вход по аккаунту


Physical pollutioneffects of gully erosion on benthic invertebrates in a tropical clear-water stream.

код для вставкиСкачать
Aquatic Conserv: Mar. Freshw. Ecosyst. 16: 733–749 (2006)
Published online in Wiley InterScience
( DOI: 10.1002/aqc.813
Physical pollution: effects of gully erosion on benthic
invertebrates in a tropical clear-water stream
Max-Planck-Institute for Limnology, Tropical Ecology Working Group, Plön, Germany
1. Many tropical streams are situated in geologically old, weathered landscapes that are prone to
erosion. Permanent or seasonally pulsing inputs of eroded material into the streams may have
significant effects on benthos and habitats, even if chemical water quality remains unaltered.
2. This study presents data from the Cerrado Region, Brazil, where the semideciduous forests have
been largely converted into agricultural areas with considerable erosion problems. In order to
quantify impacts of erosion on the benthic community, a simple, inexpensive, standardized artificial
substrate was exposed on the stream bed above and below the mouth of an erosion gully in two
experiments at the beginning and at the end of the rainy season.
3. While the abundance and biomass of most of the biota in the reference sites did not change
during the rainy season, it decreased significantly at the impact sites below the mouth of the erosion
gully. It was the combination of rainfall-driven flood pulses and increased load of suspended
particles from erosion which caused the decline of the benthic colonization in these streams rather
than the hydrological disturbance alone.
4. Site-to-site comparisons revealed a highly significant reduction in density, biomass and taxon
richness of the benthic invertebrates caused by siltation. All benthic insect taxa studied showed the
same pattern, indicating a general impact of erosion on habitat quality and food sources. Semiaquatic insects adapted to shifting habitat conditions, terrestrial food sources and aerial respiration
were the most resistant invertebrate group.
5. Restoration schemes for the stream catchments are urgently needed to reduce local population
extinctions due to impassable stream sections.
Copyright # 2006 John Wiley & Sons, Ltd.
siltation; impact; artificial substrates; aquatic invertebrates; Cerrado; Brazil
Rapid development in tropical countries in the past decades has led to large-scale land degradation and
erosion (Pimentel et al., 1995). Tropical soils are often especially susceptible to degradation because their
*Correspondence to: Karl M. Wantzen, Limnology Institute, University of Konstanz, Postfach M659, 78467 Konstanz, Germany.
Copyright # 2006 John Wiley & Sons, Ltd.
structure provides less stabilizing elements, such as soil organic matter or clay particles (Johnson and
Lewis, 1995). In the Neotropics, this is especially true for the soils of the large tertiary Guayana
and Brazilian shields. More than half of the area that was originally covered by the semideciduous
Cerrado forest in central-western Brazil has already been clearcut in the recent past in order to develop
large-scale soybean plantations and cattle ranches (Mittermeier et al., 1999). Insufficient soil conservation
led to considerable erosion in this area, especially through construction of earth roads across the contour
lines of the terrain (Couto, 1990; Wantzen, 1998a). The large erosion gullies, called voçorocas, carry high
quantities of eroded soil material into the streams, thus altering their habitat properties. Deposited
sediment can fill up or braid the channels of rivers, e.g. in the Taquari River in the Pantanal wetland
(Hamilton et al., 1998).
Physical pollution of streams by suspended solids has to be regarded as one of the global perils for
freshwater fauna (Richter et al., 1997). Siltation is caused not only by erosion but also from mining
processes (e.g. Mol and Ouboter, 2004) or flushing of reservoir outlets (Jakob et al., 2004). Increased solids
loads provoke changes in the stream ecosystem which may have synergistic effects on the biota (Chutter,
1969; Ryan, 1991; Waters, 1995). Siltation may cause an abrasion of surfaces and scour away biofilms,
algae and particulate organic matter (POM), thus leading to an impoverishment of the food sources. An
experimental test using glass slides that were pre-colonized with algae revealed that algal cover could be
completely removed from the slides by a single rainstorm event in a sand-affected zone (Wantzen, 1998b).
Increased solids loads, especially fine and slowly settling particles, disturb the optical characteristics of
streams and may reduce epilithic algal growth (Hynes, 1960; Davies-Colley et al., 1992). Increased turbidity
also reduces the success of predators with visual foraging strategies (Bruton, 1985). Fine particles have the
tendency to clog interstitial pore spaces, with negative effects on the exchange processes of the hyporheic
zone and the sediment surface and on interstitial organisms with high oxygen demand such as fish larvae
(Brunke and Gonser, 1997). If the particle size is more coarse, increased sediment deposition and coverage
of the stream bottom make superficial food sources inaccessible (e.g. leaf packs; Herbst, 1980; Mayack
et al., 1989) or reduce habitable patches and habitat diversity (Mol and Ouboter, 2004). Drifting particles
may also clog biological retention devices such as filtering nets of caddisfly larvae, or the filtering organs of
simuliid larvae or molluscs.
In the Neotropics, siltation effects have been shown for algae (Wantzen, 1998b), benthic invertebrates
(Wantzen, 1998b, Fossati et al., 2001) and fish (Power, 1984; Mol and Ouboter, 2004). These studies were
limited either to single rainstorm events (Wantzen, 1998b), or to single species (Power, 1984), or they were
descriptive and based on single samples at variably impaired sites which had different habitat characteristics
owing to long distances between sampling sites (Fossati et al., 2001) or different watersheds (Mol and
Ouboter, 2004). The composition of benthic invertebrates can vary considerably even in neighbouring
streams, making a paired-watershed design a difficult approach for this group. Another difficulty in
assessing the impact of siltation is to distinguish between the effects of human activities leading to increased
solids loads and the natural variability of substrate dynamics. Most first-order streams are characterized by
the occurrence of natural erosion and rapid substrate turnover, as they are the sediment production zones
(Petts and Amoros, 1996). Moreover, siltation occurs naturally along with increased discharge and
hydraulic disturbance (Resh et al., 1988; Bond and Downes, 2003).
