Effects of land use on vegetation in glaciated depressional wetlands in western Montana by Cynthia S Borth A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Land Rehabilitation Montana State University © Copyright by Cynthia S Borth (1998) Abstract: Human influence and biological condition in streams are commonly assessed with multimetric indices, which combine measurements at several ecological levels into an overall rating of biological integrity. Efforts to develop such biological assessment methods for wetlands are relatively recent. This study examined effects of grazing and cultivation on vegetation in 24 depressional wetlands in western Montana. The two primary objectives of this research were to (1) evaluate differences in vegetation among wetlands with different land uses and (2) identify vegetation characteristics that vary predictably with land use, and, therefore, may be useful in a multimetric index of human influence on wetlands. I examined plant community composition and vertical distribution of species in wetlands surrounded by land uses ranging from relatively undisturbed to currently grazed or cultivated. Sites were selected using physical and chemical criteria to limit natural variability among sites. Environmental data showed that wetlands differed between the two study areas selected to investigate effects of cultivation and grazing but were similar within each study area. I found limited evidence for elevated nutrient concentrations, which might alter vegetation indirectly, in areas of increased disturbance but considerable evidence of direct, physical impacts such as tillage and trampling. Patterns of differences in wetland vegetation among land uses were also consistent with direct, physical impacts of cultivation and grazing and indicated that cultivation caused greater changes in plant communities than grazing. In cultivated areas, the clearest differences were observed in groups and species related to life history, successional role or native origin: annuals, perennials, native perennials, moss, Agropyron species, Eleocharis species, and Juncus balticus. In grazed areas, dominance, Eleocharis species, and Juncus balticus differed among land uses. Vertical distributions of many species shifted to lower elevations at disturbed sites. Because land use appears to affect vascular plant communities in a measurable and ecologically rational way, vascular plants show good potential for use as metrics to assess wetland condition and human influence on wetlands. Because changes in vegetation were greater for cultivation than grazing, plant-based metrics are more likely to succeed in assessing impacts of cultivation.  EFFECTS OF LAND USE ON VEGETATION IN GLACIATED DEPRESSIGNAL WETLANDS IN WESTERN MONTANA by Cynthia S. Borth A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Land Rehabilitation MONTANA STATE UNIVERSITY-BOZEMAN Bozeman, Montana November 1998 h/sl? BU4H ii APPROVAL of a thesis submitted by Cynthia S Borth This thesis has been read by each member of the thesis committee and has been found to be satisfactory regarding content, English usage, format, citations, bibliographic style, and consistency, and is ready for submission to the College of Graduate Studies. Paul Hook Committee Chair (Signature) Date Approved for the Department of Land Resources and Environmental Sciences Jeffrey S. Jacobsen Department Head (Sig / z / ? Ar Date Approved for the College of Graduate Studies Joseph J. Fedock Graduate Dean iii STATEMENT OF PERMISSION TO USE In presenting this thesis in partial fulfillment of the requirements for a master’s degree at Montana State University-B ozeman, I agree that the Library shall make it available to borrowers under the rules of the Library. I f I have indicated my intention to copyright this thesis by including a copyright notice page, copying is allowable only for scholarly, purposes, consistent with “fair use’” as prescribed in the U.S. Copyright Law. Requests for permission for extended quotation from or reproduction of this thesis in whole or in parts may be granted only by the copyright holder. . iv ACKNOWLEDGMENTS I wish to acknowledge the many people who provided their input to this thesis. I would like to thank my advisor. Dr. Paul Hook, for his guidance, as well as committee members Dr. William Inskeep and Dr. F. Richard Hauer. Thanks to Dr. Billie Kerans and Dr. Bret Olson for insight to data analysis and Dr. John Borkowski for assistance with statistical analyses. I would not have been able to collect data required for this project without my invaluable field assistant, Melinda Ulrich. Ted Steiner and Amy Herrera also provided essential help with fieldwork and data organization. This research was funded by the Water Quality Monitoring Section of Montana Department of Environmental Quality (DEQ) and the Montana Agricultural Experiment Station (MAES Project 101199). I particularly thank Randy Apfelbeck of DEQ for his time and suggestions throughout project development. John Grant arid Mike Thompson of Montana Department of Fish, Wildlife, and Parks, Greg Neudecker of US. Fish and Wildlife Service, and Joe Brewster, Bandy Ranch manager, provided very helpful information during site selection. Finally, I would like to thank my friend, Bruce Kelling, for all of his support and patience during this endeavor. V TABLE OF CONTENTS Page INTRODUCTION.......... ......... I Research Obj ectives and Approach............................................... 4 LITERATURE REVIEW............ ........................... 8 Geomorphology, Hydrology, and Water Chemistry.....................................................8 Natural Controls on Plant Distribution...................................................................... -.13 Human Influence on Vegetation..... ................................................................... .15 Assessment of Biological Integrity..............................................................................18 METHODS.......................... .......................................... .......... :..................................... 22 Study Areas.:................................................................. 22 Study Design........................................................................................... 22 Site Selection............................................................... 23 Description of Study Sites..........................i........................................................ 24 Sampling and Analysis...... ......................................................................... 32 Hydrologic Measurements.................................................................................... 32 Water Chemistry and Physical Measurements.............................................. 32 Sediment Chemistry and Physical Measurements.................................................34 Vegetation Sampling.......................................................... 35 Topographic Measurements.................................................................................. 37 Statistical Analysis.......................................................... 38 RESULTS........................ 40 Hydrology........................................................................... 40 . Water and Sediment Characteristics........................................................................... 44 Vegetation........................................................:.........................................................53 Vertical Distribution of Vegetation................................................................. 