment from each jar was
passed through a series of 4 mm, 2 mm then 1 mm sieves and the polychaetes were preserved in
4% formalin in ambient estuary water. For benthic communities underlying waters <1.5 m deep
in the middle Swan River Estuary, Rose (1994) recorded that a 1 mm sieve captured 95.5% of
abundance and 99% of all species in sediment cores. Further, a total sampling area of 120 cm2
(5 cores) captured most species at each site, with 98% of both abundance and biomass in the top
9 cm of sediment (Rose, 1994). Consequently, the present study’s total sampling area of 240
cm2 (4 cores) to a depth of 25 cm, in each of the three sampling regions of each transect (Figure
3.2), was considered adequate to determine the abundance and biomass of nereidids at each site;
however, statistical comparison of abundance or biomass between each site was not undertaken
due to a lack of statistical power. The sampling effort required to generate significant statistical
power was not viable due to constraints imposed by the breadth of the sampling regime (12 sites
across the estuary) and the time taken to process samples (by one researcher). Regardless, the
sampling effort was considered sufficient to provide meaningful qualitative descriptions of
spatio-temporal variation in A. ehlersi and S. aequisetis distributions across the estuary. In the
laboratory, the two nereidid species were separated but no distinction was made between the
sexes.
0.7 m @ low tide1.3 m
2.0 m
Shore
20 m
25 cm
4mm mesh
2mm mesh
1mm mesh
SievesCollect
polychaetes
Determine length
Figure 4.2: Experimental design and polychaete sampling procedure at each study site.
63
4.2.3 Size determinations
Following Kristensen (1984b), total body length of nereidids fixed in formalin (formalin-length,
Lf) was determined under a dissecting microscope to the nearest 0.5 mm, and used as an
estimate of size. The relationships between the preserved formalin-length and the extant length
of MgCl2-narcotised subsamples (live-length, Ll) of each species were as follows:
A. ehlersi: Ll = 1.0502 x Lf + 0.3421 (n = 10; r2 = 0.99)
S. aequisetis: Ll = 1.0435 x Lf - 0.0876 (n = 14; r2 = 0.99)
The original length of nereidids that were fragmented during the sampling/sieving process was
estimated by the following relationships between formalin-length and width of the third
setigerous segment (W) in mm (Figure 4.3):
A. ehlersi: Lf = 19.8897 x W - 0.7571 (n = 105; r2 = 0.91)
S. aequisetis: Lf = 27.6285 x W - 1.7654 (n = 157; r2 = 0.86)
C. aequisetis A. ehlersi Linear (C. aequisetis) Linear (A. ehlersi)
S. aequisetis A. ehlersi
Figure 4.3: Relationship between width of the 3rd setigerous segment and nereidid length for preserved (4% formalin) specimens of Australonereis ehlersi and Simplisetia aequisetis. (See text for linear equations, sample sizes and r2 values.)
Size cohorts (age classes) were identified by fitting normal (Gaussian) components to monthly
length-frequency-distribution histograms using the modal progression analysis routine of the
computer program FiSAT II (FAO-ICLARM, 2005). Preliminary assessment indicated that
length modes were reasonably apparent for the A. ehlersi population at Como, but size cohorts
64
of S. aequisetis were not clearly discernible by eye (Figure 4.5). Within FiSAT II, where a fit-
by-eye was insufficient to obtain initial estimates for mode means, Bhattacharya’s (1967)
graphical method was applied initially to the data. First guess mode means were then refined
using the NORMSEP optimisation procedure (after Hasselblad, 1966). The latter method
applies the maximum likelihood concept to SEParation of the NORMally distributed
components, and provides measures of standard deviation for the estimated modes. All mode
estimates were accepted since, in every case, the separation index between mean modes was
above the critical value of 2 (Gayanilo et al., 2002). The desirability of using an optimisation
refinement procedure has been highlighted by Grant (1989), who recognised that: (i) graphical
methods alone are often unreliable unless modes are well separated and/or sample sizes are very
large; and (ii) the methods used should place confidence limits around estimates of modes (e.g.
standard errors or deviations) to warn when estimates are not reliable. In this study, measures
of standard deviation are given on the growth curve of each cohort identified (Figure 4.6[A]).
The estimated density of nereidids within each cohort every month is also given (Figure 4.6[B]).
Finally, incidental observations of the presence of gravid or brooding adults, in addition to the
dimensions of eggs or embryonic juveniles, were also noted.
4.2.4 Biomass and Production
Several narcotised individuals of various sizes from both species were measured for live-length,
then dried for 48 hours at 60°C and weighed. The relationships between dry weight (Mdw in
mg) and length (in mm) were determined as:
A. ehlersi: Mdw = 0.0085 x Ll 1.96 (n = 12; r2 = 0.92)
S. aequisetis: Mdw = 0.0100 x Ll 1.65 (n = 13; r2 = 0.94)
Utilising these regressions for each identified cohort, monthly mean lengths were converted to
monthly mean biomass. For rough comparison of Mdw with studies stating biomass in terms of
wet weight (Mww), ash-free dry weight (Mafdw) or carbon weight (Mc), calculations may be based
on the assumptions that:
Mdw = [0.130 to 0.193] x Mww
Mdw = [1.125 to 1.330] x Mafdw
Mdw = [2.466] x Mc (Cammen, 1980; Kristensen, 1984b; Sardá et al., 2000).
Monthly production of both nereidid species was estimated by using the monthly mean density
and biomass increments of each cohort as described by Crisp (1971). Where months were
sampled in more than one year (December to March), a monthly average was obtained. Annual
production (P) was determined by summing the average monthly production over 12 months. In
the same manner, mean annual biomass (B) was calculated by summing average monthly
biomass of all cohorts, and used to determine the P/B ratio.
65
4.3 RESULTS
4.3.1 Salinity, Temperature and an Unexpected Algal Bloom
The temporal variation in bottom water temperature and salinity over the 16 month
sampling period at Como is illustrated in Figure 4.4. Temperature ranged from 15 °C in
winter (August 2000) to 32 °C in summer (January 2001). Excluding the datum from
February 2000 (see below), salinity ranged from 8 g kg-1 in winter (August/September
2000) to 38 g kg-1 in summer (February 2001).
Variation in bottom water temperature and salinity throughout the 12 sampling sites in the
estuary, for both summer (February 2000) and winter (August 2000), is presented in Table
4.1. Bottom water temperature was relatively stable throughout the estuary in both summer
(26-29 °C) and winter (14-19 °C). For both seasons, there was generally an upstream to
downstream cline of increasing salinity. In winter, salinity ranged from 2 g kg-1 at
Bassendean to 36 g kg-1 at East Fremantle. In summer, salinity of the surface 2 m of water
throughout most of the estuary (Bassendean to Point Walter) was a brackish 5 to 11 g kg-1,
with salinity increasing to 32 g kg-1 nearer the mouth (East Fremantle). During the
following summer (February 2001), bottom water salinity ranged from 23 g kg-1 at
Bassendean to 36 g kg-1 at East Fremantle (Table 4.1).
The dip in salinity at Como during summer (Figure 4.4) and generally throughout the
estuary in February 2000 (Table 4.1), was due to freshwater runoff from unseasonal heavy
rain associated with the aftermath of a tropical cyclone in January (see Appendix A). The
estuary was flushed with enough freshwater to fill it five times over and the rainfall event
and resultant hydrological conditions were considered highly atypical (Atkins et al., 2001).
Additionally, the summer freshwater conditions concomitant with a large, runoff-associated
nutrient input triggered a record estuary-wide bloom of the toxic blue-green alga
Microcystis aeruginosa (see Section 2.1.3). Subsequent easing of freshwater discharge
enabled the tidal re-entrance of marine water, upstream propagation of the salt wedge and
collapse of the blue-green algal bloom, such that the system had reverted to its more typical
marine dominated state by March/April 2000 (Figure 4.4; Atkins et al., 2001). The direct
and indirect implications of both the unusual salinity regime and faunal mortality effects of
the widespread toxic algal bloom during the early stages of the study period are reviewed in
the following discussion of A. ehlersi and S. aequisetis population dynamics.
66
0
5
10
15
20
25
30
35
40
1999
D
2000
J F M A M J J A S O N D
2001
J F M
Salin
ity (g
kg
-1)
0
5
10
15
20
25
30
35
Tem
pera
ture
(°C
)
SalinityTemperature
Month
Year
*
Figure 4.4: Salinity and temperature of bottom water (depth ~ 2 m) at Como during the study period. Salinity atypically was depressed in February/March 2000 due to freshwater runoff from a large, unseasonal rain event (see text).
4.3.2 Life History – Age Distribution and Growth at Como
Australonereis ehlersi
The monthly size-frequency distributions of A. ehlersi at Como are given in Figure 4.5(i). The
mean body length and density of the cohorts identified are respectively given in Figure 4.6 [A(i)
and B(i)]. Six cohorts were identified over the 16-month study period, including: Cohort 1
(represented at the beginning of the study by medium to large-sized mature individuals; Cohorts
2 and 3 (two cohorts initially recruited as small-sized juveniles); and three cohorts M1, M2 and
M3 that immigrated as medium-sized mature individuals.
The mean body length of individuals from Cohort 1 increased from 45.8 mm (in December
1999) to 65.0 mm (in February 2000). The density of the cohort declined from 84 to 14
nereidids m-2 during the same period. The cohort was not present during March 2000.
Cohort M1 migrated to Como during autumn (March and April 2000) at a density of 84-98
nereidids m-2, with a medium-sized mean body length of 40.0-49.3 mm. The cohort was not
captured during May 2000 (likely due to lowered sampling effectiveness at lower nereidid
densities) and was last collected at the beginning of winter (June 2000) at the low density of 28
nereidids m-2 and a mean body length of 72.5 mm.
67
Cohort M2 migrated to Como during late autumn/early winter (May and June 2000) also as
medium-sized individuals (mean body length 45.0-47.0 mm) at a density of 28-70 nereidids m-2.
The density of the cohort declined over winter to be 14 nereidids m-2 in September 2000, by
which time mean body length had increased to 75 mm. During this month one large, bloated
A. ehlersi individual (73.1 mm live length) from Cohort M2, ruptured between the dorsal and
ventral parapodia when probed with forceps; the coelomic fluid contained numerous eggs, with
a mean diameter of about 310 µm.
A third migration event of medium-sized A. ehlersi in mid- to late-winter (July and August
2000) of Cohort M3 yielded a density of 42-154 nereidids m-2, with a mean body length of 43.3-
48.7 mm. The cohort was last present during September 2000 at a density of 28 nereidids m-2
and mean length of 60 mm.
The appearance of Cohort 2 in late-autumn (May 2000) marked the first recruitment event of
small-sized juvenile A. ehlersi individuals of mean body length of 17.2 mm and cohort density
of 378 nereidids m-2. A continual increase in density occurred over the winter months, peaking
at 1836 m-2 in August 2000 by which time mean body length of Cohort 2 had increased to
24.6 mm. This winter peak in density of Cohort 2 was an order of magnitude higher than the
maximum density attained by any other cohort. The density of the cohort then decreased
steadily throughout the rest of the year to 28 nereidids m-2 by December, 2000, and was absent
during January 2001. The cause of the reduction in density, whether due to mortality or
emigration, was not determined. The mean body length of Cohort 2 increased to a medium-
sized 40.0 mm over the period to December 2000.
A second recruitment event of small-sized A. ehlersi individuals occurred during the summer of
2000/01, with Cohort 3 first appearing during December at a mean body length of 17.5 mm and
a density of 224 nereidids m-2. The density of the cohort (238 nereidids m-2) remained relatively
stable through to March 2001. Over this time, the mean body length of Cohort 3 increased to
33.5 mm.
Simplisetia aequisetis
The monthly size-frequency distributions of S. aequisetis at Como are shown in Figure 4.5(ii).
The mean body length and density of the cohorts identified are given in Figure 4.6 [A(ii) and
B(ii)]. Ten cohorts were identified over the 16-month study period.
Cohorts 1, 2 and 3 were present initially in December 1999. Cohort 1 comprised larger
individuals of mean body length 38.4 mm, at a density of 322 nereidids m-2. The density of
Cohort 1 decreased to 14 nereidids m-2 by the end of summer (March 2000), whilst mean body
68
length increased to 50.0 mm. Although not detected in April/May and from July to September,
a few individuals from Cohort 1 likely remained present in low density (~14 nereidids m-2) until
June if not October. Large individuals with mean body lengths of 60.0 mm and 65.0 mm,
respectively, were captured in those latter months.
Cohort 2, present in a relatively high density of 687 nereidids m-2 during December 1999,
comprised individuals with a mean body length of 22.9 mm. The density of this cohort declined
to 14–154 nereidids m-2 during autumn/early-winter. Cohort 2 was last detected during June
2000, with a mean body length of 41.0 mm.
Initially detected in December 1999 at a low density of 56 nereidids m-2 and mean body length
of 12.7 mm, Cohort 3 exhibited a continual increase in density throughout summer to reach a
peak of 617 nereidids m-2 in March, 2000. By this time mean body length of the cohort had
increased to 24.9 mm. The density of Cohort 3 then decreased over autumn to 196 nereidids m-2
in May 2000. This was the last month in which the cohort was detected, having grown to a
mean body length of 30.1 mm.
Cohort 4 was first detected in mid-autumn (April 2000) at a density of 308 nereidids m-2, with a
mean body length of 20.1 mm. Further recruitment to the cohort was exhibited in late-autumn
(May 2000), with density doubling to 617 nereidids m-2. Over winter, cohort density declined
only slightly to 547 nereidids m-2 in September 2000. At this time mean body length of Cohort
4 had increased to 37.7 mm. The density of the cohort then decreased considerably during
spring to 14 nereidids m-2 in November 2000. At a mean body length of 55.0 mm, November
2000 was the last month in which Cohort 4 was noted.
Cohort 5 was initially detected in low densities (56-98 nereidids m-2) and at a small mean size
(11.8-15 mm) during late-autumn to early-winter (April - June, 2000). A large increase in
cohort density to 575 nereidids m-2 was then exhibited during mid-winter (July 2000), with
individuals having a mean body length of 19.6 mm. The density of Cohort 5 decreased
gradually over winter/early-spring to 168–364 nereidids m-2, but then considerably to only 14
nereidids m-2 in the last month of spring. The cohort was last recorded during November 2000
when the mean body length was 50.0 mm.
Cohorts 6, 7 and 8 were first detected in September, October and November, respectively. At
first detection, each cohort demonstrated similar average body length of ~20 mm and density
within the range 750–890 nereidids m-2. These cohorts had a longevity of only 3 - 4 months,
during which time the size of Cohorts 6, 7 and 8 attained an ultimate mean length of 43.4, 34.8
69
and 26.5 mm, respectively; whilst the density of each cohort declined to 56, 168 and 308
nereidids m-2, respectively.
Whilst Cohort 9 was first detected at a density of 70 nereidids m-2 and mean body length of
6.7 mm during late-spring (November 2000), it then exhibited a similar progression as the three
preceding cohorts. Mean body length of Cohort 9 in December 2000 was 18.7 mm, with a
cohort density of 757 nereidids m-2. During the next 4 months over summer, mean body length
of the cohort increased to 30.4 mm, whilst the density decreased to 98 nereidids m-2.
Finally, Cohort 10 was detected in early-summer (December 2000) at a density of 266 nereidids
m-2 and mean body length of 10.1 mm. Cohort density and mean length increased over summer
to be 589 nereidids m-2 and 16.4 mm during February 2001. By the last month of the study
period (March 2001), mean body length of Cohort 5 had increased to 20.5 mm, whilst density
decreased to 224 nereidids m-2.
Brooding individuals of the species S. aequisetis, ranging in live length between 31.2-53.1 mm,
were observed in cohorts 2 (May 2000), 4 (November 2000), and 7 (November and December
2000, January 2001. The individuals occasionally were still encapsulated in mucous-bound
tubes despite the sieving process. Tubes contained hundreds of 5 to 8 setiger embryonic
juveniles with body lengths of 500 to 1000 µm; the mean diameter of the eggs was about
225 µm (Figure 4.7).
70
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
5 10 15 20 25 30 35 40 45 50 55 60 65 70 75
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
0100200300400
5 10 15 20 25 30 35 40 45 50 55 60 65 70
(i) Australonereis ehlersi (ii) Simplisetia aequisetis
Body Length (mm)
December 1999
January 2000
February
March
April
May
5
6
7
8
9
10
4
3 2
1
2 M2
M1
1
July
August
September
October
November
December 2000
January 2001
February
March
Den
sity
(ner
eidi
ds m
-2)
631 631
M3
3
June
71
Figure 4.5: Monthly size-frequency distributions of (i) A. ehlersi and (ii) S. aequisetis at Como from December 1999 to March 2001. Shading delineates cohorts and bold numbers indicate cohort appearance. Two-tone shading of S. aequisetis body length intervals indicates a mixed contribution by sequential cohorts. Dashed line indicates division between sub-adult and potentially reproductive S. aequisetis (see text for explanation).
Figure 4.6: [A] Mean body length (± s.d.) and [B] density of (i) A. ehlersi and (ii) S. aequisetis cohorts at Como from December 1999 to March 2001. [C] Total monthly biomass and [D] production of combined cohorts are also depicted. Dashed line at ~ 31 mm on [A(ii)] indicates division between sub-adult and potentially reproductive S. aequisetis (see text for explanation).
11394
4992
1289 1019426
987
1734 15011055
414 1128
307 564577
2658
728
851
2579 5064
8310
6148
5935
2478
328
521 931 1040
1979
426
200
428541
701
305
65
244119
322
345
796
498
956
864
1040
226
354457
251
546
202
641
114
78
118181 166 151
393234
352
37
518
89
24
242188
396
240
499
809
455
1264
348
471
35
424
80 239
26
45
158
182
2152 10 10 11 4 4 4 6128 17 6
80
78
1
M1
M2
M3
2
3
1327
498236 110 120 125 129 133 275
1201
747
474317
706
240778
1244 1136 541
436 1276 1635
1375
1975 2178
473
778
902 475
1357
1009
979
282
1244
2138
1891
587
1129
1052
688
953
492
596
172211 274
150
328
120
1
2
3
4
5
6
7
8
9
10
1 2 3 4 5 6 7 8 9 10
Cohort No
[A] [B] [C]
[D]
(i) Australonereis ehlersi (ii) Simplisetia aequisetis M
ean
Bod
y L
engt
h (m
m)
Den
sity
(ner
eidi
ds m
-2)
Bio
mas
s (m
g dw
m-2
) Pr
oduc
tion
(mg
dw m
-2 m
onth
-1)
80
70
60
50
40
30
20
10
0
2000
1800
1600
1400
1200
1000
800
600
400
200
0
8000
7000
6000
5000
4000
3000
2000
1000
0
1400
1200
1000
800
600
400
200
0
Cohort No
Cohort No
Month D J F M A M J J A S O N D J F M D J F M A M J J A S O N D J F M
Year 1999 2000 2001 1999 2000 2001
Cohort No
1 M1 M2 M3 2 3
3874
72
[A]
2 mm
[B]
2 mm
Figure 4.7: Brooding tube of S. aequisetis [A] with eggs [B].
73
4.3.3 Biomass and Production at Como
Australonereis ehlersi
The pattern of progression in the monthly biomass (standing stock) of A. ehlersi at Como
generally reflected the changes in combined densities of the cohorts (Figure 4.6[Bi & Ci]). A
maximum monthly biomass of 11.4 g dw m-2 in winter (August 2000) resulted predominantly
from the very high density of small-sized recruits in Cohort 2, supplemented by the lower
density of large, immigrant Cohort M3 adults. Cohort 2 comprised the dominant component of
total monthly biomass throughout winter and spring (June to November, 2000), and was largely
responsible for the gradual increase and then decline in total biomass, either side of the temporal
peak. A minimum monthly biomass of 0.43 g dw m-2 in February 2000, comprised solely a low
density of large, Cohort 1 individuals. A minimum of 0.85 g dw m-2 was recorded during the
recruited small, Cohort 3 individuals. The annual mean biomass (B) for A. ehlersi at Como was
3.89 g dw m-2.
Total monthly production of combined cohorts of A. ehlersi at Como is depicted in Figure
4.6[Di]. The temporal pattern of change in monthly production largely followed that of
monthly biomass. Maximum monthly production peaked at 4.99 g dw m-2 month-1 in
September 2000, principally due to growth by the high density of Cohort 2 recruits in the
preceding month but also supplemented by the growth of migrant Cohort M3 adults. Production
of A. ehlersi was absent in March 2000, due to the disappearance of Cohort 1. During the
following summer, production was lowest in the months of November 2000 (0.23 g dw m-2
month-1) and February 2001 (0.25 g dw m-2 month-1), respectively, indicating only low monthly
increases in biomass of Cohorts 2 and 3. Total annual production (P) for A. ehlersi at Como
was estimated at 12.5 g m-2 year-1, yielding an annual turnover (P/B) ratio of 3.2.
Simplisetia aequisetis
Total monthly biomass of combined cohorts of S. aequisetis at Como in 2000 is summarised in
Figure 4.6[Cii]. Highest monthly biomass was recorded during spring (September to
December), ranging between 3.76 - 4.32 g dw m-2. During this time, total monthly biomass
comprised various components from Cohorts 4 to 9. A minimum monthly biomass of
approximately 1.49 g dw m-2 was recorded during the summer months (January to March 2000),
with components progressing from Cohort 1 to 3. Thereafter, monthly biomass gradually
increased towards the spring peak. Monthly biomass then again declined over the 2000/01
summer to be at a minimum of 0.60 g dw m-2 during March 2001, comprising similar biomass
components from Cohorts 9 and 10. Annual mean biomass of S. aequisetis at Como was
estimated as 2.58 g dw m-2.
74
For S. aequisetis at Como, a marked peak in monthly productivity of 2.22 g dw m-2 month-1
occurred during late-spring (November) 2000 (Figure 4.6[Dii]), due largely to growth
contributions from Cohort 7 recruits during the preceding month. Productivity was also high in
October 2000 = 1.57 g dw m-2 month-1 and December 2000 = 1.24 g dw m-2 month-1; with major
growth contributions from recruits in Cohorts 6 and 8, respectively. Aside from a relatively
high productivity of 0.92 g dw m-2 month-1 in January 2000 (and a large growth contribution
from Cohort 2 during the first month in the sampling period for which productivity could be
calculated), productivity was low (0.14 – 0.54 g dw m-2 month-1) during the remaining months
of the study period. Minimal values near the lower end of this range were observed during
February to April in 2000 (combined low productivity mainly of Cohorts 2 and 3) and in
January and March, 2001 (combined low productivity mainly of Cohorts 9 and 10). Total
annual production for S. aequisetis at Como was estimated to be 8.74 g dw m-2 year-1, providing
a P/B ratio of 3.4.
4.3.4 Population Structure within the Swan River Estuary
For both A. ehlersi and S. aequisetis populations of the Swan River Estuary from waters <2 m
deep, the total density and biomass at each of the 12 study sites sampled during both summer
(February 2000) and winter (August 2000) is presented in Figure 4.8. Where a species was
present at a site, the size-frequency distributions of individuals are given in Figure 4.9. Cohort
analysis was not applied to these size-frequency distributions since it is impossible to determine
congruency of cohorts both between different sites and also at the same site with 6 months
between samplings.
Australonereis ehlersi
During summer, A. ehlersi was present at only four sites in the middle estuary at low densities
(14 to 56 nereidids m-2; Figure 4.8[Ai]). The density at Attadale was at the upper end of this
range and comprised large individuals (40-70 mm live length; Figure 4.9[Ai]), which
represented a substantial biomass (1.90 g dw m-2; Figure 4.8[Bi]). Low biomass (0.25–0.45 g
dw m-2) of A. ehlersi during February 2000 occurred at Como, Pelican Point and Applecross; at
these latter two sites, however, the mean length of individuals (~50 mm and 25 mm,
respectively) were considerably smaller than at Como (~65 mm).
In contrast, during winter (August 2000) A. ehlersi was present at 7 sites, including Matilda Bay
in the middle estuary plus Point Walter and East Fremantle in the lower estuary (Figures 4.8[Aii
& Bii] and 4.9[Aii]). Excluding Como, the density of A. ehlersi at each site was low (14-168
nereidids m-2), representing a biomass at each site in the range 0.21-1.12 g dw m-2. The high
density (1991 nereidids m-2) and high biomass (11.0 g dw m-2) of A. ehlersi at Como during
August 2000 was anomalous. No other site yielded a comparable high density standing stock of
75
small-sized recruits, however a low density cohort of small individuals 10-20 mm in length was
observed at East Fremantle. In summary, the biomass of A. ehlersi at sites other than Como
during winter resulted from a low density of individuals >35 mm in length.
Simplisetia aequisetis
Simplisetia aequisetis was present at all 12 sampling sites during both the summer and winter
(Figures 4.8 and 4.9B). During February, both density (1710 nereidids m-2) and biomass (4.47 g
dw m-2) were greatest at The Spit (Figure 4.8[Ai & Bi]), located at the most downstream end of
the middle estuary. At this site, the body lengths of S. aequisetis individuals were in the range
of 10-60 mm, with a polymodal size frequency distribution which likely represented several
cohorts (age-classes) (Figure 4.9[Bi]). The 20 mm body length category was the most common
size-interval of S. aequisetis individuals. Within the adjacent lower estuary, the density and
biomass of S. aequisetis decreased sharply and were lowest at East Fremantle, i.e. nearer the
mouth of the estuary (70 nereidids m-2, 0.12 g dw m-2). At these lower densities within the
lower estuary, the size-frequency distribution remained polymodal and the 20 mm body length
category was most common; however, S. aequisetis individuals >40 mm in length were not
present. Regarding the remaining sites in the middle and upper estuary during summer, there
was a general upstream trend of decreasing density and biomass of S. aequisetis although the
species was present in only very low density and biomass at both Ascot and South Perth during
this time. Hence, sites within the main basin of the middle estuary (Attadale, Applecross,
Como, Pelican Point and Matilda Bay) supported a higher density (435-1262 nereidids m-2) and
biomass (0.85-3.65 g dw m-2) of S. aequisetis than sites within Perth Water (South Perth and
Heirisson Island) and the upper estuary (Ascot and Bassendean); wherein density and biomass
ranged, respectively, from 14–252 nereidids m-2 and 0.04–0.58 g dw m-2. The size-
constituencies of S. aequisetis individuals at sites within the middle estuary were generally
polymodal and within the body length range of 10-50 mm. Simplisetia aequisetis >50 mm in
length during summer were observed only at the Spit and Attadale. The low density of S.
aequisetis in the upper estuary during summer comprised medium-sized individuals 25-50 mm
in length.
Although the magnitude of S. aequisetis density and biomass at sites within the lower reaches of
the estuary (from Como downstream to East Fremantle) during winter were generally similar to
that observed during summer (Figure 4.8), the species’ distribution pattern upstream of Como
was markedly different from that exhibited during summer. The density and biomass of S.
aequisetis at Pelican Point and Matilda Bay (within the upper main basin of the middle estuary)
was 2-5 times higher than that during summer. Further, the density and biomass of S. aequisetis
at sites within Perth Water and the upper estuary were 1-2 orders of magnitude higher than
summer-time values. Thus, whilst S. aequisetis remained at high density (1318 – 1514
76
77
nereidids m-2) and biomass (4.73 – 5.44 g dw m-2) at the Spit and Attadale during winter (and at
low density and biomass within the lower estuary), the absolute peak in density and biomass
shifted upstream to Heirisson Island (2930 nereidids m-2) and Ascot (5.87 g dw m-2). In
summary, the estuary-wide population of S. aequisetis was greatly supplemented in winter by
the increases in density and biomass at locations upstream of Como (i.e. within the upper
estuary, Perth Water and at Matilda Bay and Pelican Point). The size-constituencies of
S. aequisetis individuals at these upstream sites were polymodal, with the mean body length of
likely cohorts (age-classes) encompassing the full range of sizes from 5-65 mm (Figure
4.9[Bii]). However, the most common/dense size-interval was always represented by the 15 or
20 mm body length category, suggesting that small-sized recruits of S. aequisetis contributed
largely to the heightened upstream density/biomass during winter.
Figure 4.8: Mean density [A] and biomass [B] of A ehlersi and S. aequisetis during (i) February and (ii) August 2000 at 12 sampling locations throughout the Swan River Estuary.
East
Fre
man
tle
Poin
t Wal
ter
‘The
Spi
t’
Atta
dale
App
lecr
oss
[A] Density [B] Biomass
Com
o
Pelic
an P
oint
Mat
ilda
Bay
Sout
h Pe
rth
Hei
rris
son
Is.
