A unified measure of the number, volume and diversity of dead trees and the response of fungal communities

A unified measure of the number, volume and

diversity of dead trees and the response of fungal

communities

Jenni Hottola*, Otso Ovaskainen and Ilkka Hanski

Metapopulation Research Group, Department of Biological and Environmental Sciences, University of Helsinki,

PO Box 65, FI-00014 Helsinki, Finland

Summary

1. Much of ecological research focuses on the responses of species and species communities to vari-

ation in the amount and quality of resources that are required for survival and reproduction. In such

research, it is critical to measure the availability of resources in a manner that is relevant in relation

to the ecological requirements of the species.

2. We have developed a measure for resource availability that integrates the contributions of

the number, volume and diversity of resource units to quantify the amount of habitat that is

available for a species community. We apply this measure to data on the occurrence of 116 species

of wood-decaying polyporous fungi in 47 study plots of boreal forest within an area of

150 · 150 km.

3. We show that species richness and pooled abundance of common species is explained well by the

number of downed logs, whereas the occurrence of 41 red-listed species is best explained by the total

volume of logs and by the abundance of large logs in particular. The occurrence of common species

is explained by the local availability of dead wood, whereas the occurrence of red-listed species is

additionally affected by the spatial connectivity of the focal forest stand to the surrounding larger

expanses of old-growth forest.

4. Our results elicit the contrasting ecologies of common and red-listed species in relation to how

the number of logs, their size distribution and diversity, and forest connectivity affect species

occurrences. The results suggest that the most cost-effective means of preventing further declines of

threatened species is to increase the amount of large downed logs through restoration and biodiver-

sity-orientedmanagement in the vicinity of existing areas of natural-like forests.

5. Synthesis.Our results illustrate that the most relevant way of measuring resource availability can

differ greatly even within a taxonomically coherent community seemingly sharing the same

resources. Our approach for modelling resource availability applies to the resources that occur as

discrete objects with variation in the size and quality of individual resource units.

Key-words: common species, community composition, connectivity, dead wood, ecological

requirements, resource availability, saproxylics, species richness, threatened species, wood-

decaying fungi

Introduction

Forests have been the dominant terrestrial ecosystem on

Earth for hundreds of millions of years, since the Late Devo-

nian (Hanski 2005). In both boreal and tropical forests, dead

wood in the form of decaying snags, logs, branches and

stumps is an important structural component and also plays a

role in ecosystem functioning. The amount of dead wood

often exceeds 100 m3 ha)1 in natural forests and may exceed

200 m3 ha)1 in both boreal (Siitonen 2001) and tropical

forests (Keller et al. 2004).

Many organisms have adapted to live in the dead-wood

microhabitat, using decaying wood as food, substrate or shel-

ter and contributing to its decomposition (Samuelsson,

Gustafsson & Ingelog 1994; Siitonen 2001; Hanski 2005; Jons-

son, Kruys &Ranius 2005). For instance, it has been estimated

that 20–25% (4000–5000 species) of all forest-inhabiting

species in Finland are saproxylic (dependent on dead wood) at

least in some part of their life cycle (Siitonen 2001). Wood-

decaying fungi are entirely dependent on dead wood and often*Correspondence author. E-mail: [email protected]

Journal of Ecology 2009, 97, 1320–1328 doi: 10.1111/j.1365-2745.2009.01583.x

� 2009 The Authors. Journal compilation � 2009 British Ecological Society

specialized to use a particular type of dead wood or prefer

particular microclimatic conditions (Niemela, Renvall &

Penttila 1995; Renvall 1995; Boddy, Frankland & van West

2008). Sufficient temporal continuity and spatial connectivity

in the occurrence of dead wood and favourable environmental

conditions are critical for long-term persistence of specialized

species (Norden & Appelqvist 2001; Siitonen, Penttila &

Kotiranta 2001; Gu, Heikkila & Hanski 2002; Penttila et al.

2006; Jonsson, Edman & Jonsson 2008). Suitable conditions

are generally not met in intensively managed forests, of which

the current Fennoscandian forest landscapes largely consist,

and consequently the populations of many species of

wood-decaying fungi have declined in Fennoscandia in recent

decades (Gardenfors 2000; Rassi et al. 2001).

Several previous studies have investigated the responses of

individual polypore fungus species and polypore communities

to spatial (Penttila, Siitonen & Kuusinen 2004; Berglund &

Jonsson 2005) and temporal variation in the occurrence of

dead wood (e.g. Groven et al. 2002; Rolstad et al. 2004).

Occurrence of fungal species is often related to the total volume

of dead wood (e.g. Junninen et al. 2006; Penttila et al. 2006),

but some studies have related the occurrence rather to the

number of dead-wood objects (e.g. Bader, Jansson & Jonsson

1995; Gilbert, Ferrer & Carranza 2002), and some to both the

volume and the number of logs (e.g. Norden et al. 2004;

Junninen & Kouki 2006). In yet other cases, also the diversity

of dead-wood objects, in terms of tree species, decay stage and

so forth, has been shown to influence polypore communities

(e.g. Penttila, Siitonen & Kuusinen 2004; Hottola & Siitonen

2008).

Here, we describe a newmeasure of dead woodwhich simul-

taneously takes into account the number, volume and diversity

of dead-wood objects and which can be used to analyse the

relative contributions of these factors to explain the occurrence

of or resource use by the consumers. The measure is calculated

for individual study plots (forest stands), and hence it can be

used to analyse variation in the occurrence of single species or

groups of species in relation to resource availability. The

measure is particularly useful for comparing the responses of

different species or groups of species, because it summarizes

the different facets of resource availability in a single measure

that has an intuitive interpretation. The measure can be

applied to any resources that occur as discrete objects with var-

iation in the size and quality of individual units. Dead-wood

objects are one example, but there are many other comparable

microhabitats (Hanski 2005), and the measure could be

applied to e.g. living plants to characterize the resources of

herbivorous insects.

