We compared the macroinvertebrate and amphibian communities of 12 excavated and 12 natural wetlands in western North Dakota, USA, to assess the effects of artificially lengthened hydroperiods on the biotic communities ofwetlands in this semi-arid region. Excavatedwetlandswere much deeper and captured greater volumes ofwater than natural wetlands. Most excavated wetlands maintained water throughout the study period (May to October 1999), whereas most of the natural wetlands were dry by June. Excavated wetlands were largely unvegetated or contained submergent and deep-marsh plant species. The natural wetlands had two well-defined vegetative zones populated by plant species typical of wet meadows and shallow marshes. Excavated wetlands had a richer aquatic macroinvertebrate community that included several predatory taxa not found in natural wetlands. Taxa adapted to the short hydroperiods of seasonal wetlands were largely absent from excavated wetlands. The amphibian community of natural and excavated wetlands included the boreal chorus frog (Pseudacris maculata), northern leopard frog (Rana pipiens), plains spadefoot (Scaphiopus bombifrons), Woodhouse’s toad (Bufowoodhousiiwoodhousii), and tiger salamander (Ambystoma tigrinum). The plains spadefoot occurred only in natural wetlands while tiger salamanders occurred in all 12 excavated wetlands and only one natural wetland. Boreal chorus frogs and northern leopard frogs were present in both wetland types; however, they successfully reproduced only in wetlands lacking tiger salamanders. Artificially extending the hydroperiod of wetlands by excavation has greatly influenced the composition of native biotic communities adapted to the naturally short hydroperiods of wetlands in this semi-arid region. The compositional change of the biotic communities can be related to hydrological changes and biotic interactions, especially predation related to excavation.
Permanent water sources in arid and semi-arid regions are rare because evaporation rates are greater than annual precipitation rates. This lack of permanent water gives seasonal water-bodies in these regions greater ecological significance than in more mesic regions (Williams, 1985) because they are often a limiting resource. The vertebrate and invertebrate communities of seasonal wetlands are composed of unique assemblages of animals that exclude most of the species of more permanent waters (Wiggins et al., 1980). Seasonal wetlands are often considered to be harsh environments because of their short hydroperiods (i.e., the period of time a wetland is flooded) separated by long periods when the wetland is dry. However, harsh environments provide unique opportunities for specially adapted species (Collinson et al., 1995). Productivity is often high in seasonal wetlands due to remineralization of nutrients during the dry period; the shallow waters of seasonal wetlands warm quickly in the spring enabling growth in species with high thermal coefficients; and most large predators are excluded from seasonal wetlands, benefiting many species vulnerable to predation (Collinson et al., 1995). However, because seasonal wetlands are inherently small and shallow, they are very vulnerable to damage by human activities (Collinson et al., 1995; Euliss and Mushet, 1999).
The Little Missouri National Grassland (LMNG) is a 415,000-ha public grassland within an expansive area of natural grassland in western North Dakota, USA, under the management of the U.S. Department of Agriculture's Forest Service (Figure 1). The LMNG and surrounding grasslands are of considerable economic value to the livestock industry, and the area has been extensively modified to enhance the distribution and availability of water resources for livestock. These human-enhanced and developed water sources have enabled ranchers to distribute livestock more evenly throughout the LMNG. Prior to these water developments, livestock were restricted in distribution and grazing pressure was exerted differentially, concentrated around available water sources such as the few permanent wetlands, streams, and springs scattered throughout the area. Recently, the environmental benefits of water developments and improvements within the LMNG have been questioned. Potential environmental concerns range from the effects of hydrologic alterations on aquatic animals and plants in wetlands and streams to impacts on native plant species composition and structure, including related impacts on wildlife that use grasslands.