Asc
ot
Bas
send
ean
6000
5000
4000
3000
2000
1000
6000
5000
4000
3000
2000
1000
0
0
78
(i) February 2000 (ii) August 2000
Den
sity
(n
erei
dids
m-2
)
Bio
mas
s (m
g dw
m-2
)
2500
2000
1500
1000
500
0
Bas
send
ean
Asc
ot
Hei
rris
son
Is.
Sout
h Pe
rth
Mat
ilda
Bay
Pelic
an P
oint
Com
o
App
lecr
oss
Atta
dale
‘The
Spi
t’
Poin
t Wal
ter
East
Fre
man
tle
UPSTREAM DOWNSTREAM
Bio
mas
s (m
g dw
m-2
)
Australonereis ehlersi
Den
sity
(n
erei
dids
m-2
)
2500
2000
1500
1000
500
0
Simplisetia aequisetis
(B) Simplisetia aequisetis
(i) February 2000 (ii) August 2000 UPSTREAM
Bassendean
(A) Australonereis ehlersi
(i) February 2000 (ii) August 2000
Ascot Heirisson Island South Perth
Matilda Bay Pelican Point Como Apple- cross Attadale
65
‘The Spit’ Point Walter East Fremantle 65
DOWNSTREAM
400
300
200
100
400
300
200
100
400
300
200
100
400
300
200
100
400
300
200
100
400
300
200
100
400
300
200
100
400
300
200
100
400
300
200
100
400
300
200
100
400
300
200
100
400
300
200
100
5 10
15
20
25
30
35
40
45
50
55
60
65
70
5
10
15
20
25
30
35
40
45
50
55
60
65
70
5 10
15
20
25
30
35
40
45
50
55
60
65
70
5
10
15
20
25
30
35
40
45
50
55
60
70
5
10
15
20
25
30
35
40
45
50
55
60
70
Den
sity
(ner
eidi
ds m
-2)
Body Length (mm)
Figure 4.9: Size-frequency distributions of A. ehlersi [A] and S. aequisetis [B] during (i) February
and (ii) August 2000 at 12 sampling locations throughout the Swan River Estuary.
645
449
631 631
547 505
785 855
79
4.4 DISCUSSION
4.4.1 Secondary Production of A. ehlersi and S. aequisetis
Sardá et al. (2000) defined secondary production in terms of the formation of heterotrophic
biomass with time, and primarily is a function of the growth of individuals, recruitment patterns
and mortality. Variation in these parameters, affected by a multitude of physicochemical and
biological factors, for example food resource quality and quantity, extent of larval dispersion,
magnitude of predation and competition, typically will lead to differences in secondary
production between separate populations. By including reference to biomass or standing stock,
the ratio of production to biomass (P/B) expresses the turnover of a population and is more
informative than comparing the productivity of populations (Kristensen, 1984b).
As far as the author is aware, this is the first study to quantify the mean annual biomass and
production for any population of A. ehlersi. Whilst the S. aequisetis population co-inhabiting
the site at Como had both lower mean annual biomass and productivity, the resultant P/B ratio
of 3.4 was slightly higher than that of A. ehlersi (P/B = 3.2). These turnover rates in the Swan
River Estuary are slightly higher than that observed by Dorsey (1981) for S. aequisetis in Port
Phillip Bay (P/B = 2.9). However, the latter population had both a high mean annual biomass
and production, based on ash-free dry weight to dry weight conversions (Section 4.2.4). These
production estimates and P/B ratios suggest that Australian nereidids are likely an important
component of total productivity in the ecosystem; offering a potential food resource for
constituent crustaceans, fish and bird species.
Turnover rates of the nereidids Simplisetia keiskama (P/B = 1.8) and S. erythraeensis7 (P/B =
1.9) in the Berg River Estuary, South Africa (Kalejta, 1992), were lower than the rates reported
for Australian nereidids; whereas derivative mean annual production and biomass estimates
were generally intermediate to those exhibited by nereidids in the Swan River Estuary and Port
Philip Bay. Similar to South African species, the Brazilian nereidid Laeonereis acuta had a P/B
ratio of 2.0 and comparable annual productivity (Omena and Amaral, 2000). The range of
turnover rates for those few southern hemisphere nereidids P/B = 1.8-3.4 is within the range
0.8-4.8 of northern hemisphere nereidids including Nereis diversicolor, Neanthes
arenaceodentata, Neanthes succinea, Neanthes virens, Lumbrinereis acicularum and
Lumbrinereis fragilis (Omena and Amaral, 2000; Medernach et al., 2000).
7 The genus of Simplisetia is presumed to replace the genus of Ceratonereis for these two South African species (see p. 27).
80
4.4.2 Australonereis ehlersi: Life History and Distribution
Life History
The first major arrival of juvenile A. ehlersi recruits (Cohort 2) occurred in late-autumn and
peaked during late-winter, a similar pattern of intensive winter-time recruitment of juveniles
recorded for Port Phillip Bay (Dorsey, 1981). Additional cohorts of A. ehlersi at Como first
appeared as medium-sized adults in low density during autumn through winter. It is not known
whether the medium-sized A. ehlersi individuals of Cohort 1 first migrated at a medium-size in
the summer of 1999/2000 or first appeared as small-sized juveniles earlier in 1999. Following
initial migration of the medium-sized A. ehlersi cohorts during autumn and winter, growth
ensued over the subsequent 2 to 4 months while cohort density concomitantly decreased. The
sudden appearance of these medium-sized groups at Como implies immigration from elsewhere,
indicative of movement within the estuary. Reproductive stages of A. ehlersi have been
observed swimming in the winter in the Swan River Estuary (McShane, 1977) and in spring and
summer in Port Phillip Bay, Victoria (Dorsey, 1981), and on both occasions were interpreted as
indicative of pelagic spawning events. The large gravid female collected from Como sediment
in September did not show morphological changes indicative of epitoky. However, large
atokous individuals about to spawn probably leave the sediment to release gametes in the
overlying water. The absence of gravid A. ehlersi specimens throughout the whole study period
supports the idea that reproductive individuals do not remain in the sediment. Thus, the decline
and eventual disappearance of cohorts M1-3 likely was due predominantly to pelagic spawning
by atokous individuals.
Pelagic spawning by epitokal stages is not uncommon in the Nereididae (Dorsey, 1981), but
atokous gravid females of A. ehlersi are more comparable to those of Nereis japonica (Dorsey,
1981). Whilst atoky may not be unusual for estuarine or brackish/freshwater nereidids, it is
usually characteristic of species which produce benthic larvae and/or brood, where limited
dispersal or confinement of larvae to an osmotically/thermally stable sediment environment,
may confer some form of reproductive advantage (see also p. 87). Hence, atokous pelagic
spawning may represent a modified reproductive strategy of estuarine nereidids facilitating
larval dispersion (Dorsey, 1981).
Prior to the likely pelagic spawning events of large individuals, Cohorts M1, M2 and M3
presumably migrated into the sediments at Como as medium-sized adults. Thus, swimming
behaviour may also facilitate migration of A. ehlersi. Indeed, the decline in this cohort’s density
over the preceding 2 months, and its ultimate disappearance, likely was due to emigration of
medium-sized adults in addition to mortality. Such emigration was not apparent in Dorsey’s
(1981) study of A. ehlersi in Port Phillip Bay, possibly because he pooled size-frequency data
between different sites.
81
The reproductive strategy and pelagic spawning potential of A. ehlersi is unknown, however the
species may be iteroparous (polytelous). Glasby (1986) suggests that the presence of
coelomoducts in both males and females of A. ehlersi (Hartman, 1954; Hutchings and Reid,
1990), which enable external transport of gametes, indicates potential for iteroparity. Nereidids
generally lack coelomoducts and are semelparous (monotelic), often spawning via rupture of the
body wall preceding death of the individual (Glasby, 1984; Wilson, 2000). Further
investigation into the reproductive behaviour of A. ehlersi, combined with histological
examination of the gametogenic cycle, is required to elucidate the frequency of spawning events
in its life history.
The mean body length of the winter recruits increased by 23 mm over the ensuing 6 month
period until the end of spring. Assuming that these medium-sized adults then emigrate and
settle elsewhere in the estuary, the life-span of A. ehlersi is estimated at 8-10 months, plus the
undetermined duration of the larval period. The length of the larval period is estimated to last
for 2-4 months, by assuming an initial embryonic length of ~2 mm and extrapolating the growth
curves of Cohorts 2 and 3. Hence, the total life-span of A. ehlersi in the Swan River Estuary is
estimated at about 10-14 months, comparable with the estimated 12-16 month total life-span for
the Port Phillip Bay population (Dorsey, 1981).
Dorsey (1981) concluded that, in Port Philip Bay, A. ehlersi spawn mainly from summer to
early-winter. Given a 2-4 month pelagic larval period, the main recruitment period of A. ehlersi
juveniles settling at Como from late-autumn to mid-winter (Cohort 2), indicates a similarly
timed major spawning period in the Swan River Estuary. However, the timing of major
spawning events may be specific to a particular location; in addition, secondary spawning
events may occur year-round. For example, Dorsey (1981) noted the beginning of a low density
juvenile recruitment event during late-summer (at the end of his sampling period), indicative of
spring-time spawning. Similarly, the low-density recruitment of juveniles to Como during the
2000/01 summer of this study (Cohort 3) indicates a spawning event during late-winter to early
spring. Uncertainty regarding the extent of larval dispersion prior to settling makes
determination of the parent cohort problematic. If these summer-time recruits were locally-
derived (from the spawn of adults located at or near Como), then candidate parents include
large-sized adults from Cohort M3 and/or medium-sized adults from Cohort 2, potentially
lending support to the hypothesis of iteroparity, with spawning at medium- and then again at
large-size. Candidate local parents of the more abundant winter-time recruits at Como (Cohort
2) include those large-sized A. ehlersi present from summer to early-winter (2000), i.e.
individuals in Cohorts 1, M1 and M2. Alternately, settling juveniles may be derived from
adults residing elsewhere in the estuary.
82
Regardless, it is apparent that A. ehlersi populations are capable of spawning throughout the
year. Furthermore, spawning and recruitment patterns may demonstrate considerable inter-
annual variability. The juvenile recruits present in the summer of 2000/01 at Como were not
observed at the same time during the preceding year. Similarly, Dorsey (1981) reported the
presence of juvenile recruits in relatively high density in the March/April of 1977 but only in
very low density during the March/April of 1978. In relation to A. ehlersi’s spatio-temporal
distribution pattern within the Swan Rive Estuary, factors potentially governing the timing and
success of spawning and settling events are discussed below.
Distribution
Evidence from this study and others (Chalmer et al., 1976; McShane, 1977; Rose, 1994;
Kanandjembo, 1998) indicates that A. ehlersi resides exclusively within the lower, and main
basin of the middle, estuary. Furthermore, of all sites sampled during summer and winter, high
density and biomass of the species were recorded at one site only (Como). The high abundance
at Como was observed only during winter and comprised mainly juveniles. Dorsey (1981)
similarly noted generally low densities of A. ehlersi at his sites throughout Port Philip Bay
except for two locations with high densities. This patchy distribution pattern possibly reflects
areas with suitable conditions for the settling and development of larval A. ehlersi. Whether a
more intensive sampling effort throughout the lower- to mid-reaches of the estuary would reveal
other locations with higher densities of A. ehlersi remains to be determined, as does the
distribution of the species into sediments underlying water depths >2 m.
Adult A. ehlersi inhabit sediment underlying water of salinity ranging between 5-37 g kg-1 (both
this study and estuaries of New South Wales (Hutchings and Murray, 1984)). Further, adult
stages of A. ehlersi regulate osmotically their body fluids (McShane, 1977). Hence, given that
adult stages theoretically could tolerate the salinity regime within the upper reaches of
Australian estuaries, it is surprising, a priori, that the species exhibits a marine euryhaline
distribution (Hartman, 1954; Dorsey, 1981; Dürr and Semeniuk, 2000). The reason they are
absent from the upper reaches likely is due to the intolerance of the pelagic larval stage of even
brief exposures to freshwater (<8 g kg-1). Juveniles settled at Como at salinities ~ 8 g kg-1, but
this is likely the lower limit of their tolerance.
Since freshwater conditions in summer will occur, albeit infrequently and unpredictably
depending on the passage of rain-bearing depressions emanating from tropical cyclones across
the catchment, it is relevant to consider, therefore, the implications of the atypical summer-time
freshwater inundation of the Swan River Estuary that occurred during this study in February
2000. Summer salinities may be expected to be within the range 23-36 g kg-1 down-stream
from Bassendean, substantially higher than the 5-11 g kg-1 recorded in 2000, during which time
83
only adult stages were collected, and at 4 locations only. These lowered salinities may have
caused considerable mortality of juvenile recruits, particularly since juveniles were present at
Como during the subsequent summer (February 2001); although recruitment of juveniles was
not occurring at Como during February 2000 prior to the freshwater flushing. Estuaries
typically experience salinity fluctuations of the water column that are substantially damped in
the sediment: larvae may find refuge from low water salinity by resting in the sediment. The
extent to which benthic populations, possibly comprising larval and adult stages, were affected
by the summer freshwater flushing in 2000 was not determined, but the impact of summer
flushing on distribution and recruitment of A. ehlersi are biologically significant issues to
elucidate in future research.
Determining the undoubtedly numerous, physical, chemical and biological factors controlling
the spatio-temporal distribution of A. ehlersi in the estuary was not an objective of this study,
but the results presented here point towards patchiness in distribution that results initially each
year from the balance between larval dispersal and post-settlement success in what is a very
Zeitzschel, 1980; Callender and Hammond, 1982; Hammond et al., 1985; Krantzberg, 1985;
Kristensen, 1988, 2000; Andersen and Kristensen, 1991; Aller and Aller, 1998; Welsh, 2003).
Benthic communities rarely comprise single-species assemblages, and elucidating the holistic
impact of nutrient recycling by benthic macrofauna will involve quantification of the dynamics of
mixed, complex macrofaunal communities (Welsh, 2003). However, a reductionist approach has
been followed here to measure the contribution to nitrogen cycling, particularly in terms of
denitrification, nitrification, ammonification and net NOx- and NH4
+ flux rates, by the two nereidid
polychaetes of the Swan River Estuary, Australonereis ehlersi and Simplisetia aequisetis. This
approach is in accordance with the suggestion of Kristensen (2000) that in studying the effects of
bioturbation a range of animal species or functional groups of infauna should be investigated.
5.2 METHODS
5.2.1 Sampling site
Nereidids (A. ehlersi and S. aequisetis) and sediment were collected on 17 October, 2001, from
Como, Swan River Estuary, Western Australia (31°59.6’S, 115°51.1’E, Figure 4.1). Physico-
chemical properties of this location are described in Table 4.1 and in De Roach et al. (2002;
89
Appendix E). At the time of sampling, water temperature was 19 °C and salinity 21 g kg-1.
Ammonium concentration [NH4+] and nitrate (+ nitrite) concentration [NOx
-] in the overlying water
were both ~2 μM. Sediment with nereidids was collected manually to a depth of 15 cm while
snorkelling. Water (60 L) used in the initial set-up of the experiment was collected in situ. For
continued replenishment of water quality (see below), additional water (60 L) was collected every
two days from the more accessible Pelican Point (Figure 4.1), until termination of the experiment at
the end of December. During this period, salinity of the estuary at Point Pelican increased slowly
from 24 to 36 g kg-1. The mean (± s.e.) values of [NOx-] and [NH4
+] of estuary water over the
period were 1.99 ± 0.05 μM and 12.7 ± 0.90 μM, respectively.
5.2.2 Sediment and nereidid preparation
In the laboratory, sediment and nereidids were transferred into V-cores (Figures 5.1 and 5.2;
Kristensen, 1984a) constructed according to De Roach et al. (2002; Appendix E). V-core
dimensions were 47 mm (length) by 16.5 mm inner diameter (i.d.) for uninhabited and S. aequisetis
inhabited cores, and 55 by 25 mm i.d. for A. ehlersi inhabited cores, reflecting spring population
densities at the study site (~1000 A. ehlersi m-2 and ~ 2300 S. aequisetis m-2), and natural burrowing
depth (De Roach, personal observation). To maintain homogeneity between V-cores, all sediment
utilised was sieved simultaneously through a 2 mm mesh and thoroughly mixed. This also
facilitated collection of the nereidids. Sediment re-establishment after sieving was chosen over
other methods of sediment collection and defaunation for reasons outlined in De Roach et al. (2002,
Appendix E).
Fifty-five V-cores were filled with sieved sediment to within 1-2 mm from the top of the cores
(Figure 5.1). A worm was introduced to 40 of these (20 with A. ehlersi ~ 200 mg wet weight, 20
with S. aequisetis ~ 10 mg wet weight) and 15 were left uninhabited. The polychaetes rapidly
buried into the sediment and each core was randomly placed into one of two aerated 20 L holding
aquaria containing water collected in situ, at room temperature (~ 22 °C). Within the aquaria, both
ends of each V-core were capped with a plastic lid containing a central aperture of 2 mm diameter
(Figure 5.1). The hole allowed solute flux between sediment and overlying water and, in the case of
A. ehlersi inhabited cores, provided a site of attachment for a silicone tube (12 x 1.6 mm i.d.,
considered an extension of the burrow) in which burrow ventilation could be measured. To avoid
the build-up of excretory products, ~ 80% of the water was replaced every two days (the remaining
20% maintained saturation of the cores), until the end of the experiment. The period of acclimation
of V-cores within the holding tanks was no less than 31 days, enabling sediment microbial
communities to re-establish after the disturbance of sediment sieving.
90
Each core was visually checked daily; any core with a dead nereidid (easily distinguished by a rapid
reduction to black sediment) was removed. For cores with A. ehlersi, only those exhibiting a
permanent burrow, continuous from one end of the core to the other, were used. Initial inspections
of cores with S. aequisetis revealed that burrow formation did not occur. Sacrifice of a few cores
revealed that the specimens were alive but living interstitially. This habit was unexpected since De
Roach et al. (2002; Appendix E) had demonstrated that larger specimens of this species construct
well-defined, U-shaped burrows. Mortality and loss of core-suitability over the period were such
that eight each of A. ehlersi and S. aequisetis inhabited cores, and nine uninhabited cores, were
utilised.
Figure 5.1: Experimental set-up depicting a sediment-filled V-core inhabited by A. ehlersi
(adapted from Kristensen et al., 1991). Other species treatments of V-cores (S. aequisetis inhabited, uninhabited) are interchangeable. Acetylene supply and flow sensor can be omitted where appropriate (for further details see text).
5.2.3 Experimental Procedure
The kinetic-fix adaptation of the acetylene block method outlined here was utilised to determine not
only denitrification rate in vitro, but also nitrification, ammonification, total NOx- flux and total
NH4+ flux rates within individual inhabited or uninhabited sediment cores. The influence of
acetylene on various microbial nitrogen cycling processes is described in Section 3.2.1 and the
theory of denitrification kinetics in Section 3.2.2. For each V-core (uninhabited, A. ehlersi and
2 cm
Acetylene Air
Water Sampler Flow sensor
Worm
91
S. aequisetis inhabited), NOx- and NH4
+ fluxes within the overlying water were measured both in the
presence (+) and absence (-) of C2H2 over a range of idealised ambient NO3- concentrations [S] ≅ 0,
20, 50, 100, 500 μM. KNO3 addition to a sample of estuary water of near 0 μM NO3- (but not
exactly known), resulted in actual values of [S] slightly deviating from idealised concentrations
(due to slight variation in initial estuary water NO3- concentration and small error in KNO3
amendments). Individual V-cores were subject to one ambient NO3- treatment only, with NH4
+ and
NOx- flux monitored first in the absence and then in the presence of acetylene. Flux rates of NOx
-
and NH4+ within water overlying V-cores were determined by the difference in initial and final
solute concentrations over a three hour period.
Initially, a randomly chosen V-core was placed into an 8 L incubation aquarium containing estuary
water amended with NO3- to an appropriate [S] (the sequence of exposure to ambient NO3
-
concentration was also randomised for subsequent cores). The aquarium was aerated and held at
20 °C by a constant temperature water bath in darkness, for a NO3- incubation period of 48 hours.
The same ambient conditions, including [S], were maintained in all experimental containers and
aquaria as described below. The volume of water in each aquarium was calculated such that the
estimated reduction in NO3- concentration of the overlying water throughout the period of V-core
exposure did not exceed 10%, even at very high denitrification rates. This limited proportion of
NO3- depreciation was necessary to maintain enzyme-driven denitrification at a linear rate, an
important requirement for Michaelis-Menten modelling of denitrification kinetics (Copeland, 1996;
Section 3.2.2).
Following NO3- incubation, the V-core was transferred to an aerated experimental set-up (Figure
5.1) in which solute fluxes could be measured in the absence of C2H2 and, in the case of A. ehlersi
inhabited cores, burrow ventilation monitored. The core was placed into a pre-moulded groove
within a plaster-of-Paris base and an initial (t = 0 min) 3 mL sample of the overlying water was
taken by syringe via a sampling port (Figure 5.1). It is assumed that the distribution of solutes
within the overlying water was sufficiently mixed by the aeration process. The water sample was
injected into a 3 mL ‘Venoject’ evacuated test-tube (Terumo Corporation, Belgium) and
immediately placed in a freezer to be stored at -4 ˚C until analysis. A final water sample was taken
after three hours (t = 180 min) using the same method. The experimental container held 500 mL
(A. ehlersi treatment) or 400 mL (uninhabited and S. aequisetis treatments) of estuary water
amended to the appropriate [NO3-]. These volumes were estimated to be minimal enough to allow
sediment N-transforming processes to impart a measurable effect upon overlying solute
concentrations over the 3 hour period, whilst being large enough not to allow a 10% depreciation in
ambient NO3- concentration.
92
Australonereis ehlersi burrow ventilation was monitored continuously by placing the tip of a FS20
micro-flow sensor (Unisense, Denmark) within the centre of the opening of the silicone tube burrow
extension (Figure 5.1). The flow sensor was positioned and held by a micromanipulator and
attached to a PA2000 Picoammeter (Unisense). An analog-digital processor (DI-705, DATAQ
Instruments, U.S.A) and data acquisition system (DI-700, DATAQ) facilitated PC recording of
ventilation waveform data using WinDaq (DATAQ) software. To minimise the need for
calibrations of the flow sensor, it was kept polarized between measurements. Calibration of the
flow sensor was conducted every two days in accordance with the manufacturer’s instructions8.
Subsequent to the three hour C2H2-free monitoring period, the V-core was transferred to an 8 L
C2H2 incubation aquarium containing ambient water but continuously purged by a mixture of 10%
C2H2, 90% air, for 20 hours (Kristensen et al., 1991). As described in Section 3.2.1, acetylene
blocks the nitrification process by inhibiting the oxidation of NH4+ to NO2
- (Hynes and Knowles,
1978; Walter et al., 1979). The lack of NO3- production and measurement of NO3
- loss in the
presence of C2H2, permits an estimate of denitrification rate (provided that assimilatory NO3- uptake
is prevented by dark incubation, and that the rate of dissimilatory NO3- reduction to NH4
+ is
negligible – both here assumed true in vitro; further see Section 3.2.1 for methodological theory and
the footnote of Section 5.3.4 for justification of the latter assumption). Also, determination of
denitrification rates over a range of [S] facilitates investigation of the apparent enzyme kinetics of
the denitrifier community within sediments of each treatment (i.e. either inhabited by A. ehlersi or
S. aequisetis, or uninhabited) (further see Section 3.2.2). If Michaelis-Menten (steady-state) type
kinetics are exhibited, then the apparent Michaelis (half-saturation) constant (Kapp) and maximum
potential denitrification rate (Vmp) can be determined (Section 3.2.2).
The V-core was subsequently returned to the experimental set-up, with a slow stream of C2H2
incorporated (Figure 5.1), and the water sampling process repeated. The experimental and C2H2
incubation set-ups were housed within a darkened fume-cupboard which collected escaping C2H2
and also flow sensor tracer (H2) gas. After the final period of +C2H2 water sampling, a visual record
of sediment redox status was taken by photographing the towel-dried V-core. The wet sediment
weight and height within both halves of the V-core were measured to ascertain sediment bulk
density.
8 Briefly, the flow sensor was placed in a calibration set-up whereby the tip was immersed in a known gravity-fed flow of estuary water exiting a from silicon tube. The velocity of the flow was controlled by clamping and calculated by measuring the time it took to fill a standard volume. The flow sensor was exposed progressively to a series of set flow rates within the range of 0.5 to 20 mL min-1 and the signal (pA) was logged when it had stabilized at each velocity. These readings were used to produce a calibration curve, through which a measured signal was converted into a velocity value.
93
For each core, the duration of the entire process totalled 3 days 2 hours; initial NO3- incubation (48
h), -C2H2 monitoring (3 h), C2H2 incubation (20 h) and +C2H2 monitoring (3 h). To enable
monitoring of two cores within this period, the NO3- and C2H2 incubation aquaria were duplicated
and the two cores were subject to the same process at a staggered interval of 5 hours (i.e. the first
core incubation was initiated in the morning and the second in the afternoon). Four days were
required, therefore, to monitor both + and - C2H2 nitrogenous fluxes in the water overlying two V-
cores, and experimental completion of all 25 cores took 52 days.
5.2.4 NH4+ and NOx
- analyses
Water samples were thawed to room temperature (~ 22˚C) and analysed for [NH4+] and [NO3
- +
NO2-] using a Skalar Sanplus Segmented Flow Autoanalyser System. The automated colourimetric
procedures for the determination of [NH4+] and [NO3
- + NO2-] are given by the American Public
Health Association (1995).
5.3 RESULTS
5.3.1 Sediment Colour and Bulk Density
The presence of A. ehlersi and S. aequisetis in sediment cores had a marked effect on sediment
colouration, visible from the second day of core acclimation until experiment termination (Figure
5.2). All cores were vertically colour-stratified into three layers of approximately the same
proportions: 0 to 3.0 cm (surface), 3.0 to 3.1-3.3 cm (mid) and 3.3 to 4.0 cm (deep). The deep
sediment colour of all cores was light grey, whilst the mid-depth interval was a band of orange
sediment that was generally 1-2 mm thicker in uninhabited cores. Uninhabited and A. ehlersi
inhabited cores contained dark brown surface sediment, whilst the same depth interval was light
brown for C. aequisetis inhabited cores. Average (± s.e.) wet bulk densities for uninhabited, A.
ehlersi inhabited and S. aequisetis inhabited cores were 1.61 ± 0.02, 1.68 ± 0.03 and 1.54 ± 0.04 g
cm-3, respectively.
Uninhabited A. ehlersi inhabited S. aequisetis inhabited
Figure 5.2: Photographs of representative V-cores, taken at termination of the experimental period, illustrating sediment colouration profiles indicative of redox status.
5 cm
94
5.3.2 Ventilation of A. ehlersi Burrows
The patterns of burrow ventilation exhibited by individual A. ehlersi specimens within sediment
cores over the experimental range of ambient NO3- concentrations, both pre- and post C2H2
incubation, are presented in Figure 5.3. Average ventilation was determined as the cumulative area
under each ventilation curve (absolute flow), standardised for time (Table 5.1 and Figure 5.3). The
degree and pattern of burrow ventilation in cores inhabited by A. ehlersi varied. Ventilation was
absent in the two cores subject to +500 μM NO3- (idealised) both prior to and during acetylene
exposure, except for two periods of low ventilation in one of the cores whilst exposed to acetylene.
Excluding the cores at +500 μM NO3-, average ventilation decreased 5-27-fold in burrows after
acetylene exposure, except for one core subject to +0 μM NO3- (idealised) that showed a 4-fold
increase in average ventilation. Average ventilation through burrows ranged from 21-578 mL h-1
prior to, and 3-113 mL h-1 following, acetylene exposure. The percentage of time spent ventilating
the burrow (i.e. flow >0 mL min-1) also decreased following acetylene exposure, and with the
exception of the core subject to +0 μM NO3-, the percentage prior to acetylene exposure was always
≥ that following exposure (Table 5.1). In the absence of acetylene, typical ventilation patterns were
assessed to be represented by Figures 5.3 [ai], [di], [ei] and [fi], and normal average ventilation was
considered the mean of the average ventilation rates exhibited by these inhabited cores, 415 mL h-1.
Table 5.1: Average burrow ventilation rate and the percentage of time for which burrow ventilation >0 ml min-1 for A. ehlersi individuals exposed to increasing initial nitrate treatments.
