ORIGINAL ARTICLE Aquatic community structure across an Andes-to-Amazon fluvial gradient Nathan K. Lujan 1 * † , Katherine A. Roach 1 , Dean Jacobsen 2 , Kirk O. Winemiller 1 , Vanessa Meza Vargas 3 , Vania Rimarach ın Ching 3 and Jerry Arana Maestre 3 1 Department of Wildlife and Fisheries, Texas A&M University, College Station, TX, USA, 2 Freshwater Biological Section, Department of Biology, University of Copenhagen, Copenhagen, Denmark, 3 Natural History Museum, University of San Marcos, Lima, Peru *Correspondence: Nathan K. Lujan, Department of Natural History, Royal Ontario Museum, 100 Queen’s Park, Toronto, ON, Canada. E-mail: [email protected]†Present address: University of Toronto, 25 Willcocks St., Toronto, ON, Canada ABSTRACT Aim Little is known about factors affecting the elevational and longitudinal zonation of tropical Andean stream communities. We investigated epilithon, macroinvertebrate and fish assemblages along a 4100-m elevational–longitudi- nal gradient in an Andean headwater of the Amazon Basin. We interpret our results within the context of environmental factors, emphasizing temperature, as well as ecological theory relating shifts in metazoan functional feeding groups to shifts in basal resources along the fluvial continuum. Location Araza-Inambari-Madre de Dios watershed, south-eastern Peru. Methods We sampled water physicochemistry, epilithon and macroinverte- brate diversity and abundance, and fish diversity at 18 main-stem and 14 tribu- tary sites from high puna grasslands (4300 m a.s.l.) to Amazon Basin lowlands (200 m a.s.l.). Results Water physicochemical parameters and the taxonomic and ecological structure of invertebrate and fish assemblages displayed mostly nonlinear responses to elevation: water temperature and percentage of macroinvertebrate taxa identified as leaf shredders had U-shaped responses; dissolved oxygen and percentage of macroinvertebrate taxa identified as grazers had hump-shaped responses. Epilithon richness increased slightly with elevation whereas macroin- vertebrate and fish richness decreased. Main conclusions Elevational gradients in physicochemical parameters are insufficient to explain abrupt and nonlinear shifts in community taxonomic and functional structure. Rather, trophic interactions, including predation and longitudinal turnover in basal food resources, seem to exert a stronger influ- ence on the distributions of Andean aquatic organisms. A steep elevational decline in relative taxonomic diversity of leaf-shredding (versus algae-grazing) insects supports the hypothesis that temperature affects the functional compo- sition of insect assemblages via its influence on microbial decomposition rates. This relationship, and the distributions of several insect and fish species across narrow elevational bands, suggests that Andean stream communities may be sensitive to global warming. Placer mining and road building impacts have already altered stream community structure, including the absence of many benthic species from low-elevation habitats. Keywords Anthropogenic impacts, fluvial gradient, global warming, longitudinal connec- tivity, mining, Neotropics, river continuum concept, sediment, shredders. ª 2013 John Wiley & Sons Ltd http://wileyonlinelibrary.com/journal/jbi 1 doi:10.1111/jbi.12131 Journal of Biogeography (J. Biogeogr.) (2013)
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
ORIGINALARTICLE
Aquatic community structure acrossan Andes-to-Amazon fluvial gradientNathan K. Lujan1*†, Katherine A. Roach1, Dean Jacobsen2,
Kirk O. Winemiller1, Vanessa Meza Vargas3, Vania Rimarach�ın Ching3 and
Jerry Arana Maestre3
1Department of Wildlife and Fisheries, Texas
A&M University, College Station, TX, USA,2Freshwater Biological Section, Department of
bidity measurements were taken, and other measurements
were taken repeatedly at 5-min intervals over c. 1 h, with
Figure 1 Map of the Araz�a-Inambariwatershed in south-eastern Peru showing
the location, stream order, and elevationalzone of survey sites S1–S32. Sites S33–S41were located in tributaries of the Madre deDios north of Puerto Maldonado (not
illustrated).
Journal of Biogeographyª 2013 John Wiley & Sons Ltd
3
Stream ecology from Andes to Amazon
Table
1Geographical,bioticandphysicochem
ical
datagathered
ateach
of18
main-stem
and14
tributary
sitesin
theAraz� a-Inam
bari-Madre
deDioswatershed,south-eastern
Peru.
Site
Elevation
(ma.s.l.)