The aim of this study was to overcome these limitations and to differentiate between the effects of spateinduced hydraulic changes and the effects of solids load on stream organisms in a single, first-order-stream.
A set of study sites was chosen which displayed very similar environmental characteristics above and below
the mouth of an erosion gully. No chemical pollution was detectable in the runoff from the gully during the
study (Instituto Ambiental do Paraná, Curitiba, pers. comm.). Potential changes in the invertebrate
assemblages due to the occurrence of different natural substrate types were ruled out by using standardized
artificial substrates. Therefore, changes in the benthic colonization could largely be attributed to the
physical changes in the habitat caused by the increased load of solids derived from the erosion gully.
Copyright # 2006 John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. 16: 733–749 (2006)
DOI: 10.1002/aqc
The study area is situated about 100 km east of Cuiabá, the capital of the state Mato Grosso, Brazil, near
the highway BR 364 (position 158 490 S, 558 530 W, altitude 750 m a.s.l.; Figure 1(A)). This area belongs to
Córrego Brilhante
Córrego Tenente Amaral (CTA)
Stream flow
Erosion gully
Gallery forest
Sand deposits
Sediment input
Figure 1. (A) Position of the Tenente Amaral Stream in Brazil (fine arrow) and the sampling sites (thick arrow). (B) Sketch of the
sampling sites above and below the erosion gully (not to scale, distances between sampling sites are given in the text).
Copyright # 2006 John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. 16: 733–749 (2006)
DOI: 10.1002/aqc
the southern rim of the tertiary Brazilian Shield and is characterized by very poor and sandy soils. In the
surroundings of the city of Jaciara, the seasonal savanna forest } the Cerrado } has been almost
eliminated in the early 1980s (Couto, 1990). The rainy season is very pronounced in this area. Intensive,
however patchy, precipitation occurs from November to March with maxima in December and January.
Natural substrates of the streams are woody debris, leaf litter, roots and bedrock. Cerrado streams occur
with different combinations of channel and riparian vegetation types (Wantzen, 2003).
Six sites were chosen in the headwater section of the stream Córrego Tenente Amaral (abbreviation:
CTA; and the sampling sites were abbreviated CTA-1, CTA-2 etc.; see Figure 1(B)) above and below an
erosion gully. The study sites were carefully chosen with respect to their substrate, flow, riparian vegetation
and nearness (distance between sites CTA-1 and CTA-6 was less than 150 m) in order to reduce withinchannel variation (as discussed in Townsend et al., 2004). Preliminary studies of the benthic invertebrate
fauna had shown that the taxa studied occurred at all sites, but with variable density and consistency
throughout the year. The discharge of the erosion gully was less than 1% of the stream discharge. Drift
measurements had shown similar quantities of organic matter and invertebrates above and below the gully
(Wantzen and Gonçalves, unpublished data).
The first site, CTA-1, was situated in an area with gallery forest where the bed sediments consisted mainly
of sand, sandstone bedrock and rounded stones (CTA-1). Ten metres below, the stream passed across an
open area with solid bedrock substrate and few sand patches (CTA-2). CTA-1 and CTA-2 were chosen as
reference sites for the natural stage of typical channel types. The surfaces of the solid sediments were
darkened by epilithic algal growth, indicating moderate scouring by suspended inorganic sediments.
Further downstream, an erosion gully released large amounts of sand and gravel into the stream, especially
during and after rainstorms. In the dry season, there was a weak but permanent input of fine sand and silt
from the baseflow of the gully. One sampling point (CTA-3) was chosen in front of the mouth of the erosion
gully, about 20 m below point CTA-2. Below the gully, the stream flowed again through a gallery forest for
a further 50 m before it passed again through native tussock vegetation. CTA-4 was in the middle of the
forest patch and had gravelly substrates; CTA-5 was 30 m downstream, just below the forest, on bedrock.
These two sampling sites (CTA-4 and CTA-5) were selected as they had habitat conditions corresponding
to those of the reference sites. A further 60 m downstream, the last site (CTA-6) was chosen at the upper
end of another gallery forest patch. Here, many trees had been killed by erosion-caused sediment
depositions in the riparian zone and the fallen trunks caused braiding of the stream channel.
Environmental measurements
In order to register changes in water quality that potentially influenced the colonization of the artificial
substrates, the following environmental variables were measured weekly: oxygen saturation, temperature,
pH and conductivity, using WTW 190 series devices (Wissenschaftlich-Technische Werkstätten, Weilheim,
Germany). Current velocity was determined using a MiniAir 2 flowmeter. Precipitation was measured daily
in a rain gauge. Discharge was calculated from continuous water-level recordings using a float gauge meter
and single discharge measurements at different gauge levels in the stream.
In order to estimate the hydraulic stress on the stream benthos caused by storm flow events in the time
prior to sampling, an index was calculated as the sum of squares of the 14 daily discharges before the
sampling day. The square function was chosen in order to stress the impact of strong but single discharge
events compared with days having a slightly increased discharge. This index was named Peak Flow Index
(PFI). In analogy, a Rain Event Index was calculated as the sum of squares of the 14 daily rainfall values
before the sampling day. These indices were used rather than direct indicators for hydraulic stress such as
Copyright # 2006 John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. 16: 733–749 (2006)
DOI: 10.1002/aqc
current velocity or shear stress (e.g. Statzner and Higler, 1986; and references therein) as rainfall and peak
flow events occurred unpredictably. Moreover, the benthos of Cerrado streams is adapted to irregular flood
pulses by short life-cycles and extended emergence patterns (Wantzen, 2003; and unpublished data) so that
an index based on recent hydrological events appeared to be more appropriate than an index based on the
annual flow regime.
Concentration of suspended solids was measured gravimetrically from water samples taken with a
sedimentological water sampler (750-ml bottle with an opening of 5 cm) in the middle of the water column.