We employ the new measure to analyse the factors that

influence the occurrence of wood-decaying fungal species

among forest stands in a large data set collected from bor-

eal forests in eastern Finland and western Russia. In addi-

tion to the local factors, we investigate the influence of the

spatial connectivity of the forest stand to the surrounding

landscape. We compare the occurrences of common and

red-listed species of polyporous fungi to identify any gen-

eral patterns in their responses to the different facets of

dead-wood abundance.

Materials and methods

STUDY AREA AND STUDY PLOTS

The study area (Fig. 1) is 150 · 150 km in size and located in a

watershed region in eastern Finland and north-western Russia in the

middle boreal vegetation zone (Ahti, Hamet-Ahti & Jalas 1968). The

forest stands represent mesic (Vaccinium myrtillus type) and subdry

(Empetrum-Vaccinium type) site types on mineral soils (Cajander

Fig. 1. The study area in northern Europe

and the locations of the individual study

plots in eastern Finland and adjacent areas

in Russia. Circles depict study plots in large

and connected forests, triangles in small and

isolated forest fragments. The forest land-

scape is more natural and continuous and

landscape fragmentation is more recent in

Russia than in Finland. Satellite image

source: Google Earth�.

A measure of resource availability for communities 1321

� 2009 The Authors. Journal compilation � 2009 British Ecological Society, Journal of Ecology, 97, 1320–1328

1926), mostly dominated by Norway spruce (Picea abies) and Scots

pine (Pinus sylvestris).

Of the 47 study plots, each with an area of 0.5 ha, 35 plots were

located in eastern Finland and 12 in the adjacent areas in Russia. The

locations of 15 Finnish study plots were randomized within two large

areas of natural or semi-natural old-growth forest. Ten other Finnish

plots were placed randomly in other natural or semi-natural old-

growth forest stands (Rassi et al. 1992; Tikkanen & Turkulainen

1997), while the remaining 10 plots were placed in randomly selected

mature managed forest stands, in which the age of the dominant trees

was at least 80 years. Satellite images and aerial photographs were

used to locate relatively continuous natural or semi-natural

old-growth forest areas in north-western Russia. The locations of the

Russian study plots were randomized within three such forest areas.

For logistic reasons, we used one, four and seven study plots per for-

est area (Fig. 1). Some of the Russian sites had been selectively logged

or used for tar extraction in the early 20th century. As visible in

Fig. 1, forest fragmentation is nowadays severe also in north-western

Russia, but greatly intensified loggings have extended to this region

only since the late 1990s. The shortest and longest distances between

any two study plots were 0.7 and 147 km, respectively.

We classified the study plots into two categories based on the

area of the respective forest stand and its connectivity to nearby

larger areas of natural or semi-natural forest. First, 11 stands

were classified as small and isolated as they were at most 3 ha in

total area and there was no or very little natural or semi-natural

forest within a radius of 5 km. Second, 36 stands were classified

as large and well-connected; these stands were mostly larger than

50 ha and were located within large continuous expanses of

natural or semi-natural forest up to tens of km2 in area. Two

intermediate forest stands with areas of 5 and 25 ha were

included in the second class, because they were located in the

vicinity of natural or semi-natural forests (>200 ha within

5 km).

STUDY SPECIES

Aphyllophorous fungi (Basidiomycota) comprise the most significant

group of wood-decaying fungi in terms of species richness and role in

decomposition (Rayner & Boddy 1988; Renvall 1995). In boreal

forests, the dominant species groups of wood-decaying fungi are

polypores (poroid Aphyllophorales) and corticioids (corticioid

Aphyllophorales). Their taxonomy and ecology have been well-stud-

ied in Fennoscandia (e.g. Renvall 1995; Kotiranta 2001; Niemela

2005).

There are altogether 230 species of polyporous fungi and 372

species of corticioid fungi reported from Finland (Kotiranta 2001;

Niemela 2005). Here, we consider all the wood-decaying polypore

species as well as 16 species of corticioid or hydnaceous fungi

(Aphyllophorales) that are threatened, near-threatened or indicator

species (Kotiranta & Niemela 1996; Rassi et al. 2001). Many of the

species are habitat specialists and their occurrence depends on the tree

species, type (downed log, standing snag, stump, etc.), size and decay

stage of the wood, the prevailing microclimate or the presence of

other fungal species. A large fraction of the species has been classified

as threatened or near-threatened in Finland (28% of all aphyllophor-

ous and 37%of polyporous fungi (Rassi et al. 2001).

We examined the responses of the species to the amount and diver-

sity of dead wood separately for common (n = 75 in our data) and

red-listed (near-threatened, vulnerable and endangered) species

(n = 41). For simplicity, we refer to the non-red-listed species as

common species, though the prevalence of some of these species can

be rather low. Among the red-listed species, those that are known to

be restricted in Fennoscandia to decay spruce were considered as a

separate group (n = 22). A species was classified as a spruce specialist

if at least 90% of its occurrences are on spruce. The classification is

based on literature (e.g. Kotiranta & Niemela 1996; Kotiranta 2001;

Niemela 2005) and on expert knowledge.

