Jason D. Ekstein and Scott E. Hygnstrom, Department of Forestry, Fisheries and Wildlife; University of Nebraska; Lincoln, NE 68583
The Rainwater Basin, located in Southcentral Nebraska, provides 1,092,000 ha of wildlife habitat and immeasurable water quality benefits with its numerous wetland depressions and interspersed agricultural land. In spring, 90% of the mid-continental population of white-fronted geese (Anser albifrons), 50% of the breeding mallards (Anas platyrhyncos), and 30% of the breeding pintails (Anas acuta) use depressions in the Rainwater Basin for staging and foraging before final flight to breeding grounds in the United States and Canada (Gersib et al. 1990). In the early 1900s, landowners started converting wetlands into additional cropland acreage, with support from state and federal agencies. Draining and filling progressed slowly during the 1920s and 1930s because of a poor economy and lack of efficient equipment. In the 1940s, efforts to convert wetlands into cropland intensified because of a prosperous post war economy and advances in earth moving equipment and farm machinery (Gersib et al. 1992). By 1965, 82% of the 3,907 major wetlands had been eliminated and nearly 65% of the 94,695 wetland acres in the Rainwater Basin were gone (NGPC 1984). By the early 1980s, an estimated 90% of the original wetlands in the Rainwater Basin of Nebraska had been destroyed or altered by draining, filling or ditching (Gersib et al. 1992).
With the demonstrated need for a stable water source to supplement rainfall for agricultural development (Smith 1924), the CNPPID initiated development of a canal system in 1936 to provide water for irrigation of farmland in Gosper, Phelps and Kearney counties of southcentral Nebraska. Water is stored in Lake McConaughy near Ogallala, Nebraska and then released where it is diverted from the Platte River at North Platte, Nebraska to the tri-county area (Fig. 1). The canal structure consists of an unlined furrow dug into the ground and built up along the sides with soil. The canal runs 85 km and ranges from 1 to 12 m wide and 1 to 5 m deep with a maximum capacity of 39 m3/sec. Lateral lines connect at various points along the canal to transport water to adjacent cropland.
Since completion of the canal system in 1941 and the subsequent widespread use of surface irrigation, the groundwater table in Gosper, Phelps and Kearney counties has increased 3 to 34 m (Steele and Wigley 1991). A result from the increased groundwater table has been the seepage into depressions and formation of permanent and semipermanent wetlands. Area residents have reported the formation of wetlands where they had not previously occurred within the past 50 years (Soil Conservation Service (SCS), Natural Resources District (NRD), Extension Service; unpublished data). In some cases this led to a loss of cropland acreage which then led to legal action, making the canal system a controversial issue in Nebraska.
Much speculation has surfaced regarding the impact the canal system has on associated wetlands. Our objective was to describe the occurrence and fate of wetlands associated with this canal system in Southcental Nebraska from 1938 to the present.
We would like to acknowledge the US Fish and Wildlife Service (USFWS) for providing support for this study from the Biodiversity Fund, secured by Nebraska Senator J. R. Kerrey.
We would also like to thank Jon Kauffeld, Jay Maher and Mark Peyton for helpful comments on the manuscript and the local SCS offices, NRD, CNPPID and USFWS for supplying aerial photos and assistance.
STUDY AREA
The study area is located in Southcentral Nebraska in the area known as the Rainwater Basin. The study area is bounded by section lines to include 559 sections (143,104 ha) that are 9.66 km north and south of the Phelps county irrigation canal. This area encompasses all lateral lines and any hydrolic effects of the canal. The area was topographically described by Condra (1939) as a Loess Plains Region. The landscape is characterized by surface depressions and gently rolling upland plains. Average rainfall is 62 cm with the majority occurring during June, July and August. About 80% of the land is cultivated, with the major crops being corn, wheat, and sorghum (Soil Conservation Service 1973). Holdrege silt loam is the dominant upland soil type with Massie, Scott, Fillmore and Butler soils characterizing the depressional areas. Depressional soils are formed by clay particles that move downward in the soil profile to form clay pans. Depressions receive water from rainfall, snowmelt, and irrigation runoff. Semipermanently flooded wetlands characterize the deepest depressions (1 - 2 m) offering the most permanent water regime. Seasonally flooded wetlands occupy the slightly shallower areas (0.5 - 1 m) and usually hold water for 3 to 5 months. Temporarily flooded wetlands occupy the most shallow depressions and usually hold water for up to 2 months. The wetlands range in size from 0.1 to 400 ha with 98% under 4 ha.