60 DISCUSSION........................................'................... ............................................. ........63 Site Characteristics........................................................... 63 Land Use Impacts.................................................................................................. . 67 Vegetation..................................................................................................... 70 Ecological Patterns Related to Land Use........................................................... 70 Vertical Distribution of Vegetation......... ......................................... ,................. 77 Conclusion......................................................................................................... .78 RECOMMENDATIONS....................... .................. ............... .........................:............ 82 Metrics................................................ ......................................................... : ....... ....85 Conclusion............!....................................... .............................................................88 TABLE OF CONTENTS—Continued vi Page REFERENCES CITED...................... .......................... : ...................... : ..........................91 APPENDICES..................... .....................................................'.......................................97 Appendix A-Water and Sediment Data.................................. ............. .....................98 Appendix B-Vegetation data.......................... ........................................................ 103 . LIST OF TABLES vii . . LIST OF TABLES ble Page 1. Environmental characteristics and land use information for the Ninepipe study area...........................................................'.................26 2. Environmental characteristics and land use information for the Ovando study area..................... .........................................................29 3. Monthly water depths in the Ninepip e and Ovando areas.............. :.................... 41 4. Direction of groundwater flow in the Ninepipe and Ovando areas.............. 42 5. Physical and chemical characteristics of water in the Ninepipe area....................45 6. Physical and chemical characteristics of water in Ovando area............................46 7. Nutrient concentrations in water in the Ninepipe area.......................................... 48 8. Nutrient concentrations in water in the Ovando area............................................ 49 9. Geochemical characteristics of sediment in the Ninepipe and Ovando areas.....:. 51 10. Nutrient concentrations in sediment in the Ninepipe and Ovandp areas...............52 11. Composition of vegetation in the Ninepipe area................................................... 55 12. Composition of vegetation in the Ovando area............................................... -58 13. Comparison of vertical distribution of plant species in the Ninepipe area............61 14. Comparison of vertical distribution of plant species in the Ovando area........ . . . 62 15. Physical and chemical characteristics of water collected in June in the Ninepipe and Ovando area sites.............................................. 99 16. Physical and chemical characteristics of water collected in July in the Ninepipe and Ovando area sites.............................................100 17 . Physical and chemical characteristics of water collected in August in the Ninepipe and Ovando area sites........................................101 vii viii LIST OF TABLES—Continued Table . 18. Physical and chemical characteristics of sediment collected in theNinepipe and Ovando area sites....... ..... .... ................................ 19. Relative abundance of plant species in the Ninepipe and Ovando area sites 20. Species group designations for vegetation analysis..................................... 21. Rationale for inclusion of a) species groups, b) species, and c) species richness in vegetation analysis... ............ .......... ............. ix LIST OF FIGURES Figure Page 1. Location of the Ninepipe and Ovando study areas in western Montana...... ..... 25 2. Location of study sites in the Ninepipe Wildlife Management Area.., ........ .....21 3. Location of study sites in the Ovando Valley.......................... 30 4. Depth to groundwater relative to ground surface elevation in the a) Ninepipe and b) Ovando areas.'.,.............................. ........................ -43 5: Chemical composition of water in the Ninepipe and Ovando areas.....................47 6: Sediment particle size distributions in the Ninepipe and Ovando areas............... 54 ' .1 X ABSTRACT Human influence and biological condition in streams are commonly assessed with multimetric indices, which combine measurements at several ecological levels into an overall rating of biological integrity. Efforts to develop such biological assessment methods for wetlands are relatively recent. This study examined effects of grazing and cultivation on vegetation in 24 depressional wetlands in western Montana. The two primary objectives of this research were to (I) evaluate differences in vegetation among wetlands with different land uses and (2) identify vegetation characteristics that vary predictably with land use, and, therefore, may be useful in a multimetric index of human influence on wetlands. I examined plant community composition and vertical distribution of species in wetlands surrounded by land uses ranging from relatively undisturbed to currently grazed or cultivated. Sites were selected using physical and chemical criteria to limit natural variability among sites. Environmental data showed that wetlands differed between the two study areas selected to investigate effects of cultivation and grazing but were similar within each study area. I found limited evidence for elevated nutrient concentrations, which might alter vegetation indirectly, in areas of increased disturbance but considerable evidence of direct, physical impacts such as tillage and trampling. Patterns of differences in wetland vegetation among land uses were also consistent with direct, physical impacts of cultivation and grazing and indicated that cultivation caused greater changes in plant communities than grazing. In cultivated areas, the clearest differences were observed in groups and species related to life history, successional role or native origin: annuals, perennials, native perennials, moss, Agropyron species, Eleocharip species, and Juncus halticus. In grazed areas, dominance, Eleocharis species, and Juncus halticus differed among land uses. Vertical distributions of many species shifted to lower elevations at disturbed sites. Because land use appears to affect vascular plant communities in a measurable and ecologically rational way, vascular plants show good potential for use as metrics to assess wetland condition and human influence on wetlands. Because changes in vegetation were greater for cultivation than, grazing, plant-based metrics are more likely.to succeed in assessing impacts of cultivation. INTRODUCTION Wetlands are transitional environments between uplands and aquatic environments. In wetlands, biological, chemical, and physical components interact in distinctive ways that generate basic ecological attributes and functions that influence adjacent ecosystems. For example, biogeochemical changes affecting chemical constituents in the water column and substratum, removal of sediment, reduction in water velocity, water storage, and provision of unique habitat