Average burrow ventilation
(mL h-1)
% time burrow ventilation
>0 mL min-1Initial Nitrate
Treatment (μM) - C2H2 + C2H2 - C2H2 + C2H2
0 578 25 100 53
0 27 113 59 80
20 21 3 72 16
50 323 69 100 100
50 373 66 98 43
100 385 14 99 22
500 0 0 0 0
500 0 11 0 24
95
0 20 40 60 80 100 120 140 160 180
500 µM
0
2
4
6
8
10
500 µM
0
2
4
6
8
10
50 µM
0
2
4
6
8
10
- C2H2[Nitrate]
0 µM
0
2
4
6
8
10
0 20 40 60 80 100 120 140 160 180
+ C2H2
0 µM
0
2
4
6
8
10
20 µM
0
2
4
6
8
10
50 µM
0
2
4
6
8
10
100 µM
0
2
4
6
8
10
Time (min)
Bur
row
ven
tilat
ion
rate
(ml m
in-1
)
Figure 5.3: Time-specific burrow ventilation pattern of A. ehlersi inhabited V-cores at increasing
initial nitrate treatments [Nitrate], in the presence (right) and absence (left) of acetylene. To reduce ‘noise’, maximal ventilation rate has been smoothed for each treatment core.
96
5.3.3 Denitrification
Potential denitrification rates (V0) in acetylene-inhibited, NO3--amended sediment cores increased
with respect to ambient NO3- concentration (S) for all three treatments (A. ehlersi and S. aequisetis
present, uninhabited), in a general Michaelis-Menten type manner (Figure 5.4[ai-iii]). The linear
Hanes-Woolf transformations used to determine the apparent kinetic constants are presented in
Figure 5.4[bi-iii]. Substrate (NO3-) loss never amounted to a >10-12.5% reduction of initial levels
in water overlying uninhabited and S. aequisetis inhabited cores; however, for NO3- amendments
less than +500 μM (idealised), NO3- loss in water overlying A. ehlersi inhabited cores typically was
11.5%-17.7%, but 25% and 31% in the +0 μM amendments.
The microbial sediment community of uninhabited cores demonstrated a higher Vmp (4.33 mmol N
m-2 h-1) and lower Kapp (133 μM NO3-) than that of S. aequisetis inhabited cores (Vmp = 3.83 mmol
N m-2 h-1; Kapp = 202 μM NO3-) (Fig’s 5.4[ai-ii], 5.4[bi-ii]). The estimated denitrification rate at
‘natural’ overlying NO3- concentration (at the sampling site during the study period - V2μM) was
higher for uninhabited sediment (64 μmol N m-2 h-1) than for S. aequisetis inhabited sediment (38
μmol N m-2 h-1).
Potential denitrification rates of A. ehlersi sediment cores measured over the experimental range of
S, as well as those values subjectively standardised to account for depressed burrow ventilation in
the presence of C2H2, are presented in Figure 5.4[aiii]. Measured denitrification rates in A. ehlersi
inhabited cores (Figure 5.4[aiii], black dots) were generally similar to or slightly greater than the
denitrification curve established for uninhabited cores (Figure 5.4[aii]), albeit in burrows where
acetylene had, for the most part, greatly retarded ventilation (see above, Ventilation results).
Further, it was generally observed that the measured V0 values of A. ehlersi inhabited cores most
similar (closest) to the uninhabited denitrification curve were those with zero to very low burrow
ventilation; and conversely that A. ehlersi cores demonstrating higher ventilation rates exhibited V0
values proportionately greater than the uninhabited denitrification curve. To estimate potential
denitrification of A. ehlersi cores at normal (typical non-acetylene affected) burrow ventilation, the
measured V0 values were therefore standardised utilising the uninhabited denitrification curve as a
baseline (representing A. ehlersi potential denitrification at zero burrow ventilation), and using the
realised average ventilation flow of each core as a scaling factor with respect to normal average
ventilation (415 ml h-1) (Figure 5.4[aiii]). Algebraically:
97
Amden = Uden + (p * (Aden – Uden)) Equation 3
where, Amden is the measured denitrification rate of the A. ehlersi inhabited core in the presence of
C2H2; Uden is the expected denitrification rate of uninhabited cores at relevant S (equivalent to
expected denitrification rate of A. ehlersi cores at zero burrow ventilation or the baseline estimate);
p is the proportion of normal, average ventilation exhibited by the A. ehlersi specimen in the
presence of C2H2; and Aden is the expected denitrification rate of the A. ehlersi sediment core at
normal ventilation in the absence of acetylene (ventilation-standardised estimate). Aden values were
estimated by rearrangement of Equation 3:
Aden = Uden + ((Amden – Uden) / p) Equation 4
The Hanes-Woolf linear transformation regression fit (r2 = 0.98) of ventilation-standardised
denitrification estimates is presented in Figure 5.4[biii]. The ventilation-standardised Vmp estimated
for A. ehlersi inhabited sediment cores (4.75 mmol N m-2 h-1) was greater than that of either
uninhabited or S. aequisetis inhabited sediment cores, whilst the associated Kapp estimate for A.
ehlersi inhabited cores (50 μM NO3-) was lower than either of the other two core treatments. At
natural overlying nitrate concentration (2 μM), the denitrification rate of A. ehlersi inhabited
sediment (standardised to normal ventilation) was estimated at 182 μmol N m-2 h-1, which is greater
than that determined for either uninhabited or S. aequisetis inhabited cores.
5.3.4 Nitrification and Net NOx- Flux
In the absence of acetylene, total NOx- flux in uninhabited, S. aequisetis and A. ehlersi inhabited
sediment cores responded to increasing initial NO3- concentration in the same generalised manner;
net efflux from sediment to overlying water at lower initial NO3- concentration, tending towards net
influx from overlying water to sediment at higher initial NO3- concentration (Figure 5.4[ci-iii],
black dots). NOx- flux ranged from 2.77 mmol N m-2 h-1 (at [S] = 24 μM) to -0.56 mmol N m-2 h-1
(at [S] = 510 μM) within S. aequisetis sediment cores (Figure 5.4[ci]); from 1.87 mmol N m-2 h-1 (at
S = 0.5 μM) to -2.21 mmol N m-2 h-1 (at [S] = 475 μM) within uninhabited cores (Figure 5.4[cii]);
and, from 2.65 mmol N m-2 h-1 (at [S] = 1.0 μM) to -1.65 mmol N m-2 h-1 (at [S] = 480 μM) within
A. ehlersi inhabited cores (Fig. 5.4[ciii]).
98
NO
x- flu
x(m
mol
m-2
h-1
)
-2
-1
0
1
2
3
(i) C. aequisetis
NO
x- flu
x [v
](m
mol
m-2
h-1
)
-5
-4
-3
-2
-1
0
(ii) Uninhabited (iii) A. ehlersi
NO
x- flux
/ In
itial
[NO
3- ](S
/v)
-200
-150
-100
-50
0S/v = -0.261S - 52.77
R2 = 0.925
S/v = -0.230S - 30.59R2 = 0.957
S/v = -0.211S - 10.56R2 = 0.983
+ C2H2
Ventilationstandardised
estimate
NH
4+ flux
(mm
ol m
-2 h
-1)
0
1
2
3
0 100 200 300 400 500
NH
4+ flux
(mm
ol m
-2 h
-1)
-2
-1
0
1
Initial NOx- concentration [S]
(µM)
0 100 200 300 400 500 0 100 200 300 400 500
(a)
(b)
(c)
(d)
(e)
+ C2H2
+ C2H2
- C2H2
- C2H2 Nitrification estimate
(i) S. aequisetis
Figure 5.4: Dissolved inorganic nitrogen (DIN) fluxes measured in the water overlying (i) S. aequisetis inhabited, (ii) uninhabited and (iii) A. ehlersi inhabited V-cores, at increasing initial concentrations of nitrate [S].
(a) Nitrate flux measured in the presence of acetylene, indicating potential denitrification rates [v]. (b) Linearisation [S/v] of (a) illustrating the regression equation used to generate the fitted curves in (a). (c) Nitrate flux measured in the absence of acetylene and estimates of nitrate flux due to nitrification (illustrating
constant average). The idealised curve illustrated represents the net product of estimated nitrification (= constant), minus the denitrification curve in (a).
(d) Ammonium flux measured in the presence of acetylene, indicating the average production from mineralisation plus, in the case of inhabited V-cores, excretion.
(e) Ammonium flux measured in the absence of acetylene, indicating the net product of ammonium produced via mineralisation/excretion minus that lost via nitrification.
99
N.B. Estimated curves of nitrate flux in water overlying A. ehlersi inhabited V-cores have been subjectively standardised to represent ‘normal’ burrow ventilation activity. The rate of dissimilatory nitrate reduction to ammonia is assumed to be negligible. See text for further details.
Total NOx- flux was assumed to be the net result of NOx
- produced via nitrification minus that
consumed by denitrification. For uninhabited and S. aequisetis inhabited sediment cores,
nitrification estimates were therefore determined by adding the expected denitrification rate
(calculated via substitution of the relevant [S] into the appropriate Michaelis-Menten equation) to
the measured total NOx- flux (Figures 5.4[ci-ii], cross-hatches). Sediment nitrification rate was
generally constant over the range of [S] and average nitrification rate was highest within
S. aequisetis inhabited cores (2.76 ± 0.17 [s.e.] mmol N m-2 h-1) and lowest in uninhabited cores
The ventilation-standardised nitrification estimates of A. ehlersi inhabited sediment cores are
presented in Figure 5.4[ciii]. The average nitrification rate (2.43 ± 0.07 mmol N m-2 h-1) was lower
than that estimated for S. aequisetis inhabited cores but higher than that of uninhabited sediment
(Figure 5.4[ci-iii]). Figure 5.4[ci-iii] also exhibits, for each treatment, the curve of expected total
100
NOx- flux estimated by subtraction of the defined Michaelis-Menten denitrification curve from the
curve (constant) of average nitrification rate. Expected total NOx- flux determined by this method
displays a very good fit to measured total NOx- fluxes for each species treatment of sediment cores9
(Figures 5.4[ci-iii]). The values of [S] at which total NOx- flux shifts from efflux to influx (zero net
NOx- flux, predicted from expected total NOx- flux curves) were estimated at 520 μM for S.
aequisetis inhabited cores, 100 μM for uninhabited cores and 52 μM for A. ehlersi inhabited cores.
5.3.5 Ammonification and Net NH4+ Flux
NH4+ production in the presence of acetylene provided an estimate of ammonification - here defined
as microbial mineralisation plus, in the case of inhabited sediment, nereidid excretion. Within each
of the species treatments of sediment cores, ammonification in the presence of acetylene was
generally constant in relation to [S] (Figure 5.4[di-iii]). Average ammonification was greatest in
A. ehlersi inhabited cores (2.20 ± 0.05 mmol N m-2 h-1, Figure 5.4[diii]), intermediate in S.
aequisetis inhabited cores (1.84 ± 0.16 mmol N m-2 h-1, Figure 5.4[di]), and lowest in uninhabited
cores (0.91 ± 0.05 mmol N m-2 h-1, Figure 5.4[dii]). Due to the uncertainty of the relationship
between A. ehlersi ventilation activity and magnitude of excretion, and also the inability to separate
the excretion and mineralisation components of ammonification, NH4+ fluxes could not be
standardised with respect to burrow ventilation within either ± C2H2 treatments of A. ehlersi cores.
Within sediment cores inhabited by A. ehlersi, net NH4+ flux in the absence of acetylene ranged
from an efflux of 0.54 mmol N m-2 h-1 to an influx of -0.43 mmol N m-2 h-1 across the range of S,
representing a slight average efflux of 0.07 ± 0.13 mmol N m-2 h-1 (Figure 5.4[eiii]). In the absence
of acetylene there was a net uptake of NH4+ from the overlying water at all ambient NO3
-
concentrations of, on average, -0.62 ± 0.08 mmol N m-2 h-1 from S. aequisetis inhabited cores and,
on average, -0.85 ± 0.12 mmol N m-2 h-1 from uninhabited cores (Figure 5.4[ei-ii]).
9 For each core species treatment, the good fit of the observed NOx
- flux curve to that expected by subtracting the denitrification curve from the observed nitrification constant (both illustrated in Figure 5.4[c]), supports the assumption that denitrification was the major source of NOx
- loss and that the rate of dissimilatory nitrate reduction to ammonium was negligible (the observed and expected curves would otherwise show substantial deviation, if the rate of DNRA was significant). The likely absence of requisite highly anoxic micro-environments is probably implicated in the observed lack of DNRA within V-cores (further see discussion of the DNRA process in Appendix D, Section D.6).
101
5.4 DISCUSSION
5.4.1 Overview
I will describe initially the occurrence and magnitude of the main nitrogen transformation processes
(denitrification, nitrification and ammonification) and net flux rates of NOx- and NH4
+ occurring
within the sediments of uninhabited V-cores (Figure 5.5[a]). Although a comparison to the relative
magnitude of nitrogenous processes and flux rates occurring within sediments of A. ehlersi and S.
aequisetis inhabited V-cores (respectively, Figures 5.5[b] and 5.5[c]) is necessary, my focus firstly
is to describe the base-line case within uninhabited sediment. Consideration of the likely reasons
for the observed effects of each nereidid species on these nitrogenous processes/fluxes is delayed
until the habit of each species is discussed. Also, in all cases the environmental conditions in vitro
entailed dark incubation of V-cores and, within the overlying water, a constant initial NH4+
concentration of about 13 µM.
Under the conditions of the measurements, the average rate of both nitrification and ammonification
were lowest in the absence of macrofauna, whilst denitrification was moderate in comparison to the
two nereidid species sediment treatments (Figure 5.5). The balance of these nitrogen cycling
processes in uninhabited V-cores resulted in a net high average rate of NH4+ influx to sediment
(Figure 5.5[a]). The flux direction of NOx- in uninhabited (and inhabited) sediments was dependent
on the NO3- concentration of the overlying water [S]: NOx
- flux shifted from an efflux at low [S] to
an influx to sediment at high [S]. The shift is due to the denitrification rate increasing in a
Michaelis-Menten type fashion in response to increasing [S] (Figure 5.4[a]); i.e. the efflux of NOx-
at low [S] switches to an influx at higher [S] because of the higher rate of sedimentary NOx-
consumption by denitrifying bacteria. Within uninhabited sediment at low [S], the average efflux of
NOx- was low in comparison to the two nereidid species sediment treatments; whilst at high [S] the
average influx of NOx- was moderate compared to the inhabited treatments (Figure 5.5). Also,
within uninhabited sediment cores, zero NOx- flux occurred when [S] was about 100 µM, which is
intermediate to the comparative zero NOx- flux [S] value of V-cores inhabited by the two nereidid
species (Figures 5.4[c] and 5.5).
My interest lies in determining the impact of A. ehlersi and S. aequisetis on these processes
(denitrification, nitrification, ammonification) and resultant net fluxes of NOx- and NH4
+. In order
to interpret the role of these two resident nereidid species on nitrogen cycling processes in
sediments of the Swan River Estuary, it is imperative to consider the potential effects of their
infaunal habit.
102
Low [S]
NH4+ NOx
-
N2
NH4+ NOx
-
N2
NH4+ NOx
-
N2
(a) Uninhabited
(b) A. ehlersi
(c) S. aequisetis
High [S]
-0 NOx flux @
Water den Sediment
nit
min
org-N
Low [S] High [S] -0 NOx flux
@ Water
den Sediment
nit
min + excr
org-N
Low [S] High [S]
-0 NOx flux @
Water den Sediment
nit
min + excr
org-N
Figure 5.5: Conceptualisation of the major nitrogen transformation and flux processes within V-core sediment in vitro (incubated in darkness) that is either: (a) uninhabited by macrofauna; (b) inhabited by individual adult A. ehlersi or (c) inhabited by individual juvenile S. aequisetis. Black arrows indicate the relative rate of denitrification (den), nitrification (nit) and ammonification (i.e. mineralisation (min) plus, in the case of inhabited sediment, nereidid excretion (excr)). Grey arrows indicate the net rate and direction of NH4
+ flux. The net rate and direction of NOx- flux (open arrows) is dependent on the
initial ambient concentration of NO3- in the overlying water [S]. Thus illustrated are NOx
- flux rate and direction at both low and high [S], in addition to the approximate concentration of [S] when NO - +
103 flux = 0. Initial ambient NHx 4 concentration in the overlying water was held constant at ~ 13 µM. The size of arrows indicates the relative magnitude of each nitrogen transformation or flux process (NB. not to scale, see Sections 5.3.3 to 5.3.5 for actual rates).
5.4.2 Nereidid Habit
Each comparatively large A. ehlersi individual (200 mg wet weight) constructed, remained within
and sporadically irrigated permanent, U-shaped burrows within the experimental V-cores. The
sediment in these cores had the highest bulk density and superficially was similar in colour (dark
brown) to that within uninhabited cores. Burrow ventilation presumably enhanced the transport of
overlying oxygenated water to deeper sediment, however either oxygen consuming processes within
the burrow were too prolific or radial diffusion of oxygen was not sufficient to produce a noticeable
discolouration of sediment at the outer edge of the V-cores (radius 12.5 mm). In cross-section,
however, a 1-2 mm ring of light brown sediment was always observed around the burrow lining
with generally wider annuli surrounding the more heavily ventilated burrows. The presence of a
burrow space likely compacted the outer sediment and increased the bulk density of A. ehlersi
inhabited V-cores. Although a decrease in interstitial spacing may restrict solute (including
oxygen) transport, it is unclear whether this compaction effect is an artefact of the presence of the
perspex core lining acting as barrier to lateral sediment movement. The habit of A. ehlersi observed
here is similar to that of many northern hemisphere Nereis spp. (e.g. N. diversicolor, N. virens, N.
succinea) which have been well studied for their effects on sedimentary nitrogen cycling; and thus
comprise a point of reference for inter-species comparisons (discussed below).
Simplisetia aequisetis, on the other hand, reworked the sediment in the present study and did not
generate permanent burrows. The V-core sediments had a lower bulk density and lighter brown
surficial colouration than both A. ehlersi inhabited and uninhabited sediments. The lighter colour is
suggestive of a more aerobic environment and it is hypothesised that the sediment reworking by
these polychaetes likely increased interstitial spacing, facilitating greater oxygenation of the entire
3 cm surface layer of sediments. The absence of permanent burrow construction by S. aequisetis in
this study contrasts to the previous study of De Roach et al. (2002), wherein permanent U-shaped
tubes were constructed and ventilated by individuals residing within similar V-cores. Dorsey
(1982) has noted previously that polychaetes of the species S. aequisetis do not inhabit permanent
burrows during most of their life-span, although reproductive individuals construct burrows wherein
eggs are deposited and larvae brooded. Within the Swan River Estuary, sexual maturity of S.
aequisetis specimens occurs at a body size within the range of ~17-39 mg wet weight
(corresponding to 30–50 mm total length; Chapter 4). Juveniles of about 10 mg wet weight were
utilised in the present study in contrast to the large (~80 mg wet weight), mature tube-building
individuals studied by De Roach et al. (2002)10. Thus, construction of permanent burrows by
specimens of S. aequisetis apparently is age-dependent: mobile immature juveniles rework 10 In the Materials and methods section of De Roach et al. (2002), the biomass ~0.8 g wet weight should read ~ 0.08 g wet weight.
104
sediment, while reproductively mature individuals inhabit permanent brooding burrows. In contrast
to the well-examined role of burrows in sedimentary nitrogen cycling, fewer studies have
investigated the effects of polychaete species that do not form burrows. Thus, for purposes of
comparison with juvenile S. aequisetis, where possible, I have considered below the impact of
sediment reworking by polychaetes and also various other species upon nitrogenous processes and
flux rates. However, the extent to which the habits of different species may represent functional
groups, i.e. sets of species or habits that impart similar effects on nitrogen-cycling processes (sensu
Chapin et al., 1992), is beyond the scope of this discussion, but is the subject of Chapter 6.
5.4.3 Denitrification
The presence of A. ehlersi considerably enhanced denitrification in comparison with uninhabited
sediment (Figure 5.5). At in situ ambient nitrate concentration, the denitrification rate (V2μM) of V-
core sediment in the presence of A. ehlersi was estimated to be 284% of uninhabited sediment. This
compares with a range of 138-500% enhancement of denitrification over defaunated sediment,
previously documented for nereidid species inhabiting permanent U-shaped burrows (Henriksen et
al., 1980; Kristensen et al., 1985; Pelegri and Blackburn, 1995a). Alternately, these U-burrowing
nereidids are responsible for between 28-80% of denitrification in the sediments they inhabit (when
considered as a single macrofaunal species addition). This effect of denitrification enhancement has
been observed for: Nereis virens (Hylleberg and Henriksen, 1980; Henriksen et al., 1980, 1983;
Kristensen, 1985; Kristensen et al., 1985, 1991; Kristensen and Blackburn, 1987; Christensen et al.,
2000), N. diversicolor (Law et al., 1991; Gilbert et al., 1995, 1997; Hansen and Kristensen, 1997,
1998; Mortimer et al., 1999; Christensen et al., 2000; Heilskov and Holmer, 2001), N. succinea
(Bartoli et al., 2000) and unidentified Nereis spp. (Binnerup et al., 1992; Pelegri and Blackburn,
1995a; Meyer et al., 2001). De Roach et al. (2002; Appendix E) similarly observed enhanced
denitrification activity surrounding the burrows of adult S. aequisetis in the Swan River Estuary.
While there may be several factors accounting for the enhanced rate of denitrification within
sediments inhabited by A. ehlersi and other U-burrowing nereidids, De Roach et al. (2002;
Appendix E) summarised the likely main contributors as: (i) the presence of a burrow provides an
increased sediment-water interface, extending the surface area (and oxic-anoxic boundary) available
for microbial inhabitation and solute exchange (Hylleberg and Henriksen, 1980; Kristensen, 1984a;
Kristensen et al., 1991); (ii) the mucus binding the burrow walls may elevate the organic matter
content and resultant microbial activity of the sediment (Aller, 1983, 1988; Aller et al., 1983;
Reichardt, 1988), and/or (iii) active ventilation of a burrow may increase nitrate supply (as a
substrate for denitrification) to the sediment; furthermore, nitrification may also be enhanced by
105
active ventilation (see Section 5.4.4), which provides another NOx- source for denitrification (Aller,
1988; Kristensen, 1988; Kristensen et al., 1991).
Each of these factors may have contributed to the enhancement of denitrification rate observed
within the A. ehlersi inhabited sediments of this study, but it was not possible to determine their
relative importance. The intermittent burrow ventilation pattern demonstrated by
A. ehlersi (Figure 5.3) likely facilitated enhanced coupling of nitrification-denitrification; since
enhanced nitrification heightened NOx- production (see Figure 5.5; Section 5.4.4) when the burrow
is being ventilated by oxygenated water (i.e. aerobic conditions within the burrow wall), which may
be consumed in surrounding sediment, probably at a heightened rate of denitrification during
periods of non-ventilation (i.e. when oxygen consuming processes allow more anaerobic conditions
to prevail) (Kristensen, 2000). Studies employing nitrogen isotope tracers within A. ehlersi
inhabited sediment are required to determine the partitioning of denitrification resultant either from
NOx- supplied by nitrification or by direct drawdown from the overlying water (e.g. see Pelegri and
Blackburn, 1995a; Bartoli et al., 2000).
The maximum potential denitrification rate of A. ehlersi inhabited sediments at non-limiting
ambient nitrate concentration (Vmp) was only 110% of the maximal denitrification rate of
uninhabited V-core sediments. Also, the denitrifying community associated with A. ehlersi
inhabited sediment demonstrated the lowest Michaelis-Menten constant (Kapp), indicating that
maximal denitrification was achieved at lower ambient nitrate concentration than in uninhabited, or
S. aequisetis inhabited, sediment. Thus, to some extent, the magnitude of denitrification
enhancement will depend on ambient nitrate concentration. This may help to explain why the
degree of enhancement has not previously appeared related to species. For example, N. virens has
been observed to stimulate denitrification by either the lower (138%) or upper (500%) end of the
above-stated range (see respectively, Henriksen et al., 1980 and Kristensen et al., 1985), whilst the
increases resultant from N. diversicolor addition are mid-range (177-300% stimulation; Hansen and
Kristensen, 1998). However, it is equally likely that the variation in the magnitude of
denitrification enhancement is due to differences and interrelations of the many other factors
controlling the rate of denitrification; for example, the specific increase in sediment-water interface
area provided by individual burrows, or the organic matter content and quality of the mucus-bound
sediment wall, and/or the actual pattern and magnitude of burrow irrigation that may markedly
influence the redox conditions which control absolute denitrification rate, amongst other factors. It
is probable that a combination of these and other regulatory factors determine the overall influence
of nereidids upon sedimentary denitrification, but elucidation of their relative magnitude of
importance largely waits for explicit testing.
106
In contrast, the presence of juvenile S. aequisetis was found to retard denitrification in comparison
with uninhabited sediment. At in situ ambient NO3- concentration, the denitrification rate within
S. aequisetis inhabited sediment was 59% of that within uninhabited sediment and only 21% of the
rate within A. ehlersi inhabited sediment. In response to increasing ambient NO3- concentration, the
denitrifying community within S. aequisetis inhabited sediment reached a maximal denitrification
rate that was 88% of uninhabited sediment and 81% of A. ehlersi inhabited sediment, but the
maximum occurred at a much higher ambient NO3- concentration (i.e. as defined by the highest
Kapp). Although comparative studies are scarce, this net inhibitory effect of inhabitation by a
sediment reworker upon denitrification activity has been observed in marine mesocosms containing
the heart urchin Brissopsis lyrifera (Widdicombe and Austen, 1998). This species ‘bulldozes’ and
aerates the top few centimetres of sediment which, the authors argue, has the effect of suppressing
anaerobic denitrification. It is likely that the degree of oxygenation caused by the sediment
reworking habit of S. aequisetis similarly inhibited denitrification within the V-cores of this study.
Conversely, freshwater oligochaetes such as Tubifex tubifex display a habit comparable to juvenile
S. aequisetis, but have been demonstrated to enhance the denitrification rate of aquatic sediments
due to tight coupling with likewise enhanced nitrification (Chartapaul et al., 1979, 1980; Pelegri
and Blackburn, 1995b; Svensson et al., 2000). In this study, juvenile S. aequisetis did enhance
sedimentary nitrification (see Section 5.4.4 below) but the implicated increase in NOx- production
was not coupled to an increase in denitrification activity. However, it is important to consider that
the observed net reduction in denitrification activity may not occur in situ, if a deeper active zone of
denitrification naturally develops (i.e. deeper than the V-cores allowed in vitro). While not
resulting in a net decrease, suppression of denitrification activity within the surface layer of
sediments and burrow walls due to oxygenation by resident fauna has often been documented
(Sayama and Kurihara, 1983; Kristensen et al., 1985; De Roach et al., 2002). Heilskov and Holmer
(2001) have suggested that the net affect of macrofauna stimulating or inhibiting anaerobic
processes [such as denitrification] probably results from the balance between counteracting effects,
namely stimulation due to an increased supply of substrates and inhibition resulting from more
oxidised conditions.
5.4.4 Nitrification
The estimated average nitrification rate of sediment within A. ehlersi inhabited V-cores was 130%
of uninhabited sediment. The magnitude of this enhancement effect is at the lower end of the range
documented previously for N. virens (138 to 228%: Henriksen et al., 1980; Kristensen and
Blackburn, 1987; Kristensen et al., 1991) and N. diversicolor (133 to 240%: Hansen and
Kristensen, 1998). Mayer et al. (1995) similarly found that potential nitrification rates within the
107
burrow walls of N. virens were 170 to 410% of that in surface sediment. Furthermore, their study
demonstrated that potential nitrification rates were stimulated considerably in the burrows walls of
11 out of 12 other infaunal species that construct definite burrow or tube structures. Factors
explaining the heightened nitrification rates of sediments containing infaunal burrows are
considered below.
The oxygen requirement of chemoautotrophic nitrifying bacteria largely limits their distribution to
aerobic layers of sediment (Henriksen et al., 1981). The presence of an infaunal burrow and
concomitant irrigation with oxygenated water considerably extends the surface sediment-water oxic
layer to depth, thus increasing the volume of sediment available for inhabitation by nitrifying
bacteria (Mayer et al., 1995). As discussed above, the ventilation rate and pattern may control both
the redox conditions and substrate supply (of NH4+ in the case of nitrification) within the burrow,
which in turn govern the process rates of aerobic/anaerobic microbial nitrogen transformations.
Significantly, Mayer et al. (1995: p. 168) suggest that “average rate of water turnover in tubes or
burrows might provide the best ‘single’ trait for predicting the availability of oxygen in tubes and
burrows and consequently for predicting species differences in the enhancement of nitrification in
these macrofaunal structures”. They further observed that those infaunal species that were strong
irrigators, i.e. with high rates of burrow ventilation (>150 mL h-1) or water turnover (>1mL min-1),
only mildly enhanced nitrification; and suggested that because higher irrigation exports NH4+ from
the sediment, enhancement of nitrification may be greatest at intermediate irrigation rates. The
results of this study lend support to the idea, since A. ehlersi exhibited a relatively high average
burrow ventilation rate (415 mL h-1) and the observed percentage enhancement of nitrification was
only at the lower end of the reported range of enhancement (see above). Nevertheless, Mayer et al.