INVENTORY OF DEAD WOOD AND FUNGAL FRUIT

BODIES

At each study site, a 50 · 100-m sample plot was delimited. All dead

trees with a minimum diameter of 5 cmwere searched thoroughly for

the focal species. Following the common practice in polypore studies

(e.g. Bader, Jansson & Jonsson 1995; Lindblad 1998; Penttila,

Siitonen & Kuusinen 2004; Junninen & Kouki 2006), we counted all

fruit bodies of a particular species on an individual tree as one

occurrence. Tree species, type (snag, log, etc.), diameter and decay

stage (classes Ia-Vb; see Hottola & Siitonen 2008) of each dead tree

were recorded.

The inventories were carried out in September 2000, July–Septem-

ber 2001 and September 2003. In the case of annually fruiting species,

the dead fruit bodies from the previous autumn have largely disap-

peared by the time the new fruit bodies are developed; hence for these

species, we counted both the dead and the living fruit bodies. Dead

perennial fruit bodies were excluded from the analyses. Monthly

mean temperature in June–September did not differ substantially

among the years, but July 2003 was unusually warm.Mean precipita-

tion varied more, and September 2000, August 2001 and June––July

2003 were drier than usual, while August 2000 and August 2003

were wetter (Finnish Meteorological Institute 2000, 2001, 2003).

Variation in weather conditions may have affected the formation of

fruit bodies of especially those species that produce annual fruit

bodies. However, fungi growing inside logs are not expected to be as

susceptible to the prevailing climatic conditions as fungi growing

inside branches and twigs or on soil (Rayner & Boddy 1988; Norden

et al. 2008).

We considered only downed logs in this study, which comprise

most of the data. Altogether we recorded 9137 downed logs and 4210

occurrences of 116 fungal species, of which 22 are classified as

near-threatened, 16 as vulnerable and three as endangered in Finland

(Rassi et al. 2001). Data for individual species are given in

Appendix S1 in Supporting Information.

A MEASURE CHARACTERIZ ING THE AMOUNT, S IZE

AND DIVERSITY OF DEAD WOOD

To characterize the amount of dead wood with a measure that

takes into account the number and volume of logs, we defined

SðxÞ ¼P

i Vxi ;where Vi is the volume of log i, and x is a parameter

that tunes the weighing between the number of logs and their volume

or size. Thus, S(0) counts the number of logs irrespective of their size,

whereas S(1) gives the total volume of logs irrespective of their num-

ber. At the limit x fi ¥, the value of S(x) is determined by the vol-

ume of the largest individual log in the sample plot.

As fungal species have dissimilar habitat requirements, the diver-

sity of different types of decaying wood is expected to affect the com-

position of polypore communities. We classified dead wood items

into n different types, and let Sj (x) measure the amount of resource in

type j. We first assigned the logs to four classes based on tree species:

spruce, pine, aspen (Populus tremula) and other deciduous trees

(mainly birch, Betula spp.). We next classified the logs in each tree

1322 J. Hottola, O. Ovaskainen & I. Hanski

� 2009 The Authors. Journal compilation � 2009 British Ecological Society, Journal of Ecology, 97, 1320–1328

species class as little decayed (decay-stage classes Ia–II) and more

decayed (decay-stage classes III–Vb) logs, yielding altogether eight

dead-wood classes.

To derive a measure accounting for the diversity of logs, we first set

x = 0, so that Sj(x) counts the number of logs of type j.Our aim was

to combine the Sj(x) values into a measureW to account for both the

total amount and the diversity of the resource. As the probability of

occurrence of a given polypore species in a sample (in our case a study

plot) will saturate asmore andmore resource is added, it is reasonable

to give more ‘value’ per log to the first logs of a given resource type

than to the additional logs. One possibility is to use a geometric series,

giving the first log the weight 1, the second log the weight y (with

0 < y < 1), the third log the weight y2, and so on. Adding up the

contributions fromS logs leads to the formula

fðS; yÞ ¼ 1� yS

1� y: eqn 1

In the neutral case of all logs having an equal weight (y = 1), eqn 1

becomes singular, but its limit coincides with the intuitive definition

f(S, 1) = S. If only the presence or absence of a resource type is

assumed tomatter (y = 0), then f(S, 0) = 1 for S > 0, andwe define

the singular case f(0, 0) = 0. Adding up the contributions from the

different resource types gives

Wðx; yÞ ¼Xn

j¼1fðSjðxÞ; yÞ: eqn 2

With this definition, themeasure has the following properties:

1 W(x, 0) counts the number of different log types indepen-

dently of parameter x.

2 W(0, 1) counts the number of logs regardless of log type.

3 W(1, 1) is the total volume of logs regardless of log type.

If x > 0 and hence Sj (x) is not simply the number of logs, the mea-

sure given by eqn 2 depends in a nonlinear manner on the unit that is

used to measure the volume of logs (e.g. m3 or mm3). To remove this

effect, we normalize the measureW(x, y) as

Wðx; yÞ ¼ SðxÞXn

j¼1fðSjðxÞ=SðxÞ; yÞ; eqn 3

where SðxÞ ¼ 1n

Pnj¼1 SjðxÞ is the average of the Sj (x) values.

To illustrate the performance of measureW(x, y), Fig. 2 shows the

values ofW(x, y) for ranges of parameters x and y in three contrast-

ing forest stands selected from the set of 47 stands included in this

study (panels a–c). As a point of reference, we used a hypothetical

average forest stand, in which both the number of logs (190) and their

volume (22 m3) per 0.5 ha correspond to the average values in our

empirical data. The contour lines in Fig. 2 show the amount of

resource as measured byW(x, y) in the focal forest stand in compari-

son with the average forest stand. Red colour indicates that the focal

forest has more resource (larger value of W(x, y)) than the average

forest, whereas blue indicates the reverse. The focal forest in panel a is

an intensively managed forest with a large number of logs (472), but

most of the logs are very small logging-residue tops and thus the total

volume is small (7 m3). Thereby this forest is ‘better than the average’

if one measures quality primarily by the number of logs (small value

of x) with little regard to their volume (Fig. 2a). On the other hand, if

forest quality is measured by the total volume of logs (x = 1), forest

A hasmuch less resources than the average forest.