Estimation of the solids load was made in a narrow, channel-like stream section (CTA-5) with high current
velocities (60–90 cm s1) that kept all sand fractions in suspension. The solids load was calculated by
multiplying averages of repetitive samples (n ¼ 5) of suspended solids by the actual discharge. Bed-load
movement in the erosion gully was estimated by multiplying the thickness and velocity of the mobile
sediment layer, the width of the erosion channel, and the average weight-to-volume ratio of five random
sediment samples taken during stormflow.
Habitat quality was assessed using an adaptation of the rapid assessment field protocol of Plafkin et al.
(1989). The original method was developed for characterizing the existence and severity of impairment and
to identify sources and causes of impairment, mainly by comparing the habitat conditions of a reference site
with an impact site. The adapted protocol used here assessed (a) the degree of impairment in the catchment,
the riparian zone, and the stream bed (especially the mobility of natural substrates and sedimentary
depositions), (b) the diversity of physical habitat patterns, e.g. sediment types and current velocity across
the channel, and (c) diversity of biotic habitat patterns, e.g. presence of debris dams, aquatic macrophytes
and fine roots from trees protruding into the stream channel. The scores were developed according to longterm and regional-scale observations by the author and by the Mato Grosso State Environment Agency
FEMA (Wantzen, 1998a). Compared with the Plafkin et al. (1989) method, this procedure was simplified,
using only three categories ranking from 1 (good habitat quality, nearly natural) to 3 (bad habitat quality,
strongly altered) in order to facilitate action plans for the improvement of the environmental state of the
streams (Wantzen et al., 2006).
Artificial substrates
Benthic invertebrates were sampled using standardized artificial substrates made from two pieces of brick
and a 10 100 cm piece of green nylon gauze folded to a package of 10 10 cm and fixed to the bricks by
rubber bands (Wantzen, 1998b; Wantzen and Pinto-Silva, 2006). Benthos data refer to one unit of artificial
substrates each composed of three sets of two bricks and one nylon gauze package. The bricks were also
used as samplers for sand and as indicators of scouring. Sand accumulated in the oval cavities which were
originally designed to take up the mortar on the upper side of bricks. Scouring was indicated by the colour
of the bricks. Under undisturbed conditions, the surfaces of the bricks darken quickly after exposure in the
streams owing to biofilm development. Darkened bricks were used that had been previously exposed at the
reference site for 4 weeks. These bricks were dried and scrubbed prior to the first exposure to ensure that no
eggs or other diaspores from the pretreatment could influence the colonization experiment. Artificial
substrates were anchored with iron stakes hammered about 30 cm into the stream bed where possible. Two
sets of three artificial substrates per site were exposed at each of the six sites with a time lag of 14 days
between the two sets. Each set was sampled after 4 weeks of exposure, a period that was found to be
sufficient to reach a steady state of colonization in earlier experiments (Wantzen and Pinto-Silva, 2006).
This procedure allowed three replicate artificial substrates to be sampled per site at 14-day intervals.
Samples were taken during two 10-week-long periods at the beginning (22 October to 16 December 1994)
and at the end of the rainy season (8 March to 3 May 1995), yielding 15 samples per site and period.
During retrieval of the artificial substrates, care was taken not to lose sample material by placing a
125-mm-mesh net below them. The degree of scouring of the surfaces of the brick substrates was estimated
Copyright # 2006 John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. 16: 733–749 (2006)
DOI: 10.1002/aqc
in the field in 20% steps from 1 (unscoured) to 5 (whole surface scoured) according to the percentage of the
surface area from which the dark algal cover had been cleaned. Then, the substrates were carefully scrubbed
with a brush. The residuals sieved through a 125-mm-mesh net were stored in 70% ethanol, and stained with
Rose Bengal. The cleaned substrates were immediately remounted and exposed in the stream again. In the
laboratory, the sieved samples were elutriated several times with water and the organic fraction (animals
and POM) taken separately. The remaining sand fraction from the brick samples was used to determine the
filling of the cavities of the artificial substrates with sand, in 20% steps from 1 (0–20%) to 5 (80–100%),
setting the maximum value (300 mL) to 100%. Volumetric determination of the sand was much less timeconsuming than drying and weighing the sand; however, as it was less precise, categories were used rather
than total values. In analogy, the POM fraction remaining after sorting out the animals was dried, weighed
and noted in 20% steps as the percentage of the maximum POM amount encountered (2.3 g). A fivecategory assessment of sand filling and POM content was used in order to keep these data compatible with
data on scouring. Aquatic invertebrates (mostly insect larvae) were identified at 10 magnification to the
lowest taxonomic level possible (family or genus). Biomass of identified animals was calculated from body
length to dry mass relationships obtained from data of similarly shaped species given in Smock (1980) and
Meyer (1989). Attribution to functional feeding groups was made according to the average ranking of the
consumed food items after gut content analysis of at least 10 individuals per taxon. In cases where gut
contents were unidentifiable or absent, the assignment was made according to data from North American
species of the same genus (Merritt and Cummins, 1996).
Data analysis
Density and biomass were not calculated per square metre because the artificial substrates were more
densely colonized than most natural substrates. Non-insects occurred in a widely varying degree on the
substrates and they had low overall abundance (2% of all organisms), therefore only insect taxa were
counted for statistical tests. Abundance, biomass and the number of benthic invertebrate taxa were
analysed by two-way ANOVA (periods/station). All biotic data were log (x þ 1) transformed to obtain
normal distribution of the data. Normal distribution and variance homogeneity was tested with Shapiro–
Wilks and Bartlett tests. Multiple Tukey post hoc comparisons allowed statements of significant differences
among stations in both periods (Zar, 1999). Influence of the physical characteristics (independent variables)
on abundance, biomass and number of taxa (dependent variables) in each sample was compared with
ANCOVA. The relationship between average of sand filling and biological variables, calculated from 15
artificial substrate samples per site, was analysed with linear regression (Zar, 1999). For most of the
statistical tests, the program XLstat 7.5 (Addinsoft, Brooklyn, NY, USA) was used; canonical
correspondence analysis of biotic and environmental data was performed with SPSS 11.0 (SPSS Inc.,
Chicago, IL).