(1995) concede that, while aspects of burrow ventilation are undoubtedly important determinants of
net nitrification rate, they are not solely responsible.
Excretion by resident fauna is a source of NH4+ that is undoubtedly of importance when considering
the enhancement of nitrification due to a heightened provision of substrate (Henriksen et al., 1983;
Welsh and Castadelli, 2004). In the present study, the rate of ammonification within A. ehlersi
inhabited V-cores was much greater than that within uninhabited sediment (Figure 5.5). While the
contribution of NH4+, either from excretion or microbial mineralisation could not be separated, it is
very likely that A. ehlersi excretion stimulated nitrification in adjacent sediment to some extent
(further see Section 5.4.6). Furthermore, intra- and inter-species variation in excretion rates may
have profound implications on the potential for nitrification within macrofaunal inhabited
sediments. Indeed, Mayer et al. (1995) found the availability of NH4+ to be correlated with
nitrification potential of sediment inhabited by the polychaete Loimia medusa. A worthwhile
108
avenue of study would be to examine whether a Michaelis-Menten type relationship exists between
ambient NH4+ concentration and nitrification rate of inhabited sediments, similar to the relationship
between ambient NO3- concentration and denitrification rate documented above. As Mayer et al.
(1995) have pointed out, the existence of substrate-dependant relationships determining the rates of
both nitrification and denitrification is not surprising, but the influence of ambient substrate
concentrations have not been recognised widely when comparing studies.
Kristensen (2000) has also pointed out that other governing factors must necessarily be considered.
For example, faecal pellet deposition may enhance nitrification, as has been shown for Macoma
balthica (Henriksen et al., 1983). Nitrification rates within burrow walls have also been correlated
with the content of mucus and fine particles (Kristensen, 2000). Despite the chemoautrophic nature
of nitrifying bacteria, Kristensen (2000) suggested that their association with the often high organic
content and fine particles of burrow walls may be due to an optimal pH environment that such
particles provide. Finally, Welsh and Castadelli (2004) recently documented nitrification due to the
activity of nitrifying bacteria occurring directly on or within the bodies of various macrofaunal
species, including an estimated 15-63% increase in nitrification rate of the benthos due to nitrifying
bacteria resident upon N. succinea. This novel aspect of nitrification enhancement and the causal
mechanisms are only just beginning to be investigated. Again, elucidation of the relative
contribution of the various factors governing the enhancement of nitrification within inhabited
sediment, largely still requires explicitly directed manipulative experiments.
The enhancement of nitrification within sediment inhabited by fauna that do not construct burrows
has been less documented. In this study, the presence of juvenile S. aequisetis enhanced the
nitrification rate of sediment in V-cores to a greater extent than adult A. ehlersi; i.e. the estimated
average nitrification rate of S. aequisetis inhabited cores was 148% of that within uninhabited cores
and 113% of that within A. ehlersi inhabited cores. A similar enhancement effect has been
demonstrated for freshwater sediment reworking tubificid oligochaetes (such as T. tubifex and
Limnodrilus spp.), wherein nitrification rates of inhabited sediment have been within the range of
123-180% of defaunated sediment (Chartapaul et al., 1979, 1980; Pelegri and Blackburn, 1995b;
Svensson et al., 2001). Obviously, the disturbance caused by this type of macrofaunal habit (at
lower worm densities, see below) does not concurrently result in a disruption to the activity of the
resident nitrifier population; on the contrary, it stimulates nitrification rate. Some of the reasons for
nitrification enhancement within macrofaunal burrows may extend to reworked sediment, e.g.
increased NH4+ supply from macrofaunal excretion and perhaps enhanced nitrification directly
associated with nitrifiers on the body of macrofauna.
109
Furthermore, it is hypothesised that the reworking habit has a similar effect on nitrification as
infaunal burrow irrigation. The constant turnover of sediment and reduction in sediment bulk
density caused by juvenile S. aequisetis likely facilitates transport of oxygenated water at a rate and
to a depth much greater than expected by diffusion, thereby considerably enhancing the volume of
sediment available for inhabitation by nitrifiers. Svensson et al. (2000) consider both sediment
reworking and burrow irrigation as forms of biological advection, but suggest that the random
nature of solute/particle transport of the former habit and directed flow of the latter may result in
very different effects on nitrification and denitrification. Certainly, it appears that the enhancement
of aerobic nitrification but suppression of anaerobic denitrification within V-cores inhabited by
juvenile S. aequisetis may be due to the homogenisation and aeration of the entire surface 3 cm
layer of sediment. In contrast, the directed flow of water in A. ehlersi inhabited V-cores and
persistence of a lower oxygen environment in the surface sediment surrounding burrows may be the
major difference facilitating the enhancement of not only nitrification, but also denitrification.
It is important to note that the studies of freshwater tubificid oligochaetes by Pelegri and Blackburn
(1995b) and Svensson et al. (2001) found that nitrification was suppressed at very high worm
densities. The authors of both studies implicated heightened tubificid net respiratory oxygen
consumption to explain the retardation of aerobic nitrification. However, it may also be that a
higher sediment turnover rate at high worm density flushes NH4+ from the sediment (similar to the
effect of high irrigation rates; see above) or causes appreciable physical disturbance to nitrifier
populations. Indeed, Svensson et al. (2001) and other authors have found high levels of sediment
disturbance (e.g. by sieving) to impact aerobic and slow-growing nitrifier populations to a greater
extent than anaerobic organisms, such as denitrifiers. Further investigation is required into the
causal mechanisms generating variability in the effect of sediment reworkers and other types of
infaunal habit upon nitrification, denitrification and other nitrogen cycling processes.
5.4.5 Net NOx- Flux
The balance of NOx- producing (e.g. nitrification) and consuming (e.g. denitrification) processes
will determine the net NOx- flux rate of a sediment. At low ambient NO3
- concentration, the net
efflux rate of NOx- from V-core sediment in vitro was enhanced by the presence of both A. ehlersi
and juvenile S. aequisetis, due to their stimulatory effect on nitrification (Figure 5.5). Within
A. ehlersi inhabited sediment this enhancement of NOx- efflux occurred even though denitrification
was also stimulated; however, the enhancement of NOx- efflux was greater within S. aequisetis
inhabited cores because not only was nitrification stimulated to a higher extent, but the consumption
of NOx- by denitrification was retarded.
110
In response to increasing ambient NO3- concentration in the overlying water, denitrification activity
increased in a Michaelis-Menten type fashion for all V-core sediment treatments, although the rate
of increase was highest in A. ehlersi inhabited sediments due to this species’ stimulatory effect on
denitrification. The heightened stimulation of denitrification caused the apparent switch from
sedimentary NOx- efflux to influx to occur at a lower ambient NO3
- concentration in A. ehlersi
inhabited sediment (i.e. at ~ 50 µM), than in uninhabited sediment (~100 µM) or S. aequisetis
inhabited sediment (~ 520 µM) (Figures 5.4[c] and 5.5). The stimulatory effect of juvenile
S. aequisetis on nitrification and concurrent inhibitory effect on denitrification entailed that NOx-
was effluxed until a very high concentration of NO3- in the overlying water (i.e. when the rate of
denitrification eventually outweighed nitrification).
Thus, although at low ambient NO3- concentration in the overlying water both species stimulated
nitrate efflux, at higher ambient NO3- concentrations the presence of A. ehlersi increased the influx
of NOx-, whilst juvenile S. aequisetis suppressed the influx of NOx
-. These observations have very
important implications when considering the effect of each species in oligotrophic (low ambient N
concentration) versus eutrophic (high ambient N) environments. That is, for the conditions in vitro,
both species would increase the NOx- available for algal uptake in the overlying water when at
lower ambient NO3- concentration; whereas, at higher ambient NO3
- concentration, the presence of
A. ehlersi in sediment would remove NOx- from the overlying water at a greater rate than in waters
overlying uninhabited sediment. Conversely, the presence of juvenile S. aequisetis at these higher
ambient NO3- concentrations would reduce the ability of the sediment to remove NOx
- from the
overlying water.
5.4.6 Ammonification and Net NH4
+ Flux
As illustrated in Figure 5.5, the ammonification rate was lowest in uninhabited sediment, higher in
S. aequisetis inhabited sediment (202% of the rate in uninhabited sediment) and highest in
A. ehlersi inhabited sediment (242% of the uninhabited rate). Whilst the excretion rates of each
species were not measured, thus disallowing segregation from the contribution made by microbial
mineralisation, uninhabited cores were obviously not subject to the former source of NH4+ supply
and it is suggested that macrofaunal excretion is a major cause of increased ammonification rates
within inhabited sediments. However, the effects of various infauna on sediment redox conditions
and resultant implications for microbial mineralisation, whether enhanced or inhibited, cannot be
discounted (for an excellent review see Kristensen, 2000). Resolution is still required of these
effects in A. ehlersi and S. aequisetis inhabited sediment.
111
Notwithstanding the inability to designate enhanced ammonification rates of inhabited cores either
to increased microbial mineralisation or excretion (or a combination of both), the net effect of
nitrogen cycling processes in sediment V-cores in vitro, resulted in: (i) a slight average efflux of
NH4+ from sediment in A. ehlersi inhabited cores and (ii) an influx of NH4
+ to sediment of cores
inhabited by juvenile S. aequisetis that was lower than the influx observed in uninhabited cores
(Figure 5.5). It is hypothesised that even though the nitrification rates within uninhabited sediments
were lower than in inhabited sediments, the high net NH4+ consumption of uninhabited sediment
was due the lack of infaunal excretion to provide a supplementary substrate source and thus a
concomitant dependence on diffusional NH4+ from the overlying water to supply nitrification. The
stimulation of ammonification within inhabited burrows (due either to excretion or enhanced
microbial mineralisation), and resultant in situ availability of NH4+ for nitrification, likely caused
the lower net influx of NH4+ to S. aequisetis inhabited sediment and slight efflux of NH4
+ from A.
ehlersi inhabited sediment (because ammonification was greatest in these sediments).
Most studies have similarly documented that the presence of macrofauna either enhances NH4+
efflux or suppresses NH4+ influx. The magnitude of the effect is often correlated with macrofaunal
density (e.g. Asmus, 1986; Law et al., 1991; Pelegri and Blackburn, 1995b; Lohrer et al., 2004) and
has been observed previously for macrofaunal communities in the Swan River Estuary (Pennifold
and Davis, 2001). The enhancement of sedimentary NH4+ efflux by a variety of species, both in the
dark and under light, is well illustrated by Andersen and Kristensen (1988), who examined the
effect of single species additions of: (i) the amphipod Corophium volutator (a 3-5 cm deep U-
burrower); (ii) the gastropod Hydrobia spp. (a sediment reworker); (iii) the nereidid polychaete
N. virens (a 10-40 cm deep U-burrower); and (iv) a combined community of all three species.
Compared to defaunated sediment, they found that NH4+ efflux was always increased by between
110 and 1000%. This range of NH4+ efflux enhancement by macrofauna is common to that often
reported by other researchers, who usually relate the effect either to the increased sediment-water
surface area provided by burrows, the presence of biological advection (either by irrigation or
sediment reworking) and/or increased ammonification due either to excretion or microbial
mineralisation. Whilst the extra provision of NH4+ via macrofaunal excretion is obvious, it is
worthwhile considering the study of Henriksen et al. (1983) to exemplify that other explanatory
factors may be very important. For both organic rich and poor sediments inhabited by similar
species to those examined by Andersen and Kristensen (1988; see above), NH4+ efflux rates were
equivalent to 60-270% of excretion rates. Efflux rates greater than 100% of excretion were thus
due to supplementary enhancement factors other than excretion.
112
Although more rarely reported, examples exist whereby the presence of macrofauna causes a
reduction in NH4+ efflux due to significant enhancement of nitrification (i.e. great enough to
outweigh any increased production of NH4+). In Long Island Sound sediments, Aller (1978)
documented this effect for the protobranch bivalve Yoldia limatula, a mobile subsurface deposit
feeder that reworks sediment to depths of about 4 cm. Within a marine sediment community off
Sweden, Enoksson and Samuelsson (1987) suggested that the presence of a brittle star (Amphiura
filiformis) can decrease the NH4+ efflux of sediment via enhanced nitrification caused both by
oxygenation of the upper layer of sediment and the ability of its calcium carbonate body structure to
buffer sediment pH to an optimal level for nitrifiers. As a final example, Caffrey and Miller (1995)
found in South San Francisco Bay that the nitrification rate within a sediment community
dominated by the tube-dwelling polychaete Asychis elongata was positively correlated with
macrofaunal biomass; and further, that net NH4+ efflux was negatively correlated with macrofaunal
biomass.
The net effect of infaunal habits and excretion rates on net sedimentary NH4+ fluxes has
implications for the regulation of the amount of NH4+ available for algal assimilation in overlying
waters. Even though the general trend documented in the literature is that the presence of fauna
generally stimulates NH4+ release from sediments, different macrofaunal species, habits and
excretion rates will undoubtedly have variant magnitudes of effect on NH4+ availability in the
overlying water. Whilst this study demonstrates the differing net effects of a mobile sediment
reworker and stationary U-burrower upon net NH4+ flux in vitro, the studies of Watson et al. (1993)
and Davey and Watson (1995) provide similar examples for in situ sediment communities of the
Tamar Estuary, England. At locations throughout the estuary, they found that the presence of
infauna resulted in effluxes of NH4+ that were 3 to 63 times greater than expected by diffusive flux.
Importantly, they observed that the degree of enhancement depended on the density and population
structure of resident macrofauna and, moreover, whether irrigating or sediment reworking fauna
predominated (with irrigators enhancing NH4+ efflux to a greater extent). The effect of different
faunal habits upon sedimentary nitrogen cycling and the availability of total dissolved inorganic
nitrogen (NOx- + NH4
+) to bloom forming phytoplankton species is further considered in Chapters 6
and 7.
5.4.7 Efficacy of the Kinetic-fix Adaptation of the Acetylene Block Method
As far as the author is aware, this is the first study in which the novel ‘kinetic-fix’ adaptation of the
C2H2 block technique (Joye et al., 1996) has been employed to examine nitrogen cycling within
sediments inhabited by macrofauna. The method is briefly re-described here to highlight its
benefits. Estimates of nitrogenous process and flux rates were determined in sediment V-cores
113
separately inhabited by individual polychaetes of the two species, and compared with uninhabited
sediment. Briefly, each V-core species treatment (A. ehlersi, S. aequisetis, uninhabited) was subject
to a range of ambient NO3- concentrations, S = 0, 20, 50, 100 and 500 μM NO3
-, at which NH4+ and
NOx- fluxes (incorporating denitrification rates) were measured first without and then with C2H2
(acetylene). This kinetic-fix adaptation of the C2H2 block technique allowed for estimates of
potential denitrification rates at saturated NO3- concentration as well as for estimates of
denitrification rates at in situ NO3- concentration. The technique also allowed resolution of
sediment nitrification and ammonification rates and determination of net NOx- and NH4
+ fluxes
within the sediment of each V-core species treatment. Although appropriate corrections and
considerations were required to account for the effect of acetylene on A. ehlersi ventilation, the
technique is cheap and technically easy, yet has been demonstrated to be powerful in resolving the
effects of macrofauna upon sedimentary nitrogen cycling.
5.5 SUMMARY
The effects of polychaete inhabitation on sediment NH4+ (ammonium) and NOx
- (nitrate + nitrite)
fluxes, including the contributing denitrification, nitrification and ammonification rates, were
assessed for two species of polychaete (A. ehlersi and S. aequisetis) within the Swan River Estuary,
Western Australia. Australonereis ehlersi reside in well-defined U-shaped burrows, whereas
juvenile S. aequisetis are sediment reworkers.
For the denitrifying communities associated with uninhabited cores and those inhabited by
A. ehlersi or S. aequisetis, the apparent Michaelis constant (Kapp) and the maximum potential rate
(Vmp) are reported. As defined by resultant Michaelis-Menten curves, the magnitude of
denitrification at in situ NO3- concentration (V2μM), and across the range of experimental S, was
greatest for sediment inhabited by A. ehlersi (V2μM = 182 μmol N m-2 h-1; Vmp = 4.75 mmol N m-2
h-1; Kapp = 50 μM NO3-), moderate in uninhabited sediment (V2μM = 64 μmol N m-2 h-1; Vmp = 4.33
mmol N m-2 h-1; Kapp = 133 μM NO3-) and lowest in sediment inhabited by S. aequisetis (V2μM = 38
μmol N m-2 h-1; Vmp = 3.83 mmol N m-2 h-1; Kapp = 202 μM NO3-). Sediment nitrification rate was
generally constant over the range of S and estimated to be higher in S. aequisetis (2.76 ± 0.17 [s.e.]
mmol N m-2 h-1) than in A. ehlersi (2.43 ± 0.07 mmol N m-2 h-1) inhabited cores, and lowest in
uninhabited sediment (1.86 ± 0.13 mmol N m-2 h-1). In the absence of C2H2, the net result of
nitrification minus denitrification was such that a transition from NOx- efflux from sediment at
lower S to sedimentary NOx- uptake at higher S occurred. The point of equilibrium (zero flux)
occurred at a lower S for A. ehlersi inhabited sediment cores (52 μM) than for uninhabited sediment
cores (100 μM), with S. aequisetis inhabited cores generally effluxing NOx- over the entire range of
114
S, up to 520 μM. For each V-core species treatment, ammonification in the presence of C2H2 was
generally constant over the range of S, and indicated the net rate of NH4+ production from sediment
mineralisation plus, in the case of inhabited cores, polychaete excretion (though these components
could not be separated). Estimated net ammonification was higher for A. ehlersi inhabited (2.20 ±
0.05 mmol N m-2 h-1) than S. aequisetis inhabited cores (1.84 ± 0.16 mmol N m-2 h-1), and lowest in
uninhabited cores (0.91 ± 0.05 mmol N m-2 h-1), within which polychaete excretion was absent. In
the absence of C2H2, and associated unblocking of nitrification, sediments of S. aequisetis inhabited
and uninhabited cores indicated a net NH4+ influx of -0.62 ± 0.08 mmol N m-2 h-1 and -0.85 ± 0.12
mmol N m-2 h-1, respectively, whereas A. ehlersi inhabited cores exhibited a slight average efflux of
NH4+ (0.07 ± 0.13 mmol N m-2 h-1) from the sediment.
The high resultant rates of nitrogenous flux determined in vitro may not reflect rates at natural, in
situ conditions. Further, the estimates of nitrification and kinetic denitrification parameters for V-
cores inhabited by A. ehlersi are more subjective than for the other two V-core species treatments
because of the requirement to extrapolate to ventilation-standardised NOx- flux estimates (due to
polychaete burrow ventilation being suppressed by C2H2). It is suggested that the difference in
magnitude of nitrogenous fluxes imparted by the two polychaete species is related to the influence
that their respective habits exert on the composition and activity of their associated sedimentary
microbial community. The interstitial habit of juvenile S. aequisetis is hypothesised to homogenise
and aerate sediment continually, thereby enhancing microbial nitrification and retarding anaerobic
denitrification. The permanence of A. ehlersi burrows is suggested to allow oxic/anoxic
stratification of sediment (both vertically and radially) which, combined with enhanced substrate
supply through burrow ventilation, provides for increased rates of microbial denitrification and
nitrification. Thus, due to the different habit of each nereidid species, the resultant net fluxes of
both NOx- and NH4
+ are variant (see above). The habits of constituent benthic faunal species may
therefore play a profound role in regulating the type and concentration of dissolved inorganic
nitrogen available for assimilation by bloom-forming algal species in the overlying water.
115
PART III– CONCLUSIONS
117
118
CHAPTER 6
POLYCHAETE EFFECTS ON NITROGEN CYCLING AT THE
SEDIMENT-WATER INTERFACE: TESTING THE
FUNCTIONAL GROUP CONCEPT
6.1 INTRODUCTION To predict, reliably, the effect of each element of the benthos exerting some biogenic
modification of sediments requires basic knowledge of the species, particularly of the
processes involved in forming, maintaining and degrading micro-environments surrounding
individual organisms (Aller and Yingst, 1985). In quantifying the impact of a benthic
species on sedimentary processes associated with nitrogen cycling, for example, the habit or
life mode, life history and population biology of members of the species need to be
considered. Within a population, intraspecific variation in habit and associated differences
in sedimentary microenvironments may have marked implications for biogeochemical
processes. Such variation may result from either: (i) ontogenetic differences in habit over a
life-cycle; (ii) short-term behavioural changes of an individual’s habit; and/or (iii) variant
predisposed habits within or between populations.
Further, spatio-temporal variation and interactions of controlling environmental parameters
hydrodynamic setting, temperature, salinity, population density/burrow spacing, for
example) may be either influential in determination of an individual’s habit, or, for a given
habit, be a source of variation in its biogeochemical effects. Thus, as part of my aim in this
study to elucidate the effect of polychaete species upon nitrogen-cycling at the sediment
water interface of the Swan River Estuary, I consider in this Chapter the influence of both
intra- and inter-specific variation in polychaete and other macrofaunal habits in the context
of their regulatory capacity of sedimentary nitrogen biogeochemistry.
6.2 CLASSIFICATION OF BENTHIC MACROFAUNAL HABITS: THE
FUNCTIONAL GROUP CONCEPT
The challenge in determining the effects of benthic fauna upon diagenetic reactions and
biogeochemical cycles, in this case of nitrogen at the sediment-water interface, is to define
appropriate working groups of animals to facilitate useful discussion instead of attempting to
reduce the analysis to consideration of individual species or even individuals. In identifying
the complex issues involved, Mermillod-Blondin et al. (2001, 2002) adopted the concept of
functional groups, a set of species that have similar effects on ecosystem processes, a
119
concept developed previously by Chapin et al., 1992; Naeem et al., 1994; Tilman et al,
1997) to identify succinctly the relationships between species diversity and ecosystem
functioning. In a study of a New Zealand soft bottom community, Lohrer et al. (2004)
suggested that the loss of a functionally important spatangoid urchin resulted in decreased
sedimentary ammonium effluxes and reduced system primary productivity; thereby
instigating biogeochemical changes that broadly impacted on ecosystem functioning. Given
the present threats to benthic biodiversity due to eutrophication, habitat destruction, over-
harvesting and climate change (Tilman, 1999; Lohrer et al., 2004), identification of
functionally important groups may be critical to our understanding of consequential shifts in
aquatic ecosystem function.
In Section 6.2 I outline a preliminary framework within which guilds of benthic fauna
defined by habit, each with bioturbative influence on sedimentary nitrogen cycling, may be
tested as functional groups. I discuss the benefits and limitations of the approach in Section
6.3 and an unavoidable lack of functional specificity within each guild is emphasised (i.e. an
inability to explicitly categorise a functional group with immutable effects). Although my
primary focus throughout my thesis concerns the effect of polychaete habits upon
sedimentary nitrogen cycling, my focus is broader in this Chapter to include a wide range of
benthic fauna taxa to facilitate both testing in future and expansion of the concept to include
these groups. This kind of holistic approach is desirable given that, within a benthic faunal
community, polychaetes represent only a sub-set of all elements that impact on sediment-
based nitrogen cycling.
Traditionally, benthic faunal species have been classified into ecologically important feeding
groups based on their role in organic matter processing (Mermillod-Blondin et al., 2002).
Five categories of benthic feeding strategy have been delineated (Rhoads, 1974; Kennish,
1990):
1) suspension feeders that consume seston suspended in the water column;
2) deposit feeders that consume deposited detritus;
3) herbivores that ingest plant matter;
4) carnivore-scavengers that eat live animals or recently dead animal tissue; and
5) parasites that obtain nutrition via the fluids of living organisms.
However, an organism’s size, microhabitat and sediment preference (Day et al., 1989) are
often used as a basis of classification. Based on their size, benthic fauna may be grouped
120
into macro-, meio- and microfauna: macrofauna and meiofauna are those animals that are
retained on sieves with a mesh gauge of 0.5-2 mm (e.g. many molluscs, polychaetes and
decapods) and 0.04-0.1 mm (e.g. most nematodes and harpacticoid copepods, many
turbellaria and juvenile macrofauna), respectively; microfauna pass through sieves of the
latter mesh dimensions (generally protozoans, for example foraminiferans and ciliates) (Day
et al., 1989; Kennish, 1990). Benthic animals are also categorised in two broad microhabitat
zonations: the epifauna and the infauna (Thorson, 1957). Infauna live below the sediment
surface and within the sediment fabric, either interstitially or within burrows and tubes;
epifauna live on the surface of sediments or hard substrates (Day et al., 1989; Kennish,
1990). Finally, the sediment substrate preference of benthic organisms may be used to
define groups such as soft bottom forms (preferring silty to clayey muds), sandy bottom
forms (preferring sand to sandy silt) and hard bottom forms (epifauna living on rocky
substrates) (Day et al., 1989). These secondary descriptors (size, microhabitat and substrate
preference) are useful at refining pertinent subdivisions within the primary (feeding) groups.
For example, delineation of soft bottom deposit-feeding meio-infauna or hard bottom
suspension-feeding macro-epifauna etc. is descriptive and informative of the size relations
and spatial positioning of benthic fauna within defined feeding groups. Variations in the
definition of functional feeding groups exist (Day et al. 1989), but the traditional emphasis
of group separation is on the trophic status of constituent fauna.
When considering faunal impacts on diagenetic and biogeochemical processes, Mermillod-
Blondin et al. (2002) have argued that trophic-based classification systems are more
restrictive than functional group definitions that include non-trophic activities such as
bioturbation (for example, see definitions by Lee and Swartz, 1980; Chapin et al., 1992).
The feeding mode of a benthic animal may be indicative of its impact through bioturbation
on a sediment (Welsh, 2003); nevertheless, as Aller (1978) noted previously, factors other
than feeding relations and including the taxonomic group, size, mobility, life habit, particle
size selectivity, depth and rate of feeding, and population densities in which the species
normally occurs are important factors in defining functional groups relevant to chemical
diagenesis at the sediment-water interface. Further, in recognising that the macrobenthic
faunal mode of feeding, locomotion and relation to the substrate affect sediment structure
and chemistry, Lee and Swartz (1980) developed a classification based on feeding-mobility
to group organisms in relation to bioturbation processes affecting the distribution of
pollutants in marine sediments. In considering the role of burrowing animals in organic
matter diagenesis in coastal marine sediments, Kristensen (2000) similarly noted that
functional groups may be defined by feeding type, life habit and mobility.
121
In Table 6.1, I propose a classification of guilds of benthic animals relevant to their role in
biogeochemical nitrogen transformations at the sediment-water interface. It is readily
acknowledged that the validity is contentious of considering each of the initially defined
guilds of benthic animals as functional groups, with members of each guild exerting similar
effects on the nitrogenous biogeochemical processes. Hence, the rationale in defining the
subdivisions firstly is discussed, followed by consideration of the functional redundancy of
each guild.
Table 6.1: Bioturbative guilds of benthic faunal habits: A preliminary classification. (See text for details on group segregation rationale).
Habit Guild Description of Habit
Epifaunal Suspension feeding
Epifaunal habits which enhance biodeposition of organic matter/nitrogen via suspension feeding. (e.g. mussels, oysters)
Non-suspension feeding
Non-suspension feeding epifaunal habits that do not directly enhance biodeposition, e.g. epibenthic grazing, carnivory and omnivory. (e.g. mud snails, crabs, prawns)
Surface Infaunal Stationary burrow Reside in the surface oxidised layer of sediment (0-3cm depth) within
permanent or relatively stationary burrows or tubes. (e.g. amphipods, spionid polychaetes, some juvenile nereidid polychaetes)
Mobile interstitial reworker
Live interstitially or in highly transitory burrows within the surface oxidised layer of sediment (0-3cm depth). (e.g. some juvenile nereidid polychaetes, many bivalves and gastropods, sea urchins, brittle stars)
Subsurface Infaunal
Stationary burrow Capable of residing in the deeper anoxic layer of sediment (0 to >10 cm depth) within permanent or relatively stationary burrows or tubes. (e.g. nereidid polychaete U-burrowers, I- and L-shaped burrowers such as maldanid polychaetes and Arenicola marina, i.e. ‘conveyor-belt’ species).
Mobile interstitial reworker
Capable of living interstitially or in highly transitory burrows within the deeper anoxic layer of sediment (0 to >10 cm depth). (e.g. many polychaetes and bivalves, freshwater oligochaetes and isopods)
An initial major division recognises the spatial distribution: epifaunal vs infaunal habits
(Table 6.1). Regnault et al. (1988) have observed, as a gross simplification, that epibenthic
macrofauna are usually associated with ammonia regeneration but have lesser effects on
other nitrogen transforming processes, but the benthic infauna stimulate either mineralisation
or nitrification/denitrification through activity of the associated microbiota. Epifaunal habits
may be divided into those which enhance biodeposition of organic matter/nitrogen via
suspension feeding, namely epibenthic suspension-feeding; and those that do not directly
122
enhance biodepostion such as epibenthic grazing, carnivory and omnivory, epibenthic non-
suspension feeding (Table 6.1).