needed by many different / lifeforms are all influenced by the wetland ecosystem (Mitsch and Gosselink 1993). These wetland functions are increasingly recognized as valuable to human society, yet wetland habitat continues to be threatened by land use either directly or in the surrounding watershed. In Montana, two of the most extensive uses , of land affecting wetlands are livestock grazing and cultivation (Montana Agricultural Statistics Service 1990). These land uses are recognized as strongly impacting wetland vegetation in the northern prairie region (Kantrud et al. 1989, Millar 1973, Stewart and Kantrud 1972, Walker and Coupland 1970, Dix and Smeins 1967). However, Kantrud et al. (1989) state that information on the impacts of different kinds of agricultural activities on wetland vegetation is not extensive and for the most part consists of incidental observations recorded as part of vegetation surveys or waterfowl studies. Changes in vegetation are among the most conspicuous and ecologically significant consequences of human activity in and around wetlands. The flora of depressional wetlands in glaciated landscapes is influenced by the water regime, salinity. and disturbance by humans, but direct and indirect disturbances are so widespread that they are often the most important factors controlling the distribution and composition of contemporary wetland vegetation (Kantrud et al. 1989). Direct physical disruption of sites by land use can alter plant community composition by removal of vegetation, . introduction of species, or physical alteration of the soil. Indirect effects of land use on vegetation may result from changes in water and sediment chemistry in wetlands due to changes in surface runoff or groundwater quality, which may alter plant growth or competition between species. Either direct or indirect impacts to wetlands can result in broad changes in vegetation. The primary legal protections for wetlands from human impacts are contained in the Clean Water Act. Section 101(a) states that the objective of the Clean Water Act is to “maintain and restore the chemical, physical, and biological integrity of our Nation’s waters.” Thejurisdiction of this Act includes wetlands (Danielson 1998). Unfortunately, although this federal mandate as well as state and local regulations are in place to protect i ' 1 - wetlands, wetlands are still altered or lost at high rates (Karr and Chu 1997). Impacts in and around wetlands continue, due in part to a limited understanding of how land use affects wetlands and how to assess the condition of wetlands prior to, during, and after human activity. In addition, while some land uses clearly alter wetlands, the long-term effects of moderate land use in wetlands, or land use in the surrounding watersheds, are less obvious, less well understood, and more difficult to quantify. Following passage of the Clean Water Act, emphasis was placed on development of chemical criteria to maintain and restore the Nation's waters. However, it has become increasingly clear that aquatic systems continue to deteriorate as a result of human 3 actions (Karr and Chu 1997). Nonpoint source pollution, habitat alteration and fragmentation, and land use within a watershed can all alter the biological components of aquatic ecosystems, yet the resulting impacts may not be detected using chemical criteria (Danielson 1998). Because chemical criteria alone are unable to measure the impacts caused by these kinds of stressors, the U.S. EPA has begun to focus on direct assessment and protection of biological condition of aquatic systems (Danielson 1998). One common approach to direct assessment of biological condition in aquatic systems is the use of multimetric indices (Karr and Chu 1997). A multimetric index combines quantitative measurements at several ecological levels into an overall rating of biological integrity. While multimetric indices have been developed for stream ecosystems in many states (Karr and Chu 1997), efforts to develop biological assessment methods for wetlands are relatively recent (Danielson 1998). Further, multimetric indices for streams have primarily used fish and macroinvertebrates, while in wetlands, vascular plants may be useful for biological assessment (Adamus and Kentula 1996). As of 1996, no state agencies in the prairie region were monitoring biological integrity of wetlands using vascular plants (Adamus and Kentula 1996), although research is now underway in some states. Researchers have documented that vascular plant community composition in wetlands changes with land use, although detailed information that could be used for quantitative monitoring is limited. Vascular plants may be especially useful for biological assessment of wetlands because water is not always present in wetlands to support other assemblages such as aquatic macroinvertebrates or fish. Furthermore, because plant populations are persistent, sessile, I 4 and respond to the soil environment, they may integrate effects of impacts over longer periods than either invertebrates or fish. Research Objectives and Approach This study examined effects of grazing and cultivation on vegetation in depressional wetlands in western Montana. The two primary objectives were to (I) evaluate differences in vegetation among wetlands with different land use and (2) identify vegetation characteristics that vary predictably in relation to land uses and, therefore, may be useful in a multimetric index of human influence on wetlands. This work was undertaken partially in response to the state of Montana’s need to develop methods for wetland assessment in compliance with the Clean Water Act. I studied effects of land use on plant community composition and species distribution relative to surface elevation in 24 depressional wetlands in western Montana. Twelve sites located in the Mission Valley were used to investigate the effects of cultivation; these sites were all within a 3-km range. Another twelve sites located in the f Ovando Valley were used to investigate the effects of grazing; these sites were within a 30-km range. Each study area of 12 sites was evenly subdivided into three groups: a reference group representing least human impact, a group currently grazed or cultivated, and a group retired from grazing or cultivation for several years. I hypothesized that wetland vegetation is impacted directly or indirectly by cultivation and grazing and, therefore, differs among sites with different land uses. Tillage can uproot plants, change soil structure and cause erosion and sedimentation. 