METHODS
The only available aerial photos of the study area were taken in 1938, 1941, 1956, 1963, 1969, 1978 and 1981. Photos from 1941, 1956, 1963, 1969 (8 in:1 mi, Department of Defense and SCS) and the 1981 National Wetlands Inventory Maps (USFWS) provided the best resolution and were therefore used for the area analysis. We viewed the series of aerial photos of Gosper, Phelps and Kearney counties by section to identify wetlands and compare the fate of wetlands from year to year. We differentiated wetlands from other habitat types by the presence of water and hydrophilic vegetation types and counted the number of wetlands in each section. We estimated the size of the wetlands to the nearest hectare by comparing them to visible agricultural land-use patterns of a known size. Our null hypothesis was that the number and total area of wetlands remained the same across the years. We used a chi-squared goodness of fit test of means to compare number and area of wetlands. To test the accuracy of our counts and measurements, we compared estimates from 4 in:1 mi photos to estimates from 8 in:1 mi photos that had twice the inherent resolution. Accuracy of wetland counts was 96%. Accuracy of wetland measurements was 92%.
To determine the immediate effect of the canal and associated increasing groundwater levels on the number and area of wetlands, we counted and estimated the area of wetlands within increasing groundwater table contours (Fig. 2, US Geological Survey, 1990) and divided the measurements by the proportional area of each contour. Our null hypothesis was that the number and area of wetlands remained the same across years among the four groundwater change contours. To reduce some environmental variability, we did not include data from the 1956 aerial photos because of dramatically reduced wetland area due to annual precipitation that was only 28 cm that year, compared to a mean annual precipitation level of 62 cm. We used a chi-squared goodness of fit test of means to compare the number and area of wetlands in the contours.
We randomly selected 15 four-section sample areas from the 559-section study area to monitor the fate of specific wetland types during 1938, 1941, 1956, 1963, 1969, 1978 and 1981. The 4 in:1 mi photos (1938 and 1978) were not incorporated in the previous analysis because of their lower resolution. Five of the sample areas were devoid of wetlands, during any of the seven study periods. Thus, we randomly selected five new areas that incorporated wetlands. Aerial photos from each sample area for all available years were scanned and computer files were formed that contained geographic information for each area. We counted and measured the size of wetland areas using digital image processing (NIH-Image 1.43, Wayne Rashand). This system discriminates habitat types by comparisons on a gray scale and determines area by counting pixels (0.5 pixels per m). Our null hypothesis was that the number and area of wetlands types remained the same across the years. We used a chi-squared goodness of fit test of the mean to compare number and area of wetlands types.
Wetland types were classified by size (small - <2 ha, medium - 2 to 10 ha, large - >10 ha) from the 1938 aerial photos for each of the 15 sample areas. We then monitored the geographic location of each wetland year by year to determine their fate. Similarly, wetlands were classified from the 1981 photos (excluding wetlands monitored from 1938) and the geographic locations of each were monitored backwards year by year to 1938 to determine if their formation was associated with the presence of the canal. We used NIH-Image 1.43 to determine the area of each wetland. Our null hypothesis was that the number and area of wetlands remained the same across the years.
We used NIH-Image 1.43 to record the locations and area of hydric soils from old soil survey maps (1 in:1 mi) from Gosper (1934), Phelps (1917) and Kearney (1927) counties and from the latest soil survey maps (4 in: 1 mi; 1981, 1973 and 1984 respectively). We overlayed the old and new geographic images for each of the 15 sample areas and compared the location and area of hydric soils on each sample area to further document the fate of wetlands over time. To reduce the effect of map scale and error on map comparisons, we used a significant difference of 100% and -50% when comparing areas of wetland soils from the different time periods. For example, a change from 4 ha to 8 ha is an increase of 100% and a change from 4 ha to 2 ha is a -50% change. We assumed that the location and area of hydric soils would remain constant across the years for the 15 sample areas. The change is symptomatic with the problem of measuring wetlands that have an inherently indeterminate boundary (Kuzila et al. 1991).
