Northern Prairie Wildlife Research Center
We installed water-level recording devices (construction details described below) in the center of each of the 36 study wetlands. We defined a wetland's center as the lowest elevation within a basin as determined with a laser level. We wired each recording device to a steel t-post driven into the wetland's substrate and installed the devices so the top of the lower indicator was flush with the wetland substrate when the indicator was at its lowest position. We installed the devices between 21 April and 13 May 1993 and removed them between 25 August and 16 September 1993. At time of removal, we measured the distance between the two slides and later used this measurement as our estimate of water-level fluctuation in a wetland during the time period each device was in place.
We restricted our analyses to wetlands that contained water at some point during the study period and to wetlands where our devices were not destroyed by livestock (n = 27 wetlands). We used two-way analysis of variance (ANOVA) techniques to assess the effects of land use (tilled versus grassland), wetland class (semipermanent, temporary, and seasonal), and the condition-by-class interaction. The response variable was the difference between the absolute maximum and minimum depth measurements divided by the total area of the catchment to correct for differences in catchment size. We used the catchment sizes determined by Freeland and Richardson (1995) for each wetland. Briefly, they measured the distance from the wetland to the catchment divide along a minimum of 4 transects and computed an area based on those measurements. We used the General Linear Models procedure of SAS (SAS Institute Inc. 1989) to perform ANOVAs. If an effect was significant, we used Fisher's Least Significant Differences (LSD) test (Milliken and Johnson 1984) to isolate differences. We used a logarithmic transformation of the response variable plus 1 to stabilize the variance prior to statistical analysis (Steele and Torrie 1980). For descriptive purposes, we calculated 68% upper (UCL) and lower (LCL) confidence limits (approximates 1 standard error above and below the mean) for our back-transformed least squares means.
|Figure 2. A water-level recording device designed to record maximum and minimum water levels of wetlands over discrete periods of time. Three, 3-cm diameter holes covered with 1-mm-mesh screen attached with silicon caulk allow water to enter the device and exclude debris. An access door cut from the wall of the pipe is held in place with plastic tie strips.|
The device we designed to record the absolute maximum and minimum water levels in wetlands over discrete time periods is similar to the device designed by Bragg et al. (1994) to monitor below-ground water tables in mires. However, our device measures absolute water levels above the bottom substrates of wetlands. Richter (1995) also designed a device for recording maximum and minimum water levels. However, his device will not record minimum water levels less than 12 cm and was not designed to withstand the harsh environment of the Northern Great Plains. Our water level recorder (Figure 2) consists of a commercially available, copper-coated steel rod (length 91 cm) that guides a large float up and down as water levels fluctuate. Indicators above and below the float mark the extent of the maximum and minimum water levels. The guide rod and float are completely enclosed in a section of 7.6-cm inside diameter (I.D.) PVC pipe that is capped at both ends. The float is a piece of 0.2-mm brass shim stock formed into a 6.5 cm diameter cone. Cones are filled with buoyant foam and a circular cap of shim stock is soldered to the base. A 6.5-cm-long piece of 0.4-cm-I.D. brass tubing is inserted through the center of the float. Two slides, one above and one below the float, are pushed by the float. Slides consist of a 2.54-cm, teflon-coated, magnetic stirring bar with in a brass housing (Figure 3). The magnets hold the slides at their last positions, so the distance between the slides minus the height of the float is the distance the water level fluctuates during the time period between installation and the final recording (Figure 4). The device can be easily reset by sliding the magnetic indicators back to positions directly above and below the current level of the float.
|Figure 3. Maximum and minimum water-level indicators are constructed by cutting a piece of 0.2-mm brass shim stock into the above shape. The shim stock is then bent into a rectangular box and soldered along joints A and B. A teflon coated stirring bar (e.g., VWR Scientific # 58948-138; use of brand names does not constitute endorsement by the U.S. Government) is inserted into the housing thus created and the indicator positioned on the guide rod (one above and one below the float).|
|Figure 4. Diagram of a prototype water-level recording device showing how changes in water levels move the float and thus the indicator, providing a measurement of water-level fluctuation.|
To test our prototype devices, we installed them in the 2 semipermanent CLSA wetlands in May 1992. We recorded the maximum and minimum water levels of the devices in September 1992 and reset them for the 1993 field season. In April 1993, we examined the devices for signs of wear and winter damage. We found some corrosion on the copper-coated steel guide rods, replaced them with new rods, and reset the water level indicators. We recorded water levels again in September 1993 and then removed the devices from the wetlands. After disassembly, we examined the devices for corrosion and other obvious problems that would limit their ability to function properly over extended time periods.
For the same May 1992 to September 1993 period, we also equipped the 2 CLSA wetlands with Telog (model WLS-2109) water-level monitoring systems (use of brand names does not constitute endorsement by the U.S. Government). These recorded water levels continuously throughout the study period. We housed the transducers inside a steel pipe sunk into the wetland sediment approximately 1 m below the water/sediment interface. Both units were standardized to read zero at this depth according to manufacturer's specifications. At the end of each year, we compared the water-level fluctuation (absolute maximum depth - absolute minimum depth) recorded by our prototype devices to the water-level fluctuation data collected by the Telog systems.