5 Grazing can stress plants by removing leaves and reducing photosynthesis and carbon reserves; defoliation may also allow other species to compete more successfully. Trampling by hooves can damage roots and shoots and compact soil, restricting plant growth. Trampling, can also alter surface microtopography, and resultant microsite changes in hydrology may cause changes in species composition. Cultivation and grazing can also affect wetland vegetation indirectly by altering water and sediment chemistry in wetlands. I hypothesized that the most likely indirect effects of these land uses would be changes in nutrient concentrations in the water column and sediment. Possible avenues for changes in nutrient concentrations in cultivated areas included increased runoff due to soil and vegetation disturbance, transport of fertilizer nutrients from surrounding cropland by runoff, and direct application of fertilizer in wetlands during drawdown periods. In grazed areas, changes in nutrient concentrations may occur due to animal defecation or urination in wetland margins, where livestock spend disproportionate amounts of time (Nader et al. 1998). The study design focused on limiting natural variation among sites in order to detect differences in land uses. All sites were depressional wetlands as defined within a hydrogeomorphic classification (Brinson 1993). Depressional wetlands occur in low areas in the landscape and receive water from surface precipitation and runoff, groundwater, or both. Based on Omernik’s (1987) ecoregion map, the two study areas were in the same ecoregion. However, the two areas were in different ecological units, according to the ecological units map developed by Nesser et al. (1997). The selected sites met criteria for hydrologic regime, salinity, and landscape setting. The study sites were as similar as possible except for the differing land uses around individual sites. Environmental data were collected to verify that sites were similar and to evaluate whether land use effects on the vegetation were most likely direct, via physical disturbance to the wetland, or indirect, via water or sediment chemistry. I compared plant community composition among land uses. Because relatively ' little research has been done on the use of vegetation to assess biological integrity of wetlands (Adamus and Kentula 1996), I examined several groups of plant species, as well as genera and individual species. These groupings primarily were based on ecological factors such as life history, lifeform, and whether species were native or exotic. Other groups were based on results of relevant research and examination of the data. I also investigated whether disturbance related to land use altered the spatial distribution of plant species with respect to elevation, and whether species distribution with respect to elevation or water depth could be incorporated as a metric. Physical •disturbance or changes in resources might affect mortality, reproduction, or competition, resulting in changes in abundance. Ifthese effects varied along the hydrologic gradient, species distributions could shift to higher or lower elevation ranges. I compared whether the distribution of each species with respect to water surface elevation differed among reference, retired, and currently impacted wetlands, and tested whether change in elevational distribution was consistent among species. Finally, data were used to identify potential variables for use in a multimetric index for depressional wetlands in western Montana. In addition to examining land use impacts generally, sampling two different land uses allowed me to examine patterns in vegetation composition and distribution specific to grazing or cultivation. Understanding 7 distinctive effects of these land uses may lead to useful diagnostic tools for assessing wetlands with unknown land use histories to determine past impacts. 8 LITERATURE REVIEW Hydrologic regime and salinity are two of the most important natural controls of plant distribution in depressional prairie wetlands (Kantrud et al. 1989, Millar 1973, Stewart and Kantrud 1972, Walker and Coupland 1970, Walker and Coupland 1968). These factors in turn are influenced by geomorphic characteristics and subsurface geology. Human disturbance also plays an important role in plant distribution (Kantrud et al. 1989). Below, I first discuss geomorphic characteristics, hydrologic regime, and salinity and how they interact in depressional wetlands formed by glaciation. Next, I describe the state of knowledge regarding the response of vegetation to these environmental factors. Third, I discuss vegetation response to land use impacts. Wetlands examined in this study are hydrogeomorphically similar to the prairie potholes of the Dakotas, where much research has been completed on depressional wetlands in response to conflicts, between waterfowl habitat preservation and agriculture. Therefore, much of the literature cited comes from this region. Finally, I outline the scientific basis for multimetric indices and how vascular plants may be useful in a multimetric index. Geomorphologv. Hydrology, and Water Chemistry Hydrology is the driving force behind the existence of wetlands. “A wetland is an ' ■’ . ecosystem that depends on constant or recurrent, shallow inundation or saturation at or 9 near the surface of the substrate. The minimum essential characteristics of a wetland are recurrent, sustained inundation or saturation at or near the surface and the presence of physical, chemical, and biological features reflective of recurrent, sustained inundation or saturation.” (National Research Council 1995). Wetland hydrology is the most important determinant of a wetland’s structure and function (Mitsch and Gosselink 1993). Geomorphic characteristics and subsurface geology influence hydrologic regime, water source, and water and sediment chemistry in wetlands, all of which influence the biological attributes of wetlands. Depressional wetlands are common in the glaciated northern plains of North America; they are referred to as prairie potholes in the United States and as sloughs in Canada (Kantrud et al. 1989). These wetlands are formed in glacial deposits: till, glaciolacustrine sediments, and outwash. All these types of glacial deposition result in generally similar surface topography: a relatively young landscape of gently rolling to hummocky topography with very little drainage development, pockmarked with small, shallow, and often temporary bodies of water in slight depressions in the landscape. The shallow depressional topography and physical characteristics of the glacial material influence water source, hydrologic regime, and water and sediment chemistry. The physical and chemical composition of the parent material of the glacial deposits also play a role in characteristics of these wetlands (Fetter 1994, Van Voast and Novitzki 1968). Glacial till is deposited directly by glacial ice without significant sorting by running water. Generally, till has a low hydraulic conductivity, especially if it is clay- rich (Fetter 1994). Permeability of till depends upon the clay content, mode of deposition and degree of weathering (Fetter 1994). Hydraulic conductivity of the unfractured till 10 matrix is IO'9 cm/s for in the prairies of Canada and North Dakota (Winter and Rosenberry 1995). Other researchers report hydraulic conductivity of till ranging from IO"5 to IO'9 cm/s (Fetter 1994). Fracturing affects hydraulic conductivity. The unffactured till matrix generally has low hydraulic conductivity, but the presence of fractures due to desiccation, particularly near the surface, gives a secondary permeability, which may increase the hydraulic conductivity of the overall matrix by orders of magnitude (Woo and Rowsell 1993). Fetter (1994) states that glaciolacustrine sediments have hydraulic conductivity values similar to those of tills, ranging from IO"6 to IO"8 cm/s. Glacial outwash sediment / is deposited by meltwater from the glacier, and therefore is well-sorted, giving it a higher hydraulic conductivity. Fetter (1994) gives a general range for outwash of IO"3 to I cm/s, but states that some silty outwash materials have hydraulic conductivities as low as IO"5 cm/s. Low hydraulic conductivity slows the downward movement of water, causing ponding in wetlands for extended periods. According to Stewart and Kantrud (1971), permanence of surface water in prairie potholes is modified by permeability of bottom soils. Shjeflo (1968) reports that average outflow seepage ranged from 3x10'7 cm/s to 3x10"6 cm/s in 10 semi-permanent and seasonal potholes. Sloan (1970) observed changes in pothole water levels indicating that outflow seepage rates are much