Northern Prairie Wildlife Research Center
We collected soil samples from 19 wetland pairs located throughout the PPR of North Dakota (Figure 1) during a single visit to each site; we sampled 14 wetland pairs during the summer of 1995 and an additional 5 pairs during the summer of 1996. Wetland pairs were identified from 40 km² hexagons that were established by the EPA to monitor wetland condition for EMAP. Each wetland pair consisted of a randomly selected temporary wetland and the closest temporary wetland of the opposite condition class. The condition classes we identified for this study were 1) temporary wetlands with catchment basins (i.e., area immediately surrounding each wetland from which surface water flows into the basin) within intensively farmed agricultural fields (hereafter referred to as "cropland wetlands") and 2) temporary wetlands with catchment basins entirely within U.S. Fish and Wildlife Service Waterfowl Production Areas or similar grassland habitat (hereafter referred to as "grassland wetlands"). The spatial location and classification of each wetland was verified in early spring when basins were flooded from snow melt.
From each study wetland, we collected ten 500-ml soil samples. First, we located the lowest elevation of the basin using a Spectra-Physics Model 650 Laserplane. Here, we established a 2 × 2 m quadrat comprised of a 20 × 20 grid of four hundred 100-cm² cells. We collected one 250-ml sample of soil from each of 20 randomly selected cells within the quadrat. Samples were taken from the top 5 cm of soil using a hand trowel after loose vegetative litter was removed from the soil surface by hand. All soil samples from a single wetland were combined in a large, plastic mixing bowl and homogenized using a hand trowel. Homogenized samples were subdivided into ten 500-ml subsamples, placed in polypropylene jars, frozen, and stored at 0°C until processed.
|Figure 1. The prairie pothole region of North Dakota showing locations of EMAP hexagons and wetland pairs sampled. Solid hexagons contained the paired wetland sites sampled in this study. Numbers next to hexagons indicate number of wetland pairs in hexagons with more than one pair.|
We sorted 5 of the 10 soil samples collected from each wetland by hand to estimate the abundance of cladoceran resting eggs (ephippia) and other invertebrate remains. Soil samples were thawed for 24 hours before processing. We then passed samples through a 0.5-mm standard soil sieve and rinsed them with water to remove small soil particles and other material; residues larger than 0.5 mm (e.g., ephippia, snail shells, ostracod shells) were retained on sieve screens. Residues were then placed in a 23 × 36 × 5 cm Pyrex® baking dish, covered with 1-2 cm of water, and examined on a light table. We sorted all residues and transferred ephippia and all other invertebrate remains to petri dishes; every fifth sample was reexamined to verify completeness of sorts. Ephippia and other invertebrate remains were identified, enumerated, and recorded.
To estimate relative abundance of small invertebrate eggs that were easily missed by hand-sorting (i.e., those of anostracans, conchostracans, notostracans, and ostracods), the remaining 5 soil samples from each wetland were incubated for 8 weeks in 37.8-L glass aquaria filled with water maintained under standardized conditions of light, temperature, and water chemistry described below. Many authors have shown that light, temperature, oxygen, and osmotic conditions are the most important factors influencing diapause release and development of resting eggs (e.g. Ogi et al. 1951, Panchella and Stross 1963, Proctor 1964, Stross and Hill 1965, Stross 1966, Shan 1970). Resting eggs collected from the field during dry conditions have usually been exposed to dynamic fluctuations in temperature and moisture and will hatch under optimal water and chemistry regimes (Moritz 1987).
Before samples were placed in aquaria, specific conductance of the water was adjusted to 700 µS cm-1 by diluting well water with distilled water. We illuminated each aquarium using a full-spectrum flourescent tube set on a timer to provide 12 hr of light each day. Water temperatures were maintained at 10° C for the initial 28 days of the incubations and then increased and maintained at 20° C for an additional 28 days to simulate early spring and late summer thermal regimes. After hatching, invertebrates were counted and removed as soon as accurate identifications could be made. All invertebrates were removed before they reproduced.
We used analysis of variance (ANOVA) techniques to assess effects of wetland condition class (grassland or cropland) on the number and viability of cladoceran ephippia and the number and taxon richness of other invertebrate eggs and remains in soil of the temporary wetlands. We conducted the ANOVAs using the general linear models procedure (PROC GLM) of SAS (SAS Institute, Inc. 1989) with a randomized block design; each wetland pair represented a block. Counts from individual samples were summed for each wetland prior to analysis. We calculated viability of cladoceran ephippia as the ratio of the number of cladocerans successfully incubated in aquaria to the number of ephippia found by hand-sorting. We transformed all data, except taxon richness, to stabilize variance (Steel and Torrie 1980); a square root transformation was used on the viability data, and remaining variables were converted to their natural logs. Here, we report back-transformed means and confidence intervals. Due to a lack of cladoceran ephippia in most cropped wetlands (only 5 of 19 contained ephippia), we restricted analysis of our viability data to the 5 wetland pairs that had paired values for both grassland and cropland wetlands. More precise estimates of viability resulted when more ephippia were present (i.e., 75 of 100 ephippia hatching was a better estimate that 75% of the ephippia in a wetland were viable than if 3 of only 4 hatched); thus, in our ANOVAs, we weighted our viability estimates by the number of ephippia found by hand-sorting to give more weight to the better estimates.