Water development projects completed over the past several decades in the LMNG have been initiated mostly to enhance water availability for livestock. Stock tanks associated with springs are one of the older types of water developments and are still common. Additionally, earthen stock dams have been constructed in drainages and wetland basins have been excavated to extend the period water is available to livestock. More recently, extensive networks of underground pipelines have been developed that supply water to stock tanks from wells placed within the LMNG. In addition, private and federal conservation agencies have constructed a number of wetlands and shallow impoundments that serve the dual purpose of providing brood-rearing habitat for waterfowl and water for livestock. In all situations, the result has been to put more water on the landscape than was historically present. Here, we compare the hydroperiod and biotic communities (plants, macroinvertebrates, and amphibians) of natural and excavated palustrine wetlands in the LMNG.
The topographic features of the LMNG consist of mostly moderate-relief grassland with steeper gradient areas near the Little Missouri River. The erodability of soils in this region is well documented (Bluemle, 2000) and it is the main factor responsible for the physical appearance of the steeper gradient areas, commonly known as ‘Badlands.’ Average annual precipitation in the LMNG is approximately 38 cm or less (Figure 1). National Wetlands Inventory (NWI) basin coverage for the LMNG identified 2599 palustrine wetlands fully or partially located on federal lands (Table 1). Of these, approximately 21% were identified as having a temporary water regime (i.e., usually go dry in late spring or early summer), 24% as having a seasonal water regime (i.e., usually go dry by mid to late summer), and 55% as having a semipermanent water regime (i.e., only go dry during periods of extended drought). Additionally, 83% of the palustrine wetlands were identified as being altered by dams, excavations, or a combination of the two (Table 1). In fact, the NWI basin coverage identified only 239 temporary wetlands, 184 seasonal wetlands, and 23 semipermanent wetlands on federal lands in the LMNG that have not been altered by impoundment or excavation.
We limited our sample universe to include only wetlands with seasonal water regimes located entirely on federal land. From these wetlands, we randomly selected 12 wetlands modified by excavations (hereafter referred to as excavated wetlands) and 12 wetlands unmodified by any type of water development (hereafter referred to as natural wetlands). To insure that communities had time to adjust to new ecological conditions, we did not include recently excavated wetlands (i.e., wetlands excavated within the last 6 years) in our sample. To represent wetlands of various sizes in our sample, we arranged all potential study wetlands into four groupings based on wetland size, and randomly assigned a number from 1 to n to each wetland within each grouping. We then sorted our list of potential study wetlands by their randomly assigned numbers and selected the first 3 excavated wetlands and the first 3 natural wetlands from each size grouping for a total of 24 wetlands. We assessed the suitability of potential sample wetlands initially with high-altitude aerial photographs supplied by the Custer National Forest followed by inspection by field crews. If any anomalies (e.g., another livestock water source within the wetland basin or catchment) that would preclude a wetland's use in this study were identified, we rejected the site and the next wetland within that size group was selected and evaluated as a potential study site. We repeated this process until 24 wetlands (12 excavated and 12 natural) were selected (Table 2).
We evaluated the hydrology and basin morphometry of study wetlands using monthly (May to October) readings from staff gauges and topographic survey data. During each monthly sampling visit, we recorded the water level and specific conductivity (a proxy for salinity) of each wetland using staff gauges installed in each wetland and an Orion model 126 conductivity meter, respectively. We conducted a detailed topographic survey of each wetland and the adjacent upland using a Nikon, model DTM 750 total station. Topographic data from our total station survey were used to develop three-dimensional models that we used to calculate water volumes and surface areas of wetlands at water levels observed throughout the summer (Tripod Data Systems, 1997).
In mid-July, we sampled plant communities by placing two 1 m × 1 m plots (one in the center of the wet-meadow zone and one in the shallow-marsh zone; Stewart and Kantrud, 1971) along randomly placed transects in each wetland. We recorded species composition, percent cover by species (Daubenmire, 1959), and percent unvegetated area within each plot. To facilitate our ability to identify species under intense grazing pressure, we installed two circular cattle exclosures 10 m away from each sample plot. Exclosures were constructed of 1.2 m by 4.9 m commercially available livestock panels with the ends joined together to form a circle approximately 1.5 m in diameter. Each exclosure was held in place by steel T-posts driven into the ground. At each study wetland, we placed one exclosure in the shallow-marsh zone and a second in the wet-meadow zone. Plant specimens collected from exclosures were identified to species using Flora of the Great Plains (Great Plains Flora Association, 1991 ) and used to aid in the identification of plants exposed to grazing in our plots.