The depth of inhabitation can be utilised further to classify the infauna into surface vs.
subsurface habits (Table 6.1). Surface fauna typically inhabit the surficial oxidised layer of
sediment (~0-3 cm deep), whereas subsurface fauna are capable of penetrating and living in
the deeper anoxic sediment (0->10 cm) by maintaining a connection with the overlying
oxygenated water and/or sediment via transitory movements or the construction of irrigated
burrows. Henriksen et al. (1983), in investigating the effect of four species of infauna on
sediment-water dissolved inorganic nitrogen fluxes within a Danish coastal sediment, found
that two shallow/surface dwellers, the amphipod Corophium volutator and the bivalve
Macoma balthica, each increased the efflux of nitrate from the sediment (likely due to a
stimulation of nitrification over denitrification). Conversely, the polychaetes Nereis virens
and Arenicola marina, both deeper sub-surface infauna, each caused an influx of nitrate into
the sediment (likely due to an enhancement of denitrification over nitrification).
A final category recognises the mobility of infaunal habits with regard to the capacity for
transport of sediment particles (sediment reworking) vs. water irrigation (Table 6.1). The
two habits are not mutually exclusive (Section 6.3), but sediment re-workers generally tend
to be more mobile and live interstitially or in highly transitory burrows; whereas, irrigating
fauna tend to reside in relatively stationary and more permanent burrows or tubes within
which water is actively or passively injected to facilitate respiration and/or suspension
feeding. The functional importance of each habit has been identified by Watson et al.
(1993) in their study on the Tamar Estuary, Cornwall or Devon, England, when they found
that enhanced fluxes (of nitrate and ammonium) resulted from the activities of benthic
macrofauna, but differences between sites depended on which habit was predominant,
irrigation or sediment reworking.
Given the paucity of information concerning the biology of polychaete species likely to
impact on their biogeochemical roles, I have assigned the polychaetes to guilds in Table 6.1
based on minimal information, primarily regarding the mobility and depth of sediment
inhabitation of each polychaete species. Thus, the classification may be undertaken with
some confidence because detailed information regarding a species habit is not required.
Most polychaete species for which information concerning their effects on nitrogen cycling
is available, are assigned to one of two broad bioturbative guilds, either subsurface
stationary burrow dwellers or subsurface mobile interstitial dwellers. Investigations
conducted have been limited to a few species only, with most belonging predominantly to
123
the former guild. Of these, nereidids that construct U-shaped burrows, predominantly Nereis
virens and N. diversicolor, constitute the most commonly studied forms. A functional
distinction may be made between N. diversicolor which may irrigate burrows vigorously to
suspension-feed (given appropriate overlying phytoplankton concentration), and N. virens
which does not (Christensen et al., 2000; Section 6.3). A second sub-category of subsurface
stationary burrowers comprises polychaetes that reside within L- or I-shaped burrows, for
example Arenicola spp., the maldanids Clymenella spp. and Asychis spp., and the capitellid
Heteromastus. spp. These polychaetes and similar taxa have been described as conveyor-
belt species by Rhoads (1974) since typically they ingest sediments at the deeper end of their
burrows and deposit ingested particles as faeces at the sediment-water interface (Knox,
1977). The subsurface mobile interstitial habit for which information is available is almost
solely represented by polychaetes belonging to the Nephtys genus. Polychaetes belonging to
the surficial guilds of either stationary burrowers (some spionids, for example) or mobile
interstitial dwellers (for example Capitella spp.) no doubt constitute an important component
of the benthic fauna, however there are few studies elucidating their effect on nitrogen
cycling. Juvenile polychaetes may be important functional representatives of these surficial
groups. For example, Watson et al. (1993) suggested that a large efflux of nitrate during
winter at their study site was associated with a high abundance of juvenile N. diversicolor
that enhanced oxygenation and nitrification of surface sediments.
6.3 TESTING THE FUNCTIONAL GROUP CONCEPT
Mermillod-Blondin et al. (2001) argued that if two taxa belong to the same functional group,
then according to the redundancy model (sensu Walker, 1992), they should generate similar
effects on ecosystem processes. The six guilds of faunal habits identified above (Table 6.1)
represent distinct groups of bioturbative influence in terms of sedimentary nitrogen
transformations but they are not necessarily functional groups. The classification system of
Lee and Swartz (1980) recognises three groups of benthic organisms, epifaunal/infaunal,
deposit/suspension feeding and mobile/stationary, but based on their criteria, 12 guilds are
recognised. Further refinement (subdivision) of either system possibly could increase the
functional redundancy of each guild identified and also the number of groups. Lee and
Swartz (1980) tolerated a trade-off between maintaining functional redundancy and
minimising the number of groups for practical purposes; for example, tubiculous and non-
tubiculous groups were not separated to minimise the number of guilds. Similarly, the
classification presented here could include, for example, recognition of suspension feeding
and non-suspension feeding burrowers/irrigators; to do so may have some functional
relevance, however only the six guilds defined above are initially retained to facilitate a
more simplified consideration.
124
Moreover, although the functional group concept is useful for defining relevant bioturbative
guilds of benthic animals, the task of delineating ideal functional groups, i.e. those with
members of each group exerting immutable effects on biogeochemical processes at the
sediment-water interface (i.e. with absolute functional specificity), may be theoretically, but
not practically, possible. The inherent intraspecific variation in life histories, habits and
feeding modes of benthic fauna makes it difficult (if not impossible) to segregate exclusive
groups of species or individuals on such bases.
Juvenile and adult animals may exert varying effects upon nitrogen cycling. Lewin et al.
(1979) found that, although the excretion rate of ammonium is positively correlated to the
length of razor clams (Siliqua patula), the weight-specific rate of excretion (i.e. per gram of
animal) is negatively related to clam size. Therefore, a juvenile population of razor clams
will excrete a larger amount of ammonium than an adult population of the same biomass.
Similarly, juvenile cockles (Cardium edule) will excrete 93-96% of absorbed organic
nitrogen as ammonium and retain 4-7% for somatic growth while adult cockles excrete, as
ammonium, only 60-64% of absorbed organic nitrogen, retaining 36-40% for somatic
growth and the production of gametes (Loo and Rosenberg, 1989). Appositely, it is
documented in Chapter 5 that juveniles of the nereidid polychaete Simplisetia aequisetis
suppressed in vitro anaerobic sedimentary denitrification activity, likely due to the
homogenisation and aeration of sediments generated by the juvenile’s mobile interstitial
burrowing habit. On the other hand, De Roach et al. (2002) established that adult S.
aequisetis are capable of constructing permanent U-shaped burrows. Converse to the effect
of juveniles, these authors reported that actively irrigating adult specimens significantly
enhanced the denitrification activity of proximal burrow sediment in vitro. It is
hypothesised that, whilst similar to the juvenile habit, the adult creation of a permanent
burrow allows the provision of oxygen (and nitrate) at depth, it contrastingly maintains and
extends a stratified anoxic layer within the sediment in which denitrification is enhanced.
Information on age structure of a population may therefore be integral to any predictive
attempt at ascertaining a species net effect on sedimentary nitrogen cycling.
Variation in habit and resultant nitrogen cycling effects may also occur within age-specific
cohorts. In an in vitro study of the effects of bioturbation on freshwater midge larvae
(Chironomus riparius) on benthic nitrogen cycling, Stief and de Beer (2002) recognised two
depth-specific life modes: approximately 90% of larvae were surface sediment reworkers but
10% of larvae made permanent and irrigated subsurface U-shaped burrows. Sediment
ingestion was supposed to be predominant in the former habit and was hypothesised to
reduce the abundance and activity of nitrogen-transforming microbes; conversely, the
125
additional sediment-water interface area for microbial colonisation provided by the latter
burrow habit, combined with an enhanced substrate (nitrate, oxygen, organic nitrogen)
supply via irrigation, was suggested to increase the abundance and activity of nitrogen
cycling micro-organisms (Stief and de Beer, 2002). As an adult, the New Zealand
spantangoid urchin (Echinocardium) may graze on epibenthic microphytobenthos,
potentially reducing oxygen production and thereby favouring anaerobic processes such as
denitrification; alternatively, it may deposit-feed, possibly reducing particulate organic
substrates and the abundance or activity of denitrifying bacteria (Lohrer et al., 2004).
Variation in behaviour associated with alternative habits by an individual may occur in
accordance with spatial or temporal differences in local environmental parameters such as
hydrodynamic setting, temperature, salinity, population density/burrow spacing, for
example. Consequently, the behavioural habit of benthic residents may change under
different environmental settings. This potential for change is well exemplified by the
polychaete N. diversicolor that inhabits and irrigates sub-surface U-shaped burrows. Under
conditions of high phytoplankton concentration in the overlying water column (~ 1-3 μg chl
a l-1), the polychaete tends towards a suspension feeding habit of vigorous irrigation, whilst
at lower algal concentrations a habit of deposit-feeding (on burrow and surficial sediment)
predominates, with less intensive respiratory burrow ventilation (Riisgård et al., 1996;
Vedel, 1998; Nithart et al., 1999; Christensen et al., 2000). For deposit feeding
N. diversicolor habituated to a low phytoplankton load, Christensen et al. (2000)
documented a three-fold increase in ammonium efflux and a halving of nitrate efflux
compared to defaunated sediment. At comparatively high phytoplankton concentration, the
polychaete’s shift to suspension feeding resulted in an almost twelve-fold stimulation of
ammonium efflux and a two-thirds increase in nitrate efflux (compared to defaunated
sediment under high phytoplankton load). The increased stimulation of ammonium efflux
and shift from a reduction to increase in nitrate efflux when suspension feeding was due
mainly to enhanced biodeposition of organic matter/nitrogen, i.e. to a rate 30-times faster
than passive sedimentation (Christensen et al., 2000). Therefore, variation in the intensity of
irrigation and related microbial provisioning of pertinent substrates markedly influences the
magnitude, direction and composition of nitrogenous fluxes at the sediment-water interface.
Another example is given by the intertidal soldier crab (Mictyris longicarpus) which, over
the period of a tidal cycle, will alternate its mode of habit between epibenthic grazing (at low
tide) and subsurface burrowing/deposit feeding (at high tide); each of the habits have explicit
and differing effects on sedimentary nitrogen cycling (Webb and Eyre, 2004a). Thus, the
representation of an ideal functional group (or indeed of a bioturbative guild with lower
126
functional redundancy) is not necessarily composed of a group of species but rather of a
group of habits. A single, benthic species may display more than one habit, either
ontogenetically, due to intra-population variation in life habits, and/or via short-term
behavioural shifts, and therefore belong to more than one functional group or bioturbative
guild. Consequently, it is the habit of a benthic species, rather than the species per se, that
determines membership to a defined functional group or guild. For any given species,
variance in expression of alternate habits (life modes) of individuals and/or populations may
result in net nitrogen cycling effects that are highly dissimilar.
Further complexity is introduced from the effect of a defined functional group changing
paradoxically with variation in environmental settings (i.e. which comprise an array of
natural habitats within a constituent individual’s or population’s residing range). The habit
of benthic fauna may not necessarily change in response to local environmental variation,
but differing impacts upon sedimentary nitrogen cycling may result. Christensen et al.
(2000) provide an elegant example describing the effects of both N. diversicolor and N.
virens on sediment-water nitrogen fluxes under two ambient phytoplankton loads. In
contrast to the change in habit of N. diversicolor from deposit feeding to suspension feeding
and greater stimulation of ammonium efflux under higher phytoplankton concentration, N.
virens maintained a deposit feeding habit irrespective of phytoplankton load. However, N.
virens also stimulated ammonium efflux (compared to that from defaunated sediment) four-
fold under high, compared to two-fold stimulation at low, ambient phytoplankton
concentrations (Christensen et al., 2000). Thus, environmental factors, in this case
phytoplankton (organic matter) loading or degree of eutrophication, can affect the magnitude
of stimulation of nitrogenous effluxes by benthic fauna.
Although some allowance for differences in effect intensity may be possible in the definition
of a functional group (Mermillod-Blondin et al., 2001), variations in the composition and/or
direction of sediment-water nitrogenous fluxes, as influenced by a given benthic faunal
habit, are not permissible. Yet such deviations in the effect of habit do occur, depending on
variation in local environmental parameters. For example, following identical additions of
chironomid larvae (Chironomus plumosus) to sediments of three freshwater Swedish lakes,
Granéli (1979) found the initial effect upon sediment-water inorganic nitrogen flux to be
different for each lake (although the manipulation of chironomid numbers had little long-
term influence on the exchange): ammonium efflux increased initially in sediments of Lake
Årungen and Lake Vombsjön, but ammonium was consumed in sediments of Lake
Trummen; nitrate was initially consumed in sediments of Lake Årungen but nitrate efflux
was increased from sediments of Lake Trummen and Lake Vombsjön. Despite similarities
127
in the burrowing habit of the chironomids in each lake, it was hypothesised that the variation
in direction and composition of resultant nitrogenous fluxes among the lakes resulted from
lake-specific differences in ambient conditions, particularly of redox status and alkalinity
concentration, that affected microbially mediated reactions, especially those controlling the
balance between aerobic nitrification and anaerobic denitrification (Granéli, 1979).
Similarly, diurnal shifts in redox status and inorganic nitrogen availability caused by
photosynthetic oxygen production and nitrogenous assimilation of benthic primary
producers greatly influences the net result of faunal habit effects upon nitrogen cycling. In
the dark, both ammonium and nitrate are often effluxed from sediment inhabited by benthic
fauna; whereas light-associated assimilation of each nitrogen species due to benthic
phototrophic growth typically reduces the efflux of ammonium and may cause a shift to
nitrate influx within inhabited sediments (for example by amphipods and polychaetes
(Andersen and Kristensen, 1988) and decapods (Webb and Eyre, 2004a,b)). For a given
benthic faunal habit, variation in light intensity is thus capable of altering the magnitude,
composition and/or direction of nitrogenous fluxes, both temporally (over the period of a
day) and also spatially: light intensity and periodicity will vary with depth and latitude. The
same is true for variation in redox status which will also change diurnally, as will many
other regulatory environmental factors such as population density, burrow spacing,
temperature and water movements, for example.
A change in habit effect, aside from slight changes in effect intensity, defeats the very
definition of a functional group, the constituents of which must exert similar influence on
ecosystem processes. The possibility that a defined benthic habit may exert a variable effect
on ecosystem functioning, albeit in accordance with natural variation in regulatory
environmental parameters, challenges the usefulness and even validity of the functional
group concept. The qualification must be made that a functional group of similar habits is
only relevant to a particular environmental setting, wherein the habit effect is constant.
Thus, the functional group definition must be relaxed to: A group of habits which, within a
defined environmental setting, impart similar effects on ecosystem functioning – in this case,
on nitrogen cycling. The usefulness of this refined concept in delineating practical
functional groups depends on the robustness of the habit effect in comparison to the effect
and natural variation of other regulatory environmental parameters. If the effect of the
benthic habit is a major determinant of nitrogen cycling processes, overriding or even
determining the influence of other environmental parameters, then recognition of that habit
as a functional group will have broad application in many environmental settings. However,
if variation in other environmental parameters (such as organic and inorganic nitrogenous
128
substrate availability, redox status, photoperiod, for example) is independently more
important in causing change in nitrogen cycling processes, then the effect of habit may only
be constant within a restricted and defined range of those parameters within which the
functional group concept would apply.
For a given habit, the multiplicity of functional groups then depends on the extent to which
environmental variation results in habit effects that are appreciably different. As a
hypothetical example, if magnitude and quality of organic matter supply to sediments largely
determines nitrogen cycling effects, and this parameter is highly variable across a range of
habitats, then many functional groups may need to be recognised for a given habit, because
the habit effect is likely to change in accordance with variation in organic matter supply.
However, if supply of organic matter is invariable across habitats, then the effect of a
specific habit is likely to remain constant (provided there is limited influence from other
factors), and represents only one functional group. Many other environmental parameters
control the magnitude, composition and direction of nitrogenous fluxes at the sediment-
water interface, which may or may not dominate habit effects.
Resolution of the potential number of functional groups for any given habit (which
determines practicality of the concept), thus depends on a concurrent consideration of the
influence and natural variation of regulatory environmental parameters. Since these
parameters interact with each other and with the effect of habit in complex relationships, the
task is complicated. Expanding on the previous example, organic matter supply may control
the potential rate of production of remineralised ammonium. Even if supply is invariant, the
intensity and direction of ammonium flux may be altered by interactions with benthic
primary producer activity (ammonium assimilation, oxygen production) which, in turn, may
be dependent on benthic faunal habit (for example grazing vs deposit feeding) or variation in
light intensity/periodicity, amongst other factors (e.g. see Lohrer et al., 2004). The
culmination of such interactions comprises a distinct benthic environmental setting, which is
represented by a particular balance of nitrogenous fluxes.
Thus, to ascertain the potential for marked deviations in the effect of a given faunal habit
upon the balance of sedimentary nitrogen cycling (i.e. identifying the potential number of
functional groups) requires, firstly, knowledge of the degree of heterogeneity in regulatory
environmental parameters within and between benthic environmental conditions. To some
extent this depends on the scale of environmental setting being considered. When
encompassing marine, estuarine and freshwater sedimentary environments, i.e. adopting a
global view of the benthos, a defined benthic habit is expected to be subject to a large range
129
of variation in parameters and their interactions that influence nitrogen cycling. Variation in
regulatory parameters is also expected within each of these broadly defined environments,
not necessarily on a smaller scale: for example (i) between near-shore and deep sea marine
sediments (subject to differences in terrestrially-derived organic matter supply, light
availability, temperature and pressure, for example); (ii) along an estuarine gradient (subject
to variation in salinity and a multitude of related and influential environmental parameters);
and (iii) between lakes and rivers subject to differing hydrological regimes and associated
effects. Indeed, considerable environmental variation is expected at even smaller scales, due
to the multitude of possible feedbacks coincident with slight variation of just one regulating
parameter. Thus, from a biogeochemical viewpoint, patchiness or heterogeneity within
benthic environments is the norm rather than the exception.
6.4 CONCLUSION
Discerning the effects of polychaete habits on sedimentary nitrogen cycling may be a very
useful test of the functional group concept in its broad application to the general effect of
habit upon ecosystem processes. Regarding high versus low usability of the functional
group concept, whether only a few or a plethora of groups need to be recognised for any
given faunal habit, the pertinent question becomes:
Is the effect of habit robust enough to remain (relatively) constant, despite variation in
environmental setting; or does such variation result in too much multiplicity of habit
effects?
In an attempt to explore the above question, either by future research or reinterpretation of
previous studies, I am suggesting that:
(i) polychaete habits initially be assigned to one of the six bioturbative guilds
defined in Table 6.1; then,
(ii) assess the variation in polychaete habit effects within each guild; and,
(iii) relate such variation to either:
(a) finer-scale differences in habit that may comprise functional groups; and/or,
(b) differences in regulatory parameters (and their interactions) which comprise
particular environmental settings, perhaps making it harder to define
functional groups (or making them specific to defined environments).
130
It is not suggested a priori that the six preliminary guilds defined in Table 6.1 represent
practical functional groups, rather a lack of functional specificity is acknowledged. For
example, the subsurface stationary burrow habit as represented by N. diversicolor has been
demonstrated already to exert a multiplicity of effects depending on organic matter supply
and irrigation intensity). The relevant question of whether it is feasible to subdivide further
each bioturbative guild, either in recognition of alternate effects of a particular habit for
different environmental settings, and/or in relation to finer-scale definition of that habit to
represent functional groups, or such division would lead to too many groups for the concept
to be practicable, awaits exploration. Importantly, the bioturbative guilds defined above
offer a hierarchical starting point to examine the potential for structuring benthic faunal
habits with respect to their effects on ecosystem functions.
131
CHAPTER 7
GENERAL DISCUSSION: INTEGRATING BENTHIC FAUNAL
BIOLOGY WITH SEDIMENTARY NITROGEN CYCLING
In this thesis I have addressed one aspect of the regulation of nitrogen availability to primary
producers in aquatic environments: namely, the role of the benthos and its constituent
macrofauna. I have presented information concerning the life history, geographical population
structure and production of the polychaetes A. ehlersi and S. aequisetis within the Swan River
Estuary before describing the marked and varying effects of individuals from each species on
microbial denitrification and nitrogen cycling within sediments of the estuary. Further, I have
suggested an exploratory framework by which guilds of benthic fauna, each with similar
designated habits, may be tested for predictable bioturbative influence on nitrogen cycling, i.e.
whether particular habits may be considered ‘functional groups’. In this concluding discussion
chapter I integrate my results towards defining the fine-scale effects of A. ehlersi and
S. aequisetis on microbial nitrogen cycling within a broader-scale understanding of their
population dynamics. Within this context I emphasise the functional role of benthic fauna in
estuarine nitrogen cycling and the important implications for eutrophication management.
Whilst benthic fauna have long been recognised to influence the biogeochemistry of sediments
(e.g. see reviews by Petr, 1977; Aller, 1982; Krantzberg, 1985; Kristensen, 1988), acceptance of
the pervasive importance of the benthos as an actual component of sedimentary biogeochemical
cycling has been more recent and not yet fully realised. The biogeochemical effects of well-
studied individual species are becoming known (e.g. Nereis virens and Corophium volutator;
Henriksen et al., 1983; Kristensen et al., 1985, 1991; Andersen and Kristensen, 1988,
Christensen et al., 2000) and the net effect of some benthic communities on certain processes
have been determined empirically (e.g. via the use of bulk cores or in situ benthic chambers;
Hammond et al., 1985; Enoksson and Samuelsson, 1987; Watson et al., 1993; Caffrey and
Miller, 1995; Gilbert et al., 1997); however, the integral mechanisms of variation in species
effects and ecological interactions that drive observed community biogeochemistry are less well
understood. As such, prediction of spatial and temporal differences in biogeochemical cycling
due to variation in benthic species composition has been difficult to achieve. There is a great
need to investigate both inter- and intra-specific effects of benthic fauna on sedimentary
nitrogen cycling, but this task is often problematic due to a lack of detailed information
regarding the biology of individual species and/or general community ecology. In this regard,
use of the functional group concept to classify habit effects may offer a starting framework to
explore these issues (Chapter 6), but an assessment of the basic biology and ecological niche of
a species is considered requisite to developing understanding of their biogeochemical role in the
133
environment. Given the importance of sediment biogeochemistry to local and global nutrient
budgets (Barbanti et al., 1992), it is worth striving for a better understanding of the population
biology of benthic fauna and interactions with nutrient cycling.
In Table 7.1, my results describing the distribution of A. ehlersi and S. aequisetis in the Swan
River Estuary (Chapter 4) have been combined with their effects on nitrogen cycling process
and flux rates determined in vitro (Chapter 5). My intention is to illustrate the potential
magnitude of effect of the presence of each nereidid species when extrapolated and considered
at an estuary-wide scale. The scaled-up annual nitrogenous flux rates are indicative of A.
ehlersi and S. aequisetis occurring throughout their potential habitat at the experimental size and
density cited in Chapter 5 (i.e. adult A. ehlersi, 60-70 mm in length at 1,000 m-2, juvenile
S. aequisetis, 15-20 mm in total length at a density of 2,300 m-2). The total area of intertidal
flats and tidal sand banks within the Swan River Estuary is 2.46 km2 (Geoscience Australia,
2005). The estimated flux rates provided in Table 7.1 assume that this whole area comprises
potential habitat for the true euryhaline S. aequisetis, while only a portion of this (approximately
50%) potentially will be inhabited by the marine euryhaline A. ehlersi, which is restricted in its
distribution to the lower and middle estuary. Since A. ehlersi was found at 7 out of the 12 sites
examined within the estuary (Chapter 4), this is an adequate estimate of its percentage cover of
the intertidal area.
Table 7.1: In vitro nitrogen cycling processes and flux rates (tonnes N yr-1) extrapolated to an estuary-wide annual basis (see text for further explanation and assumptions).
Sediment species
treatment
Den
itrifi
catio
n at
am
bien
t* N
O3-
Max
imum
Po
tent
ial
Den
itrifi
catio
n
Nitr
ifica
tion
Net
NO
x- Flu
x*
Net
NH
4+ Flu
x
Net
Dis
solv
ed
Inor
gani
c N
itrog
en
(NO
x- + N
H4+)
Fl
ux*
Uninhabited
19 1307 561 542 -257 286
A. ehlersi
27 717 367 339 11 350
S. aequisetis
11 1156 833 822 -187 634
*At an ambient NO3- concentration of 2 µM.
It is acknowledged that the resultant flux rate estimates should not be interpreted as indicative of
conditions in situ, since experimental manipulation necessarily involves constraining
during winter (phase 3), flushes the surficial salt water from the upper estuary (Stephens and
Imberger, 1996). However, high salinity water and thus stratification will remain in the deeper
areas unless discharge exceeds 23-50 m3 s-1, which is sufficient to flush all saline water from the
upper estuary (Stephens and Imberger, 1996; Kurup et al., 1998).
Phase 1 – Salt wedge dominated ~ January to May
Phase 2a – Salt wedge waning ~ June to July
Phase 3 – Salt wedge absent ~ August to October
Phase 3 – Salt wedge propagation ~ November to December
0 5 10 15 20 25 30 35 40Distance from mouth (km)
Dep
th (m
)
Figure A1: Phases of salinity-depth contours within the Swan River Estuary throughout the
course of a typical year of freshwater discharge. (Adapted from Kurup et al., 1998; Swan River Trust, 2004a).
170
Deep waters (>5 m) of the main basin typically remain marine year-round, but are overlain by
seaward flowing fresh to brackish water in late winter to early spring, i.e. during phase 3
(Stephens and Imberger, 1996). When the depth of freshwater discharge exceeds the depth of
the Fremantle Sill (~5 m), upstream tidal movement of marine water into the main basin is
limited or nonexistent (Figure A2; Stephens and Imberger, 1996). The presence of the halocline
at a depth below that of the sill entombs the deeper high salinity water of the main basin, and
disallows marine exchange. Persistence of such stratification and lack of mixing, combined
with sedimentary biochemical oxygen demand, often causes significant deoxygenation of
bottom waters in the main basin during late winter to early spring (Figure A2; Hodgkin, 1987;
Stephens and Imberger, 1996). Similarly, locally stratified deep saline pockets of the upper
estuary may experience deoxygenation during the same period of freshwater overlay. As
freshwater discharge recedes and the depth of out-flowing water over Fremantle Sill is reduced,
richly oxygenated marine water is tidally forced into the deeper part of the main basin, and
eventually upstream as a wedge, thereby relieving anoxia (Figure A2; Stephens and Imberger,
1996). Additionally, lack of mixing within the highly stratified leading edge of the saline
wedge may also cause transitory yet significant deoxygenation of bottom waters during phase
2a (early winter) and phase 2b (late spring to early summer), as the salt wedge respectively
propagates and wanes (Figure A3).
Dep
th (m
)
Deoxygenated basin water
Surface discharge layer
Marine water
2 km
(a) (b)
(c) (d)
Figure A2: Seasonal entombment, deoxygenation and recovery of deep water within the main
basin of the Swan River Estuary. (a) Freshwater overflow in late winter to early spring at a depth greater than that of the Fremantle Sill; (b) subsequent decrease in freshwater discharge and depth of overflow allows tidal intrusion of oxygenated marine water; (c) further subsidence in freshwater discharge and resultant tidal flushing of the deep basin with oxygenated water; (d) summer-time absence of freshwater influence and upstream migration of marine water.
171
Dep
th (m
)
0 5 10 15 20 25 30 35 40
Distance from mouth (km)
(a)
(b)
Figure A3: Salinity (a) and dissolved oxygen (b) depth profiles in the Swan River Estuary during July 1997. Lack of circulation in the bottom waters of the waning salt wedge, including the upstream deeper pockets, causes marked deoxygenation. (Adapted from Swan River Trust, 2004a).
The seasonality of rainfall and resultant freshwater discharge governs both the salt wedge and
other stratification dynamics of the Swan River Estuary. Kurup et al. (1998) consider that
freshwater inflow is the most important factor affecting salt wedge position in the estuary.
Surficial freshwater flow is also the causal agent of winter/spring stratification and associated
anoxia in both the main basin and, when present, in the upper estuary. It is therefore apparent
that inter-annual variation in rainfall and freshwater discharge has profound potential to alter the
hydrodynamics of the estuary. To highlight the dependence of the hydrology of the estuary on
freshwater discharge, Kurup et al. (1998) considered salt wedge dynamics within the Swan
River Estuary during very wet (1974) and very dry (1993) years.