Panels b and c represent two natural-forest stands, with higher than

average number of logs (B: 226, C: 161) and total volume (B: 31 m3,

C: 27 m3). Logs in forest C are somewhat larger and more diverse

than in forest B. These two forests are ‘better than the average’ almost

independently of exactly how we measure the amount and diversity

of resources. As forest C is characterized by large and diverse logs, it

attains an especially high value of W(x, y) in comparison with the

average forest when x is large (much weight for large size) and y is

small (much weight for diversity). In the case of forest B, the highest

values of W(x, y) in comparison with the average forest are attained

along a ridge running from situations with many small logs belonging

to only a few classes (small x and large y) to some large logs of many

types (large x and small y). These examples illustrate how W(x, y)

depends in a complex manner on parameters x and y, hence the very

same forest stand has a low or a high amount of resources depending

on exactly how the amount and diversity of resource types are mea-

sured. In other words, the same forest stand can have different ‘value’

for species with dissimilar resource requirements.

DATA ANALYSIS

We usedGLMs to examine the effects of the explanatory variables on

species richness and pooled abundance of fungi in the forest stands.

The model link function was log. We used the quasi-Poisson error

distribution to account for over-dispersion in the data (Venables &

Ripley 2002).

We examined how species number and pooled abundance of fungi

depend on the measure W(x, y) for different values of x and y. The

0 0.25 0.5 0.75 1 1.25 1.5 0.2

0.4

0.6

0.8

1

1.2

0 0.25 0.5 0.75 1 1.25 1.5 0.2

0.4

0.6

0.8

1

1.2

0 0.25 0.5 0.75 1 1.25 1.5 0.2

0.4

0.6

0.8

1

1.2

Parameter x Parameter x Parameter x

Par

amet

er y

(a) (b) (c)

Fig. 2. The behaviour of themeasureW(x, y) in three contrasting forest sites, selected among the 47 sites included in this study. The contour lines

show how much more (red) or less (blue) resources (dead wood) the particular forest has compared to the average of the 47 forest stands, mea-

sured usingW(x, y) for a range of parameters x and y. The forest in panel a is an intensively managed forest stand with abundance of small-sized

logs with low total volume. The forests in panels b and c are in natural state and have a high volume of deadwood. The letters indicate the param-

eter combinations that best explained the occurrence of common (C), red-listed (R), and red-listed spruce-specialist (S) species in the full material.

For further explanation see ‘Material andmethods’ and ‘Results’ sections.

A measure of resource availability for communities 1323

� 2009 The Authors. Journal compilation � 2009 British Ecological Society, Journal of Ecology, 97, 1320–1328

amount of dead wood as measured byW(x, y) was log-transformed,

because relative change in the amount of resource is biologicallymore

meaningful than the absolute change. We used the residuals from the

GLMs as a response variable in regression analyses to examine

whether the average decay stage or forest category (small and isolated

vs. large and connected) had an effect on species number or pooled

abundance of fungi in addition to the effect of W(x, y). The average

decay stage was calculated by weighing each of the seven decay-stage

classes by the total volume of logs in the corresponding class. In the

regression analyses, we included also the second-order term for the

average decay stage to allow for a nonlinear response. We calculated

semi-variograms (Cressie 1993) to examine whether the residuals

from the GLMs were spatially correlated. Spatial correlation struc-

ture could be expected because some of our study sites were located in

clusters (Fig. 1).

All statistical analyses were carried out using version 2.6.2 of the

statistical programming environment R (R Development Core Team

2008).

Results

SPECIES RICHNESS AND POOLED ABUNDANCE OF

FUNGI EXPLAINED BY THE NUMBER VS. VOLUME OF

LOGS

The volume of logs, measured by W(1, 1), explained a

relatively small proportion (17.2%) of the variance in the num-

ber of common species among the study plots and even less of

the variance in their pooled abundance (5.9%; Fig. 3a,b). In

contrast, the volume of logs accounted for more than half of

the variance in the number (53.7%) and pooled abundance

(50.9%) of red-listed species (Fig 3c,d). Considering the red-

listed spruce-specialist species, the volume of spruce logs

explained especially well the variation in species number

(59.3%) and pooled abundance (73.2%) (Fig. 3e, f).

We next examined how well species richness and pooled

abundance were explained by the measure S(x) = W(x, 1),

which characterizes the effects of the number and size of logs

but does not take into account their diversity. The results were

strikingly different for the common and red-listed species

(Fig. 4). Species number of common species was best explained

by weighing primarily the number of logs (the value of x at the

optimum, x* = 0.24; 55.0% of the deviance explained),

whereas for red-listed species the best-explaining measure was

close to the total volume of logs (x* = 0.87; 54.2% of the

deviance explained). In the case of red-listed spruce-specialist

species, the optimal measure gave disproportionate weight to

the volume of the largest logs (x* = 1.36; 59.4% of the

deviance explained). The patterns were essentially the same for

the number of species and their pooled abundance (Fig. 4).

