|
METHODS FOR MACROINVERTEBRATE COMMUNITY
DATA
Field collection of macroinvertebrate samples
Macroinvertebrates were collected at 60
locations distributed throughout the watersheds between 1
and 18 May 2000, 30 April and 10 May 2001, and 5 and 16 May 2002.
The sampling protocol was designed to characterize riffle-inhabiting
macroinvertebrates in a reach that included several riffles (i.e.,
for additional habitat and biotic diversity) rather than the approach
of characterizing macroinvertebrates from a single riffle or part
of a riffle. Reach length varied among streams and rivers, but
generally included 20-50 m of riffle. Random sampling locations
were chosen based on their longitudinal (e.g., along the length
of the study reach) and lateral positions. For example, a sampling
location in a stream might be designated as 17-25, which would
represent 17 m upstream and 25% across the stream from the bank.
The sampling protocol called for collecting a total of four composite
samples representing 16 samples at each site. The sampling design
was modified for several sites each year in response to limited
riffle habitat availability. In most cases this resulted in eight
samples being collected and/or sampling being collected where
possible (i.e., partially-random) rather than at the random sites.
Only four samples could be collected at a few sites.
Benthic
macroinvertebrates were collected in riffle habitats with a Surber
sampler (1 ft2 or 0.093 m2; 250-µm
mesh) using a quantitative composite sampling regime that was
modified from Stroud SOP S-04-09. Sampling started at the downstream
end of the sampling area and proceeded in an upstream direction.
The operator identified the location of each sampling area based
on the longitudinal and lateral position. If boulders or large
woody debris interfered with sampling at the designated sampling
location, the location was moved slightly until there was no obstruction.
If it was impossible to obtain a good sample from this location,
an alternative sampling site that was also randomly chosen was
used for this sample. To collect the macroinvertebrate sample,
the back edge of the Surber Sampler is set on the stream bottom
so that there is a tight seal across the substrate to prevent
animals from migrating under the sampler. The square bottom frame
is then laid out on the stream bottom to delimit the 1 ft2
sample area. Rocks that were under the frame were included in
the sample if more than half of the rock was inside the frame;
if more than half of the rock was outside of the frame it was
not included in the sample. Larger rocks (> 65 mm in longest
dimension) were removed individually, and scrubbed with a soft
bristled brush under the water in front of the net. Scrubbing
removed most attached organisms while the water current moving
through the sampler carried these dislodged organisms into the
sample net. Each scrubbed rock was placed in a plastic bucket
(held by a second person) for subsequent counting. The minimum
rock counted and/or measured was = 65 mm on the longest axis.
Large rocks that could not be moved were scrubbed in place. After
all rocks were scrubbed and removed, the enclosed benthic area
was rapidly stirred and agitated for at least 20 seconds to suspend
any residual organisms in the water column and subsequently into
the sample net. The sampler was then removed from the bottom and
stream water splashed onto the outside of the net in order to
wash clinging animals into the bottom of the net. Each sample
was randomly assigned to one of four composite samples so the
net for a sample was inverted and the contents washed into a plastic
bucket designated for that composite sample.
Composite
samples resulted from combining four- 1 ft2 samples
(if possible) into one composite sample (i.e., containing macroinvertebrates
from 4 ft2) and then subsampling the combined samples
in the field such that a subsample equaled one sample (i.e., macroinvertebrates
representative of 1 ft2). After all samples (usually
16) had been collected and combined into four composite samples,
each composite sample was split into subsamples (each representing
1 ft2), with one of the subsamples being preserved
and brought back to the laboratory for analysis. Each composite
sample was washed into a large sample splitter that was placed
in a large plastic barrel half filled with water. The mixture
of macroinvertebrates, detritus, and sediments was homogenized
and resuspended by stirring, agitating, and pushing water into
the subsampler. The material then resettled across the bottom
of the subsampler while slowly drawing the subsampler out of the
barrel. If the material did not appear evenly distributed, the
resuspension and settling process was repeated. The net-covered
bottom (250-µm mesh) was separated from the rest of the
subsampler, and a plastic separator was pushed into the sample
material, dividing the material into four equal parts. A spatula
and scissors was used to separate subsamples and transfer material
to a labeled sample jar filled with 5% buffered formalin, which
was then transported to the laboratory. If the composite sample
contained four samples, then 1/4th of the composite
material represented macroinvertebrates from 1 ft2.
If only eight samples were collected, then each composite sample
contained the contents of two samples (i.e., macroinvertebrates
from 2 ft2), and the composite sample was split into
two subsamples (each representing 1 ft2).
Sample
compositing has advantages over standard (non-compositing) macroinvertebrate
sampling. For example, compositing increases the accuracy of the
desired description by increasing the number of samples collected
and therefore the area sampled in these riffles without increasing
the number of samples processed. At the same time, compositing
homogenizes spatial variation when these samples are combined,
which reduces variance among samples in statistical analyses.