Latitude
(DD)
Longitude
(DD)
M/T/O
Taxon
Richness
MI#
Turb.
(FTU)
Tem
p.
(ºC)
pH
SpCond
(lScm
�1)
Sal.
(ppt)
DO
(%sat.)
DO
(mgmL�1)
DIN
(mgL�1)
SRP
(mgL�1)
Chlor.a:
water
(mgm
�3)
Chlor.a:
benthic
(mgm
�2)
EMI
F
Puna
S14304
�13.6293
�71.0705
M48
60
540
2.7
11.9
6.6
87.4
0.03
46.0
5.0
0.34
0.18
0.89
1.59
S23958
�13.6047
�71.0517
M52
110
886
0.0
11.8
7.5
104.5
0.04
61.9
6.7
0.66
0.17
1.16
3.77
S33516
�13.5748
�71.0264
M36
90
3030
0.2
9.9
7.5
98.7
0.04
64.3
7.3
0.39
0.07
1.34
4.15
Upper
cloud
forest
S43051
�13.5830
�70.9871
M44
110
525
2.2
8.1
7.5
0.0
0.03
66.9
7.9
0.24
0.10
0.62
8.67
S52630
�13.5914
�70.9565
M40
80
495
25.8
12.2
7.8
178.4
0.08
73.6
7.9
0.44
3.04
1.07
14.30
S62630
�13.5911
�70.9570
T43
92
375
7.9
14.6
7.7
135.5
0.06
80.2
8.2
1.05
0.37
0.62
10.40
S72630
�13.5914
�70.9559
T37
130
1395
–9.8
8.1
126.0
0.05
77.3
8.8
––
––
Lower
cloud
forest
S82218
�13.5572
�70.8996
M31
82
960
33.6
13.7
7.7
220.4
0.10
82.6
8.6
0.65
3.12
0.71
23.39
S91966
�13.5281
�70.8967
T–
–1
593
–13.0
8.1
109.8
0.04
83.2
8.8
––
––
S10
1721
�13.5051
�70.8998
T42
162
593
0.0
14.9
7.6
84.9
0.03
87.7
8.9
0.37
0.98
0.53
38.51
S11
1587
�13.4813
�70.8887
M–
121
178
84.0
16.4
8.1
164.3
0.07
91.4
9.0
0.61
0.19
1.78
1.73
S12
1334
�13.4411
�70.9024
T26
124
315
0.3
15.0
7.6
83.6
0.03
90.1
9.1
0.09
0.66
0.53
0.09
S13
1253
�13.4271
�70.9043
M46
123
1232
33.8
15.1
8.0
149.7
0.06
86.3
8.7
1.06
0.22
2.14
7.04
S14
1118
�13.3964
�70.9000
T–
–4
––
16.9
7.8
57.5
0.02
87.6
8.5
––
––
Piedmont
S15
842
�13.3247
�70.8428
M48
147
712
28.1
19.5
7.8
93.3
0.03
88.2
8.1
0.06
1.82
2.37
11.49
S16
767
�13.2957
�70.7933
T50
169
194
0.7
18.8
7.6
29.1
0.00
85.4
8.0
0.51
0.17
1.07
13.49
S17
700
�13.2634
�70.7801
M44
1713
741
98.0
22.5
7.6
85.4
0.03
84.4
7.3
0.66
0.33
0.59
7.49
S18
666
�13.2244
�70.7601
T–
2117
225
––
––
––
––
––
–
S19
604
�13.2182
�70.7212
M32
1417
488
118.0
22.1
7.9
91.2
0.03
88.8
7.8
0.83
1.20
2.97
88.00
S20
482
�13.1857
�70.6266
T39
1210
164
13.0
24.9
7.6
65.3
0.02
71.9
6.0
0.25
0.15
0.09
3.13
S21
478
�13.1860
�70.6187
T37
1810
907.9
22.7
8.3
67.6
0.02
91.4
7.9
0.80
0.90
1.36
6.13
S22
472
�13.1915
�70.5543
T–
194
297
––
––
––
––
––
–
S23
429
�13.2090
�70.5449
T41
2116
208
2.4
––
––
––
0.07
0.20
35.39
37.83
S24
419
�13.2090
�70.5445
M36
1911
387
64.6
24.4
7.8
95.6
0.04
80.8
6.8
1.05
1.76
–122.74
Lowland
(Inam
bari)