Subsequent to April 1974, the estuary received greater than average river inflow which persisted
until the end of the year (Kurup et al., 1998). As a result, the salt wedge had been flushed from
the upper estuary by July (i.e. early initiation of phase 3) and the estuary remained freshwater
dominated until the following summer (i.e. salt wedge emplacement, phase 2b, was inhibited
and the estuary remained in phase 3 throughout the remainder of the year). Additionally, it was
noted that the magnitude of inflow by July (>50 m3 s-1) was sufficient to flush high salinity
water from the deeper upstream pockets, thereby diminishing the potential for salinity
stratification and deoxygenation in these areas. Since Kurup et al.’s (1998) study area did not
172
include the main basin, the effect of the extended period of heightened river inflow upon
stratification in this area is uncertain. Presumably, the period of salinity stratification and
perhaps anoxia within the main basin would likewise have been extended. In a similar year of
above average rainfall (1992), Stephens and Imberger (1996) demonstrated that freshwater
discharge maintained stratification and deoxygenation within the main basin at least until early
October.
In contrast, river inflow by July 1st of the dry year 1993 was insufficient (<20 m3 s-1) to flush the
salt wedge from the upper estuary (Kurup et al., 1998). A subsequent burst in river discharge
(>100 m3 s-1) flushed the salt wedge to downstream of the Narrows by mid-September (i.e. late
initiation of phase 3, freshwater dominance); however, a resumption of low inflow resulted in
salt wedge emplacement 10 km upstream of Heirisson Island by mid-spring (i.e. earlier than
usual initiation of phase 2b, salt wedge propagation in the upper estuary). By early summer
(October) the salt wedge had migrated beyond the upstream boundary of their study, i.e. the
estuary’s confluence with Jane Brook (~50 km upstream of the mouth). As such, the Swan
River Estuary was dominated by marine conditions (phase 1) for most of this very dry year.
The effect of the low freshwater inflow upon upper estuary and main basin stratification is
uncertain. While the late winter burst of freshwater discharge was enough to flush the upper
reaches of saline water, the preceding period of low inflow may have stratified the deeper
pockets of this area and caused substantial deoxygenation. Presumably, the lack of freshwater
influence over the main basin throughout most of the year would have allowed greater than
usual tidal exchange, thereby alleviating winter stratification and deep water anoxia.
Aside from deviations in winter rainfall, large-scale unseasonal summer rainfall events can also
alter the typical hydrology of the Swan River Estuary. For example, heavy rain in early January
2000 saturated much of the Avon catchment (Atkins et al., 2001). One week later this was
followed by further atypical heavy rainfall associated with the aftermath of a northern tropical
cyclone (Atkins et al., 2001). As a result, the Swan River Estuary was flushed with enough
freshwater to fill it five times over (Atkins et al., 2001). Whilst the estuary had settled
previously into its usual marine dominated summer-time state (phase 1), by the end of January
the entire estuary had reverted to a winter-like state of freshwater dominance (phase 3), with
marked stratification of the main basin (Atkins et al., 2001). These freshwater conditions,
combined with a large runoff-associated nutrient input, triggered a record bloom of toxic blue-
green freshwater algae (see Chapter 2, Section 2.1.3). Subsequent easing of freshwater
discharge allowed the tidal re-entrance of marine water and upstream propagation of the salt
wedge, such that the system had reverted back to its typical marine dominated state by the
beginning of March.
173
The influence of seasonality upon the hydrology of the Swan River Estuary also extends to
water temperature. Within the Mediterranean climate, average temperature of surface water
(<5 m deep) overlying the main basin typically ranges from 12-14 °C in winter to 22-24 °C in
summer (Figure A4; Hodgkin, 1987). The annual range in average water temperature of the
shallow upper estuary is greater, from 10-13 °C in winter to 28-30 °C in summer (Figure A4;
Hodgkin, 1987). Very shallow areas of the estuary (<2 m deep; typically where benthic fauna is
most abundant), such as nearshore and at several upstream locations, may experience even
greater seasonal variation in water temperature (Hodgkin, 1987). The upper estuary and
shallower areas are also subject to diurnal fluctuations in temperature of 4-5 °C (Kurup et al.,
1998). Deeper water of the main basin is less variant, seldom falling below 15 °C in winter or
rising above 22 °C in summer (Figure A4; Hodgkin, 1987). Haloclines generated by the
trapping of high salinity water under cold freshwater winter flow (i.e. in the deep basin, deep
pockets of the upper estuary and at the leading/waning edge of the salt wedge), are also subject
to reversed thermoclines during the period of freshwater influence. The difference in salinity,
and thus density, maintains the warmer marine water at depth (Figure A4). Salinity, not
temperature, overwhelmingly drives the density variations within the Swan River Estuary
(Kurup et al., 1998).
Summer
Winter
Temperature(ºC)
Figure A4: Typical temperature-depth profiles of the Swan River Estuary in winter (July) and summer (December). Adapted from Swan River Trust (2004a).
174
APPENDIX B
ALGAL BLOOM SUCCESSION IN THE SWAN RIVER ESTUARY
The annual succession of population growth of various algal species within the Swan River
Estuary has been an on-going topic of relevance to the residents of Perth since settlement.
Hodgkin and Vicker (1987) detail the history of algal blooms in the Swan River Estuary, which
is briefly summarised below prior to a consideration of the present day annual succession of
bloom-forming phytoplankton species.
Pre-dating the problems associated with blooming phytoplankton species in recent history, the
Swan River Estuary was subject to recurrent, large populations of macroalgae. The following is
recounted from Hodgkin and Vicker (1987). From at least the 1870s, shore-line accumulation
and decay of dense mats of macroalgae were a major source of complaint by Perth residents.
During spring and summer, green algae (e.g. Cladophora, Chaetomorpha and Rhizoclonium
spp.) and filamentous diatoms (e.g. Melosira spp.) grew rapidly in the warm, calm and shallow
waters of the main basin and Perth Water. Subsequent detachment and drift allowed
accumulation and then degradation upon the estuary’s beaches and sand banks. Newspaper
articles of the 1870s record of swans avoiding “large masses of algae floating near the
causeway” and a “large collection of seaweed in the river near Mill Street Jetty”. At the turn of
the century, “rotting weed was being collected and carted away from beaches as far away as
Claremont” (some distance from the main settlement for the period). In 1913 it was recorded
that “some half mile of foreshore is covered with filthy green vegetable growth”. Not only did
these accumulations provide a blanket cover of the estuary’s beaches and a barrier to
recreational enjoyment, their ensuing decay produced hydrogen sulphide and other malodorous
gases that were the major cause of complaint. The stench together with the associated flies,
maggots and other detritivores were also perceived to be a public health threat. Up to 400
tonnes of seaweed was manually removed from the estuary per year, until the 1960s.
Reclamation of sheltered areas and an increase in river flow (resultant from channel dredging
and catchment changes) are most likely the causes of a subsequent and significant decrease in
annual macroalgal accumulations. Some collection of algal weed still occurs, but in contrast to
the magnitude of the problem prior to the 1970s, the accumulations are small and only of minor
public nuisance. In summation, large algal accumulations within the Swan River Estuary pre-
date its current eutrophic status and perhaps even European settlement.
It is the increasingly eutrophic status of the estuary and shift from productive macroalgal
species towards both toxic and non-toxic phytoplankton species, together with the consequences
of an intensification of frequency and severity of blooms, which is presently driving
environmental and public concern. The modern annual succession pattern of phytoplankton
within the estuary is described below, in addition to the conditions necessary for bloom-forming
species.
175
Twomey and John (2001: p. 2656) state that “inter-annual variation of phytoplankton succession
can be large, but in general the patterns are broadly repeated from one year to the next”. The
‘normal’ succession and blooming of phytoplankton species within the Swan River Estuary is
defined for those years of typical rainfall, freshwater discharge and salt-wedge migration (see
Chapter 2, Section 2.1.2), and is as follows. When freshwater discharge is high during winter,
the upper estuary is dominated by dislodged epiphytic and periphytic diatoms (e.g. Synedra
Kanandjembo, 1998). In this Appendix I present a brief review of these studies and the current
status of knowledge pertaining to benthic macrofaunal community composition and trophic
relations in the Swan River Estuary.
A review conducted by Rose (1994) identified that 134 benthic macrofaunal species have been
noted in the Swan River Estuary, including both continuous and temporary inhabitants. At a
high taxonomic level and similar to other estuaries, the Polychaeta, Mollusca and Crustacea
together comprise the vast majority of species and individuals. T he best-studied regions are the
shallows (overlying water <2 m deep) of the main basin and upper estuary wherein polychaetes
of the families Nereididae, Spionidiae, Capitellidae, Sabellidae and Orbiniidae contribute 46-
47% of the total abundance (Rose, 1994; Kanandjembo et al., 2001). Molluscan bivalves and
gastropods together contribute 14-43% towards abundance, while amphipod, decapod and
isopod crustaceans contribute 7-39%, depending on the region (Rose, 1994; Kanandjembo et
al., 2001).
Regarding mollusc species, Chalmer et al. (1976) document that the community of the lower
estuary is dominated by species of marine affinity, both temporary (64 stenohaline species) and
continuous (25 marine euryhaline species). They argue that the high molluscan diversity (89
species) demonstrates a distinctiveness of the lower estuary and gives weight to the three-part
biotopic subdivision of the Swan River Estuary (see Section 2.2.1). The group is characterised
by great attenuation above the lower estuary, with only 6 of the marine euryhaline species
capable of living continuously in the middle estuary and only two species having made
‘temporary incursions’ into the upper estuary (during sustained periods of low river discharge)
(Chalmer et al., 1976; Hodgkin, 1987). The molluscan fauna of the middle and upper estuary is
generally characterised by 7 true euryhaline species (4 of which may extend into the lower
estuary during years of strong winter floods). At any one time, abundance and diversity of these
true euryhaline species typically decreases in the upper reaches (Chalmer et al., 1976; Hodgkin,
1987). The only true oligohaline mollusc is a freshwater gastropod, which is restricted to the
upper estuary. Stenohaline and marine euryhaline species from other taxonomic groups
181
(including the Polychaeta) have not been so well studied, but similar broad patterns of
distribution are expected for the lower and middle estuary. For example, Chalmer et al. (1976)
further documented that two polychaete species (A. ehlersi and Diopatra sp.) and two
crustaceans (the crab Halicarcinus bedfordi and barnacle Balanus sp.) of marine affinity, each
resided in the lower estuary but at few or no locations in the middle and upper estuary.
The distinction between the middle and upper estuary may be further elucidated by comparing
the macrobenthic community studies of Rose (1994) and Kanandjembo (1998) (Table C1).
Succeeding thorough sampling of shallow water sites (<2 m deep) over a comparable period, the
former found 53 macrobenthic species within the middle estuary, while the latter revealed only
37 species from the upper estuary. During a brief sampling period in autumn, Wallace (1977)
similarly found a greater species richness in the middle estuary than the upper estuary, both for
shallow and deep waters (depth >2 m). Kanandjembo (1998) suggested that the lower species
richness of the upper estuary was due to a more variable, extreme and sometimes very low
salinity environment, which fewer euryhaline species were able to tolerate. For example, two
euryhaline crustaceans (the amphipod Grandidierella sp. and the isopod Tanais dulongii) were
abundant in the middle estuary but not present in the upper estuary (Table C1). In general, the
abundance of small crustaceans was much higher in the middle than upper estuary.
Nevertheless, some species were characterised as upper estuary specialists (e.g. the spionid
polychaetes Pseudopolydora sp. and Desdemona ornata, as well as oligochaete and chironomid
species), that were somewhat abundant in the upper estuary but not present in the middle
estuary (Table C1). Further, the abundance of the euryhaline bivalves Arthritica semen and
Xenostrobis securis was much greater in the upper than middle estuary. In summary, while
many euryhaline species are capable of inhabiting both the middle and upper estuary, other
species are less tolerant of the lower and more variable salinities of the upper estuary, leading to
a lower species richness. Oligohaline species and some euryhaline species have a lower salinity
preference, residing exclusively or in greater abundance in the upper estuary, which results in an
appreciable contrast in community composition between the upper and middle estuary.
While many species may be capable of inhabiting a certain region of the estuary, only a few
species will typically dominate the benthic community, both in terms of total abundance and
biomass (Table C1). For example, in rank order the four most abundant species within the
shallows of both the middle estuary (S. aequisetis, Grandidierella sp., Capitella ‘capitata’ and
T. dulongi) and upper estuary (A. semen, Pseudopolydora sp., S. aequisetis and X. securis),
together comprise approximately 70% of total abundance in each locale (Rose, 1994;
Kanandjembo et al., 2001). Further, the ten most abundant species together encompass greater
than 90% of total abundance in each area (Table C1). It is pertinent that abundance is only one
of many population descriptors that provide information about the dominance or role of
182
community constituents. For instance, within shallows of the middle estuary the constituency of
the top four ranked species in terms of biomass dominance is much different to those species
dominating abundance. That is, while polychaetes and crustaceans dominate abundance (see
above), 70% of total biomass is dominated by three bivalve molluscs (Sanguinolaria biradiata,
X. securis and Musculista senhousia) (Table C1; Rose, 1994). Thus, dominance is a subjective
term unless the context of consideration is clearly defined, i.e. dominance of taxa in terms of
abundance, biomass, productivity or other functional role in ecology (e.g. trophic effects and
community structuring) or biogeochemical effects (e.g. upon sediment stability, reworking,
particle size distribution or nutrient cycling).
Table C1: The rank dominance of the most common macrobenthic inhabitants of the upper and middle Swan River Estuary, either in terms of percentage contribution to total abundance [A & B] or biomass [C]. Adapted from Rose (1994) and Kanandjembo et al. (2001).
17 Tatea preissi M, G 0.4 Australonereis ehlersi ##
P, N 0.2
18 Chironomid sp. 1 #
I 0.4 Heteronemertean sp. ##
N 0.2
183
Table C1: (continued)
[A] Upper Estuary [B] Middle Estuary
Ran
k
Spp Group*
% o
f tot
al
abun
danc
e
Spp Group*
% o
f tot
al
abun
danc
e
19 Arachnid sp. #
A 0.2 Irus crenata
M, B 0.2
20 Amarinus laevis #
C, D 0.2 Ficopmatus enigmaticus ##
P, P 0.2
21 Nassarius burchardi
M, G 0.2 Ericthonius sp.
C, A 0.1
22 Melita matilda
C, A 0.1 Halicarcinus bedfordi
C, D 0.1
23 Velacumanthus australia
M, G 0.1 Tellina deltoidalis
M, B 0.1
24 Chironomid sp. 2 #
I 0.1 Munna brevicornis ##
C, I 0.1
25 Marphysa sanguinea
P, E 0.1 Tatea preissi
M, G 0.1
26 Tellina deltoidalis
M, B <0.1 Nodilittorina unifasciata ##
M, G 0.1
# Not found within the middle estuary. ## Not found within the upper estuary. * 1st letter: P = Polychaeta, O = Oligochaeta, M = Mollusca, C = Crustacea, I = Insecta, A = Arachnida, N = Nemertean.
2nd letter: for polychaetes, S = Spionidae, P = Serpulidae, N = Neredidae, O = Orbiniidae, E = Eunicidae; for molluscs, B = Bivalvia, G = Gastropoda; for crustaceans, A = Amphipoda, D = Decapoda I = Isopoda.
Geographic variation in salinity regime has above been argued to be the main delineator of the
major biotic zones within the estuary. Similarly and perhaps most importantly, it is seasonal and
inter-annual variations in the dynamics of the characteristic salt-wedge of the estuary (driven by
variation in rain and freshwater discharge, Section 2.1.2 and Appendix A), that largely govern
spatio-temporal variability in community structure. Firstly on a seasonal basis, the boundaries of
the ranges of the species, as defined by salinity tolerance, within the lower, middle and upper
estuary may be expected to shift upstream during summer and autumn, concomitant with upstream
emplacement of higher salinity water, i.e. the salt-wedge (Phase 1). Conversely, the distribution of
species may be expected to shift downstream during late winter to early spring, as freshwater
discharge increases and the salt-wedge wanes (Phase 3). Thus, the theoretical upper and lower
boundary of each species’ range as determined by salinity tolerance, will seasonally fluctuate
concomitant with progression of the salt-wedge. From year to year, the absolute magnitude of these
upstream and downstream shifts in biotic zones will be dependant upon inter-annual variation in
rainfall and freshwater discharge which ultimately governs the progression of the salt wedge.
These boundaries are based hypothetically only on the influence of salinity. Within these bounds,
many other regulatory factors such as faunal dispersal capability and habitat/substrate suitability
and availability will regulate the true expression and dynamics of a macrobenthic species
distribution. Salinity may primarily govern the broad distribution and composition of the biota
184
within the Swan River Estuary (Hodgkin, 1987); however, at a finer scale, a multitude of
physicochemical and biological factors, and their interactions, will regulate where particular species
live and how abundant or productive they are. Some of the major determinants in fine-scale
patterns of species distribution and resultant community composition are briefly discussed below.
Fauna of sediments that underlie shallow waters have been more comprehensively studied than
deeper benthic communities within the Swan River Estuary, likely due to differences in the ease
and accessibility of sampling. However, water depth and correlating parameters (such as light
availability, primary productivity and temperature, for example) have been documented to
largely influence community composition and abundance of constituent fauna at a local scale.
Those studies that have sampled deeper sediments have found that both species richness and
abundance are generally greater in shallow than deeper waters. For the upper estuary,
Kanandjembo et al. (2001) established that the difference in composition was mainly due to the
greater presence and abundance of the nereidid S. aequisetis and bivalve X. securis in shallow
waters. The brief sampling regime of Wallace (1977) also found that species densities were
generally greater in the shallows than in deeper water of both the upper and middle estuary.
Kanandjembo et al. (2001) offer the explanation that the aforementioned shallow-water species
either: (i) prefer and are proliferating on an increased abundance of photosynthetic food material
in shallower water, or (ii) coarser sediments in the shallows buffer against hypoxia which is a
more conducive environment for burrowing invertebrates, or (iii) a combination of both these
reasons results in the observed discrepancy between shallow and deep water environments.
Contrastingly, the spionid polychaete Prionospio cirrifera and sabellid polychaete Desdemona
ornata were more abundant in deeper than shallow water; it was argued that these species more
exclusively feed on detrital material that settles out in deep waters (Kanandjembo et al., 2001).
Manipulative studies are required to test these hypotheses and determine which factor, or
combination of factors, are responsible for the differences in community composition with
depth.
The observation that regulatory factors likely do not act unilaterally, but rather synergistically, is
pertinent to an understanding of population and community dynamics. For example, the seasonal
ebbing and waning of the salt-wedge will be accompanied by changes not only in salinity but also
in temperature, dissolved oxygen, nutrients, organic matter supply, water flow rates and
stratification events (Section 2.1.2 and Appendix A). To reiterate, while salinity may determine the
theoretical distribution of a species, it will be variations in these other factors that result in the
tangible expression of population structure and community composition. An eloquent example to
note is again provided by the study of Kanandjembo et al. (2001), wherein community composition
at any given time did not differ significantly between sites of the upper estuary; however,
community composition did vary significantly between seasons of a year and also between seasons
185
of successive years. An argument may be invoked that similar salinities in the upper reaches at any
one time resulted in comparable community composition and that seasonal shifts in salinity
likewise imparted change on the distribution of constituent species. Yet, this does not explain the
variance in community composition from year to year, since salinities between seasons of
successive years were reasonably similar. Kananjembo et al. (2001) argue that while salinity
regime was largely responsible for the seasonal, and perhaps some of the inter-annual, variation in
community composition, water flow rates associated with discharge events also affected species
richness and densities. It was proposed that extremely high rates of discharge during one of the
winters caused much scouring of substrates, particularly in the shallows, which resulted in
dislodgement of fauna and both depressed species richness and overall abundance. Somewhat
lower flow rates in other winters, while resulting in similarly low salinities in the upper estuary, did
not appear to cause this scouring effect to shallow substrates; wherein species richness and
abundance remained comparatively high. The recovery phase from a disturbance scouring event
led to community composition of subsequent seasons being appreciably different to seasons of the
previous, non-impacted year.
Finally, although previous examples have placed emphasis on physicochemical regulators of
species distribution and community composition, biological determinants may be just as
important (Day et al., 1989). To some extent, biological influences on community dynamics –
e.g. competition, predation, other trophic relations, dispersal/reproductive mechanisms,
intraspecific interactions and biogeographical constraints etc. – are often overlooked, since they
are usually not as easily quantified as physicochemical determinants that only require a quick
measurement or sample (e.g. salinity, dissolved oxygen, substrate type etc.). As such, empirical
evidence of biological controls on the distribution of species within the Swan River Estuary are
largely lacking. Nevertheless, certain hypotheses have been proposed. As alluded to above,
food resource availability such as plant material in shallow waters may be an important
determinant of species presence and abundance (Kanandjembo et al., 2001). Hodgkin and
Lenanton (1981) suggested that the most important item in the food chain of the Swan River
Estuary is detritus. Many benthic invertebrates are detrital feeders and the distribution of
detritus, in terms of both abundance and quality, may influence species distributions. Further
many resident fish species are either detritivores or feed on benthic macroinvertebrates
(Hodgkin and Lenanton, 1981). As such, fish predation may exert considerable influence on the
abundance and even presence of sedimentary fauna. Kanandjembo (1998) studied the diet of
four fish species (Acanthopagrus butcheri, Amniataba caudavittata, Pseudogobius olorum and
Leptatherina wallacei) and found that benthic macrofauna were a component of each, but that
the composition of macroinvertebrates was different for each fish species. Thus, the
composition of the fish community may in turn regulate benthic community composition.
186
As a final example of biological and synergistic effects on the distribution and abundance of
macrobenthic fauna, it is salient to reconsider the potential influence of algal blooms. In
Section 2.1.3 I observed that the decomposition of collapsed algal blooms may result in
localised anoxia or toxic effects, which have profound implications for the mortality and/or
abundance of both relatively sessile macrofauna and mobile fish species. Of course, nutrient
availability may ultimately regulate the extent and severity of algal blooms and the resultant
potential impacts on the macrobenthos. This illustrative, if not simplistic, example of the effects
of eutrophication and the interaction between nutrients, algal blooms, dissolved oxygen
concentration and toxic effects on both benthic fauna and fish (a potential source of predation),
clearly highlights the type of synergistic effects that may govern macrobenthic faunal
community composition. The contention of this thesis that macrobenthic fauna may themselves
regulate nutrient availability, conceptually completes a circular linkage whereby the effects of
macrofauna may govern factors which influence their own distribution.
187
The reaction is catalysed by nitrogenase, a complex molybdo-iron enzyme (Capone, 1988).
Nitrogen fixation is energetically expensive requiring about 15 adenosin triphosphate molecules
(ATP) per N2 reduced, with the metabolic cost being met by ATP production from
photosynthesis or chemosynthesis (Capone, 1988; Heathwaite, 1993).
189
APPENDIX D
COMPONENT PROCESSES OF THE NITROGEN CYCLE
D.1 OVERVIEW
It is not the intent of this Appendix to comprehensively review nitrogen cycling in aquatic
environments, and the reader is referred to Brezonik (1972), Keeney (1973), Carpenter and
Azospirillum spp., Campylobacter sp. and Beggiatoa spp.), facultative anaerobes (e.g.
Enterobacter spp., Klebsiella spp. and Vibrio spp.) and strict anaerobes (e.g. Desulfovibrio spp.
and Clostridium spp.) (Capone, 1988).
190
NO3-
NH4+
Organic N
NO2-
N2O N2Organic N
NH4+ NO3
- NO2-
N2O N2
NH4+
NO3-
NO2-
Organic N
Export
NH4+
NO3-
NO2-
Organic N
Import
NH4+, NO3
-, NO2-
Groundwater Import
Ammonification / Mineralisation
Nitrification
Denitrification
Sediment / Water Column / Atmosphere Solute Flux
ATMOSPHERE
WATER COLUMN
SEDIMENT
Nitrogen Fixation
Ammonium Assimilation
Assimilatory Nitrate Reduction
Dissilmilatory Nitrate
Reduction to Ammonium
LEGEND
Figure D1: The estuarine nitrogen cycle. A conceptualisation of the main transformations occurring in the sediment and water column. For simplicity many intermediates have been omitted. See text for further details.
Perhaps of most importance to estuarine systems are the bloom-forming nitrogen fixing
cyanobacteria. In the absence of available (dissolved inorganic) nitrogen, many species are
capable of exploiting a phosphorous nutrient source (together with an ambient dinitrogen gas
supply) to form dense microbial or algal mats either in the plankton (some species have a gas
vacuole buoyancy system) or on the sediment bed (Capone, 1988; Heathwaite, 1993). Some of
the most notorious (i.e. toxin-producing, anoxia-generating) nuisance algal bloom species are
within an aquatic environment, unlike the gas-liberating process of denitrification (Kuenen and
Robertson, 1988). For unclear reasons, ammonium conservation by DNRA has historically
been thought quantitatively unimportant (or even worse, sometimes ignored) in the nitrogen
budget of natural systems, and has therefore traditionally received little scientific attention
(Cole, 1990). Whilst the presumption may be justified in many environments, there is evidence
that DNRA plays a significant role in the nitrogen cycling of numerous systems. DNRA is
generally an anaerobic process that competes with denitrification for nitrate/nitrite as a substrate
(Cole, 1988). Since many denitrifying bacteria are facultative anaerobes that use nitrate only
when oxygen is not available (see Section D.7), it appears that denitrification is the dominant
process in environments that favour a respiratory mode of life (rich in nitrate but relatively poor
in organic carbon), whilst DNRA may predominate in more highly anaerobic carbon-rich
systems favouring colonisation by fermentative bacteria (Cole, 1988, 1990; Capone, 2002). In
support of this notion, Cole (1990) cites an early study by Sørensen (1978b) (amongst others)
who demonstrated that the denitrification capacity of a coastal marine sediment decreased
rapidly with depth, but the potential for DNRA in deeper anaerobic sediment was better
maintained. Further, ammonium was the dominant product of nitrate reduction in a highly
anaerobic digested sludge (Kaspar and Tiedje, 1981; Cole, 1988). Favourable conditions for
DNRA are generally not found in natural soils (typically 1 to 5% of total nitrate reduction is due
to dissimilation to ammonium); however, exceptions exist and DNRA was found recently to
account for 15% of total nitrate reduction in an organically-rich anaerobic Australian paddy soil
(Cole, 1990; Yin et al., 2002). Whilst understudied, DNRA in freshwater systems has been
found important when nitrate is available and may account for a moderate proportion of total
nitrate reduction (approximately 10% in one system) (Jones et al., 1982; Cole, 1990). Tobias et
al. (2001) similarly found that in a fringing marsh, DNRA was quantitatively more important
when nitrate supply was high during freshwater discharge in spring.
Within marine and estuarine sediments, the proportion of total nitrate reduced to ammonium by
dissimilatory pathways has often been assumed negligible, but under appropriate circumstances
it may in fact be considerable. For example, proportions of 35 to 42% (Buresh and Patrick Jr.,
1981), 32% (Goeyens et al., 1987) and 4% to 21% (Jørgensen, 1989) have been cited. In a
recent study of a shallow estuary in Texas, An and Gardner (2002) found that DNRA accounted
202
for between 15 and 75% of all radio-labelled nitrate reduced in the sediment, while the
equivalent range for denitrification was only between 5% and 29%. In this case, the authors
argued that high sulphide concentrations inhibited nitrification and denitrification but enhanced
DNRA by providing an electron donor. Consequentially, ammonium was recycled efficiently
within the system and was inferred to allow the persistence of a long-lasting algal bloom.
However, as Buresh and Patrick Jr. (1981) have pointed out, care should be taken when
extrapolating results from laboratory-based results, especially where nitrate has been added
and/or sediments have been (perhaps artificially) anaerobically incubated under an inert gas
atmosphere. Whilst intensely-anaerobic sediment conditions do exist in nature, it is unclear if
much nitrate is ever exposed to such environments (Buresh and Patrick Jr., 1981). When more
natural conditions were considered in the study of Buresh and Patrick Jr. (1981), whereby
nitrate moved downward from the overlying water into the sediment, the estimate of total nitrate
dissimilatorily reduced to ammonium decreased from 35-42 to 15%. Nitrate supplied either
from nitrification or the overlying water and diffusing into anaerobic zones is quickly reduced
(Buresh and Patrick Jr., 1981). The sequential diffusion of nitrate is such that anaerobic zones
of denitrification may often be more closely related (spatially and/or temporally) to areas of
nitrate supply than the more highly anaerobic (e.g. deeper) fermentative DNRA zones. It is very
likely that this type of reasoning, sometimes justifiably, has lead to ignorance of the DNRA
process.