higher from ephemeral and temporary potholes. Rates of outflow influence hydrologic regime, as discussed below. Due to low hydraulic conductivity, which prevents rapid equilibration of groundwater levels, transpiration by phreatophytic vegetation can cause pronounced I. 11 diurnal depressions in groundwater directly surrounding a pothole. Water levels in the zone of phreatophytic vegetation adjacent to potholes rise and fall in response to variations in climate and groundwater flow in much the same way as the pothole water surfaces do, but with much greater amplitude, resulting in groundwater levels that are often significantly above or below the pothole water surface (Sloan 1970). Low hydraulic conductivity causes water to follow the surface topography of glacial deposits closely. Sloan (1970) notes that closely spaced potholes are often separated by a large vertical distance and a steep water-table gradient, which result from the low permeability conditions characteristic of the glacial till surrounding the potholes. The relation of the water table to the wetland position in the landscape influences groundwater flow direction. This relationship causes wetlands to vary in water source. Water source is one of the most important factors controlling wetland hydrologic regimes. Hydrologic regime refers to the depth and duration of inundation during the growing season (Kantrud et al. 1989). Stewart and Kantrud (1971) state that water permanence is influenced by permeability of bottom soils, as explained above, and by groundwater. Groundwater can be an important contributor to a wetland’s surface water. - Wetlands typically fall into three categories according to water source: (i) “recharge wetlands” recharge groundwater with water derived primarily from overland flow and snow-catch, (ii) “throughflbw wetlands” receive both water from and yield water to the groundwater system, and (iii) “discharge wetlands” obtain some water from overland flow but receive the majority as groundwater discharge (Arndt and Richardson 1989). Stewart and Kantrud (1971) developed a wetland classification system for natural ponds and lakes (prairie potholes) in the glaciated prairie region. The hydrologic regime of the vegetation zone in the central or deeper part of a wetland is used to distinguish several classes. Major classes of hydrologic regime are ephemeral ponds, temporary ponds, seasonal ponds and lakes, semi-permanent ponds and lakes, permanent ponds and lakes, alkali ponds and lakes, and fens. In the prairie pothole region, recharge wetlands tend to have ephemeral or seasonal hydrologic regimes because the main water source is from precipitation, largely as spring runoff, and water is lost by percolation as well as evapotranspiration. Duration of ponding can be modified by the hydraulic conductivity of bottom soils. Flow-through wetlands tend to be semi-permanent because their water source is shallow groundwater discharge. Discharge wetlands tend to be semi-permanent or permanent because their water source, groundwater, is more constant than seasonal runoff. Water chemistry is related to hydrologic regime, water source, and underlying substrate, which is derived from the geologic source material of glacial deposits. Water chemistry is largely influenced by the configuration of the water table surrounding the wetland (Sloan 1970) and subsurface geology (Van Voast and Novitzki 1968). Salinity is the concentration of total dissolved solids in the soil or water column (Kantrud et al. 1989). Total salinity of water in a pothole is largely controlled by the rate of groundwater flow and the relationship of seepage inflow to outflow: generally, potholes with water relatively low in dissolved solids have net seepage outflow, and potholes with high concentration of dissolved solids have net seepage inflow (Eisenlohr and Sloan 1968). Total dissolved solids are low in rainfall, but high in discharging groundwater. Few dissolved solids are removed from potholes, other than by seepage outflow or infrequent pothole overflow. Evapotranspiration further concentrates 13. dissolved solids (Eisenlohr 1972). Electrical conductivity is often measured to estimate salinity (Cowardin et al. 1979, Stewart and Kantrud 1972). The chemical composition of freshwater in recharge wetlands can vary due to subsurface geology. Richardson et al. (1994) state that several chemical processes other than mineralogical controls explain subtle variations in the dilute solutions characteristic of recharge wetlands. These controls include weathering and lithological differences, differences in ionic mobility, exchange relationships between the aquifer media and solute, and biological cycling. Natural Controls on Plant Distribution Water depth and duration largely determine distribution of plant species (Kantrud et al. 1989). In the prairie potholes, the gradient of depth and duration of water in each pothole creates visually conspicuous zonation, with each zone dominated by a single plant species (Kantrud et al. 1989). The Stewart and Kantrud (1972) classification uses the central vegetation zone of each wetland to identify hydrologic regime. Zones corresponding to less permanent wetlands may also occur as peripheral bands around more permanent wetlands. These zones and corresponding hydrologic regime are as follows: a wetland-low-prairie zone associated with ephemeral ponds; a wet-meadow zone associated with temporary ponds; a shallow marsh associated with seasonal ponds and lakes; a deep marsh associated with semipermanent ponds and lakes; and permanent open water associated with permanent ponds and lakes (Stewart and Kantrud 1972). As long-term hydrologic cycles occur, peripheral vegetation zones may shift in elevation or the central vegetation zone may change altogether as the hydrologic regime of a wetland is altered. Consequently, classification of a wetland may change. Plant species have differing tolerances for salinity. Researchers in the prairie pothole region have found that salinity has a profound effect on the plant community composition and the distribution of species (Kantrud et al. 1989). Stewart and Kantrud (1971) used salinity to define subclasses of vegetation types in prairie potholes. Within any hydrologic regime, the dominant and subdominant species may vary with salinity. As salinity increases, the number of species diminishes considerably (Kantrud et al. 1989). Walker and Coupland (1968) report most of the species in sloughs to be strongly affected by the hydrologic regime and somewhat less affected by salinity. Long-term climatic changes can alter the wetland hydrologic regime and vegetation zonation. It may take a year or more for species composition within a zone to adjust to environmental change. These lags sometimes result in abnormal zonation patterns, particularly after a change in water level (Kantrud et al. 1989). Kantrud et al. (1989) caution that because hydrologic regime is a principal factor in controlling distribution and composition of vegetation, it follows that hydrologic fluctuation will induce changes in the composition and structure of wetland vegetation. Consequently, any attempt to study the composition, classification, or dynamics of the region’s wetland vegetation must recognize the impact of long-term climatic fluctuation. 