S25
397
�13.1755
�70.4945
T9
208
433
0.2
––
––
––
0.57
5.56
0.00
7.40
S26
358
�13.1736
�70.3829
T18
1412
193
168.7
21.2
5.0
7.7
0.00
18.9
1.7
––
0.00
0.82
S27
348
�13.1934
�70.3900
M–
–11
––
25.6
8.0
340.0
0.17
78.7
6.4
––
––
S28
279
�12.8904
�70.3395
M–
728
30178.0
––
––
––
0.13
5.46
4.45
0.45
S29
231
�12.7867
�70.0536
M–
–2
–313.0
23.1
7.1
49.4
0.01
62.0
5.3
0.16
6.94
0.00
3.59
S30
221
�12.7689
�70.0000
M–
–18
––
––
––
––
––
––
S31
220
�12.7851
�70.0099
M–
–13
––
––
––
––
––
––
S32
200
�12.7608
�69.8476
M–
819
15479.3
22.6
6.7
47.9
0.01
70.0
6.0
0.66
2.06
34.27
0.00
Journal of Biogeographyª 2013 John Wiley & Sons Ltd
4
N. K. Lujan et al.
average values being reported. To minimize diel variation in
solar insolation, most measurements were taken between
12:00 and 16:00 h. Dissolved nutrient concentrations were
measured by collecting water samples in polyethylene bottles
that had been repeatedly rinsed with stream water, filtering
the water through Whatman GF/F filters (Whatman Inc.,
Piscataway, NJ, USA), and then using Hach colorimetric test
kits and a Hach DR 2800 mass spectrophotometer (Hach
Co., Loveland, CO, USA) to obtain estimates. Water-
column and benthic chlorophyll a were measured following
the methods of Wetzel & Likens (1991). Briefly, triplicate
water samples were collected in polyethylene bottles that had
been rinsed with stream water and filtered through Whatman
GF/C filters. Triplicate sediment samples were taken using a
small plastic Petri dish (5 cm diameter 9 1.3 cm height)
and a spatula. For extraction, filters were immediately placed
into individual dark vials with 90% ethanol for 24 h. Chlo-
rophyll a concentration was measured spectrophotometrically
and corrected for pheophytin by subtracting absorbances
after addition of 0.1 n HCl.
Epilithon assemblages were sampled qualitatively by using
a toothbrush to dislodge detritus and attached algae from
the surface of a rock arbitrarily selected from flowing water,
then rinsing the loosened epilithon into a jar with 10% for-
malin. Macroinvertebrates were sampled quantitatively using
a Surber sampler (200 cm2, 0.200 mm mesh, five replicates
per site for a total of 1000 cm2), and qualitatively using a
dipnet and targeted searches of all obvious habitat types: the
samples were then preserved in 70% ethanol. Fishes were
sampled using combinations of electrofishing, seines, dipnets,
gillnets and hook-and-line. Fish sampling effort was similar
across sites (i.e. 1–2 h of electrofishing at all sites with fish,
combined with similar periods of gill-netting at lowland
sites) but was not standardized, with the exception of sites
S28–S32 where these methods were supplemented by 58
person-hours of fishing with hook and line. Specimens were
euthanized by immersion in a 1% solution of tricaine
methanesulfonate (MS-222), preserved in 10% formalin, and
later transferred to 70% ethanol and catalogued at museums
affiliated with universities in Peru (Museo de Historia Natu-
ral, Universidad Nacional Mayor de San Marcos) and the
USA (Auburn University Museum). Specimens were counted
and identified using microscopes and a variety of keys and
checklists.
Data analysis
We used regression analysis to examine how physicochemical
parameters changed with elevation for main-channel (par-
tially affected) and tributary (relatively unaffected) sites.
Effects of both elevation and stream type (tributary versus
main channel) on species richness of epilithic algae, macroin-
vertebrates and fishes were examined using analysis of
covariance (ANCOVA). We also compared assemblage struc-
ture between tributary and main-channel sites using non-
metric multidimensional scaling (NMDS) analysis of eitherTable
1Continued Site
Elevation
(ma.s.l.)
Latitude
(DD)
Longitude
(DD)
M/T/O
Taxon
Richness
MI#
Turb.
(FTU)
Tem
p.
(ºC)
pH
SpCond
(lScm
�1)
Sal.
(ppt)
DO
(%sat.)