We used sweep nets (Swanson et al., 1974) to collect three aquatic macroinvertebrate samples monthly (May–October) from random locations within each study wetland. The 20.3 cm × 45.7 cm nylon sweep net had 8 meshes per cm with 1 mm apertures. We standardized the horizontal distance covered by each sweep to 61 cm with the aid of a U-shaped rod to provide a sample from 57 l of wetland water. The small sizes of the wetlands and the relatively uniform vegetation within wetlands negated the need to stratify sampling by plant communities. We preserved samples from each sweep in 80% ethanol until we were able to identify taxa and estimate abundance. After we identified and counted each taxon, the samples were archived at the Northern Prairie Wildlife Research Center. We compared taxonomic richness, relative abundance of total and individual taxa, and foraging guild (Merritt and Cummins, 1984) ratios between excavated and natural wetlands using analysis of variance and compositional analysis (Aitchison, 1986).
We sampled amphibians along wetland edges and in the adjacent upland area using visual encounter surveys (VES) (Crump and Scott, 1994) and from within wetlands by seining each month (May–October) with a 15-m minnow seine (6-mm mesh). We standardized the area of each wetland seined by anchoring one end of the seine to a pole located at the wetland edge. We then pulled the other end of the seine to make a 180° arc through the wetland, starting and ending on the shore. We identified larval amphibians to the lowest taxon possible and adults to species or subspecies using Conant and Collins (1991). For each seine sample, we counted the number of individuals of each taxon. We then compared amphibian populations between excavated and natural wetlands.
Excavated wetlands had greater water depths, held greater volumes of water, and had longer hydroperiods than natural wetlands (Table 3). Maximum water depth of all but three excavated wetlands was greater than 1 m when we first sampled in May. The water depth of each wetland gradually decreased throughout the summer, declining by an average of 0.68 m (sd = 0.37 m) by October. Six of the 12 excavated wetlands still had water deeper than 1 m during our last sampling in October; only a single excavated wetland went dry during 1999. Water volumes during May in excavated wetlands ranged from 54 to 2849 m3 (Table 3). The average volume of water retained in excavated wetlands ranged from 1006 m3 (sd = 959 m3) in May to 387 m3 (sd = 472 m3) by October. In contrast, all 12 natural wetlands had a maximum water depth less than 1 m in May and most (8 of 12) were dry on our second visit in June; all 12 wetlands were dry by August. One natural wetland, dry in June and July, refilled by August from isolated thundershowers and contained standing water through our October sampling period. Water volume in the natural wetlands in May ranged from 3 to 370 m3. The average volume of water held in the natural wetlands in May was 81 m3 (sd = 101 m3).
In addition to being deeper and containing greater volumes of water, excavated wetlands also had higher concentrations of dissolved salts than natural wetlands (Table 3). Specific conductance readings in May in all but one excavated wetland were greater than 100 μS cm− 1. In contrast, only four natural wetlands had specific conductance values greater than 100 μS cm− 1. Specific conductance of wetlands increased seasonally in response to water loss from evapotranspiration. One excavated wetland (1E, Table 3) had higher specific conductance than other wetlands that appeared to result from extremely salty groundwater entering the basin from groundwater seeps around the basin's periphery.