DIVERSITY OF DEAD WOOD

Figure 5 illustrates how the deviance in the number of com-

mon and red-listed species depends on the values of x and y

5

10

15

20

25

30(a)

0

50

100

150

200

Abu

ndan

ce o

f spe

cies

Abu

ndan

ce o

f spe

cies

(b)

0 10 20 30 40 50 60 0 10 20 30 40 50 60

0 10 20 30 40 50 60 0 10 20 30 40 50 60

0

2

4

6

8

10

Num

ber

of s

peci

esN

umbe

r of

spe

cies

Num

ber

of s

peci

es

(c)

0

5

10

15

20

25

Abu

ndan

ce o

f spe

cies

0

5

10

15

20

25

(d)

0 5 10 15 20 25 30 3501234567

Volume of logs (m3)

0 5 10 15 20 25 30 35

Volume of logs (m3)

(e) (f)

Fig. 3. Species number (left) and pooled abundance (right) of common (panels a–b), red-listed (c–d) and red-listed spruce-specialist species (e–f)

against the volume of all (a–d) or spruce logs (e–f). Circles are for large and connected forests, triangles for small and isolated forest fragments.

Lines indicate the fitted response based on theGLMand the 95% confidence limits.

1324 J. Hottola, O. Ovaskainen & I. Hanski

� 2009 The Authors. Journal compilation � 2009 British Ecological Society, Journal of Ecology, 97, 1320–1328

when explained by W(x, y). The measures that best explained

the numbers of common, red-listed and red-listed spruce-spe-

cialist species were W(0.29, 0.42) (63.1% of the deviance

explained), W(1.05, 0.72) (57.6%) and W(1.39, 0.25) (60.2%).

Thus, taking into account the diversity of dead wood helped

somewhat to explain the variation in the number of common

species, but not that of red-listed species. The corresponding

percentages of the deviance explained in the case of pooled

abundance of fungi were 70.7%, 56.2%and 76.2%.

To elucidate the results concerning the measure W(x*, y*),

we have indicated in Fig. 2 the optimal values (x*, y*) that

best explained the numbers of common (C), red-listed (R) and

red-listed spruce-specialist species (S). The intensively man-

aged forest stand A (Fig. 2a) was judged to be of better-than-

average quality for the common species, but of poor quality

for the red-listed species. This is consistent with the observed

species numbers in this particular stand, which had 21 com-

mon species (compared with the average of 18.6 among all the

stands) but not a single red-listed species. Forest stands B and

C were judged to be of higher-than-average quality for all

species, but especially for the red-listed ones. In these stands,

the observed number of common species equalled the average

or was somewhat higher (18 for B and 24 for C), while the

numbers of red-listed species (8 and 9 for B and C) and red-

listed spruce-specialist species (6 and 5) weremuch greater than

the average (4.5 and 1.9, respectively).

Species number and pooled abundance of common species

attained their maxima in forest stands in which the average

decay stage was low. The corresponding maxima for red-listed

and red-listed spruce-specialist species were in intermediate

average decay stages, implying a continuum of dead wood in

all decay-stage classes. None of these results was significant,

however, and the effect sizes were small.

AREA AND CONNECTIV ITY OF THE FOREST STANDS

We examined the effect of forest stand category, whether iso-

lated or connected, after removing the effect of the amount

and quality of the resource with the model involving the

best-explaining measureW(x*, y*). In the case of the common

(a) (b)

Fig. 4. Deviance in species number (panel a) and pooled abundance (panel b) explained by the measure SðxÞ ¼P

i Vxi for different values of

parameter x (see text). The solid line is for common species, the dashed line for red-listed species and the dotted line for red-listed spruce-specialist

species.

(a) (b)

Fig. 5. Isoclines of the deviance in species number of common (a) and red-listed (b) species explained by the measure W(x, y) in relation to the

values of parameters x (weighing for the size of logs) and y (weighing for the diversity of logs).

A measure of resource availability for communities 1325

� 2009 The Authors. Journal compilation � 2009 British Ecological Society, Journal of Ecology, 97, 1320–1328

species, there was no difference between the small and isolated

vs. large and connected forest stands (adjusted r2 = )0.022;F1,45 = 0.0, P = 0.976). In contrast, there were significantly

more red-listed species (adjusted r2 = 0.206; F1,45 = 13.0,

P = <0.001) and especially red-listed spruce-specialist species

(adjusted r2 = 0.171; F1,45 = 10.5, P = 0.002) in large and

connected than in small and isolated forest stands (Fig. 6; see

also Fig. 3).

There was no statistically significant spatial correlation

structure in the model residuals for species number or pooled

abundance of fungi in any of the species groups.

Discussion

The measureW(x, y) can be applied to types of resources that

occur as discrete objects with variation in the size and quality

of individual resource items as is often the case with resources

for most omnivorous consumers. Previous studies on single

consumer species have employed resource selection functions

to quantify which components of habitat or resource quality

are relevant for a particular species (e.g. Lemaıtre & Villard

2005). Given the availability of sufficient data, resource selec-

tion functions can provide more detailed information of the

resource requirements of a particular species. Our measure

W(x, y) can also be applied in the studies of single species, but

it is particularly suitable for integrating species-specific

resource selection functions for multiple species and can thus

be used to study how the composition of a species community

depends on the amount and characteristics of the available

resources. The measure W(x, y) describes the structure of

resources for a community in a concise and simple way, and it

is applicable also to situations where the data are insufficient

for detailed species-specific analyses.

In much of the existing literature, the amount of resources

for wood-decaying fungi has beenmeasured simply as the total

volume of deadwood (e.g. Berglund& Jonsson 2005; Junninen

et al. 2006), partly because volume is the traditional quantity

used in forestry, timber trading and forest research. One could

also argue that volume is a natural measure from the viewpoint

of wood-decaying fungi, as it describes the amount of biomass

available for decomposition. However, our results clearly dem-

onstrate that volume is not the best measure for all groups of

species. In particular, non-red-listed species, most of which are

common, were most numerous in forests with lots of logs

regardless of their total volume. In contrast, the number of

red-listed species was much influenced by log size, and hence

by the volume of dead wood. These results are consistent with

previous findings suggesting that threatened species are often

confined to specific ecological conditions and are often found

only on large logs (e.g. Bader, Jansson & Jonsson 1995;

Renvall 1995; Lindblad 1998; Kruys et al. 1999; Tikkanen

et al. 2006).