DO
(mgmL�1)
DIN
(mgL�1)
SRP
(mgL�1)
Chlor.a:
water
(mgm
�3)
Chlor.a:
benthic
(mgm
�2)
EMI
F
Lowland(M
dD)
S33
272
�11.7833
�69.1852
O–
–15
––
21.6
7.3
94.8
0.04
78.4
6.9
––
––
S34
248
�12.2771
�69.1524
O–
–29
––
22.7
4.9
6.0
0.00
61.6
5.3
––
––
S35
268
�11.4090
�69.4738
O–
1723
193
––
––
––
––
––
–
S36
264
�11.4321
�69.3415
O–
–3
––
––
––
––
––
––
S37
262
�11.5050
�69.3008
O–
–8
––
––
––
––
––
––
S38
256
�11.4631
�69.3064
O–
–10
––
––
––
––
––
––
S39
242
�11.0106
�69.5557
O39
2917
282
6.8
––
––
––
0.05
2.59
0.00
0.05
S40
225
�10.9454
�69.5723
O–
–21
––
22.2
7.0
381.2
0.19
25.9
2.3
––
––
S41
206
�12.7851
�70.0099
O–
1831
456.9
22.8
6.2
26.0
0.00
61.0
5.3
0.91
1.40
30.97
0.00
DD,decim
aldegrees;DIN
,dissolved
inorganic
nitrogen;DO,dissolved
oxygen;E,epilithon;F,fish;M,mainstem
oftheAraz� a-Inam
bari;MdD,Madre
deDios;MI,macroinvertebrate;MI#,number
of
macroinvertebrate
individualsper
1000
cm2;m
a.s.l.,
metresabove
sealevel;O,outsideofAraz�a-Inam
bariwatershed;Sal.,
salinity;
SpCond,specificconductivity;
SRP,soluble
reactive
phosphorous;T,
tributary
oftheAraz� a-Inam
baririver;Tem
p.,temperature;Turb.,turbidity.
Journal of Biogeographyª 2013 John Wiley & Sons Ltd
5
Stream ecology from Andes to Amazon
absolute (macroinvertebrates and fish) or relative (algae)
abundance data with Sørensen (Bray–Curtis) distance mea-
sures. All data were log10(n + 1)-transformed in order to
approximate a normal distribution. Analysis of similarity
(ANOSIM) was used to test for significant differences in
epilithon assemblage structure between sites without (S1–S7)
and with (S8–S26) fish; macroinvertebrate assemblage struc-
ture between sites without (S1–S7) and with (S8–S32) fish;
and macroinvertebrate and fish assemblage structure (sepa-
rately) between sites upstream (S1–S26) and downstream
(S27–S32) of the onset of intense placer mining. For sites
where both assemblage and environmental data were col-
lected, canonical correspondence analysis (CCA) was used to
explore relationships between epilithon, macroinvertebrate
and fish assemblage structure and physicochemical data.
Multivariate analyses were conducted using the programs
pc-ord 5.21 and primer-e 6.
Predictions made by the RCC regarding distributional
responses of macroinvertebrate functional feeding groups
(FFGs) to availability of alternate food sources (i.e. coarse al-
lochthonous detritus versus autochthonous benthic algae)
were tested by comparing the taxonomic richness of shred-
ders versus scrapers as a percentage of total macroinverte-
brate richness.
RESULTS
Physicochemical gradients
Water temperature had a left-skewed U-shaped response to
elevation (R2 = 0.93 for main-channel sites), decreasing
slightly from 12 °C at 4300 m a.s.l., and then increasing to
around 25 °C at 200 m (Fig. 3a). Dissolved oxygen (DO)
showed a strongly centre-weighted hump-shaped response to
elevation (R2 = 0.89 for main-channel sites), increasing from
5.0 mg L�1 at 4300 m a.s.l. to c. 9.0 mg L�1 between 2700
and 1200 m, then decreasing again to 200 m (Fig. 3b).
Across all sites, pH ranged from 4.9 to 8.3 (Table 1), and
dissolved inorganic nitrogen (DIN, the sum of ammonium,
nitrate and nitrite) concentrations ranged from 0.05 to
1.06 mg L�1 (Fig. 3g). Neither pH nor DIN appeared to be
related to either elevation or stream type. Water-column
chlorophyll a (range: 0.0–35.4 mg m�3) also showed no
strong relationship, although higher values generally were
recorded at lower-elevation sites (Fig. 3j). Maximum benthic
chlorophyll a concentrations (range: 0.0–122.7 mg m�2)
were also measured at lower elevations (Fig. 3k).