15 Human Influence on Vegetation ( Direct and indirect disturbances by human activity are so widespread that they are often the most important factors controlling the distribution and composition of contemporary wetland vegetation (Kantrud et al. 1989). Cultivation, grazing, mowing, and burning have altered the composition of pothole vegetation (Kantrud et al. 1989, Millar 1973, Stewart and Kantrud 1972, Walker and Coupland 1968). Kantrud et al. (1989) and Stewart and Kantrud (1972) documented differences in species found in wetlands of the same hydrologic and salinity regime but surrounded by contrasting land uses of grazing, haying, and cultivation or. idle land. Walker and Coupland (1968) rated the intensity of disturbance by mowing, grazing, and cultivation and were able to distinguish dominant species associated with varying intensities of these land uses. They concluded that cultivation is by far the most drastic type of disturbance, completely overriding the effects of natural gradients. The Stewart and Kantrud (1971) classification recognizes a completely separate vegetation phase for tilled wetlands. Land use impacts vary with hydrology and climate. Relatively permanent wetlands are exploited for native hay and as pastures, and less permanent wetlands are used extensively for crop production during dry years (Kantrud et al. 1989). Wetland- low-prairie, wet-meadow, and shallow-marsh zones are frequently cultivated when they become dry, which creates early successional plant communities that differ markedly from those of undisturbed sites (Millar 1973). Following cultivation, such communities often retain early serai characteristics for a few years. Subsequently, the number of 16 species decreases progressively as small, rapidly growing annuals are excluded by competition; succession plays a significant role during extended periods without disturbance (Walker and Coupland 1970). Although farmers often cultivate as far into depressions as possible, or even across them when they are dry, these sites are usually too wet to seed crops (Dix and Smeins 1967). Whether seeded to a crop or not, species may be introduced by tillage. Cultivated depressions typically become mudflats after disturbance, providing excellent seedbeds for weedy species, both native and introduced (Dix and Smeins 1967). Mudflat communities often develop on such sites after disturbance. During dry years, however, some shallow wetlands are planted to crops. Small grain or row crops are often present in the cropland tillage phase of prairie potholes (Stewart and Kantrud 1972). Agriculture may result in indirect impacts to vegetation. Agricultural activities have affected many wetlands through runoff that contains large quantities of sediment, nutrients, and pesticides (Kantrud et al.1989). Writing in 1989, Kantrud et al. reported that indirect effects of runoff water quality on vegetation had not been studied in prairie- potholes. However, impacts of nutrients on wetland vegetation have been investigated in other parts of the United States. For example, increased agricultural runoff, laden with phosphorus, is believed to have promoted the spread of Typha spp. in conservation areas the Everglades in Florida (Gunderson and Loftis 1993, Koch and Reddy 1992). There is conflicting information on the relative importance of water and sediment nutrient levels to wetland vegetation change. Numerous experimental studies have shown that submerged plants take up nutrients both from the water through the leaves and from the substrate via foots (Bristow 1975, Carignan 1982). In contrast, Anderson and KalfF (1988) state that it is now generally agreed that rooted aquatic plants derive almost all their nutrients from sediments. Because surface water and sediment chemistry are related through sedimentation and solute exchange, this distinction may be more important to water quality monitoring than vegetation change.. Grazing has effects similar to cultivation. Dominance types in wetlands that are grazed are often very different from nongrazed wetlands (Kantrud et al. 1989). Marshes often show greater vegetation response to grazing than upland communities (Bassett 1980). Species composition changes in grazed areas because soil is physically disturbed, plants are stressed by defoliation, and light competition is reduced (Archer and Smeins 1991). Changes in vegetation composition are likely related to the exposure of soil that can result from grazing (Millar 1973). Effects of defoliation on species dominance may be related to differences in the palatability of emergent species, which results in preferential grazing, and to opening of the otherwise dense canopy, which allows establishment or increase of species that are not abundant under closed canopies (Kantrud 1986). Research in riparian wetlands shows a wide range of plant community responses to grazing (Kauffman and Krueger 1984). Hansen et al. (1995) interpret several Montana riparian community types as resulting from excessive grazing. ' Effects of grazing may not all be negative; prairie wetlands probably were grazed frequently by native herbivores before European settlement, and had grazing-tolerant species. Moderate grazing may promote greater species diversity, complex spatial patterns, and sharp boundaries between vegetation zones in prairie wetlands (Bakker and Ruyter 1981). However, overgrazing can reduce primary production, increase water turbidity and even eliminate vegetation (Kantrud et al. 1989). 18 Grazing may affect wetland water and sediment chemistry, though probably less severely than cultivation. Runoff can be increased by trampling, soil compaction, and reduced vegetative cover, which may accelerate sediment and nutrient delivery to wetlands (Spaeth et al. 1996). Churning of water and sediment by trampling and defecation and urination may also affect water quality (Nader et al. 1998). Assessment of Biological Integrity . The ITS. Environmental Protection Agency (EPA) has adopted the policy that measures of biological condition are necessary to maintain biological integrity as mandated by the Clean Water Act (Danielson 1998). Biological integrity is “the capability of supporting and maintaining a balanced, integrated, adaptive community of organisms having a species composition, diversity, and functional organization comparable to that of natural habitat of the region” (Karr 1987). Continuing degradation of aquatic systems suggests that chemical measures of water quality are not sufficient to reflect all impacts that can occur to aquatic systems (Karr and Chu 1997). Simple biological measurements such as toxicity bioassays and diversity indices do not provide integrated assessments of biological condition and, therefore, have not adequately protected aquatic resources (Karr and Chu 1997). “Multimetric” indices attempt to provide more integrative ratings of biological integrity by including measurements at several ecological levels. A “metric” is a measurable biological attribute that reflects specific and predictable responses of organisms to changes in landscape condition, is sensitive to a range of factors that stress biological systems, and is relatively easy to measure and interpret (Karr and Chu 1997). A multimetric index combines metrics of different attributes of an assemblage (e.g. aquatic invertebrates, fish, or vascular plants) into an overall score of integrity, or condition, of an aquatic ecosystem. Development of a multimetric index for wetlands starts with classification of wetland types and regional environments (Danielson 1998). Systems for wetland classification include the Army Corps of Engineers’ hydrogeomorphic classification (Brinson 1993) and the Fish and Wildlife Service’s Classification of Wetlands and Deepwater Habitats of the United States (Cowardin et al. 1979). Stewart and Kantrud’ s classification of natural ponds and lakes in the glaciated prairie region (1971) is a regional classification of depressional