The relative importance of the number and size (volume) of

logs can be expected to depend on the life cycle and the ecologi-

cal requirements of particular species. In the dispersal phase,

the probability of a spore hitting a given log can be assumed to

be proportional to surface area, which scales as volume raised

to power 2 ⁄3. In terms of the establishment of mycelia, the role

of the log size is twofold. First, the size of a log affects the

physiochemical conditions and the rate of decay, larger logs

providingmore stable conditions and a larger variety of micro-

habitats required by specialized and slow-growing species

(Rayner&Boddy 1988;Norden et al. 2004; Boddy, Frankland

& van West 2008). Second, log size is likely to make a differ-

ence through interspecific competition, which is most intensive

in the largest logs (Boddy 2000; Woodward & Boddy 2008).

We hypothesize that interspecific competition is the likely rea-

son for some common species being less numerous and abun-

dant in species-rich natural forests than in species-poor

managed forests. Finally, considering the reproductive phase

of the life cycle, the amount of resources available for the pro-

duction of sufficient mycelial biomass to enable the formation

(a) (b) (c)

Fig. 6. Residuals for number of common (a), red-listed (b) and red-listed spruce-specialist (c) species after taking into account the effect of

resource availability as measured by W(x*, y*). The forest stands have been divided into small and isolated (n = 11) vs. large and connected

stands (n = 36).

1326 J. Hottola, O. Ovaskainen & I. Hanski

� 2009 The Authors. Journal compilation � 2009 British Ecological Society, Journal of Ecology, 97, 1320–1328

of reproductive organs and to induce spore production can be

considered to be roughly proportional to the volume of the log

(Moore et al. 2008).

Kruys& Jonsson (1999) andNorden et al. (2004) have eluci-

dated the importance of small-diameter logs for species rich-

ness of wood-decaying fungi. Comparing the equal volumes of

small- and large-diameter logs, the high number and large

pooled surface area of small-diameter logs may support many

fungal species with a short-life cycle, which need only a small

mycelial biomass for fruiting and which are inferior in compe-

tition. The majority of such wood-decaying species presum-

ably consists of ascomycetes and corticioid basidiomycetes.

Species able to grow on small logs have generally declined less

than species that are specialized on large logs, thus the contri-

bution of small logs to total species richness of basidiomycetes

is mainly through common species.

Red-listed polypore species have been shown to use mostly

intermediate or well-decayed logs rather than little-decayed

logs (Bader, Jansson & Jonsson 1995; Renvall 1995; Lindblad

1998; Kruys et al. 1999; Tikkanen et al. 2006). We did not

detect a strong signal of dissimilar responses of common and

red-listed species to average decay stage, most likely because

stand-level average values embrace little of all the relevant

dead-wood features in a forest.

A more diverse set of dead wood means a greater range of

resources and niches, which is expected to increase the number

of species present in a fungal community.We found that diver-

sity of dead wood clearly increased the number of common

species, but had only aminor effect on the number of red-listed

species. In the case of common species, a species that is specific

to a given resource type is likely to be present when even a

small amount of that resource is available, and hence the num-

ber of species is largely determined by the range of different

resource types. In contrast, the prevalence of red-listed species

was so low in our small study plots that their number was

mainly determined by the availability of large logs, while the

type of logs influenced which particular red-listed species were

present. Furthermore, different host-tree species are not

equally important as substrates, but spruce and pine host more

red-listed polypores and corticioids than deciduous tree species

in Finland (Kotiranta & Niemela 1996; Ministry of the Envi-

ronment 2000; Rassi et al. 2001). Our result on the diversity of

logs explaining little of the occurrence of red-listed species is

consistent with observations concerning red-listed species

requiring certain kinds of substrates instead of a diverse

set of substrates. In the case of red-listed spruce-specialist

species, lack of response to diversity of logs is additionally

explained by our classification including only two relevant

resource classes, little and more decayed spruce logs, for these

species.

Our results indicate a clear difference between the common

vs. red-listed species in relation to how the number of logs,

their size distribution and diversity influence the number of

fungal species and their pooled abundance. The common and

red-listed species differ also in terms of the effect of the sur-

rounding forest landscape on the occurrence of species. The

quality of the surrounding landscape had no effect on the num-

bers of common species, but there was a significant and strong

effect on the numbers of red-listed species. Apparently, the

threshold condition for the persistence (Hanski & Ovaskainen

2000) of many red-listed species is not met regionally in forest

landscapes with little natural forests remaining, hence these

species are likely to be absent from small fragments of favour-

able habitat represented by our small and isolated study for-

ests. We hypothesize that the most important traits making it

difficult for red-listed species to occur in isolated fragments of

natural forest are dispersal limitation and ecological specializa-

tion, the latter leading to low overall prevalence which elevates

the effect of demographic stochasticity. Our results suggest

that the most cost-effective means of preventing further

declines of the red-listed species are to increase the numbers of

large downed logs through restoration and biodiversity-ori-

entedmanagement in the vicinity of existing areas of natural or

natural-like forests.