Nonetheless, scenarios exist whereby highly anaerobic organic-rich sediments are supplied with
substantial quantities of nitrate such that DNRA may be prolific (for example where nitrate-
laden groundwater forces a substrate supply from below or in situations where denitrification is
inhibited, by sulphide for instance). The recent identification of nitrate accumulating sulphur
bacteria that are capable of DNRA in microaerobic coastal marine and continental shelf
sediments provides an interesting example of the potential importance of the DNRA process
(Sayama, 2001; Zopfi, 2001). Three closely related genera (Thioplaca, Beggiatoa and
Thiomargarita) have been isolated from coastal upwelling zones off Peru, Chile, Namibia and
in Tokyo Bay (Sayama, 2001; Zopfi, 2001). Each species stores nitrate intracellularly in large
vacuoles at high concentrations (40 to 800 mM; Sayama, 2001). Therefore, whilst a large
nitrate pool exists in these sediments it is neither a pore-water constituent nor available for
denitrification (Sayama, 2001). Rather, these sulphur bacteria dissimilatorily reduce the stored
nitrate to ammonium which is released to the environment, potentially exacerbating
eutrophication (Sayama, 2001). The microbiology, ecology and general quantitative importance
of DNRA bacteria are still not well understood (Yin et al., 2002). However, the consequences of
DNRA, including the potential for nitrogen conservation within aquatic environments, are such
that the process warrants detailed study and at the very least deserves consideration in any
nitrogen cycling budget.
203
D.7 DENITRIFICATION
The potential importance of denitrification in removing inorganic nitrogen from eutrophic
systems is of major consequence to this thesis. An explanation of the bacteria and inherent
enzymes that mediate and define denitrification, together with regulatory factors controlling this
process and typical ambient denitrification rates will therefore be discussed in some detail.
Thereafter, studies examining the kinetics of denitrification will be reviewed to place my thesis’
kinetic consideration of denitrification (Chapter 5) within the context of prior research.
D.7.1 The Bacterial and Enzymatic Process of Denitrification
Denitrification may be defined as the reduction of soluble nitrogen oxides, nitrate (NO3- ) and
nitrite (NO2-), to the gases nitric oxide (NO), nitrous oxide (N2O) and dinitrogen (N2), by
bacteria using the pathway as a respiratory alternative to oxygen (Kuenen and Robertson, 1988;
Groffman et al., 1999) (Figure D1):
NO3- NO2
- NO N2O N2
The major implication of denitrification in aquatic systems is that dissolved inorganic nitrogen
(DIN) atoms (within nitrate and nitrite) are removed from the ecosystem as they are returned to
the atmosphere as nitrogenous gases (nitric oxide, nitrous oxide and dinitrogen). Denitrification
therefore has great potential to reduce the availability of DIN to primary producers. Both
facultative anaerobes and strictly anaerobic bacteria from a wide range of taxa, including
Pseudomonas, Bacillus, Thiobacillus, Propionibacterium amongst others, are capable of
denitrification (Brezonik, 1972; Hattori, 1983; Wrage et al., 2001). Denitrifiers are
predominantly heterotrophic, facultative anaerobes respiring nitrate (instead of oxygen) as an
electron acceptor in anoxic or low-oxygen conditions, and the terminal product is usually
dinitrogen gas (Tiedje, 1988; Wrage et al., 2001). However, chemoautotrophic and
photosynthetic denitrifying bacteria exist, and whilst in all groups the end product of
denitrification must by definition be gaseous, not all reductive steps may occur (Knowles,
1982). The maximum possible reactions of denitrification by certain bacteria are given in Table
D2. For a more comprehensive consideration of denitrifying microflora the reader is referred to
Payne (1973, 1981), Jeter and Ingraham (1981), Knowles (1982) and Kuenen and Robertson
(1988). Recently, some autotrophic and heterotrophic nitrifying bacteria have been
demonstrated to be capable of denitrification in oxygen-limited or even aerobic environments
(Jetten et al., 1997). The consequences of potential denitrification by these organisms may be
great; however, the process of denitrification by nitrifying bacteria will not be considered
further here, and the reader is referred to Section D.8. This section pertains only to conventional
anaerobic denitrification. Further, although nitrate and nitrite may be reduced photochemically,
in water these processes are restricted to a very thin layer at the surface (Hamilton, 1964). In
204
aquatic ecosystems only biological processes are of quantitative importance in the
transformation of nitrate (Hattori, 1983; Tiedje, 1988).
Table D2: Examples of bacterial genera which exhibit various reactions of nitrate respiration and denitrification. (Adapted from: Knowles, 1982; Kuenen and Robertson, 1988)
Like all other biological components of the nitrogen cycle, denitrification is an enzyme driven
process and each of the four inherent steps is catalysed by a specific reductase: NO3- to NO2
- by
nitrate reductase (NAR); NO2- to NO by nitrite reductase (NIR); NO to N2O by nitric oxide
reductase (NOR); and, N2O to N2 by nitrous oxide reductase (NOS) (Jetten et al., 1997; Bothe et
al., 2000; Moura and Moura, 2001). The reactions involved in each respective step are detailed
below (Bothe et al., 2000):
NO3- + 2 H+ + 2 e- NO2
- + H2O
NO2- + 2 H+ + e- NO + H2O
NO + 2 H+ + 2 e- N2O + H2O
N2O + 2 H+ + 2 e- N2 + H2O
As was noted in Section D.6 above, the first step of nitrate reduction to nitrite may be common
to the DNRA process and may generally be termed nitrate respiration (Table D2), but further
reduction to a gaseous end-product must occur for the process to be deemed denitrification.
Bacterial genes encoding at least four different classes of NAR have been described (Bothe et
al., 2000; Moura and Moura, 2001). Denitrifying bacteria lacking NAR but capable of reducing
nitrite are defined as nitrite-dependant (Table D2; Knowles, 1982). Two types of NIR have
205
been found in denitrifiers and for unclear reasons their presence in individual bacteria is
mutually exclusive (approximately ⅔ contain a heme-cd1 NIR whilst ⅓ contain a copper NIR)
(Jetten et al., 1997; Bothe et al., 2000; Moura and Moura, 2001). The production of nitric oxide
as a free intermediate during denitrification has historically been questioned. However, there is
now strong evidence that NO is a kinetically competent intermediate, with its steady-state
concentration ranging between 1 to 65 nM, depending on bacterial species (Jetten et al., 1997).
There is evidence of two genes encoding nitric oxide reductase (Bothe et al., 2000) but the
structure of NOR has yet to be confirmed (Moura and Moura, 2001). One type of nitrous oxide
reductase is currently defined and its structure has recently been characterised (Bothe et al.,
2000; Moura and Moura, 2001). Denitrifiers lacking NOS have nitrous oxide as their terminal
product and have been dubbed nitrous oxide producers (Table D2; Knowles, 1982). On the
other hand, certain bacteria (nitrous oxide reducers) can not produce N2O from nitrate or nitrite
(and technically are not denitrifiers), but contain NOS and therefore are capable of producing N2
(Table D2; Knowles, 1982).
An increasing build-up of both nitric oxide and nitrous oxide within the atmosphere, potentially
related to anthropogenic impacts on the global nitrogen cycle, has implications for ozone
depletion and greenhouse warming (see Section D.8; Colliver and Stephenson, 2000; Bonin et
al., 2002). Dinitrogen gas may not be the ultimate product of denitrification, and release of the
gaseous intermediates of NO and N2O may occur for a variety of reasons. Firstly, the genetic
lack of nitric oxide reductase (NOR) or nitrous oxide reductase (NOS) in certain denitrifying
bacteria physically prevents conversion to dinitrogen gas and the intermediates are in fact end-
products of the denitrification process (see Table D2). Within the majority of bacteria that
possess the full enzymatic apparatus to produce dinitrogen gas, sudden changes in the
environment may result in termination of the denitrification process at an intermediate stage.
Jetten (2001) notes that in response to environmental perturbations, bacterial regulation of the
various enzymes involved in denitrification can be immediate, and that the transition between
oxic and anoxic conditions often results in the release of intermediate gases. A limitation of
electron donors (e.g. organic carbon) or the presence of toxic and/or inhibitory compounds (e.g.
sulphide; Sørensen, 1980) may also stimulate the emission of gaseous intermediates (Jetten,
2001).
D.7.2 Factors Regulating Denitrification
Denitrification is dependent primarily upon availability of nitrate, anoxic conditions and the
presence of an energy (electron) source (Kristensen, 1988). Indeed, the supply of nitrate, the
rate of oxygen supply and consumption, and the presence of organic matter (carbon) as an
electron source are the chief regulating factors of denitrification at a particular habitat microsite
(Tiedje, 1988). These major regulatory factors are discussed below. For simplicity,
206
denitrification is preposed to occur across a flat two-dimensional sediment-water interface,
described as ambient sediment, without the presence of bioturbating macrofauna. The influence
of bioturbation and the creation of burrow micro-environments considerably alters the potential
for denitrification in comparison with ambient sediment (e.g. Aller and Yingst, 1978; Henriksen
et al., 1983; Kristensen et al., 1985; Aller and Aller, 1986; Aller, 1988; Reichardt, 1988;
Kristensen et al., 1991; Svensson, 1997; De Roach et al., 2002). The impact of bioturbation,
creation of burrows and ventilatory activity by benthic infauna in regulating denitrification rates
is the subject of Chapter 5.
To a great extent, rates of denitrification appear to be limited by nitrate availability in estuarine
and coastal marine sediments. Diffusional nitrate supply from the overlying water and from
sediment nitrification, may largely regulate denitrification in these systems (Kristensen, 1984;
Henriksen and Kemp, 1988). Diffusional processes are of great importance in exchange across
the sediment-water interface and the amount of nitrate in overlying water is essential in
determining both the rate and direction of nitrate exchange (Kristensen, 1984). Kristensen
(1984) demonstrated that an equilibrial (no net flux across the sediment-water interface)
concentration of nitrate existed in overlying waters with 10-15 μM of nitrate. Nitrate diffused
into the sediment (for subsequent denitrification) at overlying concentrations above equilibrium,
whereas nitrate was released from the sediment (due to nitrification) at overlying concentrations
below equilibrium (Kristensen, 1984). The equilibrial concentration will vary for differing
sediment environments, but nitrate concentrations in overlying water will obviously affect the
diffusional rate into sediment and considerably regulate the degree (and even presence) of
denitrification activity.
In situ nitrification in sediments may also considerably regulate denitrification activity (Jenkins
and Kemp, 1984; Henriksen and Kemp, 1988) and the process has been discussed previously in
Section D.5. Generally, nitrification provides the denitrification process with a substrate
(nitrate) and may also provide denitrifiers with an anoxic environment as the aerobic nitrifiers
consume oxygen (Heathwaite, 1993). The relationship between supply of nitrate from
nitrification and consumption of those nitrates by denitrification is by no means simple. This is
primarily because the two processes require opposite redox conditions; nitrification and
denitrification are characteristically oxic and anoxic processes, respectively (Henriksen and
Kemp, 1988). Several models and quantitative studies have demonstrated that in ambient
sediment the ‘available nitrate zone’ typically is 1 to 5 cm deep (depending upon oxygen
regulatory factors, i.e. supply and consumption); and the maximum nitrate concentration is in
the uppermost 0.5 to 1 cm due to nitrification in the oxic zone, with decreasing concentration
below 1cm due to denitrification in the anoxic zone (Vanderborght and Billen, 1975;
Vanderborght et al., 1977; Kristensen, 1988; Henriksen and Kemp, 1988). The nitrate produced
207
by nitrification in the oxic layers will pass either to the overlying water or enter the anoxic
sediment where it is reduced to nitrogenous gas by denitrification (Kristensen, 1988). Evidence
suggests that not all reduced nitrate resulting from nitrification is accounted for by
denitrification; the remainder is reduced to ammonium by dissimilatory pathways (Section D.6).
It is also found that temporal rather than spatial separation of nitrification and denitrification is
an effective means of relating the two processes (Henriksen and Kemp, 1988). For example,
nitrification may be a more important process in the daytime when primary producers (e.g.
microphytobenthos) are releasing oxygen for the aerobic respiration of nitrifiers, whilst
denitrification may occur at a greater rate at night under lower oxygen concentrations.
Generally, denitrification takes place at aerobic-anaerobic interfaces (Bonin and Raymond,
1990). However, there is on-going debate in the literature concerning the lowest oxygen
concentration that permits denitrification (Bonin and Raymond, 1990). Much early confusion
probably related to ignorance of aerobic nitrifier-denitrification processes (see Section D.8).
Conceptual separation of conventional anaerobic denitrification from these novel processes
should help to settle the debate in the future. Bonin and Raymond (1990) have demonstrated
that in ambient sediment denitrification was found to occur when the oxygen concentration was
increased to 5 mg l-1, but the activity was only 10% of that obtained under anaerobic conditions.
In addition, the enzymes associated with each step of denitrification were affected differently
with respect to oxygen concentration (Bonin et al., 1989). Nitrate reductase is the least
sensitive toward oxygen and its activity was completely blocked at oxygen concentrations
above 4.05 mg l-1, whereas nitrite and nitrous oxide reductases were more sensitive, the activity
of each was blocked at oxygen concentrations above 2.15 and 0.25 mg l-1 respectively (Bonin et
al., 1989). The reactions that these enzymes catalyse will therefore exhibit corresponding
sensitivity and the loss of the products dinitrogen gas, nitrous oxide and nitrite will be lost in
that order following increasing oxygen levels. Additionally, when ambient sediments are
restored to anoxic conditions following a period of oxygen enrichment, denitrification rates are
never fully restored to original anoxic rates (Bonin et al., 1989; Bonin and Raymond, 1990).
The oxygen status of a habitat microsite is controlled by the rate of oxygen supply to that site
and the rate of oxygen consumption (respiration) (Tiedje, 1988). Sediments or detritus exposed
to aerated water can be almost anoxic at the surface, providing they have sufficiently high rates
of oxygen uptake (Jørgensen and Revsbech, 1985). This can explain the occurrence of
anaerobic bacteria on exposed sediments where fully oxic conditions would intuitively be
expected (Jørgensen and Revsbech, 1985). Oxygen generally penetrates no more than a few (3
to 10) millimetres in ambient sediment (Revsbech et al., 1980; Jørgensen and Revsbech, 1985).
The thinness of this layer is due in part to the slow diffusive rate of oxygen into sediment and
also to the high sedimentary respiration rate (Jørgensen and Revsbech, 1985; Tiedje, 1988).
The extent and concentration gradient of the oxic zone in ambient sediment will regulate the
208
occurrence and reaction rates of denitrifying bacteria. In response to decreasing oxygen levels,
an elevated activity and abundance of denitrifiers is generally expected with depth.
The role of carbon in the denitrification process is to provide the electron donor for nitrate
reduction (Tiedje, 1988). Nitrate will accept five electrons when reduced all the way to
dinitrogen gas, and four when reduced to nitrous oxide (see equations above; Revsbech and
Sørensen, 1990). In ambient sediment these electrons are primarily provided by the carbon
present in the detrital organic component of the sediment profile (Tiedje, 1988; Seitzinger,
1990). As detrital matter generally accumulates and cycles on the top surface layer of ambient
sediments, carbon content is expected to decrease with depth. If the available carbon used to
drive the respiration of aerobic nitrifiers (and other aerobic electron consuming processes) in the
surface layer of sediments is diminished (either in time or space), before becoming available to
denitrifying populations, then net denitrification rates may become carbon limited (Tiedje,
1988). Sediments containing low amounts of organic matter are likely to exhibit lower
denitrification rates than highly organic sediments.
In summary, the gradient profiles of oxygen, nitrate and available carbon usually define the
denitrification zone (Tiedje, 1988), as illustrated in Figure D2. Generally, denitrification rates
increase as a function of nitrate concentration but are inhibited by high oxygen concentrations.
Additionally, when there is a low availability of organic carbon to the denitrifying population,
the denitrification process may become limited by the supply of electrons.
Figure D2: Conceptualised gradient profiles of oxygen, nitrate and available carbon, and relative zone of denitrification, in sediments. (Adapted from Tiedje, 1988).
Aerobic Fermentation and sulphate reduction Denitrification respiration
O2
Con
cent
ratio
n
Sediment Depth
NO3-
Available carbon
209
D.7.3 Ambient Rates and Efficiencies of Sedimentary Denitrification
An extensive review of the literature on denitrification rates measured in sediments from lakes,
streams, rivers, estuaries, coastal marine systems, continental shelf and deep sea environments
was given by Seitzinger (1990). Since that time, continued problems of nitrogenous
eutrophication and concern over increasing emissions of the ozone-depleting / greenhouse gas
intermediates of denitrification, has lead to an explosion in the number of studies addressing
denitrification in aquatic environments. It is beyond the scope of this thesis to summarise here
all of this work. A consideration of Seitzinger’s (1990) review as well as more recently
reported denitrification rates pertaining to estuarine sediments, in addition to the current state of
knowledge regarding denitrification in Australian sediments, will suffice.
Seitzinger (1990) gives the following comparisons of denitrification rates between differing
aquatic environments. Lowest denitrification rates (0.03 to 2.4 μmol N m-2 h-1) generally occur
in deep sea sediments. Denitrification rates in continental shelf sediments are in the range of 0
to 54 μmol N m-2 h-1 and are generally lower than those in estuarine and coastal environments.
Denitrification rates in oligotrophic to moderately eutrophic lakes range from 0.3 to
56 μmol N m-2 h-1, whilst generally higher rates exist in eutrophic lakes (20 to 292 μmol N m-2
h-1). Most denitrification measurements from rivers and streams are from temperate systems
that receive considerable inputs of anthropogenic nutrients, and as such, estimates of
denitrification in these environments are generally very high (40 to 2121 μmol N m-2 h-1). More
information is still required on natural rates of denitrification in unpolluted aquatic
environments. The differences in denitrification rates within and between each specified
environment are predominantly due to variation in broad-scale regulatory factors (nitrate/carbon
availability and redox conditions), and the reader is referred to Seitzinger (1990) for specific
details.
Seitzinger (1990) noted that highly polluted estuarine sediments exhibit some of the highest
sedimentary denitrification rates of all aquatic systems (e.g. >500 μmol N m-2 h-1 within the
Tama Estuary in Japan and the Tejo Estuary in Portugal); however, rates in most estuaries and
coastal marine systems range from 5 to 250 μmol N m-2 h-1. Although several more estuarine
denitrification studies have been conducted in the fifteen years since Seitzinger’s important
review, her summary is still remarkably valid. An inclusion to the list of polluted estuaries that
exhibit very high ambient denitrification rates is the River Thames Estuary in England; an
extraordinary upstream ambient denitrification rate of up to 19,916 μmol N m-2 h-1 may occur in
sediments adjacent to where London’s major sewage works discharge, but this tapers off to
between 0 and 70 μmol N m-2 h-1 at the mouth of the estuary (Trimmer et al., 2000a). The upper
rate of denitrification is by far the highest ambient activity ever reported for an aquatic
sediment, and is probably related to the fact that the Thames’ catchment has the highest rate of
210
nitrate export per unit area (10 tonne N m-2 yr-1) known in the world (Trimmer et al., 2000a).
The Tamar Estuary, also in England, has been characterised as having a relatively high ambient
rate of denitrification, and may be within the range of 423 to 541μmol N m-2 h-1 during late
spring to summer (Law et al., 1991). It has also been reported that the Westerschelde Estuary in
the Netherlands had a very high ambient denitrification rate in the 1970s (640 μmol N m-2 h-1),
although following eutrophication intervention this rate declined to 335 μmol N m-2 h-1 in the
1980s (Dettman, 2001). In his review of 11 North American and European estuaries, Dettman
(2001) noted a range of average denitrification rates between 18 and 94 μmol N m-2 h-1
(excluding the Westerschelde Estuary). In fact, most estuarine denitrification rates reported
since Seitzinger’s (1990) review are within this lower range; however, notable and recently
reported exceptions of 154 μmol N m-2 h-1 (for maximal ambient day-time rates in Galveston
Bay, Texas; An and Joye, 2001) and 160 μmol N m-2 h-1 (within Waquoit Bay, Massachusetts;
LaMontagne et al., 2002) may deem that the original and more conservative range of 5 to
250 μmol N m-2 h-1 is still appropriate for typical estuarine denitrification rates. Ambient
denitrification data for estuaries outside of Europe, North America and Japan are largely
missing.
An increasing effort to quantify the proportionate contribution of denitrification to estuarine
nitrogen budgets has occurred over the last decade. Estuarine denitrification efficiencies, i.e. the
percentage of an estuary’s total nitrogen load that is denitrified, as low as 0.1 to 1.3% (Welsh et
al., 2000), 0.3 to 1.7% (Trimmer et al., 2000b) and 3 to 9% (Cabrita and Brotas, 2000), have
respectively been reported for the Bassin d’Arcachon in France, Langstone and Chichester
Harbours in England, and the Tagus Estuary in Portugal. These systems are characterised by
very high primary production by seagrasses, macroalgae and/or benthic microalgae, and
assimilatory pathways outweigh denitrification by one or two orders of magnitude.
LaMontagne et al. (2002) similarly demonstrated decreased denitrification rates in macroalgal
beds compared to bare sediments of Waquoit Bay, and suggested that increasing algal cover
could create a positive feedback as assimilation continues to proportionately out-compete
denitrification in dissolved inorganic nitrogen consumption. Nitrogen fixation may also be
prevalent in systems of proportionately lower denitrification (particularly in seagrass meadows),
and nitrification may be negligible and/or not coupled to denitrification (Welsh et al., 2000).
Meyer et al. (2001) found that denitrification was not coupled to denitrification in sediments of
Randers Fjord, Denmark. Further, Andersen et al. (1984) demonstrated that day-time oxygen
production by benthic microalgae was inhibitory to anaerobic denitrification within sediments
of Lendrup Vig, a shallow estuary in Denmark. Lowered denitrification efficiencies may also
be partly caused by short water residence times in estuaries, i.e. increased flushing and reduced
physical contact between water and sediment (Svensson et al., 2000).
211
The depressed influence of denitrification in the above-mentioned systems is in contrast to the
69 to 75% denitrification efficiency range reported by Dettmann (2001) for 11 estuaries in
Europe and North America. Within these estuaries, higher denitrification efficiencies were
mainly related to longer freshwater residence times, i.e. decreased flushing (Dettmann, 2001).
While important, water residence time is not the only factor governing denitrification
efficiencies. As noted above, the presence of primary producers is of pervasive importance, and
their effect may often be to stimulate rather than depress denitrification. In their study of
Galveston Bay, Texas, An and Joye (2001) reported that denitrification efficiency (>50%) was
greater than that expected by water residence time alone. They found that oxygen production by
benthic microalgae actually enhanced rates of, and coupling between, nitrification and
denitrification (and warned that denitrification estimates of shallow estuaries may be
underestimated if the influence of benthic primary production is discounted, e.g. in dark
incubations of sediment). Oxygen production by the microphytobenthos of the River Colne
estuary, England, enhanced rates of denitrification coupled to nitrification but inhibited
denitrification of nitrate sourced from the overlying water column (Dong et al., 2000).
Denitrification efficiency in this estuary is moderate (18 to 27%; Ogilvie et al., 1997; Dong et
al., 2000). Where ammonium is the major source of dissolved inorganic nitrogen, stimulation
of nitrification may be very important in providing a substrate for denitrification. Jenkins and
Kemp (1984) demonstrated almost complete coupling of nitrification with denitrification in
sediments of the Patuxent Estuary, Chesapeake Bay, as more than 99% of nitrogen lost as
dinitrogen gas was sourced from ammonium. Tight coupling of nitrification and denitrification
has been reported for many estuaries, including the River Colne estuary (see above), Nauset
Marsh Estuary, Massachusetts (Nowicki et al., 1999), Tokyo Bay (Kuwae et al., 1998) and the
Tama Estuary, Japan (Nishio et al, 1983), amongst many others. While the relationship usually
results in at least moderate denitrification efficiencies (e.g. in the River Colne estuary, see
above), competing pathways of nitrogen consumption and other regulatory factors may lessen
the proportional importance of denitrification. For example, the denitrification efficiency of
Waquoit Bay is estimated to be between 32 and 37%, and the high importance of macroalgal
nitrogen assimilation in this system has already been described (LaMontagne et al., 2002).
Within this system, denitrification efficiency was lowered in summer and was inversely
correlated with oxygen availability and ammonium flux (LaMontagne et al., 2002). Further,
while organic matter addition did not affect denitrification efficiency in Waquoit Bay, it did
increase net rates of denitrification (LaMontagne et al., 2002). A similar positive relationship
between organic carbon and denitrification rate was also demonstrated for the nearby Nauset
Marsh Estuary (Nowicki et al., 1999).
There is obviously a great range (<1 to 75%) in the efficacy of estuaries to unload nitrogenous
inputs via denitrification. The interactions between the competing or intrinsically dependant
212
nitrogen-transforming processes (e.g. ammonification, nitrification, denitrification), and
between the parameters governing these processes (e.g. nitrogenous substrate, oxygen and
carbon availability), are often complex and result in denitrification efficiencies largely specific
to individual estuarine systems. Resolution of the factors determining estuarine denitrification
efficiency requires careful consideration.
There are scarce data on the importance and rates of denitrification in Australian estuaries.
Harris (2001) recently reviewed both the scientific and grey literature for a limited number of
Australian systems. He found strong evidence that sediment denitrification becomes of greater
quantitative importance as the water residence time of estuaries increases. When residence
times are short, the ammonification and nitrification products of the bottom waters of Australian
estuaries are frequently discharged downstream (Harris, 2001). However, the long water
residence time of other estuaries and coastal lagoons can lead to denitrification efficiencies at
the high end of the global scale (Harris, 2001). For a shallow sub-tropical coastal embayment
(Moreton Bay), Eyre and McKee (2002) determined an average ambient denitrification rate of
36 μmol N m-2 h-1 and found that about 56% of the total nitrogen load was lost through
denitrification and 41% was physically exported to the ocean. Of the nitrogen not exported, this
equates to a 95% loss of nitrogen via denitrification whilst the remainder (5%) was recycled
within the system. For the heavily studied Port Phillip Bay (a temperate marine embayment),
ambient denitrification rates are about 54 μmol N m-2 h-1, with denitrification efficiencies
around 75% to 85% (Heggie et al., 1999; Nicholson and Longmore, 1999; Harris, 2001). It was
found that most of the nitrate utilised in denitrification was sourced from in situ nitrification,
and the importance of ammonification and subsequent coupled nitrification-denitrification was
highlighted (Heggie et al., 1999). Further, for several warm-temperate Australian lagoons, Eyre
and Ferguson (2002) found that ambient rates of denitrification were in the range of 8 to 69
μmol N m-2 h-1 and that denitrification efficiency was mainly regulated by carbon loading. This
latter point was also true of the cooler Port Phillip Bay sediments (Heggie et al., 1999). As
carbon load rates increased, denitrification efficiency decreased in response to either: (i) a lack
of oxygen and therefore coupled nitrification-denitrification; (ii) elevated sulphide levels and
inhibition of nitrification and denitrification; and/or, (iii) an increased proportion of
dissimilatory nitrate reduction to ammonium (Eyre and Ferguson, 2002). Furthermore the
diurnal activities of dominant primary producers in each lagoon, characteristic of different
stages of eutrophication, (i.e. seagrasses, phytoplankton, benthic microalgae and macroalgae),
were highly important in determining overall nutrient fluxes (Eyre and Ferguson, 2002). A
characteristic pattern of sedimentary N-efflux during dark conditions and N-influx during light
conditions was observed (Eyre and Ferguson, 2002). As estuarine primary producer
assemblages change in response to nutrient loading, so too does their impact on the net ratio of
productivity to respiration (p/r) and therefore on N-fluxes (Eyre and Ferguson, 2002). In fact,
213
change in p/r was considered one of the key indicators of eutrophication in these estuaries. Eyre
and Ferguson’s (2002) study highlights the managerial importance of limiting carbon loading,
and maintaining a balance of autotrophy and heterotrophy, so as to facilitate better
denitrification and nitrogen removal within eutrophic systems.
D.7.4 Studies of Denitrification Kinetics
To place this thesis’ kinetic consideration of denitrification (Chapter 5) into the context of prior
research, an extensive review of the literature examining the kinetics of denitrification, nitrate
and/or nitrite reduction in various systems (estuarine/marine, freshwater, soil, pure cultures and
cell-free reductase extracts) has been undertaken. The maximal potential denitrification rates
(Vmax or Vmp) and half-saturation constants (Km or Kapp ) of reported studies, together with
methods utilised, relevant location and sediment/soil depth (where appropriate) have been
summarised in Table D3. The reader is referred to this table for reference in the following
discussion. Pertinent results of this thesis are included in the table but their implications and
relevance are considered in the discussion of Chapter 5.