wetlands. Classifications of regional environments include the EPA’s Ecoregions (Omernik 1987), and the US. Forest Service’s Ecological Units (Nesser et al. 1997). Within each wetland class, biological attributes of one or more assemblages are measured over the range of human influence, from “reference,” or relatively unimpacted, to severely impacted sites. The biological measurements that show a clear response to this gradient of impact are identified as potential metrics. Ideally, several metrics reflecting different ecological levels of the assemblage will be identified. Each metric is scored individually, and all metric scores are combined into a final index score. The index is then tested and validated on new wetlands. Adamus and Kentula (1996) recommended analysis of a broad array of community level and individual species measurements during initial phases of bioindicator development; subsequently, the most useful measurements should be refined and combined. Barbour et al. (1995) also advocated combining several biological measurements to arrive at an overall score for environmental health. A multimetric index should be composed of measurements from the individual, population, community, and ecosystem levels into a multimetric index (Karr and Chu, 1997). Danielson (1998) suggests that within a multimetric index, some metrics should be more sensitive to chemical alteration, others to physical alteration, and still others to biological alterations in order to discern many types of human influence. Multimetric indices have been developed primarily for stream systems and have been based on fish or macroinvertebrates. Forty-two states now use multimetric assessments of biological condition of streams (Karr and Chu 1997), but efforts to develop multimetric indices for wetlands are relatively recent. Several states, including Florida, Maine, Minnesota, Montana, North Dakota, and Ohio, are currently developing biological assessment methods for monitoring wetlands (Danielson 1998). In Montana, development of wetland multimetric indices utilizing macroinvertebrates and diatoms has been initiated (R. Apfelbeck, 1997, personal communication). Indices based on vegetation may be useful to assess wetland biological condition, either alone or in conjunction with another assemblage. Plants may be easier for non­ specialist field personnel to quantify, may integrate water quality over longer time scales, and may respond somewhat differently to water quality and disturbance, thus providing additional information about water quality and biological integrity. Chemicals from high-intensity, occasional events or low-level but chronic impacts may be transient in the water column, but remain and accumulate in wetland substrates. Because plants take up chemicals primarily from sediment, they may show the effects of these types of impacts 21 more effectively than aquatic organisms. Perhaps most importantly, vegetation can be sampled when no water is present, which is not possible with most other biological assemblages used in wetland bioassessment. Adamus and Kentula (1996) suggested that vascular plants may be the best biological indicators of ongoing or recent changes in water regime or salinity, good indicators of excessive nutrients and anoxia, and fair to poor for monitoring herbicides, insecticides, and heavy metals. Small et al. (1996) sampled many riparian wetlands in the Chesapeake Bay watershed and found that differences in stream water quality could be identified by analyzing vascular plant communities.. Small et al. (1996) noted that vegetation is ideal for an initial biosurvey because plants are easier to identify and quantify than other organisms used in biomonitoring of stream impairment. The Minnesota Pollution Control Agency is developing a multimetric index for biological assessment for depressional wetlands using emergent and submerged vascular plants, mosses, and liverworts. Initial development has shown effective identification of impaired and unimpaired sites (M. Gemes, 1997, personal communication). Because the use of vascular plants for assessment of biological condition of wetlands is not yet well established, particularly for glaciated depressional wetlands (Adamus and Kentula 1996), there are no clear conventions in the selection and use of vegetation groups for quantitative monitoring. 22 METHODS Study Areas Study Design I studied effects of land use on plant community composition and species distribution relative to elevation in 24 depressional wetlands in western Montana. The study design included two major land uses, three levels of impact within each land use, and four replicate sites for each level of impact. Twelve sites located in the Mission. Valley, at the Ninepipe Wildlife Management Area, were sampled to investigate the effects of cultivation and 12 sites located in the Ovando Valley were sampled to investigate the effects of grazing. The 12 wetlands in each of these valleys will each be referred to as the “ Ninepipe area” or “Ovando area.” Each area contained three land use groups with four replicates: a reference group representing least human impact, a group currently grazed or cultivated, and a group retired from grazing or cultivation for several years. These three land use groups will be referred to as “reference,” “retired,” and “current” or “currently impacted” sites. Because land use impacts were not interspersed or randomly assigned among wetlands, there is potential to confound land use effects and inherent site differences; thus caution is required when extrapolating results. Site Selection The study was limited to depressional wetlands, as defined by a hydrogeomorphic classification system (Brinson 1993). Study site selection took place in May 1997, when over 50 candidate sites were surveyed over a 240 km range, in western Montana to determine the general study location. Two study areas were selected based upon public access and the range of land use impacts present. After screening to select general study areas, an additional 50 sites were sampled within those areas during the final site selection process in June 1997. Study sites were selected using physical and chemical criteria to limit natural variability among sites. Two primary selection criteria related to hydrology and salinity. All sites had semipermanent or extended seasonal inundation, based on information from land managers, past research data, presence on topographic maps, and presence of vegetation known to be tolerant of these hydrologic regimes. Surface water at all sites had conductivity under 800 pmhos/cm during site selection in June. Secondary site selection criteria included absence of surface water connections with other wetlands, moderate slopes around wetlands, similar depths among reference and impacted sites, size between 0.1 and 1.2 hectares, similar topographic settings, and / isolation from other human influences. Overall, the emphasis in applying secondary criteria was to keep hydrogeomorphic characteristics similar among sites. Description of Study Sites The study of cultivation impacts took place in the Mission Valley of Montana, r approximately 30 km south of Flathead Lake (Figure I). Environmental characteristics and land use are summarized in Table I . A large portion of the Mission Valley is cultivated. The study area was on the Ninepipe Wildlife Management Area, which is managed for waterfowl habitat. Parcels of land in the Ninepipe area that were managed at differing intensities of land use were used in the study design. Sites were all within 3 km of one another (Figure 2). Some of the Nine Pipe Wildlife Management