Acknowledgements

We thank Mariko Lindgren and Juha Siitonen for their help in planning the

field work, Olli Manninen and Timo Kosonen for assisting in the field, and

Timo J. Hokkanen, Raimo Heikkila and Olli Turunen for help with practical

arrangements. Part of the data was collected while J.H. was employed by the

Finnish Forest Research Institute in a project led by Juha Siitonen. We thank

Jari Oksanen for his advice on semi-variograms. Evgeniy Meyke and Sami

Ojanen assisted in the preparation of Fig. 1. Comments byBengt-Gunnar Jons-

son helped improve the manuscript. J.H. was financially supported by Ella and

Georg Ehrnrooth Foundation and the Finnish Society of Forest Science, O.O.

and I.H. by theAcademy of Finland (grants 213457 and 211173).

References

Ahti, T., Hamet-Ahti, L. & Jalas, J. (1968) Vegetation zones and their sections

in northwestern Europe.Annales Botanici Fennici, 5, 169–211.

Bader, P., Jansson, S. & Jonsson, B.G. (1995) Wood-inhabiting fungi and sub-

stratum decline in selectively logged boreal spruce forests. Biological Conser-

vation, 72, 355–362.

Berglund, H. & Jonsson, B.G. (2005) Verifying an extinction debt among

lichens and fungi in northern Swedish boreal forests. Conservation Biology,

19, 338–348.

Boddy, L. (2000) Interspecific combative interactions between wood-decaying

basidiomycetes. FemsMicrobiology Ecology, 31, 185–194.

Boddy, L., Frankland, J.C. & vanWest, P., eds (2008) Ecology of Saprotrophic

Basidiomycetes. Elsevier, Amsterdam.

Cajander, A.K. (1926) The theory of forest types. Acta Forestalia Fennica, 29,

1–108.

Cressie, N.A.C. (1993) Statistics for Spatial Data. John Wiley & Sons,

NewYork.

FinnishMeteorological Institute (2000). Ilmastokatsaus, Rep. No. 6-9. Finnish

Meteorological Institute, Helsinki.

FinnishMeteorological Institute (2001). Ilmastokatsaus, Rep. No. 6-9. Finnish

Meteorological Institute, Helsinki.

FinnishMeteorological Institute (2003). Ilmastokatsaus, Rep. No. 6-9. Finnish

Meteorological Institute, Helsinki.

Gardenfors, U., ed. (2000) The 2000 Red-List of Swedish Species. Swedish Spe-

cies Information Centre, Uppsala.

Gilbert, G.S., Ferrer, A. & Carranza, J. (2002) Polypore fungal diversity and

host density in a moist tropical forest. Biodiversity and Conservation, 11,

947–957.

Groven, R., Rolstad, J., Storaunet, K.O. & Rolstad, E. (2002) Using forest

stand reconstructions to assess the role of structural continuity for late-suc-

cessional species. Forest Ecology andManagement, 164, 39–55.

Gu, W.D., Heikkila, R. & Hanski, I. (2002) Estimating the consequences of

habitat fragmentation on extinction risk in dynamic landscapes. Landscape

Ecology, 17, 699–710.

Hanski, I. (2005) The Shrinking World: Ecological Consequences of Habitat

Loss. International Ecology Institute, Oldendorf.

A measure of resource availability for communities 1327

� 2009 The Authors. Journal compilation � 2009 British Ecological Society, Journal of Ecology, 97, 1320–1328

Hanski, I. & Ovaskainen, O. (2000) The metapopulation capacity of a frag-

mented landscape.Nature, 404, 755–758.

Hottola, J. & Siitonen, J. (2008) Significance of woodland key habitats for poly-

pore diversity and red-listed species in boreal forests. Biodiversity and Con-

servation, 17, 2559–2577.

Jonsson, M.T., Edman, M. & Jonsson, B.G. (2008) Colonization and extinc-

tion patterns of wood-decaying fungi in a boreal old-growth Picea abies for-

est. Journal of Ecology, 96, 1065–1075.

Jonsson, B.G., Kruys, N. &Ranius, T. (2005) Ecology of species living on dead

wood – lessons for deadwoodmanagement. Silva Fennica, 39, 289–309.

Junninen, K. & Kouki, J. (2006) Are woodland key habitats in Finland hot-

spots for polypores (Basidiomycota)? Scandinavian Journal of Forest

Research, 21, 32–40.

Junninen, K., Simila, M., Kouki, J. & Kotiranta, H. (2006) Assemblages

of wood-inhabiting fungi along the gradients of succession and natural-

ness in boreal pine-dominated forests in Fennoscandia. Ecography, 29,

75–83.

Keller, M., Palace, M., Asner, G.P., Pereira, R. & Silva, J.N.M. (2004) Coarse

woody debris in undisturbed and logged forests in the eastern Brazilian

Amazon.Global Change Biology, 10, 784–795.

Kotiranta, H. (2001) The Corticiaceae of Finland. Publications in Botany from

the University of Helsinki, 32, 1–29.

Kotiranta, H. & Niemela, T. (1996) Uhanalaiset Kaavat Suomessa, 2nd revised

edn. Finnish Environment Institute, Helsinki.

Kruys, N. & Jonsson, B.G. (1999) Fine woody debris is important for species

richness on logs inmanaged boreal spruce forests of northern Sweden.Cana-

dian Journal of Forest Research-Revue Canadienne De Recherche Forestiere,

29, 1295–1299.

Kruys, N., Fries, C., Jonsson, B.G., Lamas, T. & Stal, G. (1999) Wood-

inhabiting cryptogams on dead Norway spruce (Picea abies) trees in

managed Swedish boreal forests. Canadian Journal of Forest Research,

29, 178–186.

Lemaıtre, J. & Villard, M.A. (2005) Foraging patterns of pileated woodpeckers

in amanagedAcadian forest: a resource selection function.Canadian Journal

of Forest Research, 35, 2387–2393.