Depending on the methods utilised, kinetic constants (Vmp and Kapp) reported in the literature
can pertain to either denitrification, nitrate reduction or nitrite reduction. Where nitrous oxide
or nitrogen gas production has been measured the resultant kinetic constants are specific to
denitrification, whilst measurement of the rate of nitrate or nitrite consumption refers to the
more general case of nitrate or nitrite reduction (either to nitrogenous gases or to ammonium,
perhaps incorporated into organic biomass). In many cases the rates of dissimilatory nitrate
reduction to ammonium (DNRA) and assimilatory nitrate uptake are assumed or demonstrated
to be negligible, and nitrate/nitrite consumption is approximated to be the rate of denitrification.
However, it must be noted that even under these special circumstances, the actual rate of nitrate
reduction (presumed denitrification) may be underestimated if nitrate production via
nitrification is not accounted for. Generally, the distinction must be made between reported
kinetic constants for nitrate/nitrite reduction whereby end products may not necessarily be
gaseous, and kinetic constants specifically pertaining to denitrification. In kinetic studies
relating to bacterial populations in sediment or soil environments it is also essential to consider
reported sediment/soil depth increments or total bulk. Where Vmp is reported on a per area
basis, the depth to which nitrate reduction/denitrification is obviously important. A deeper
sediment/soil column equates to a larger volume where bacterial activity can occur, and
therefore a potentially greater population size and value of Vmp. However, it is often found that
denitrification and nitrate/nitrite reduction activity tapers off very quickly in deeper sediment
where substrate supply is limited (Table D3). It is assumed that most studies incorporate the
relevant zone where the majority of denitrification or nitrate/nitrite reduction occurs, but this
may not always be the case. The affinity of bacteria for nitrate/nitrite (as measured by Kapp) in
redox-stratified sediment/soil is also expected to vary for different depth increments.
214
The maximum potential rate of denitrification (Vmp) in estuarine or marine sediments reported
by studies utilising a kinetic approach ranges between 0.04 and 3.33 mmol N m-2 h-1
respectively within the Potomac and Choptank Tributaries of Chesapeake Bay (Table D3;
Twilley and Kemp, 1986). Variation in salinity, nitrate and organic carbon availability was
proposed to account for the wide range in maximal denitrification rates (Twilley and Kemp,
1986). Kinetic studies of sediments off the coast of California (Oremland et al., 1984; Joye et
al., 1996), France (Raymond et al., 1992), Denmark (Oren and Blackburn, 1979) and England
(Dong et al., 2000) have demonstrated maximal denitrification rates intermediate to the reported
range (Table D3), thus highlighting the capacity for denitrification potentials to vary greatly
within a limited geographical range (i.e. within the vicinity of Chesapeake Bay). The reported
range of maximum potential rates of general nitrate reduction is broader (Table D3). Absence
of nitrate reduction was reported for sediments of the North Sea off the Belgian coast (Billen,
1978) whereas a maximal nitrate reduction rate of 12.9 mmol N m-2 h-1 was demonstrated for an
intertidal salt marsh in the United Kingdom (Nedwell, 1982). Within the North Sea sediment,
Billen (1978) contends that nitrate reduction is reflective of denitrification rate and that activity
is absent in the oxidised layer (up to 8 cm deep) of certain sediments. The high rate of maximal
nitrate reduction in the deep (30 cm) salt marsh sediment is argued mainly to result in
denitrification, although Nedwell (1982) estimates an 8% contribution by DNRA and a
potentially small assimilatory consumption of nitrate.
In estuarine or marine environments, the affinity of denitrifying populations for nitrate as
judged by the concentration required for half-maximal activity (Kapp), ranges from less than
2 μM (high affinity) measured during September in the top 3 cm of sediment of Tomales Bay,
California (Joye et al., 1996), to 1818 μM (low affinity) determined in the 2 to 4 cm depth
interval of sediment from Carteau Cove, France (Table D3; Raymond et al., 1992). Reported
Kapp values for both denitrification and more general nitrate reduction for other estuarine and
marine areas, including off the coasts of Fiji (Nedwell, 1975), Japan (Koike et al., 1978; Koike
and Hattori, 1979), France (Esteves et al., 1986), Denmark (Oren and Blackburn, 1979),
Belgium (Billen, 1978), England (Nedwell, 1982; Dong et al., 2000) and North America (Koike
et al., 1978; Capone and Taylor, 1980; Iizumi et al., 1980; Oremland et al., 1984; Twilley and
Kemp, 1986), are intermediate to these extremes (Table D3). The variation in Kapp values has
often been correlated with natural nitrate availability, whereby communities exhibiting a high
affinity (low Kapp) are adapted to an environment of low nitrate concentration or supply, whilst a
low affinity (high Kapp) is more reflective of a community adapted to an ample or abundant
nitrate supply (Nedwell, 1975; Joye et al., 1996; García-Ruiz et al., 1998). Stated conversely,
nitrate reduction would be inefficient by high affinity denitrifying bacteria in an environment of
high nitrate concentration, and low affinity bacteria would likewise be inefficient in a low
nitrate concentration environment (García-Ruiz et al., 1998). Elevated nitrate availability may
215
alter either cell-specific enzyme activity (i.e. increased activity of individual denitrifying
bacteria) or enhance growth rates (i.e. an increased number of bacteria), both of which may lead
to a greater Kapp; however, separation of denitrifier biomass and enzyme expression is
technically challenging (Joye et al., 1996). Variations in Kapp have also been observed for
similar ambient nitrate concentrations, and related to differing composition or physiology of the
denitrifying community, variation in organic carbon availability (as an electron donor),
dissimilar redox conditions and/or differences in transport rate of nitrate (via diffusion or active
processes) to active sites of denitrification (Nedwell, 1975; Joye et al., 1996; Strong and Fillery,
2002). A combination of these factors (natural ambient nitrate, oxygen and organic carbon
concentration; abundance, composition and physiology of the adapted denitrifying community;
rate of transport to active denitrifying zones) including a consideration of relevant temporal
changes in these parameters, explains the large variation of reported Kapp values both within and
between geographical regions (large and fine scale) and over time.
Furthermore, Myrold and Tiedje (1985) suggest that dissimilatory nitrate reducers may possess,
and be able to switch between, both high and low affinity nitrate uptake mechanisms. This
explanation would account for the apparent discrepancy between reportedly high Km / low
affinity values (250 – 1300 μM) for enzyme extracts of nitrate reductases (Myrold and Tiedje,
1985), and the generally much lower Km / high affinity values (1.7 – 15 μM) reported for pure
cultures of nitrate reducers (Table D3; Betlach and Tiedje, 1981; Edwards and Tiedje, 1981;
Parsonage et al., 1985; Christensen and Tiedje, 1988). The exception of a low affinity (Km =
260 μM) Pseudomonas pure culture (Brown et al., 1975) is important, and Myrold and Tiedje
(1985) argue that pure culture studies have generally not been conducted at high enough nitrate
concentrations to detect a low affinity system.
The number of kinetic studies conducted within estuarine, marine and other environments is
somewhat limited and it is insightful to consider kinetic constants reported for freshwater
sediments and soil. Within lake and river sediments of Denmark (Anderson, 1977), England
(García-Ruiz et al., 1998) and the U.S.A. (Messer and Brezonik, 1984), maximum potential
rates of denitrification or nitrate/nitrite reduction (Vmp) are generally within the range
determined for estuarine/marine sediments (Table D3); although higher nitrate reduction Vmp’s
were reported for farm pond sediments in Kentucky, U.S.A (12.4 to 26.6 mmol N m-2 h-1;
Murray et al., 1989), and were even higher in sediments of Lake Vechten, Denmark (24.0 to
120 mmol N m-2 h-1; Hordijk et al., 1987). Likewise, reported Kapp values of freshwater
sediments are within the range of marine/estuarine sediments, although very high (low affinity)
half saturation constants have been reported for nitrate reduction in lakes Kvind sø (Kapp = 2856
μM) and Kalgaard sø (Kapp = 6024 μM) of Denmark (Anderson, 1977).
216
Whilst reported Kapp values of soils are similarly variable and generally within the range
determined for sediments (estuarine, marine or freshwater), even higher (lower affinity) half-
saturation constants may exist in soil environments (Table D3). For example, a fine sandy soil
from Coachella, California had a nitrate reduction Kapp of 12,137 μM (Bowman and Focht,
1974). The lower affinities of bacterial communities for nitrate within agricultural soil
presumably largely relates to higher nitrate supply and ambient concentration due to
fertilisation. Higher fertiliser-nitrate availability may also partially explain the generally higher
maximal rates of denitrification, nitrate and/or nitrite reduction (Vmp) in soils of New Zealand
(Schipper et al., 1993), Denmark (Ambus, 1993; Maag et al., 1997), Sweden (Klemdtsson et al.,
1977), Russia (Ryzhova, 1979), U.S.A (Bowman and Focht, 1974; Kohl et al., 1976; Myrold
and Tiedje, 1985; Abdelmagid and Tabatabai, 1987; Murray et al., 1989) and Canada
(Yoshinari et al., 1977) compared to estuarine/marine sediments (Table D3). The Vmp range of
soils is typically between 1.38 mmol N m-2 h-1 (for nitrite reduction in a riparian agricultural soil
in Copenhagen, Denmark; Ambus, 1993) to 512 mmol N m-2 h-1 (for nitrate reduction in soil
under a forb-grass meadowland within Moscow Oblast, Russia; Ryzhova, 1979). The only
reported exceptions, with much lower denitrification Vmp’s (0.003 to 0.89 mmol N m-2 h-1), have
been determined within south-western Australia for poorly nutrient-retentive, sandy soils under
arable crop or pasture (Strong and Fillery, 2002). The implication is important within the
regional context of this thesis, since the comparatively low maximal rate of terrestrial
(catchment) denitrification suggests that a potentially higher load of nitrate is available for
export to the aquatic systems of south-western Australia. These low terrestrial denitrification
Vmp’s therefore may be a contributing factor to the recent eutrophication of many of these
systems. To the best of the author’s knowledge, apart from this thesis, no studies examining the
kinetics of nitrate reduction or denitrification have been undertaken for Australian estuarine,
marine or freshwater sediments.
217
Table D3: Half saturation constants (Km or Kapp) and maximal rates (Vmax or Vmp) of denitrification, nitrate and nitrite reduction of various systems.
Location Process Sediment or
soil depth range (cm)
Km or Kapp (μM NO3
-)
Vmax or Vmp (converted# to
mmol N m-2 h-1) Vmax or Vmp
[reported units] Method Reference
Marine/Estuarine Sediment Swan River Estuary, Western Australia - A. ehlersi inhabited sediment - uninhabited sediment - C. aequisetis inhabited sediment
Nitrate Reduction 0 – 5.5 0 – 4.7 0 – 4.7
50
133 202
4.75 4.33 3.83
[mmol N m-2 h-1] 4.75 4.33 3.83
C2H2 block, NO3-
consumption above in vitro sediment cores.
This study
Vatuwaga River Estuary, Fiji (polluted mangrove sediment) - upstream, adjacent to effluent outfall - mouth, 2.5 km from effluent outfall
Nitrate Reduction
NR NR
180 600
3.65 1.09
[mmol N m-2 h-1]
3.65 1.09
In situ NO3-
consumption of overlying water.
Nedwell 1975
Tokyo Bay, Japan Mangoku-Ura, Japan
Denitrification Denitrification
NR
NR
24
27 – 42
–
–
[μmol N g-1 h-1] 0.11
0.017 – 0.019
15NO3- 15N2.
Sediment slurries. Koike et al. 1978
Bering Sea, Japan
Denitrification 0 – 2
~ 4
–
NR
15NO3- 15N2.
Sediment slurries. Koike and Hattori 1979
Port Cros Island, France Nitrate Reduction 0 – 1.2a
78
1.47
[μmol N cm-3 d-1] 2.94
NO3- consumption
of in vitro sediment cores.
Esteves et al. 1986
West Mediterranean, France
- Lavera - Berre - Cassis - Carry le Rouet - Port Cros - Carteau
Denitrification
0 – 2 0 – 2 0 – 2 0 – 2 0 – 2
0 – 2 0 – 4 0 – 6 0 – 8
0 – 10
228 256 499 539 560
1428 1818b
740b
1666b
1428b
0.39 0.32 0.25 0.25 0.25
0.50 0.66 0.68 0.71 0.81
[μmol l-1 (sediment) d-1]
464 386 302 298 300
600 192b
20b
37b
120b
C2H2 block, N2O production. Sediment slurries.
Raymond et al. 1992
218
Table D3: (continued)
Location Process Sediment or
soil depth range (cm)
Km or Kapp (μM NO3
-)
Vmax or Vmp(converted# to
mmol N m-2 h-1)
Vmax or Vmp[reported units] Method Reference
Kysing Fjord, Denmark
Denitrification 0 – 1
344
0.18
[nmol N cm-3 d-1] 422
15NO3- 15N2.
Sediment slurries. Oren and Blackburn 1979
North Sea, Belgium (coastal ‘zone’) - Station M1230 (April 1975) - Station M1283 (March 1976) - Station M01 (August 1975) - Station M01 (March 1976) - Station M11 (April 1975)
Nitrate Reduction 10 – 15
0 – 5 0 – 3
20 – 25 6 – 11 0 – 5 0 – 8
50 1.17 0.58 0.49 0.81 0.14
0 0
[μmol N cm-3 s-1] 6.5 x 10-6
3.2 x 10-6
4.5 x 10-6
4.5 x 10-6
0.75 x 10-6
0 x 10-6
0 x 10-6
NO3- consumption
Sediment slurries. Billen 1978
Colne Point, Colchester, U.K. (intertidal salt marsh sediment)
Nitrate Reduction 0 – 30
319 ± 114
12.9 ± 2.11
[mmol N m-2 h-1] 12.89 ± 2.11
NO3- consumption
above in vitro sediment cores.
Nedwell 1982
River Colne Estuary, Colchester, U.K. (sewage affected intertidal muddy sediment)
NR - Not Reported. * - Nitrite concentration. # - For comparative purposes and where possible, reported Vmax or Vmp units were converted to mmol N m-2 h-1 by utilising available information on soil/sediment depth and/or bulk density
and/or water content (see References for further data):
a - Sediment height estimated from Figure 1a of Esteves et al. (1986).
b - Km and Vmax values are for the deepest 2 cm increment within the depth range.
c - Estimates of Kapp and Vmp for each 3 cm depth interval were respectively taken from Figures 3 and 4 of Joye et al. (1996). Values are for the deepest 3 cm increment within the depth range.
d - September values are underestimates since they do not include the Vmp data for the top 3 cm of sediment (not reported).
e - Denotes range of Km or Vmax values within depth interval (incomplete information regarding values at specific depths within interval).
f - Lower estimate assumes lowest Vmax throughout entire depth interval; vice versa for higher estimate.
g - Assuming a wet sediment bulk density of 1 to 1.15 g cm-3 (which must be true for the reported sediment water content of 87%, whereby the collective sediment is more dense than water).
h - Sediment assumed to be sampled to a depth of 10 cm.
i - Assuming sediments have a ‘fine-silty’ bulk density of 1.15 g cm-3, i.e. similar to the surrounding (often flooded) soil (see ‘s’).
j - Assuming September 1987 (0 – 2 cm) estimate of Vmax ≅ ½ June 1988 (0 – 2 cm) value (as are the values for the 0 – 4 cm and 0 – 6 cm depth increments in September 1987 compared to June 1988).
k - Km and Vmax values are for the deepest 5 cm (l), 10 cm (m), 25 cm (n) or 50 cm (o) increments within the depth range. Converted Vmax values are based on estimates of soil water content (1 – dry matter) taken from Figure 1 of Maag
et al. (1997); and assume that the collective soil is more dense than water. For the depth increments where Vmax values were not reported (p), the converted Vmax estimate assumes that the missing increment equals the average of the
Vmax values immediately above and below that increment.
q - Assuming 12.8 to 17.6% of slurry is original wet soil, by volume. (Slurry = 15g original wet soil + 70 ml buffer; reported original wet soil water content = 68.6 to 81.6%; and, whereby the collective soil is more dense than water).
224
225
Notes (continued)
r - Given the slurry mix of [20 g dry soil : 100 ml solution], and where soil type is undefined the lower estimate assumes a ‘clayey’ bulk density (1.05 g cm-3), whilst the upper estimate assumes a ‘sandy’ bulk density (1.55 g cm-3).
s - Assuming all soils are sampled to a depth of 10 cm and have a ‘fine-silty’ bulk density of 1.15 g cm-3.
t - Given the slurry mix of [75 g dry soil : 125 ml solution], and assuming a typical in situ silt loam bulk density of 1.15 g cm-3.
u - Range assumes a bulk density between 1.1 and 1.4 g cm-3.
v - Given the slurry mix of [50 g soil : 50 ml solution], assuming soil is taken from the surface 5 cm of the profile and dried, and has an in situ ‘fine-sandy’ bulk density of 1.3 g cm-3.
D.8 NOVEL PROCESSES OF NITROGENOUS TRANSFORMATION AND
GAS PRODUCTION
The biological processes of nitrogen transformation described above conventionally are
considered to be the major pathways of nitrogen conversion in natural environments.
Traditionally, denitrification has been assumed to be the major source of gaseous nitrogen (NO,
N2O and N2) from inorganic derivatives, and thereby an all-important nitrogen sink in aquatic
environments. Nitrous oxide is a 200-fold more effective greenhouse gas than carbon dioxide,
whilst nitric oxide is implicated in a set of key reactions that cause depletion of the ozone layer
(Colliver and Stephenson, 2000; Bonin et al., 2002). Concern over increasing concentrations of
N2O and NO in the atmosphere (an annual rate of 0.2 to 0.3% for N2O,) potentially linked to
anthropogenic activities (such as mass global fertilisation), in addition to discrepancies in
nitrogen budgets of waste-water treatment and natural systems, has triggered more thorough
research into the sources of gaseous nitrogen production (Colliver and Stephenson, 2000; Bonin
et al., 2002). Recent identification of a variety of novel microbial nitrogen conversion
pathways, including nitrifier-denitrification, anaerobic ammonia oxidation (anammox) and
heterotrophic nitrification linked with aerobic denitrification, has challenged conventional
thought and forced a re-assessment of possible nitrogen transfers and sinks in aquatic
environments. When rationalising ecosystem nitrogen budgets, a consideration of the potential
for each of these processes to occur is now necessary. The characterisation and general nitrogen
cycling implications of these novel microbial pathways have been extensively reviewed by
Jetten et al. (1997), Jetten (2001) and Zehr and Ward (2002), and are summarised below.
D.8.1 Nitrifier-denitrification and Chemodenitrification
Autotrophic ammonium oxidisers (nitrifiers) have been implicated in the production of the
gases dinitrogen (N2), nitrous oxide (N2O) and nitric oxide (NO) in many marine environments
(Capone, 2000; Wrage et al., 2001). While the mechanisms of gaseous nitrogen production by
ammonium oxidisers are still unclear, they are thought to produce mainly the nitrogenous gases
in a process termed nitrifier-denitrification (Bonin et al., 2002). In nitrifier-denitrification,
ammonium oxidisers are capable of reversing what is considered to be their natural reaction
process by reducing nitrite in the presence of low levels of oxygen (Colliver and Stephenson,
2000). The oxidation of ammonium to nitrite (NH4+ NO2
-) is followed by the reduction of
nitrite to nitric oxide, nitrous oxide and/or dinitrogen gas (NO2- NO N2O N2) (Wrage et
al., 2001). The end product is dependant on the species involved and the presence of a suitable
electron acceptor, but importantly the nitrification and denitrification path both occur in the
same organism (Colliver and Stephenson, 2000; Wrage et al., 2001). Ammonium oxidisers that
are capable of undertaking this kind of denitrification (so far only isolates from the genus
Nitrosomonas) are thought to possess and use essentially the same enzymes that are involved in
denitrification by conventional denitrifying bacteria (Colliver and Stephenson, 2000; Wrage et
226
al., 2001). Since the denitrification step usually only occurs under conditions of
microaerophily, nitrifier-denitrification is sometimes also termed oxygen-limited autotrophic
nitrifier-denitrification or OLAND (Kuai and Verstraete, 1998; Philips et al., 2002). A
particular point should be made that nitrifier-denitrification by autotrophs is distinct from the
two-or-more organism process of coupled nitrification-denitrification, whereby conventional
denitrifying bacteria reduce nitrate and nitrite (produced by nitrifying bacteria) to N2O and N2
(Jetten, 2001; Wrage et al., 2001). Autotrophic nitrifier-denitrification is also different to
heterotrophic nitrification linked with denitrification in the presence of oxygen (see below)
(Wrage et al., 2001).
Alternatively, N2O may be formed from decomposition of the intermediates of ammonium
oxidation (e.g. hydroxylamine - NH2OH, see Section D.5) or nitrite itself (Kaplan, 1983; Wrage
et al., 2001). These processes are usually regarded as a special case of chemodenitrification and
probably only constitute a small fraction of the nitrogenous gases produced by ammonium
oxidisers (Colliver and Stephenson, 2000; Wrage et al., 2001). Additionally, the anaerobic
oxidation of ammonia and amino-level nitrogen in organic compounds to dinitrogen gas via
inorganic (non-biological) catalysis by manganese oxide has been demonstrated in some
sediments (Ward, 2000); however, the quantitative significance of this process is questionable
(Thamdrup and Dalsgaard, 2000).
The contribution of nitrifier-denitrification and other processes within sediments, surface waters
and the marine environment to atmospheric flux of and NO, N2O and N2, is only just beginning
to be quantified. Together with incomplete conventional denitrification by anaerobic
denitrifying bacteria (see Section D.7), nitrification by ammonium oxidisers is thought to be a
major biological nitrous oxide producing process (Colliver and Stephenson, 2000; Bonin et al.,
2002). Bonin et al. (2002) found for marine surface layers of the Mediterranean Sea that
nitrification is the dominant process that produces nitrous oxide, whilst incomplete
denitrification accounted for more production in surface waters near the mouth of a river and
near the bottom. Estuaries are thought to be an important source of atmospheric N2O, whereby
production is associated with high rates of both nitrification and denitrification (de Bie et al.,
2002). Net fluxes of N2O to the atmosphere will depend on the factors controlling these
processes, e.g. ammonium and nitrate supply, respectively (de Wilde and de Bie, 2000; de Bie
et al., 2002).
D.8.2 Anaerobic Ammonium Oxidation (Anammox)
It was discovered recently that ammonium oxidation can also occur in the absence of oxygen
(Mulder et al., 1995). Under anaerobic conditions, ammonium is oxidised with nitrite serving
227
as the electron acceptor with an end-product of dinitrogen gas (Kuypers et al., 2003). The
process has been coined ‘anammox’ (anaerobic ammonium oxidation) and can be denoted as:
NH4+ + NO2- N2 + 2 H2O
Bacteria that are capable of this reaction belong to the order Planctomycetales. Isolates from
wastewater treatment plants and natural environments have been provisionally named and
include two freshwater species (Brocadia anammoxidans and Kuenenia stuttgartiensis) and one
marine species (Scalindula sorokinii) (Strous et al., 1999; Jetten et al., 2003). All three species
contain a distinctive organelle, the anammoxsome, which houses a unique set of enzymes to
catalyse the characteristic reaction (Jetten et al., 2003). Like all planctomycetes, anammox
bacteria grow very slowly, dividing only once every 11 to 14 days (Strous et al., 1999; Jetten et
al., 2003). The rate of anaerobic ammonium oxidation may be substrate dependent. Ammonium
typically is plentiful in anoxic ecosystems, whilst nitrite may be provided by nitrate reducing
bacteria (nitrate respirators including denitrifiers and bacteria capable of dissimilatory nitrate
reduction to ammonium) or microaerophilic ammonium-oxidising bacteria (nitrifiers) (Jetten et
al., 2003). In systems with limited oxygen supply it is thought that these bacteria may occur in
close association with anammox bacteria (Schmidt et al., 2002; Jetten et al., 2003).
Additionally, anaerobic ammonium oxidation also has been recently exhibited in the
conventionally aerobic (or microaerophilic) nitrifier Nitrosomonas eutropha (Schmidt and
Bock, 1997; Schmidt et al., 2002). The process is similar to that described for aerobic nitrifier-
denitrification; however, nitrogen tetroxide (N2O4, the dimeric form of nitrogen dioxide, NO2) is
used as an electron acceptor (instead of O2) in the conversion of ammonia to nitrite, with NO
being released as an additional product (Schmidt et al., 2002):
NH3 + N2O4 NO2- + 2 NO + 3 H+ + 2 e-
The nitrite produced is used partly as an electron acceptor leading to the formation of dinitrogen
gas (Schmidt et al., 2002):
NO2- + 4 H+ + 3 e- ½ N2 + H2O
Nitrogen di/tetroxide (NO2/N2O4) is not readily available in anaerobic environments and the
process is thus limited by transport of the substrate/s from oxic layers (Schmidt et al., 2002).
The term anammox is restricted to the reaction defined by the above-mentioned planctomycete-
like bacteria; anaerobic ammonium oxidation by N. eutropha may perhaps be termed anaerobic
nitrifier-denitrification. Nitrite production by this latter process may also provide a substrate for
228
annamox, and the two types of bacteria may be closely related in anaerobic environments
(Schmidt et al., 2002; Jetten et al., 2003). To complicate matters further, under anaerobic
conditions some nitrite-oxidisers (strains of the genus Nitrobacter) have been reported to have a
similar nitrifier-denitrification process as that described for autotrophic ammonium oxidisers
above, producing N2O via reduction of nitrate but using pyruvate as an electron acceptor
(Wrage et al., 2001). Very little information is available regarding this pathway (Wrage et al.,
2001). As Schmidt et al. (2002) have acknowledged, we are far from understanding the
prevalence, variations and complexities of all these novel nitrogen transforming processes.
The most striking implication of the anammox process is its capacity to facilitate removal of
both ammonium and nitrite from aquatic systems as they are converted to dinitrogen gas,
thereby providing an alternative inorganic nitrogen sink to denitrification. Justifiably, the
application of anammox to the treatment of nitrogenous waste has received much attention and
the process has been patented (European Patent 0327184A1, US Patent 427849[5078884];
Mulder, 1995; Jetten et al., 1997; Trimmer, 2003). Indeed it was in a denitrifying fluidised bed
reactor treating effluent from a methanogenic reactor that anammox was first discovered
(Mulder, 1995; Jetten et al., 1997). It is also becoming obvious that anammox may play a
marked role in nitrogen turnover in natural sediments, estuaries and marine environments.
Risgaard-Petersen et al. (2003) suggest that annamox accounts for less than 6% of N2
production in estuarine sediments. In sediments of the Thames Estuary, anammox accounted for
between 1% and 8% of total N2 production (the remaining percentage attributable to
denitrification), respectively at the mouth of the estuary and near the river head (Trimmer et al.,
2003). Anammox was found to be insignificant in the sediment of a coastal eutrophic bay
(Aarhus Bay), whilst it accounted for between 24% and 67% of total N2 production in Baltic-
North Sea continental shelf sediments (Dalsgaard and Thamdrup, 2002; Thamdrup and
Dalsgaard, 2002). In reviewing studies of the annamox process in anoxic waters of the Black
Sea (Kuypers et al., 2003) and off the coast of Costa Rica (Dalsgaard et al., 2003), Devol
(2003) suggests that anammox could account for between 30% to 50% of all biological N2
production in the world’s oceans, and that a major revision of the relative importance of
denitrification in the global marine nitrogen budget may be necessary. However, research is
required before the quantitative importance of anammox and similar processes in estuarine,
marine and other environments can be accurately ascertained.
D.8.3 Heterotrophic Nitrification and Aerobic Denitrification
In addition to autotrophic nitrifying bacteria (see Section D.5 and Table D1), there is a
heterogeneous range of prokaryotes (bacteria) and eukaryotes (fungi and some unicellular algae)
that are capable of oxidising both ammonium and organic nitrogen compounds in the process
called heterotrophic nitrification (Sprent, 1987; Kuenen and Robertson, 1988; Šimek, 2000).
229
For an expansive list of these species, the reader is referred to Kuenen and Robertson (1988).
The enzymes utilised in heterotrophic nitrification are argued to be quite different to
comparative autotrophic nitrifier enzymes (Ward, 2000; Wrage et al., 2001). In contrast to
autotrophic nitrification, oxidation of reduced nitrogen compounds by heterotrophic nitrifiers
requires energy, generally resulting in net energy loss (Sprent, 1987; Ward, 2000; Jetten, 2001).
In fact, the process can actually serve as an electron sink (Jetten, 2001). Many aspects of the
physiology of heterotrophic nitrification are still unclear (Šimek, 2000), although it has been
suggested that nitrogenous substrates are oxidised to inactivate toxins (including nitrite) and to
metabolise substrates which may be required for the growth of other competing organisms,
amongst other reasons (Sprent, 1987).
The existence of heterotrophic nitrifiers has long been known; however, their rates of
nitrification traditionally have been considered insignificant or very low in comparison to