Our survey data suggest that the plant communities of natural and excavated wetlands are very different. We observed 64 plant species in our study wetlands. Twelve species were common to both excavated and natural wetlands, while 22 species were found only in excavated wetlands and 30 were found only in natural wetlands. Many wet-meadow and shallow-marsh plant species (Stewart and Kantrud, 1971) were conspicuously absent from the excavated wetlands but were found in natural wetlands. These species included needle spikerush (Eleocharis acicularis), common spikerush (Eleocharis macrostachya), baltic rush (Juncus balticus), and several sedge species (Carex spp.). Additionally, submergent species such as pale watermilfoil (Myriophyllum spicatum), sago pondweed (Potamogeton pectinatus), baby pondweed (Potamogeton pusillus), and claspingleaf pondweed (Potamogeton richardsonii), and the deep-marsh emergent hardstem bulrush (Scirpus acutus) were absent from the natural wetlands. Of special significance was the difference in the percentage of unvegetated area between excavated and natural wetlands. Eight of the 12 excavated wetlands lacked vegetation in greater than 90% of the area in our shallow-marsh plots; four excavated wetlands did not have any visible vegetation inward from the wet-meadow zone. In contrast, only four of the 12 natural wetlands had unvegetated area in our shallow-marsh plots.
The number of aquatic macroinvertebrate taxa we observed in each wetland from May until October, 1999, differed between excavated and natural wetlands (F1,22 = 13.22, P = 0.0015). We observed an average of 35.3 (SE = 3.0) taxa in excavated wetlands and only 20.1 (SE = 3.0) taxa in natural wetlands. Eight natural wetlands had 21 or fewer taxa while all 12 excavated wetlands had greater than 21 taxa (Figure 2). We did not identify any significant differences in foraging guild composition between excavated and natural wetlands (λ = 0.706, F5,18 = 1.50, P = 0.2385). However, some taxa such as the mosquito Aedes, the fairy shrimp Eubranchipus, and the snails Promenetus umbilicatellus and Stagnicola caparata occurred only in the natural wetlands. The leeches Erpodella and Helobdella, the damselfly Enallagma, the dragonfly Aeshna, and the pill clams Sphaerium lacustre, Pisidium casertanum, and Pisidium compressum occurred only in the excavated wetlands.
We observed the boreal chorus frog (Pseudacris maculata), northern leopard frog (Rana pipiens), plains spadefoot (Scaphiopus bombifrons), Woodhouse's toad (Bufo woodhousii woodhousii), and tiger salamander (Ambystoma tigrinum) in our survey of excavated and natural wetlands in the LMNG. We also encountered the plains garter snake (Thamnophis radix) and the bullsnake (Pituophis melanoleucus sayi). While adult boreal chorus frogs and northern leopard frogs were present in almost all wetlands in the spring, they only successfully reproduced (i.e., we observed tadpoles) in wetlands that were void of tiger salamanders. We encountered tiger salamanders in all 12 excavated wetlands, but in just a single natural wetland (Figure 3); the observation in the natural wetland was represented by a single adult. The plains spadefoot occurred only in 3 wetlands, all of which were natural. We encountered the Woodhouse's toad, plains garter snake, and bullsnake infrequently around the edges of both excavated and natural wetlands.
Our data indicate that seasonal wetlands in the LMNG have been substantially altered by excavation. Consistent with their design and objectives, excavated wetlands ponded more and possessed deeper water, and contained water for longer periods than natural wetlands. However, in addition to providing a more reliable water source for livestock, these changes have significantly impacted the native plant and animal communities of wetlands in the LMNG. The impact of hydrologic alterations alone is sufficient to change the aquatic invertebrate composition (Euliss et al., 1999) and data from our study suggest significant alteration in the plant and amphibian community as well.
Excavation has greatly altered the plant community of wetlands in the LMNG. The natural wetlands had two well-defined plant zones populated by typical wet-meadow and shallow-marsh species, whereas the excavated wetlands were largely unvegetated or contained submergent and deep-marsh plant species typical of the deeper waters and extended hydroperiods resulting from excavation. Vegetative structure has a significant influence on aquatic macroinvertebrate communities (Euliss et al., 1999). Macrophytes increase habitat structural complexity within wetlands, providing food and living space for species that would not otherwise be present (e.g., Berg, 1949, 1950; Krull, 1970; Gilinsky, 1984; Bergey et al., 1992). They also function as sites for oviposition (Sawchyn and Gillott, 1974a, b, 1975), emergence (Sawchyn and Gillott, 1974a, b), respiration (Batzer and Sjogren, 1986), attachment (Campbell et al., 1982), and pupation (Butcher, 1930) and can modify predator-prey interactions in wetlands (e.g., Rabe and Gibson, 1984; Gilinsky, 1984; Batzer and Resh, 1991).