Associated
with each sample, water depth was measured to the nearest cm and
current velocity was measured with a current meter set at a point
0.6 of the distance from the bottom to the water surface. The
number of large rocks (= 65 mm in longest dimension) that had
been in that sample was also recorded. Periphyton biomass (as
chlorophyll a and ash free dry mass [AFDM]) was measured for each
composite sample by collecting a small algae-covered stone (3-5
cm in diameter) near where each sample was collected and placed
in labeled plastic Tupperware containers associated with each
composite sample (i.e., 2 or 4 rocks per composite sample). The
plastic Tupperware containers were stored on dry ice (in field)
or in a freezer (in laboratory) until chlorophyll a and AFDM analyses
were completed in the laboratory (< 28 d for chlorophyll a).
Laboratory
processing of macroinvertebrate samples
Benthic
materials (i.e., macroinvertebrates and detritus) were transferred
from the sample jar into a 250-µm mesh sieve and rinsed
thoroughly with water to remove fine particles. Because macroinvertebrates
were abundant (hundreds to thousands per sample), each sample
was split into four subsamples, and then one of those subsamples
was split into four subsamples (i.e., 1/16th of a sample).
Actual subsample size processed varied among samples (1/16, 1/8,
3/16, 1/4) and reflected the number of macroinvertebrate per sample.
Our target was to identify 100-300 macroinvertebrates per subsample.
Macroinvertebrates were separated from detritus by taking a small
portion from the subsample and placing it in a plastic sorting
tray partially filled with 80% ethanol. This material was then
carefully examined with the aid of a dissecting microscope (12
X magnification).
All
macroinvertebrates were removed from the detrital material collected
in the subsample, and the detrital material was transferred to
an aluminum weigh boat (see Benthic Organic Matter below). All
macroinvertebrates were identified to the lowest taxonomic level
possible. For aquatic insects, this was generally genus or species;
other macroinvertebrates (e.g., crustacea, mites, flatworms, oligochaetes,
and nematodes) were commonly left at higher taxonomic levels (e.g.,
order, family). Specimens that were damaged or extremely small
were identified to the taxonomic levels possible, but these were
higher than species and even genus. Chironomids were subsampled
before identification, and the number examined represented the
percentage of chironomids in that sample. For example, if a sample
contained 300 macroinvertebrates and 40% of them were chironomids,
then 40 chironomids were identified to genus/species and these
identifications were applied proportionally to the remaining 80
chironomids. Identified macroinvertebrates were placed in vials
containing 80% ethanol and a permanent label. Macroinvertebrate
specimens (sorted and unsorted material) are archived by the Stroud
Water Research Center (SWRC) for at least 10 years after the collection
date. After verification, selected voucher specimens may be incorporated
into the permanent macroinvertebrate collection at SWRC.
Periphyton
chlorophyll a and biomass was estimated for rocks collected
in association with each composite sample. For chlorophyll a
analyses, rocks were extracted overnight in alkaline acetone and
optical densities determined at 665 nm and 750 nm (for turbidity)
before and after acidification with a drop of 1 N HCL. Optical
densities were used to determine chlorophyll a concentrations
with correction for phaeophytin (Lorenzen 1967). These rocks were
then scrubbed with small brushes to remove attached organic material
(i.e., the biofilm of algae, fungi, and bacteria). This organic
material was captured on a pre-ashed GF/F filter, dried at 60°C
for >48 h, weighed (dry mass of organic and inorganic matter
on rock surfaces), ashed at 550°C for 5 hours, and then weighed
again (dry mass of inorganic materials). Weight loss during ashing
represents the organic content of the periphyton expressed as
mg or g AFDM/m2. Periphyton chlorophyll a and
biomass are measures of the biofilm that represents macroinvertebrate
food attached to rocks.
Benthic
Organic Matter (BOM) is also a measure of macroinvertebrate food,
but in the form of medium and coarse organic particles (i.e.,
captured by a 250-µm mesh sieve) intermixed among rocks
and finer substrates in the stream bed. BOM was estimated as the
detrital material associated with each processed subsample. After
macroinvertebrates were removed, wet detritus (organic and inorganic
material) was transferred to an aluminum weigh boat and dried
at 60°C for >48 h. The sample was weighed (dry mass of
organic and inorganic materials), ashed at 550°C for 5 hours,
and then weighed again (dry mass of inorganic materials). Weight
loss during ashing represents BOM expressed as mg or g AFDM/m2.
QA/QC
of macroinvertebrate data
Errors
for macroinvertebrate data were measured three ways: sorting errors,
identification/count errors, and identification accuracy. Sorting
error (or efficiency) was measured on at least 5% of samples in
each year by resorting through processed detrital material looking
for macroinvertebrates not found during the first sort. Sorting
error (expressed as a percentage of the total number of macroinvertebrates
found for a sample) averaged 6% across the three years. Error
in macroinvertebrate identifications and counts was assessed by
reexamining the specimens identified in at least 5% of the samples
in each year. Error in macroinvertebrate identifications and counts
(expressed as a percentage of the total number of macroinvertebrates
identified) averaged 1% across the three years. Errors arose due
to incorrect identifications or counts or placing an individual
in the wrong vial. Identifications for all taxonomic groups were
verified by sending voucher specimens for each genus and/or species
the Aquatic Resource Center, 6604 Third Street, College Grove,
TN 37046.
|