Rainfall and discharge
During the sampling period, single rainfall events occurred from September to October 1994 and regular
rainfall and heavy rainstorms occurred from November 1994 to April 1995. The nearest long-term
observation site (158 410 5600 S, 568 080 0100 W) had a 5-year mean (1995–1999) of 1397 mm, minimum
758 mm, and maximum 2318 mm (according to a report from the Mato Grosso State Environmental
Protection Agency, FEMA-MT, Runoff
development in the stream was clearly rainfall-driven. At the beginning of the rainy season, recurring
rain events caused an increase in the discharge from baseflow values of about 500 to about 800 L s1 from
Copyright # 2006 John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. 16: 733–749 (2006)
DOI: 10.1002/aqc
L s–1
May 95
Apr 95
Mar 95
Feb 95
Jan 95
Dec 94
Nov 94
Oct 94
Sep 94
Figure 2. Discharge of the Tenente Amaral Stream. Dark bars indicate experimental periods. Modified after Wantzen (2003).
September to December 1994. Heavy rainfall in late December further increased the runoff to about
1200 L s1 until April 1995 (Figure 2). Thus, both study periods experienced rainfall-driven events } the
first was at the beginning and the second at the end of the rainy season.
Event-driven changes
Temperature, pH, oxygen concentration and conductivity varied little throughout the experiments
(Figure 3). Rainfall events had less influence on temperature variation than daily or seasonal changes
(Wantzen, 2003). Concentrations of dissolved solids in the stream water did not show significant changes
during the rain events. The conductivity was generally low (below 5 mS cm1). Weekly measurements of
oxygen, conductivity, and pH showed very stable patterns with averages plus or minus standard deviation
of 6.9 0.5 mg O2 L1, 2.7 0.3 mS cm1 and 4.6 0.6, respectively.
Concentrations of suspended solids increased during rainfall events. During a single 62-mm rain event,
suspended solids increased upstream of the mouth of the erosion gully from 0.006 to 0.236 g L1 for a short
period and returned to normal shortly after the rain stopped (Figure 4). Conductivity changed very little
during the event. In the erosion gully, the discharge and the suspended solids concentrations increased very
rapidly with the rainfall from 0.5 to nearly 10 g L1 in 1 hour. The water below the mouth of the gully
turned very turbid (650 NTU) and remained much more turbid than the reference site (2.5–12.8 NTU) until
the end of the measurements. The bedload that was carried from the erosion gully into the stream was
estimated at 10 kg of coarse sand and gravel per second during maximum flow. Large amounts of solids
were deposited in the stream channel and on the banks during strong storm events. These deposits impeded
the runoff in the upstream part and caused a deepening of the stream bed (Figure 1(B)).
Physical characteristics during the experiments
All environmental variables measured showed significant differences between the reference sites (CTA-1
and CTA-2) and the impact sites (CTA-3 to CTA-6, ANCOVA for sand filling as independent variable
F: 263.5, p50:0001; for Habitat Quality Index: F: 198.4, 0.0001, for organic matter: F: 126.6, p50:0001;
Figure 5). Sand and organic matter content revealed a complementary pattern. While the reference sites
showed little or no sand filling, the impact sites often became completely filled with sand, especially the site
below the erosion gully (CTA-3). At the reference sites, black leaf particles were commonly found in the
crevices of the artificial substrate samplers. The sand found in the bricks from the impact sites was washed
Copyright # 2006 John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. 16: 733–749 (2006)
DOI: 10.1002/aqc
µS cm-1
pH units
mg O L-1
May 95
Apr 95
Mar 95
Feb 95
Jan 95
Dec 94
Nov 94
Oct 94
Sep 94
Figure 3. Monthly means of oxygen, pH and conductivity at the Tenente Amaral Stream. Error bars are standard deviations (n ¼ 4).
and lacked visible organic particles. The large amounts of eroded material washed from the gully into the
stream (maximum amount recorded during rain events at CTA-5: 50 t d1) covered the stream bottom after
rain events. Often, the position of the artificial substrates could only be found by the metal poles that were
used to fix the substrates. The sediment layers, which were 10–20 cm thick at CTA-3 and CTA-4, and up to
50 cm at CTA-6, remained for 1–10 days before they were washed downstream; however, recurrent rainfall
events quickly refilled them. At site CTA-5, high current velocities greater than 60 cm s1 and solid
sandstone stream bed did not allow sediment deposition; however, at this site the scouring of the artificial
substrates was most intensive, as evidenced by complete removal of the dark varnish from the outer
surfaces of the bricks. In all sites where metal poles were used, oxidation of the poles showed that at least
the upper 10 cm of the sediments were well-oxygenated. Only at CTA-6 did dark coloration of the poles at
15 cm depth indicate scarcity or lack of oxygen.
The habitat quality was worst at the confluence with the erosion gully that had the lowest substrate
diversity and the highest sediment mobility. Further downstream, these variables recovered slightly,
yielding HQI averages of 2.7 (Figure 5). At all impact sites, habitat quality worsened from the beginning to
the end of the rainy season. At the sites below CTA-3, some rocky patches were still covered with aquatic
macrophytes or algal biofilms at the beginning of the first experiment, but they were completely scoured off
at the end of the second experiment.
Benthic invertebrates
Density of aquatic invertebrates on the artificial substrates was relatively sparse and differed greatly from
site to site (from 0 to 303 individuals per substrate unit, Figure 6). Invertebrate numbers on bricks were on
Copyright # 2006 John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. 16: 733–749 (2006)
DOI: 10.1002/aqc
Precipitation (mm)
Discharge (L s–1)
Supended solids (ppm)
erosion gully
Conductivity (µS cm–1)
Turbidit y (NTU)
Time (h)
Figure 4. 16-hour time-series of changes in the Tenente Amaral Stream caused by a single rainfall event of 62 mm. (A) rainfall, (B)
discharge, (C) suspended solids, (D) conductivity and (E) turbidity, above, within and below the erosion gully.