Additionally, aquatic plants can alter the physical and chemical environment of the water. Aquatic plants influence water circulation patterns and contribute to gradients in light, temperature, and dissolved oxygen (e.g., Kollman and Wali, 1976; Carpenter and Lodge, 1986; Rose and Crumpton, 1996). Further, anoxic conditions that can impact macroinvertebrate abundance, movement, and behavior (e.g., Murkin and Kadlec, 1986; Murkin et al., 1992) can also be altered by stands of emergent vegetation (Suthers and Gee, 1986; Rose and Crumpton, 1996). All of these factors interact to result in different macroinvertebrate communities often being associated with different plant communities (e.g., Voigts, 1976; McCrady et al., 1986; Wrubleski, 1987; Olsen et al., 1995).
In addition to vegetative composition, hydroperiod also directly influences the invertebrate community of wetlands. In order to persist in temporary and seasonal wetlands, aquatic macroinvertebrate species must possess two essential life cycle adaptations: an ability to withstand the period when the wetland is dry and an ability to synchronize life cycles with the occurrence of water (Williams, 1985). Wiggins et al. (1980) placed animal inhabitants of seasonal waters into four groups based on these lifecycle adaptations:
Group I. Overwintering Residents. These permanent residents are incapable of active dispersal. They exist in a dormant state during dry periods as drought-resistant cysts and eggs or as juveniles and adults.
Group II. Overwintering Spring Recruits. These overwinter in the dry basin mainly as eggs or larvae but actively disperse in the spring.
Group III. Overwintering Summer Recruits. These active dispersers are not dependent on the presence of surface water for reproduction. Oviposition occurs in the dry wetland basin and they overwinter mainly in the egg stage or as larvae.
Group IV. Non-wintering Spring Migrants. These active dispersers enter seasonal pools in the spring but must leave the pool before it dries, overwintering mainly in permanent waters.
Given the potential pool of Group I, II, III, and IV macroinvertebrates in the LMNG, hydroperiod is the primary factor influencing aquatic communities in freshwater wetlands. Schneider (1999) presents a model of community organization for wetland invertebrates that demonstrates how pond duration acts as a filter; only taxa with life history strategies compatible with pond duration are able to pass through the filter and inhabit the pond. Invertebrates with long life cycles such as dragonflies and predaceous diving beetles are filtered out and are unable to exist in the shortest hydroperiod wetlands. However, as hydroperiod lengthens, the filter becomes more porous and an increasing number of aquatic invertebrates can occupy the wetland. In our study, we observed that excavated wetlands had the greatest richness of aquatic macroinvertebrate taxa (Figure 2). It is important to note that while short-hydroperiod wetlands may support relatively fewer aquatic invertebrates, it does not mean that they are species poor for the entire annual cycle. Terrestrial invertebrates may be a significant component of the fauna of seasonal wetlands during the dry phase (Collinson et al., 1995).
While we did not detect any significant differences in the foraging guild ratios, we encountered several predatory taxa (e.g., the leeches Erpodella and Helobdella, the damselfly Enallagma, and the dragonfly Aeshna) only in excavated wetlands. Other non-predatory taxa also occurred only in the excavated wetlands likely due to the artificially lengthened hydroperiods. Schneider and Frost (1996) found that predators from long-hydroperiod ponds fed preferentially on prey taxa such as Aedes mosquitoes and the fairy shrimp Eubranchipus that are most characteristic of short hydroperiod ponds. Aedes and Eubranchipus were not observed in the excavated wetlands we sampled.