RESULTS AND DISCUSSION
We observed no overall change in the number of wetlands in the 559-section study area from 1941 to 1981 (Fig. 3). This finding is notable, considering the overall loss of 90% of wetlands throughout the entire Rainwater Basin during the same time period (Gersib et al. 1992). The total number (X2 = 141.7772, df = 4, P < 0.001) and area (X2 = 3946.6558, df = 4, P < 0.001) of wetlands did change from year to year during the 1941 to 1981 time period. No clear pattern of influence is apparent for the canal system because of differences in rainfall, land use and the time of year when aerial photos were taken. The year 1941, was an average year for precipitation (62 cm). The number of wetlands is high but area is low because rainfall came early in the spring and the photo was taken in September when the landscape was dry and entire basins could be hayed. The drop in number and area in 1956 can be attributed to the low annual precipitation of 28 cm. The year 1963 had an average annual precipitation (62 cm) and the increase in number and area of wetlands can be attributed to an increase in the use of irrigation water (400% increase from 1941 to 1963, Gersib et al. 1990) which aides in increasing number and size of wetlands through runoff. The total area of wetlands in the study area in 1969 is high because of an annual precipitation of 84 cm. Over the entire study area in 1969, we observed small adjacent wetlands combining into larger wetlands, thus increasing the area yet decreasing the number of wetlands. In 1981, the number of wetlands remained high but total area declined. The 1981 NWI photos were taken in May, before the seasonal rainfall and irrigation had occurred. The total number may have not changed because the increased groundwater table, due to the canals, may have provided supplemental water to the depressions adjacent to the canals. While the depressions were not filled by the supplemental water, they did retain some water when they otherwise would have been dry.
In the increased groundwater table contours, we observed differences (Fig. 4) in the number of wetlands in the 12-m (X2 = 10.3632, df = 3, P = 0.0157) contour and no difference in the 18-m (X2 = 3.7639, df = 3, P = 0.2881), 6-m (X2 = 1.3908, df = 3, P = 0.7077) and 0-m (X2 = 6.9106, df = 3, P = 0.0708) contours. The difference within the 12-m contour can be attributed to the fact that 40% of the study area and 49% of the canal was located within the 12-m contour. In addition, draining, filling and irrigation runoff likely added to the variability in numbers of wetlands in all contours. The number of wetlands in the 18-m contour was consistent only because 8% of the total area is contained within the contour. The number of wetlands in the 6-m and 0-m contours show no difference, indicating that groundwater recharge and irrigation runoff may have supplemented water to some depressions. Differences in area are observed in all contours (18-m (X2 = 21.8043, df = 3, P < 0.001), 12-m (X2 = 682.9291, df = 3, P < 0.001), 6-m (X2 = 252.6948, df = 3, P < 0.001) and 0-m (X2 = 250.9792, df = 3, P < 0.001)). The area of wetlands within the 18-m contour were stable until 1981. Photos in 1981 were taken before seasonal rainfall and irrigation occurred, thus decreasing the area in all the contours and causing the difference in the 18-m contour. Area in the 6-m and 0-m contours increased across the years until the 1981 photos, due to irrigation runoff and groundwater recharge from the canal system. The 12-m contour had the largest increase in area in 1963 which also dropped in 1969. The large increase is an indication of supplemental water from the groundwater and/or canal. The drop in 1969 may be attributed to an 83% increase in landleveling activity from 1963 to 1969 (Gersib et al. 1990) which drained and filled wetlands and shaped the landscape to increase irrigated cropland acreage. By this time, center pivot irrigation was developed and additional landscaping was done to allow full circle irrigation.
Detailed observations made on the 15 sample areas (Fig. 5) indicate a difference in the number of temporarily flooded (X2 = 86.3873, df = 6, P < 0.001) and seasonally flooded (X2 = 20.9966, df = 6, P = 0.0018) palustrine wetlands during 1938 to 1981, with a significant increase in concentration pits (X2 = 394.4689, df = 6, P < 0.001) from 1969 to 1981. The variability in the number of temporarily flooded wetlands across the years is expected. They are unstable because they occupy very shallow depressions and are dependent on rainfall and/or runoff. The decline in the number of temporarily flooded wetlands is in part, a response to the increase in land leveling and the number of concentration pits which can hold water year round. The number of seasonally flooded wetlands increased across the years with the supply of a stable water source from the canals, even with the implementation of concentration and reuse pits to drain wetlands and capture irrigation runoff.
