Effect of reduced hydroperiods on seasonal woodland pool invertebrate communities

Sixteen artificial woodland pools were constructed in the red maple dominated forest south of the shore of Grapevine Point at the University of Michigan Biological Station in Pellston, MI. The pools were placed in natural depressions when possible and lined with an approximately 1.0 x 1.3 m blue Polyethylene tarp. Pools were filled partially with well water and then to capacity with two subsequent rains. Each pool was seeded for organisms with 200 mL of randomly selected, typical decaying leaf litter from two large, natural seasonal woodland pools on Grapevine Point. The artificial pools therefore started with random populations based on what organisms were in the leaf litter.

Decreased hydroperiods were induced by pumping water out of each pool using a hand pump with a 154 micron zooplankton net as a filter. Each pool was subjected to two dry-down events. The dry-downs took place in two stages, with a maximum depth of approximately eight centimeters for the first stage and all water being removed for the second stage. Natural evaporation was allowed to take place for the remainder of the dry-down event and any water added by rain was removed by pumping back to the eight centimeter depth. Pools were randomly assigned to the control or experimental group, with eight pools in each group. Control pools remained drawn down for two days (drawn down to eight centimeter on the first day), while experimental pools remained dry for nine days during the first dry down period and eight days during the second dry down period (drawn down to eight centimeters after four days). At the end of each dry-down event, the pools were refilled, once again using well water. Invertebrates were collected were collected at the beginning and after a short recovery period after the refilling of the pools for both dry-down events.

To sample invertebrates suspended in the water column, a turkey baster was used to remove one liter of water from each pool by running random transects across the diameter just above the leaf litter. The water sample was run through a 153 micron zooplankton net and 125ml of sample was stored in a zooplankton bottle for transfer back to the laboratory. In the laboratory, 70 ml of each sample was filtered again through a 153 micron nytex filter. The filter was then rinsed into a petri dish and the invertebrates were counted using a grid under a dissecting scope. The size of the filter and the magnification of the scope limited identification to those organisms larger than 153 microns. Invertebrates were classified to family level when possible, but are referred to using the established aquatic invertebrate common name categories. Abundance of invertebrates was then recorded as number of organism per volume of water.

To sample benthic macroinvertebrates, man-made leaf packs were placed in the pools (Dobson, 1994; Brooks, 2000). Forty-eight leaf packs (three per pool) were constructed, each using ten leaves from the surrounding forest leaf litter. Quercus rubra, Populus grandidentata, Fagus grandifolia, and Acer rubrum leaves were included relative to their estimated natural abundance in the area. Each leaf pack was wrapped twice with 15-mm-mesh black plastic garden netting, tied with kite string, and marked with plastic flagging. One leaf pack was collected during each sample: before the first dry-down event, after the first dry-down event, and the last at the conclusion of the second dry-down event. In addition, algae tiles that were present during the last week were collected during the last sample. Both the leaf packs and algae tiles allowed a known surface area to be observed and the number of attached invertebrates recorded. Because only two invertebrate groups were present, the abundance of organisms from leaf packs and algae tiles was added into the data collected with the water column sampling.