Copyright # 2006 John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. 16: 733–749 (2006)
DOI: 10.1002/aqc
Mean sand filling (rank)
Mean organic matter (rank)
Mean HQI (rank)
Figure 5. Box-and-whisker plot of environmental data for six sampling sites in the Tenente Amaral Stream during the early (left
bars: I) and late (right bars: II) rainy season. (A) sand filling of the artificial substrates; (B) organic matter content; (C) Habitat Quality
Index (HQI).
average 12% higher than on nylon nets; however, these differences were not significant. The community
composition was dominated by chironomid larvae (45%), caddisfly larvae (13%), beetle larvae and nonchironomid midge larvae (each 10%), mayfly larvae (9%), stonefly larvae (5%), dragonfly nymphs (2%),
adult beetles (2%), heteropterans (1%) and neuropterans (1%). The non-insects made up 2%, mainly
aquatic mites (Hydracarina), harpactoid copepods and some oligochaetes. Mollusca and decapod shrimps
were lacking. Collectors were the most frequent functional feeding group, representing 49–56% of the total
abundance, predators comprised 11–50%, and scrapers and shredders 0–5% each. The latter two groups
were only found in the reference sites.
Copyright # 2006 John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. 16: 733–749 (2006)
DOI: 10.1002/aqc
Number per SAS
(B) 0
(D) 0
(F) 0
(H) 0
Figure 6. Box-and-whisker plots of benthic invertebrate data on standardized artificial substrates (SAS) in the Tenente Amaral Stream
during the experiments at the beginning of the rainy season (left bars: I) and late rainy season (right bars: II). (A) overall density,
(B) taxon richness, and densities of larval (C) Chironomidae, (D) Ephemeroptera, (E) Plectoptera, (F) Trichoptera, (G) Coleoptera
and (H) Odonata.
All taxa and functional feeding groups reacted to the siltation impacts in a similar way (Figure 6).
Plecopterans, mostly Anacroneuria (Bispo et al., 2005), were the most sensitive organisms; at the end of the
rainy season they completely disappeared at sites CTA-4 and CTA-6, and only a single individual was
found at CTA-3. Chironomids were still present at all sites during the second experiment but they also
Copyright # 2006 John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. 16: 733–749 (2006)
DOI: 10.1002/aqc
Table 1. Results of ANOVA and Tukey post hoc analysis (HSD) of density (individuals per substrate unit), biomass (mg dry weight
per substrate unit) and number of invertebrate taxa (n per substrate unit) at the reference sites (CTA-1, CTA-2) and their
corresponding impact sites (CTA-4, CTA-5) in the Tenente Amaral Stream. Groups with different letters (A–E) differed significantly
Fisher’s F
Taxa number
HSD group
HSD group
HSD group
showed greatly reduced abundances. Filter-feeders (e.g. Trichoptera), visually foraging predators (e.g. most
Odonata) and even large dobsonfly larvae (Corydalidae) which occur in the most current-exposed habitats
of undisturbed streams, were greatly reduced or eliminated at the sites influenced by siltation. A few taxa
were less severely affected, e.g. burrowing dragonfly larvae of the family Gomphidae. The most common
genus, Progomphus, colonizes organic-rich sediment deposits in pools and debris dams in undisturbed
streams of the area (Wantzen and Priante, unpublished data).
Density of invertebrates and number of taxa were highest at the two reference sites (CTA-1 and CTA-2)
and lowest at the site at the mouth of the erosion gully (CTA-3) while the other impact sites showed
intermediate values (Figure 6). When comparing reference sites with their corresponding impact sites, the
differences become obvious both for forest sites (CTA-1 and CTA-4) and sites in open stretches (CTA-2
and CTA-5). ANOVA and post hoc Tukey analysis grouped the reference site data from both experiments
for density and biomass, but taxon number differed between CTA-1 and CTA-2. The corresponding impact
sites were clearly separated (Table 1). Differences in density and taxon richness between reference and
impact sites were less pronounced at the beginning of the rainy season than during the late rainy season
(Figure 6). Both reference and impact sites showed lower abundance after heavy rains, but the abundance
at the impact sites was more strongly reduced and subsequently recovered less than at the reference sites, or,
in some cases, did not recover not at all.
During the second experiment at the end of the rainy season, the differences between reference sites and
impact sites were even more pronounced. The reductions of the density from the first to the second
experiment were 25% and 24% for the reference sites (CTA-1 and CTA-2), 36% for CTA-3, and ranged
from 58% to 79% for the impact sites CTA-4, CTA-5, and CTA-6. In the same way, the number of
identifiable taxa remained stable at the two reference sites (CTA-1 and CTA-2), and became reduced by
22% to 57% at the impact sites (CTA-4 to CTA-6). At CTA-3, reductions in abundance and taxon
numbers were less pronounced (4% and 7%, respectively), as this site was very sparsely colonized
throughout the experiments.
The sand content in the artificial substrates had negative effects on mean density, biomass and taxon
numbers. ANOVA did not reveal significant differences in these relationships between the two experiments
in 1994 and 1995, therefore data from both sets were pooled for linear regressions (Table 2). When
comparing the impacts of sand on abundance, biomass and taxon number in the reference sites (CTA-1 and
CTA-2) with the corresponding impact sites (CTA-3 to CTA-6), the squared regression coefficients were
much smaller in the reference sites than in the impact sites.
Copyright # 2006 John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. 16: 733–749 (2006)
DOI: 10.1002/aqc
Table 2. Linear regression between sand filling of the standardized artificial substrates and abundance, biomass and number of taxa of
invertebrates in the Tenente Amaral Stream. Data were log(x þ 1) transformed and pooled from two experiments at the beginning and
at the end of the rainy season
CTA-1, 2
CTA-4, 5
CTA-3 to 6
CTA-1 to 6
Number of taxa
Figure 7. Canonical correspondence analysis (CANOCO) of biotic and environmental data in the Tenente Amaral Stream (data
pooled from both experiments). SAND: sand filling of the artificial substrates, RAIN: rain index, OM: organic matter content, HQI:
Habitat Quality Index, PFI: Peak Flow Index.