Differences were also observed in total area of temporarily flooded (X2 = 170.2897, df = 6, P < 0.001) and seasonally flooded (X2 = 239.4944, df = 6, P < 0.001) palustrine wetlands, with a significant increase in the total area of concentration pits (X2 = 146.1286, df = 6, P < 0.001). The total areas of temporarily and seasonally flooded wetlands fluctuated together, indicating that each responded similarly to the effects of changes in land use and environmental variability. While pits do supply surface water, they have altered the natural wetland hydrology and vegetation of surrounding wetlands (Gersib et al. 1992) by concentrating the water in the pits and stopping the flow of runoff into nearby depressional areas. After 1969, pits were located throughout the study area. Without the presence of these pits, the number and size of naturally-occurring wetlands would likely have been greater.
We monitored 168 small, 30 medium and 11 large wetlands from 1938 to 1981 (Fig.6) on the 15 sample areas to determine the fate of individual wetlands over time. In 1938, the 168 small wetlands ranged in size from 0.1 to 2.0 ha. By 1941, 134 of the 168 wetlands had disappeared but the remaining 34, ranged in size from 0.1 to 18.2 ha. By 1981, only 26 of the original 168 small wetlands remained. In 1938, 30 medium-sized wetlands ranged in size from 2 to 7.5 ha. By 1941, 13 of the medium wetlands remained and ranged in size from 0.1 to 24.2 ha. In 1938, 11 large wetlands ranged in size from 10.4 to 78.0 ha. By 1941, 8 large wetlands remained with a range of 0.1 to 95.7 ha. As individual wetlands disappeared, we observed the growth and decline of existing wetlands and the formation of new wetlands. The reduction in number of small wetlands after 1938 is attributed to temporary wetlands only being present when conditions are favorable. While the number and size of small wetlands fluctuated, the mean area stabilized from 1963 to 1981, indicating that supplemental water from the canals are supplying depressions with water. The size of these wetlands, however, decreased as wetland draining and filling continued. The number of medium-sized wetlands was stable across the years and the mean area fluctuated with annual rainfall. The end result is an increased stability in the number and area of medium-sized wetlands by 1981. The number of large wetlands appeared to be stable across the years, however, 4 of the 11 original large wetlands were less than 2 ha by 1981. The other 7 increased in size and kept the mean area relatively high across the years. These 7 wetlands appear to have received supplemental water from the canal and/or groundwater. The 4 that decreased were drained or ditched.
We monitored an additional 165 small, 7 medium and 3 large wetlands from 1981 back to 1938 (Fig. 7) on the 15 sample areas to determine if they were naturally-occurring or if they may have been the result of surplus groundwater from the canal. In 1981, the 165 small wetlands ranged in size from 0.1 to 2 ha. In 1978, only 51 wetlands existed and ranged in size from 0.1 to 10.8 ha. None of the wetlands selected in 1981 existed in 1938. In 1981, 7 medium-sized wetlands ranged in size from 2.6 to 8.6 ha. In 1978, only 5 medium-sized wetlands existed and ranged in size from 1.4 to 4.6 ha. All but one of these existed back to 1941, and none were present in 1938. Three large wetlands, ranging in size from 16.5 to 20.9 ha were present from 1981 back through to 1963, although the sizes of each varied considerably. Only one of the three was present from 1956 back to 1941, and even that one was not present in 1938. Wetlands monitored from 1981 back to 1938 show that these wetlands formed at a time coinciding with the development and use of the Central Nebraska Irrigation canal system. We speculate that the increased groundwater table caused by the canal resulted in the formation of several small, medium and large wetlands in the area.
Old soil survey maps (1 in:1 mi) represent the location of hydric soils for that period of time and the new soil survey maps (4 in:1 mi) give a more detailed representation of where hydric soil locations exist today. We overlayed the old and new soil survey maps in the 15 sample area and, when comparing aerial delineation, we found a match of 716 ha, a loss of 982 ha and a gain of 825 ha of hydric soils. A survey of the Edgar Northwest quadrangle in nearby Clay County, Nebraska similarly showed a match of 758 ha, a loss of 1,147 ha and a gain of 477 ha (Kuzila et al. 1991). Four sites in the study area (2, 3, 5 and 14) increased in hydric soils and two sites (8 and 10) decreased (Fig. 8). Sites 9, 11, 12 and 13 showed declining trends in hydric soils but the differences were not significant. Five sites (1, 4, 6, 7 and 15) showed relatively little difference in the amount of hydric soils. We feel this change in hydric soil location is more a function of mapping conventions and personnel change than of a location to the canals or groundwater increase.