Regression analyses indicated that turbidity increased with
declining elevation among Inambari main-stem sites
(a) (b) (c)
(d) (e)
Figure 2 Examples of representative habitats (a–c) and anthropogenic impacts (d, e) in the Araz�a-Inambari watershed, south-easternPeru: (a) source of the Araz�a River (site S1, 4304 m a.s.l.); (b) the upper Araz�a main stem (site S4, 3051 m a.s.l.); (c) the lower
Inambari River (site S34, 200 m a.s.l.); (d) construction of the Interoceanic Highway along the Araz�a main stem (site S11, 1587 ma.s.l.); (e) placer mine targeting gold along the banks of the Inambari River (near site S33, 220 m a.s.l.).
Journal of Biogeographyª 2013 John Wiley & Sons Ltd
6
N. K. Lujan et al.
(a) (b) (c)
(d) (e) (f)
(g) (h) (i)
(j) (k) (l)
Figure 3 Physicochemical and biological trends along a 4100-m elevational range in the Araz�a-Inambari watershed, south-eastern Peru.
In plots (a–k), open symbols represent tributaries (t) and closed symbols represent main-channel sites (m). Subsets of sites weresampled for physicochemical and biological parameters: (a) temperature (°C); (b) dissolved oxygen (DO) (mg L�1), (c) epilithon
species richness; (d) turbidity (FTU); (e) specific conductivity (lS cm�1), (f) macroinvertebrate taxon richness; (g) dissolved inorganicnitrogen (mg L�1); (h) soluble reactive phosphorous (mg L�1); (i) fish species richness, vertical dashed lines represent (1) the upper
elevational limit of piscivorous fishes, (2) the upper elevational limit of herbivorous-detritivorous fishes, and (3) the upper elevationallimit of fishes (Fig. 5); (j) water column chlorophyll a (mg m�3); (k) benthic chlorophyll a (mg m�2); and (l) percentage total richness
of macroinvertebrates identified as scrapers (open symbols) and shredders (closed symbols). In (l) model for scrapers:y = �0.000004x2 + 0.017x + 24.47; shredders: y = 0.000005x2�0.016x + 16.91. Regression equations and R2 values for temperature
and dissolved oxygen calculated for main-stem data only (model for temperature: y = 0.000001x2�0.0095x + 26.85;DO: y = �0.0000007x2 + 0.003x + 5.77).
Journal of Biogeographyª 2013 John Wiley & Sons Ltd
7
Stream ecology from Andes to Amazon
(R2 = 41.6, P < 0.05) but not among tributaries (R2 = 8.1,
P = 0.26), indicating that anthropogenic impacts within the
main channel were likely to be the primary source of sus-
pended sediment. Anthropogenic activities associated with
sediment fluxes shifted from road construction (Fig. 2d) at
intermediate elevations (c. 2600–400 m) to placer mining
(Fig. 2e) at lower elevations (c. 400–200 m) where turbidity
reached 479 FTU (Fig. 3d). Erosion also appeared to have
altered specific conductivity, which was markedly higher in
main-channel sites than in tributary sites (Fig. 3e), and solu-
ble reactive phosphorus concentrations, which were highest
at lower-elevation sites where river sediments were being
mined (Fig. 3h).
Taxonomic diversity gradients
The species richness of epilithic algae (228 total species; Dryad
data repository: http://datadryad.org, doi:10.5061/dryad.
28mb1) showed a slightly positive overall relationship with
elevation (R2 = 0.15, P = 0.08), but this pattern was not
observed in main-stem or tributary sites when viewed sepa-
rately (Fig. 3c). Macroinvertebrate taxon richness (79 total
taxa; Dryad data repository, doi:10.5061/dryad.28mb1)
decreased (R2 = 0.38, P < 0.0001) and abundance increased
with elevation (R2 = 0.37, P < 0.0001) across both habitat
types (Fig. 3f; Dryad data repository, doi:10.5061/dryad.
28mb1). Mean macroinvertebrate abundance at the highest six
sites (1145 individuals 1000 cm�2), all of which lacked fish,
was more than twice that of the next highest six sites having
fish (502 individuals 1000 cm�2). Although elevational pat-
terns in macroinvertebrate richness and abundance spanned
both habitat types, relationships were not as significant or as
consistent in the main channel (richness: R2 = 0.22, P = 0.09;
abundance: R2 = 0.29, P = 0.05) when compared with tribu-