The canonical correspondence analysis displayed the differences between reference sites and impact sites
very clearly (Figure 7). The first two axes explained more than 80% of the variances. The axes were
composed by the following environmental variables: particulate organic matter content of the artificial
substrates, filling of the bricks with sand, peak-flow index, rain intensity, date and habitat-quality index
(HQI). The sand content and HQI were inversely related to the organic matter content; i.e. those sites,
where substrates were filled with large amounts of sand, revealed a bad habitat quality and a low organic
matter content. Along this axis, the sites were separated into reference sites (CTA-1 and CTA-2, almost
coinciding) and impact sites (CTA-3 to CTA-6). Date, rainfall and PFI were in a perpendicular position to
this axis, separating different events within the groups, although they did not influence the distribution of
the groups themselves.
Peak flow events combined with increased solids loads caused by erosion reduced benthic fauna
considerably, whereas the same events did not have strong effects at the reference sites with lower solids
Copyright # 2006 John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. 16: 733–749 (2006)
DOI: 10.1002/aqc
loads. Stream insects are highly adapted to individual, non-predictable flood pulses (e.g. Matthaei and
Townsend, 2000; Wantzen and Junk, 2000) but these data show that they are apparently less adapted to the
co-occurrence of flood pulse and ‘sediment-pulse’. This combination affected taxon richness, abundance
and biomass.
Various studies suggested that natural hydraulic disturbance increases the contribution of predators and
reduces the numbers of shredders in low-order streams (Winterbourn et al., 1981; Boulton et al., 1988;
Wantzen and Junk, 2000). Siltation effects may increase this trend even further. Mol and Ouboter (2004)
described shifts in feeding guilds of neotropical fish towards predators because of the increased exposure of
prey organisms following the destruction or obstruction of sheltering habitats by siltation. While this effect
is limited to a transitional period from an undisturbed to a disturbed system (i.e. until the sensitive animals
have been driven out of their habitats), an additional bias of feeding guild proportions towards predators
may be caused by the use of terrestrial invertebrates as a food source while instream food sources have been
reduced by siltation. This view is supported by the observation that the first organisms to colonize erosion
gullies and the last organisms found in completely braided stream sections were predators (e.g.
belostomatid, nepid and gelastochorid bugs; Wantzen, pers. obs.). These organisms are also independent
of aquatic respiration that might be disturbed by increased concentrations of suspended solids.
Fossati et al. (2001) identified groups of taxa that were sensitive to siltation by different mechanisms;
however, they also found an overall reduction in abundance and species richness as in the present study.
The reasons for this ‘universal’ pattern may be indicated by the field observations from the artificial
substrates. Habitat structures and food sources became damaged and/or reduced at the same time. Statzner
and Higler (1986) considered the hydraulically induced occurrence of fine sediments as the most important
factor explaining the distribution of benthic invertebrate species in a South African stream system.
Apart from these general patterns, the frequency of disturbance appears to be crucial for the survival of
benthic invertebrates in streams disturbed by erosion (Resh et al., 1988). Hydraulic disturbance reduces
species richness of benthic invertebrates in streams (e.g. Matthaei et al., 1996; Mathooko, 1999; Lake, 2000;
and see review by Vinson and Hawkins, 1998); however, recolonization quickly restores the species
assemblages (Boulton et al., 1988; Mackay, 1992; Townsend et al., 1997). The period for recovery is
generally reported from 8 to 30 days (Boulton et al., 1988; Lake, 2000), but extreme events may have a
2-year recovery time (Collier and Quinn, 2003). Experimental studies in tropical African streams
(Mathooko, 1999) and in temperate streams (Matthaei and Townsend, 2000; McCabe and Gotelli, 2000)
have shown decreasing diversity with increasing disturbance frequency. The hydrograph presented in this
study showed a superposition of repetitive disturbance events in the Tenente Amaral stream (Figure 2).
Results from the present study suggest that the combination of hydraulic and sediment stress requires much
longer recovery periods than hydraulic stress alone because the effects from siltation persist longer than the
hydraulic pulse due to backslumping of riparian sediment deposits. Apparently, the frequency of rainfall
events was too high for recovery of the sites affected by sand. Experiments are needed to analyse recovery
times of biofilms, POM, and aquatic metazoans in detail. If sufficient resources for recolonization
(propagules, drifting invertebrates, etc.) are present, algal colonization requires at least 1 week (Wantzen,
1998b) and benthic invertebrate colonization about 4 weeks (Wantzen and Pinto-Silva, 2006) in Cerrado
streams. However, the real world situation in these streams is that habitats that could provide resources for
recolonization are dramatically decreasing.
Removal of riparian forests increased the recovery time after pulse disturbances in stream invertebrate
assemblages (Collier and Quinn, 2003). Gallery forests play a central role in the habitat quality of Cerrado
streams as they deliver POM as food and substrate, provide solid substrates such as logs and roots, and
stabilize the bed structure. All streams in the area investigated showed siltation impacts and progressive dieback of the gallery forest vegetation. We observed that the roots protruding into the stream become scoured
and that sand cover in the riparian zone suffocates the trees. In the first years of erosion impacts, die-back
of riparian trees leads to an increase in organic matter inputs from tree logs and an increase in the retentive
Copyright # 2006 John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. 16: 733–749 (2006)
DOI: 10.1002/aqc
structures that accumulate organic debris. Later, the dead logs are carried away by the current, and the
stream channel begins to braid. The lack of in-channel woody debris reinforces erosive effects in the lower
stream sections, causing a downstream cascading of the process. Long, braided stretches are inhabitable
and impassable for a large number of benthic invertebrates and fish, leading to an interruption of the
genetic exchange of populations and local extinctions, e.g. for stream fish. We observed the extinction of an
unidentified fish population (Characidae, Tetragonopterinae) in a neighbouring stream (C. Brilhante)
during the study. Loss of longitudinal connectivity may cause ‘sinks’ for species which cannot leave specific
habitats any more (Pringle, 1997). In parts of the streams that are still forested, we also observed deepening
of the stream channel and upstream erosion which extended up to the stream source. Especially in the
erosion-prone hillslope wetlands of the Cerrado (Oliveira-Filho et al., 1989), it is very common for the
stream sources to be completely devastated by erosion so that recolonization of stream stretches further
downstream from springs (Mackay, 1992), e.g. after catastrophic floods, is impossible. Melo and Froehlich
(2001) have shown that low-order streams can be more diverse than medium-order streams in the
Neotropics. These low-order sections are especially threatened by erosion.