Lindblad, I. (1998) Wood-inhabiting fungi on fallen logs of Norway spruce:

relations to forest management and substrate quality.Nordic Journal of Bot-

any, 18, 243–255.

Ministry of the Environment (2000) Forest Protection in Southern Finland and

Ostrobothnia. The Finnish Environment 437. Edita, Helsinki.

Moore, D., Gange, A.C., Gange, E.G. & Boddy, L. (2008) Fruit bodies: their

production and development in relation to environment. Ecology of Sapro-

trophic Basidiomycetes (eds L. Boddy, J.C. Frankland & P. van West), pp.

79–103. Elsevier, Amsterdam.

Niemela, T. (2005) Polypores, Lignicolous Fungi. Helsinki University Press,

Helsinki.

Niemela, T., Renvall, P. & Penttila, R. (1995) Interactions of fungi at late stages

of wood decomposition.Annales Botanici Fennici, 32, 141–152.

Norden, B. & Appelqvist, T. (2001) Conceptual problems of ecological

continuity and its bioindicators. Biodiversity and Conservation, 10, 779–

791.

Norden, B., Ryberg, M., Gotmark, F. & Olausson, B. (2004) Relative

importance of coarse and fine woody debris for the diversity of wood-

inhabiting fungi in temperate broadleaf forests. Biological Conservation,

117, 1–10.

Norden, B., Gotmark, F., Ryberg, M., Paltto, H. & Allmer, J. (2008) Partial

cutting reduces species richness of fungi on woody debris in oak-rich forests.

Canadian Journal of Forest Research-Revue Canadienne De Recherche

Forestiere, 38, 1807–1816.

Penttila, R., Siitonen, J. & Kuusinen,M. (2004) Polypore diversity in managed

and old-growth boreal Picea abies forests in southern Finland. Biological

Conservation, 117, 271–283.

Penttila, R., Lindgren, M., Miettinen, O., Rita, H. & Hanski, I. (2006) Conse-

quences of forest fragmentation for polyporous fungi at two spatial scales.

Oikos, 114, 225–240.

RDevelopment Core Team (2008)R:A Language and Environment for Statisti-

cal Computing. R Foundation for Statistical Computing, Vienna.

Rassi, P., Lindholm, T., Salminen, P. & Tanninen, T., eds (1992) Vanhojen

Metsien Suojelu Valtion Mailla Etela-Suomessa. Ministry of Environment,

Helsinki.

Rassi, P., Alanen, A., Kanerva, T. &Mannerkoski, I., eds. (2001) Suomen Laj-

ien Uhanalaisuus 2000. Ministry of the Environment and Finnish Environ-

ment Institute, Helsinki.

Rayner, A.D.M. & Boddy, L. (1988) Fungal Decomposition of Wood: Its Biol-

ogy and Ecology. JohnWiley & Sons, Bath.

Renvall, P. (1995) Community structure and dynamics of wood-rotting Basid-

iomycetes on decomposing conifer trunks in northern Finland. Karstenia,

35, 1–51.

Rolstad, J., Saetersdal, M., Gjerde, I. & Storaunet, K.O. (2004) Wood-decay-

ing fungi in boreal forest: are species richness and abundances influenced by

small-scale spatiotemporal distribution of dead wood? Biological Conserva-

tion, 117, 539–555.

Samuelsson, J., Gustafsson, L. & Ingelog, T. (1994) Dying and Dead Trees: A

Review of Their Importance for Biodiversity. Swedish Threatened Species

Unit, Uppsala.

Siitonen, J. (2001) Forest management, coarse woody debris and saproxylic

organisms: Fennoscandian boreal forests as an example. Ecological Bulle-

tins, 49, 11–41.

Siitonen, J., Penttila, R. & Kotiranta, H. (2001) Coarse woody debris,

polyporous fungi and saproxylic insects in an old-growth spruce forest

in Vodlozero National Park, Russian Karelia. Ecological Bulletins, 49,

231–242.

Tikkanen, O.-P. & Turkulainen, S., eds (1997) Natura 2000 Pohjois-Karjalan

ita – ja Pohjoisosissa. Pohjois-Karjalan Luonnonsuojelupiirin Esitys. Osa 1:

Kartat. North Karelia district of the Finnish Association for Nature Con-

servation, Joensuu.

Tikkanen, O.P., Martikainen, P., Hyvarinen, E., Junninen, K. & Kouki, J.

(2006) Red-listed boreal forest species of Finland: associations with forest

structure, tree species, and decaying wood. Annales Zoologici Fennici, 43,

373–383.

Venables, W.N. & Ripley, B.D. (2002) Modern Applied Statistics With S., 4th

edn. Springer, NewYork.

Woodward, S. & Boddy, L. (2008) Interactions between saprotrophic fungi.

Ecology of Saprotrophic Basidiomycetes (eds L. Boddy, J.C. Frankland & P.

vanWest), pp. 125–141. Elsevier, Amsterdam.

Received 18April 2009; accepted 25August 2009

Handling Editor: Jeremy Burdon

Supporting Information

Additional supporting information may be found in the online ver-

sion of this article:

Appendix S1. The focal species, the number of sites they occurred in

and the total number of occurrences.

As a service to our authors and readers, this journal provides support-

ing information supplied by the authors. Such materials may be

re-organized for online delivery, but are not copy-edited or typeset.

Technical support issues arising from supporting information (other

thanmissing files) should be addressed to the authors.

1328 J. Hottola, O. Ovaskainen & I. Hanski

� 2009 The Authors. Journal compilation � 2009 British Ecological Society, Journal of Ecology, 97, 1320–1328