While our data demonstrates several impacts of excavation on wetland invertebrates, there are likely additional impacts hidden in the coarse taxonomic resolution of our evaluation. We rarely identified individuals to species and usually only to genera or family. One exception was for the gastropods, which we identified to species. Of the eight snail species, only one (Gyralus circumstriatus) occurred in both excavated and natural wetlands. Gyralus parvus, Helisoma trivolvis, Physa gyrina, Physa jennessi, and Stagnicola elodes occurred only in excavated wetlands while Stagnicola caparata and Promenetus umbilicatellus occurred only in natural wetlands. At the genera level, we would have shown Stagnicola as occurring at equal frequencies in both excavated (3) and natural (3) wetlands. Other macroinvertebrate groups likely had many more species represented by just a single genus or family. A more detailed assessment to a finer taxonomic resolution will likely reveal that other invertebrates adapted to short hydroperiod wetlands have been negatively affected by excavation.
Animal species adapted to survive in seasonal waters are often rare with a restricted distribution. Clarke (1981) describes Promenentus umbilicatellus as being rather uncommon, restricted to vernal pools, marshes, and spring-time flooded margins of intermittent streams. The Eubranchiopoda (which includes the fairy, tadpole, and clam shrimps) are among the most characteristic inhabitants of seasonal ponds and pools (Pennak, 1989). While some Eubranchiopoda are widely occurring, many are quite restricted, and some are only known from a single pool or pond (Pennak, 1989). Currently, five species of fairy shrimp are listed on the federal list of threatened or endangered species (U.S. Fish and Wildlife Service, 2001). Given the often restricted nature of aquatic invertebrates adapted to survive in seasonal waters and their vulnerability to increases in predatory and competitive pressures, care should be taken to avoid artificially extending the hydroperiods of wetlands in semi-arid regions. Additional consideration should be given to restoring natural hydroperiods in altered wetlands where possible.
Semlitsch (2000) provides a checklist of questions useful in developing biologically based management plans for amphibian communities. The first two questions Semlitsch poses are: 1) “What species currently inhabit your region?” and 2) “What was the historical species richness and relative abundance of amphibians in your region?” While the first question can be easily answered, the second is more difficult, but it is of special significance to amphibians in the LMNG. While we do not know the historical species richness and abundance of amphibians in the LMNG, it is clear that the hydroperiods of most wetlands are inherently short. Amphibian species existing in the seasonal waters of LMNG wetlands likely possessed physiological tolerances and life histories compatible with the hydrology of this semi-arid region. The larval period of spadefoot toads, which can be as short as 21 days (Semlitsch, 2000), is an example of a life history adaptation allowing this species to capitalize on the short hydroperiods of wetlands in the region. Spadefoot toad tadpoles are extremely vulnerable to predatory insects and salamanders. Thus, they are dependent on wetlands with hydroperiods too short to support predatory populations (Pechmann et al., 1989). By contrast, the tiger salamander can produce metamorphs only in wetlands that do not dry for at least 3 to 4 months (Semlitsch, 1983). Because both adult and larval tiger salamanders are carnivorous and frequently occur at high densities, they can exert considerable predation pressure on other amphibian species (Morin, 1981). Regulation of amphibian community structure in wetlands occurs through the interaction of hydroperiod, competition and predation (Wellborn et al., 1996); changes to any of these three factors will result in changes to the community structure. The increase in hydroperiods of most wetlands in the LMNG now allows tiger salamanders and predatory insects to exert pressure on anuran tadpoles at levels that did not previously exist.
Historically, wetlands with hydroperiods long enough to support tiger salamander populations were scarce in the LMNG. However, current NWI basin coverage data identified that over half (1441) of the 2599 palustrine wetlands fully or partially on federal lands in the LMNG currently have semipermanent water regimes (Table 1). NWI also identified that 98% (1418) of these basins have been altered by excavations and/or dams to extend their hydroperiods. Historically, most of these wetlands would likely have had temporary or seasonal water regimes consistent with the precipitation events that characterize this semi-arid region. Additionally, of the LMNG temporary and seasonal wetlands identified by NWI (Table 1), 55% and 71%, respectively, have been altered to increase hydroperiods through excavation or impoundment. This shift from a landscape dominated by rangeland interspersed with isolated temporary and seasonal wetlands, and very few semipermanent or permanent water sources, to one in which most wetlands have semipermanent water regimes likely had a negative influence on native amphibian populations.