There is an urgent need for catchment conservation programmes in the Neotropics. These should
include environmental education in order to impart knowledge about the ecological and economical
value of gallery forests to the local population, as well as the identification and planting of tree species
that are appropriate to re-establish typical gallery forests and stabilize erosion gullies (Wantzen et al.,
This paper resulted from the cooperation of the Bioscience Institute of the Federal University of Mato Grosso
(UFMT), Cuiabá, Brazil, and the Tropical Ecology Working Group of the Max-Planck-Institute for Limnology
(MPIL), Plön, Germany.
Financial and technical support was provided by the German Ministry of Science and Technology (BMBF) (project
no. 0339373B), the Brazilian Research Council (CNPq, reg. no. 690001/97-5) and the Deutsche Gesellschaft für
Technische Zusammenarbeit (GTZ, Tropical Ecology Support Program).
Special thanks to Wolfgang J. Junk, and Rüdiger Wagner for critical discussions on the study design, Guillermo
Rueda-Delgado for support with statistical analysis, Hinnerk Boriss for the development of the Peak Flow Index, and
my students Aecio Moraes de Paula, Gabriela Priante and Cristina Menezes Butakka for their help with the fieldwork
and sample processing. Odile Fossati and Jan de Mol gave valuable comments on an earlier version of the manuscript.
I am indebted to the farmers Arno and Carlos Schneider for allowing us to work on the Fazenda Santa Fé.
Bispo PC, Neves CO, Froehlich CD. 2005. Two new species of Perlidae (Plecoptera) from Mato Grosso State, Western
Brazil. Zootaxa 795: 1–6.
Bond NR, Downes BJ. 2003. The independent and interactive effects of fine sediment and flow on benthic invertebrate
communities characteristic of small upland streams. Freshwater Biology 48: 455–465.
Boulton AJ, Spangaro GM, Lake PS. 1988. Macroinvertebrate distribution and recolonization on stones subjected to
varying degrees of disturbance: an experimental approach. Archiv für Hydrobiologie 113: 551–576.
Brunke M, Gonser T. 1997. The ecological significance of exchange processes between rivers and groundwater.
Freshwater Biology 37: 1–33.
Bruton MN. 1985. The effects of suspensoids on fish. Hydrobiologia 125: 221–242.
Chutter FM. 1969. The effects of silt and sand on the invertebrate fauna of streams and rivers. Hydrobiologia 34: 57–76.
Collier KJ, Quinn JM. 2003. Land-use influences macroinvertebrate community response following a pulse disturbance.
Freshwater Biology 48: 1462–1481.
Couto EG. 1990. O uso da terra e o garimpo na bacia do Rio São Lourenço, Mato Grosso: reflexos no ambiente.
Technical Report, FEMA, Mato Grosso, Brazil.
Copyright # 2006 John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. 16: 733–749 (2006)
DOI: 10.1002/aqc
Davies-Colley RJ, Hickey CW, Quinn JM, Ryan PA. 1992. Effects of clay discharges on streams: 1. Optical properties
and epilithon. Hydrobiologia 248: 215–234.
Fossati O, Wasson JG, Cecile H, Salinas G, Marin R. 2001. Impact of sediment releases on water chemistry and
macroinvertebrate communities in clear water Andean streams (Bolivia). Archiv für Hydrobiologie 151: 33–50.
Hamilton SK, Souza OC, Coutinho ME. 1998. Dynamics of floodplain inundation in the alluvial fan of the Taquari
River (Pantanal, Brazil). Verhandlungen der Internationalen Vereinigung für theoretische und angewandte Limnologie
26: 916–922.
Herbst GN. 1980. Effects of burial on food value and consumption of leaf detritus by aquatic invertebrates in a lowland
forest stream. Oikos 35: 411–424.
Hynes HBN. 1960. The Biology of Polluted Waters. University of Liverpool Press: Liverpool.
Jakob C, Robinson CT, Uehlinger U. 2004. Longitudinal effects of experimental floods on stream benthos downstream
from a large dam. Aquatic Sciences 65: 223–231.
Johnson DL, Lewis LA. 1995. Land Degradation: Creation and Destruction. Blackwell: Oxford.
Lake PS. 2000. Disturbance, patchiness, and diversity in streams. Journal of the North American Benthological Society
19: 573–592.
Mackay RJ. 1992. Colonisation by lotic macroinvertebrates: a review of processes and patterns. Canadian Journal of
Fisheries and Aquatic Sciences 49: 617–628.
Mathooko JM. 1999. Effects of differing inter-disturbance intervals on the diversity of mayflies recolonizing disturbed
sites in a tropical stream. Archiv für Hydrobiologie 146: 101–116.
Matthaei CD, Townsend CR. 2000. Long-term effects of local disturbance history on mobile stream invertebrates.
Oecologia 125: 119–126.
Matthaei CD, Uehlinger U, Meyer EI, Frutiger A. 1996. Recolonization by benthic invertebrates after experimental
disturbance in a Swiss prealpine river. Freshwater Biology 35: 233–248.
Mayack DT, Thorp JH, Cothran M. 1989. Effects of burial and floodplain retention on stream processing of
allochthonous litter. Oikos 54: 378–388.
McCabe DJ, Gotelli NJ. 2000. Effects of disturbance frequency, intensity, and area on assemblages of stream
macroinvertebrates. Oecologia 124: 270–279.