Determination of the effects of the Central Nebraska Irrigation canal system on associated wetlands is confounded by variation in the environment and land use over time and by the timing of aerial photos relative to rainfall events. Stabilization of the number and area of wetlands within the study area is in stark contrast to the dramatic loss of wetlands documented in other areas of the Rainwater Basin. Despite draining and filling of wetlands, the stabilization is attributed to the increased groundwater table which is a direct result of the canal system. Wildlife benefits of the canal system are expected to be additive, but have yet to be studied.
REFERENCES
Condra, G. E. 1939. An outline of the principal natural resources of Nebraska and their conservation. University of Nebraska Conservation and Survey Division. Lincoln, Nebraska. Bulletin Number. 20.
Gersib, D., J. Cornely, A. Trout, J. Hyland, and J. Gabig. 1990. Concept plan for waterfowl habitat protection, Rainwater Basin Area of Nebraska, Category 25 of the North American Waterfowl Management Plan. Nebraska Game and Parks Commission, Lincoln, Nebraska. 71pp.
Gersib, R. A., K. F. Dinan, J. D. Kauffeld, M. D. Onnen, P. J. Gabig, J. E. Cornely, G. E. Jasmer, J. M. Hyland, and K. J. Strom. 1992. Rainwater Basin Joint Venture Implementation Plan. Nebraska Game and Parks Commission, Lincoln, Nebraska. 56pp.
Gersib, R. A., R. R. Raines, W. S. Rosier, and M. C. Gilbert. 1990. A Functional Assessment of Selected Wetlands Within the Rainwater Basin Area of Nebraska. Nebraska Game and Parks Commission, Lincoln, Nebraska. 41pp.
Kuzila, M. S., D. C. Rundquist and J. A. Green. 1991. Methods for estimating wetland loss: The Rainbasin region of Nebraska, 1927-1981. Journal of Soil and Water Conservation, November-December 1991:441-446.
Nebraska Game and Parks Commission. 1972. Survey of Habitat Work Plan K-71, W-15-R-28. Nebraska Game and Parks Commission, Lincoln, Nebraska. 52pp.
Nebraska Game and Parks Commission. 1984. Survey of Habitat Work Plan K-83, W-15-R-40. Nebraska Game and Parks Commission, Lincoln, Nebraska. 23pp.
Smith, F. F. 1924. Report on the Tri-County Project Irrigation and Power Possibilities, Gosper, Phelps, Kearney, and Adams counties Nebraska. Department of the Interior, Bureau of Reclamation. 1pp.
Soil Conservation Service. 1973. Soil survey of Phelps County, Nebraska. Washington, DC: Soil Conservation Service, U. S. Department of Agriculture, and Conservation and Survey Division, University of Nebraska-Lincoln. 1-10pp.
Steele, G. V. and P. B. Wigley. 1991. Groundwater levels in Nebraska 1990. Nebraska Water Survey Paper No. 69. University of Nebraska Conservation and Survey Division, Lincoln, Nebraska. 33pp.
LIST OF FIGURES
Figure 1. Central Nebraska Irrigation canal system, wetland study area and 15 sample areas in Southcentral Nebraska.
Figure 2. Contour areas of significant groundwater level increase (m) in Gosper, Phelps, and Kearney counties from 1940-1990 (US Geological Survey, 1990), relative to the Central Nebraska Irrigation canal system.
Figure 3. Total number and area (ha) of wetlands in the study area from 1941-1981.
Figure 4. Total number and area (ha) of wetlands in the study area from 1941-1981 relative to the change in the groundwater level over time.
Figure 5. Total number and area (ha) of palustrine wetland types from 1938-1981 in 15 sample areas in Southcentral Nebraska.
Figure 6. Total number and area (ha) of small, medium and large wetlands within 15 sample areas monitored from 1938 to 1981.
Figure 7. Total number and area (ha) of small, medium and large wetlands within 15 sample areas monitored in 1981 and going backward to 1938.
Figure 8. Area (ha) of hydric soils shown in old (1917, 1927 and 1934) and new (1973, 1981 and 1984) soil survey maps in the 15 sample areas in Southcentral Nebraska.
Return to 1995 Platte River Basin Ecosystem Symposium 
Last updated by Darren A. Jack on 9/16/97