We observed an apparent range extension of the barred subspecies of the tiger salamander (Ambystoma tigrinum mavortium) in the LMNG. Current range maps show the blotched tiger salamander (Ambystoma tigrinum melanostictum) as being the subspecies occurring throughout most of South Dakota and in southwestern North Dakota with the barred tiger salamander occurring only as far north as northern Nebraska (Conant and Collins, 1991). However, we encountered both barred and blotched tiger salamanders in the LMNG wetlands. While the reasons for this apparent range expansion are unclear, we hypothesize that the creation of numerous wetland habitats suitable for colonization by salamanders across what were historically unfragmented grasslands inhospitable to salamanders may have allowed these two subspecies to intermix. In addition to the numerous wetlands altered and created in North Dakota, over 70,000 retention and pit-retention reservoirs have been created in western South Dakota (Ruwaldt et al., 1979) that may have provided a suitable habitat corridor from Nebraska. In addition, transportation of barred tiger salamanders as bait for sport fishermen may have greatly increased the dispersal rate of this subspecies. Thus, while the natural isolation of permanent water sources in dry landscapes can promote the genetic divergence of populations (Thomas et al., 1998), reduction in this isolation through the creation of additional permanent water bodies can lead to the intermixing of these once isolated populations.
It is not surprising that when the LMNG was first settled by Europeans, grazing was identified as the principal land use. Nor is it surprising that ranchers have enhanced water availability to livestock in this semi-arid region. Excavation of short hydroperiod wetlands has been a common method of enhancing water availability in the LMNG for livestock because it is cost effective. More recently the benefits of wetlands in south-western North Dakota to waterfowl have been used to promote the creation and alteration of additional wetlands. Currently, there is a wildlife program to create over 4000 ha of wetlands in southwestern North Dakota for the mutual benefit of wildlife and livestock (Willis, 2001). The benefits of putting additional water on the landscape in semi-arid regions for waterfowl are well documented (Bue et al., 1964; Lokemoen, 1973; Forman et al., 1996). However, as we found for excavated wetlands, such actions can have significant impacts on native aquatic macroinvertebrate, amphibian, and plant communities. The changes we observed were largely the result of deeper water depths, increased hydroperiods, increased salt concentrations, decreased vegetative structure, and the presence of a diverse assemblage of predatory macroinvertebrates and salamanders in excavated wetlands. Populations of species vulnerable to the changes we documented need to be identified for periodic status assessment, and programs need to be developed to restore the natural hydroperiods and biotic communities of wetlands to increase the availability of these unique habitat types in the LMNG.
Financial support for this project came from the U.S. Department of Agriculture's Custer National Forest and the U.S. Geological Survey. We thank Gary Foli, Clinton McCarthy, and Cindy Ragland for their help identifying the need and for funding this study. We thank the staff of both the Dickinson and Watford City Forest Service Field Offices and the Dakota Prairie Grasslands Office in Bismarck for a wide variety of assistance with this project. We thank Susan Rinehart for verifying plant identifications and Eric Routman for verifying our identification of the barred subspecies of the tiger salamander (Ambystoma tigrinum mavortium). Lastly, we thank all the technicians who assisted with field and laboratory data collection, Deborah Buhl for performing our statistical analysis, and Jane Austin, Robert Gleason, Jay Hestbeck, Douglas Johnson, Wayne King, Murray Laubhan, Mike McEnroe, Jim Ringleman, Dan Svingen, Ralph Tramontano, and Kevin Willis for providing comments on the final report from this project and/or earlier drafts of this manuscript. Any use of trade, product, or firm names in this report is for descriptive purposes only and does not imply endorsement by the U.S. Government.