Melo AS, Froehlich CG. 2001. Macroinvertebrates in neotropical streams: richness patterns along a catchment and
assemblage structure between 2 seasons. Journal of the North American Benthological Society 20: 1–16.
Merritt RW, Cummins KW (eds). 1996. An Introduction to Aquatic Insects of North America. Kendall/Hunt Publishing:
Meyer E. 1989. The relationship between body length parameters and dry mass in running water bodies. Archiv für
Hydrobiologie 117: 191–203.
Mittermeier RA, Myers N, Mittermeier CG. 1999. Hot Spots } Earth’s Biologically Richest and Most Endangered
Terrestrial Ecoregions. CEMEX, Conservation International: New York.
Mol JH, Ouboter PE. 2004. Downstream effects of erosion from small-scale gold mining on the instream habitat and
fish community of a small neotropical rainforest stream. Conservation Biology 18: 201–214.
Oliveira-Filho AT, Shepherd JG, Martins FM, Stubblebine WH. 1989. Environmental factors affecting physiognomic
and floristic variation in an area of cerrado in central Brazil. Journal of Tropical Ecology 5: 413–431.
Petts GE, Amoros C. 1996. The fluvial hydrosystem. In Fluvial Hydrosystems, Petts GE, Amoros C (eds). Chapman &
Hall: London; 1–12.
Pimentel D, Harvey C, Resosudarmo P, Sinclair K, Kurz D, McNair M, Christ S, Shirtz L, Fitton L, Saffouri R,
Blair R. 1995. Environmental and economic costs of soil erosion and conservation benefits. Science 267:
Plafkin JL, Barbour MT, Porter KD, Gross SK, Hughes RM. 1989. Rapid Bioassessment Protocols for Use in Streams
and Rivers (RPBS). United States Environmental Protection Agency: Washington, DC.
Power ME. 1984. The importance of sediment in the grazing ecology and size class interactions of the armored catfish,
Ancistrus spinosus. Environmental Biology of Fishes 10: 173–181.
Pringle CM. 1997. Exploring how disturbance is transmitted upstream: going against the flow. Journal of the North
American Benthological Society 16: 425–438.
Resh VH, Brown AV, Covich AP, Gurtz LHW, Minshall GW, Reice SR, Sheldon AL, Wallace JB, Wissmar RC. 1988.
The role of disturbance in stream ecology. Journal of the North American Benthological Society 7: 443–455.
Richter BD, Braun DP, Mendelson MA, Master LL. 1997. Threats to imperiled freshwater fauna. Conservation Biology
11: 1081–1093.
Ryan PA. 1991. Environmental effects of sediments on New Zealand streams: a review. New Zealand Journal of Marine
and Freshwater Research 25: 207–221.
Smock LA. 1980. Relationships between body size and biomass of aquatic insects. Freshwater Biology 10: 375–384.
Copyright # 2006 John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. 16: 733–749 (2006)
DOI: 10.1002/aqc
Statzner B, Higler B. 1986. Stream hydraulics as a major determinant of benthic invertebrate zonation patterns.
Freshwater Biology 16: 127–139.
Townsend CR, Scarsbrook MR, Doledec S. 1997. Quantifying disturbance in streams: alternative measures of
disturbance in relation to macroinvertebrate species traits and species richness. Journal of the North American
Benthological Society 16: 531–544.
Townsend CR, Downes BJ, Peacock K, Arbuckle CJ. 2004. Scale and the detection of land-use effects on morphology,
vegetation and macroinvertebrate communities of grassland streams. Freshwater Biology 49: 448–462.
Vinson MR, Hawkins CP. 1998. Biodiversity of stream insects: variation at local, basin, and regional scales. Annual
Review of Entomology 43: 271–293.
Wantzen KM. 1998a. Analysis of environmental impacts of man-made soil erosion on running water ecosystems using
biomonitoring in Mato Grosso, Brazil. Deutsche Gesellschaft für Technische Zusammenarbeit (GTZ) GmbH,
Eschborn, Germany.
Wantzen KM. 1998b. Effects of siltation on benthic communities in clear water streams in Mato Grosso, Brazil.
Verhandlungen der Internationalen Vereinigung für theoretische und angewandte Limnologie 26: 1155–1159.
Wantzen KM. 2003. Cerrado Streams } characteristics of a threatened freshwater ecosystem type on the tertiary
shields of South America. Amazoniana 17: 485–502.
Wantzen KM, Junk WJ. 2000. The importance of stream-wetland-systems for biodiversity: a tropical perspective. In
Biodiversity in Wetlands: Assessment, Function and Conservation, Gopal B, Junk WJ, Davies JA (eds). Backhuys:
Leiden; 11–34.
Wantzen, KM, Pinto-Silva V. 2006. Uso de substratos artificiais para macroinvertebrados bentônicos para a avaliação
do impacto de assoreamento em nascentes dos tributários do Pantanal do Mato Grosso, Brasil. Revista Brasileira de
Recursos Hı´dricos 11: 99–110.
Wantzen KM, Sá MFP, Siqueira A, Nunes Da Cunha C. 2006. Conservation scheme for forest-stream ecosystems of
the Brazilian Cerrado and similar biomes in the seasonal tropics. Aquatic Conservation: Marine and Freshwater
Ecosystems 16: 713–732.
Waters TF. 1995. Sediments in Streams } Sources, Biological Effects, and Control. Monograph no. 7, American
Fisheries Society, Bethesda, MD.
Winterbourn MJ, Rounick JS, Cowie B. 1981. Are New Zealand stream ecosystems really different? New Zealand
Journal of Marine and Freshwater Research 15: 321–328.
Zar JH. 1999. Biostatistical Analysis, 4th edn. Prentice-Hall: Englewood Cliffs, NJ.
Copyright # 2006 John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. 16: 733–749 (2006)
DOI: 10.1002/aqc
Без категории
Размер файла
307 Кб
physical, water, pollutioneffects, invertebratesвђ, stream, gull, erosion, tropical, cleary, benthic
Пожаловаться на содержимое документа