Area is cultivated every other year with a cereal grain crop, a portion of which is left unharvested for waterfowl food and cover. Wetlands sampled in these areas will be referred to as the “current” group to I reflect continuing cultivation. Cultivation historically occurred as close to water’s edge as soil conditions would allow farm machinery to operate effectively. Impacts to vegetation associated with cultivation include physical disturbance and the introduction of species during drawdown periods. The local topography, featuring low slopes of 2- 8%, allows some runoff to occur. Because crop fertilization was infrequent and application of herbicides was only occasional (J. Grant, 1997, personal communication), chemicals carried by runoff probably did not affect these wetlands greatly. A nearby area of land on the Nine Pipe Wildlife Management Area was once cultivated in a similar manner, but was retired from cultivation and seeded with tame grasses and legumes 10 to 12 years ago (I. Grant, 1997, personal communication). These sites will be referred to as the “retired” group. If past and current agricultural practices 25 Nincpipc study area FOHT e tMTON THOMFfON FAU-I Ovando study area MISSOULA WWlT l SU tPH U H T H H tC FOHKS oOZEMi V lH C lN ia CITY Figure I. Location of the Nincpipc and Ox ando study areas in western Montana. Table I. Environmental characteristics and land.use information for the Ninepipe study area. All sites were within a 3-km range. Characteristic Reference sites Retired sites Current sites Number of sites 4 4 4 Land use idle retired from cultivation currently cultivated Legal description1 NW 1/4 S36, T20N, R20W NW 1/4 S26, T20N, R20W NE 1/4 S27, T20N, R20W Aspect1 SW SW SW Elevation1 [m (ft)] 930 - 933 (3,050 - 3,060) 930 - 933 (3,050 - 3,060) 930 - 933 (3,050 - 3,060) Mean annual precip.2 [cm (in)] 41-46(16-18) 41-46 (16 - 18) 41-46(16-18) Mean annual temp.2 [0C (0F)] 8 (46) 8(46) 8(46) Soil Iandform3 lake plains / moraines lake plains / moraines lake plains / moraines Soil parent material3 glaciolacustrine deposits/till glaciolacustrine deposits glaciolacustrine deposits Land use history grazed up to 50 years ago; never cultivated4 retired from cultivation 10-12 years ago and pasture seeded w/ tame grasses and legumes4 cultivated field; evidence of recent tillage through wetland edges Current influences undisturbed pasture, low-density roads and houses nearby infrequent herbicide application4 cultivation, infrequent fertilizer/ herbicide application 4 1 U.S.G.S. Charlo and Fort Connah quadrangles, 7.5 minute series. 2 MAPS Atlas-A land and climate information system (Caprio et al. 1994). 3 Soil Survey of Lake County Area, Montana (NRCS 1997). 4 John Grant, Montana Department of Fish, Wildlife, and Parks, Wildlife Area Manager, 1997, personal communication. ----1;-; 7% vr-'e IL v Figure 2. Location of study sites in the Ninepipe Wildlife Management Area. Adapted from U.S.G.S. Charlo and Fort Connah quadrangles, 7.5 minute series. are assumed to be similar, the retired sites can be treated as intermediate with respect to disturbance and recovery between currently cultivated and reference groups. However, because retired sites and adjacent fields were seeded, they do not represent completely passive successional development. Some land on the Nine Pipe Wildlife Management Area has never been cultivated and contains the least impacted sites available for this study. These sites will be referred to as the “reference” group, to which the retired and cultivated sites are compared. This land was grazed from settlement until approximately 50 years ago. These uncultivated pastures are small (a quarter section or less) and adjacent to a highway, and therefore are likely to have some low-level human impacts. Rock piles created historically by farmers aided site selection. Farmers removed rocks from cultivated areas and piled them into wetlands (J. Grant, 1997, personal communication). These 0.3 to 0.6 meter high rock piles are now under water. Presence of rock piles in wetlands indicates that they were wet regularly enough that farmers thought they were not losing arable ground and that water levels were sometimes low enough to permit farmers to drive equipment well into the wetlands. Presence of rocks implies fluctuations between periods of extended seasonal or semi-permanent ponding and periods dry enough to allow cultivation well inside current wetland boundaries. Rock piles were present in all current and retired wetlands sampled. Grazing impacts were studied in the Ovando area, east-northeast of Missoula, Montana (Figure I). Environmental characteristics and land uses are summarized in Table 2. Because sites in the Ovando area were further away from one another than those at Ninepipe, background variation among sites was likely to be greater (Figure 3). In Table 2. Environmental characteristics and land use information for the Ovando study area. All sites were within a 30-km range. Characteristic Reference Retired Current Number of sites 4 4 4 Land use idle, game range retired from grazing currently grazed Legal description1 ' E 1/2 S21, N 3/4 S27, T15N, R14W E 1/2 S21, T14N, RllW Wl/2 S15, T15N, R13W Aspect1 S. S S Elevation1 [m (ft)] 1,268 - 1,280 (4,160-4,200) 1,305 - 1,329 (4,280 - 4,360) 1,250- 1,262.(4,100-4,140) Mean annual precip.2 [cm (in)] 36-41(14-16) ■ 46 -51 (18 -20) 41 - 46 (16 -18) Mean annual temp.2 [0C (0F)] 6(42) 4 (39) - 5 (41) Soil Iandform3 moraines alluvial fans / stream terraces mountains / moraines Soil parent material3 alpine till alluvium / colluvium / outwash till Land use history grazed up to 50 years ago4 pasture retired 15 years ago from heavy grazing without rest5 tussocks, pockmarked soil visible pasture grazed through 1996 using moderate stocking rates and seasonal pasture rotation6 Current influences elk refuge wildlife use livestock grazing, wildlife use 1 U.S.G.S. Woodworth, Ovando, and Marcum Mountain quadrangles, 7.5 minute series. 2 MAPS Atlas-A land and climate information system (Caprio et al. 1994). 3 Soil Surveys of Missoula and Powell County Areas, Montana (NRCS 1995, NRCS undated). . 4 Mike Thompson, Montana Department of Fish, Wildlife, and Parks, Wildlife Biologist, 1997, personal communication. 5 Greg Neudecker, U S. Fish and Wildlife Service, Wildlife Biologist, 1997, personal communication. 6 Joe Brewster, Bandy Ranch manager, 1997, personal communication. Figure 3. Location of study sites in the Ovando Valley Reference sites are located in the Blackfoot Clearwater Game Range, retired sites are located in Uie Blackfoot Waterfowl Production Area, and currenUy grazed sites are found on the Bandy Ranch. Adapted from Lolo National Forest map. addition to eating wetland plants, cattle often congregate around water, trample vegetation and soil, and impact water quality (Nader et al. 1998, Kauffman and Krueger 1984). Thus effects on vegetation may include both physical impacts and nutrient inputs. Currently grazed sites were on a half-section pasture of the Bandy Ranch, a research area managed by University of Montana. Pastures were grazed annually using a seasonal rest-rotation system through the summer of 1996, the year prior to data collection (I. Brewster, 1997, personal communication). These are referred to as the “current” sites. Retired sites were on a USFWS Waterfowl Production Area, which had not been grazed by cattle for 15 years. Although specific records were not available, USFWS land managers believed use was without rest and heavier than the land could sustain (G. Neudecker, 1997, personal communication). Similarity of historical grazing impacts on retired and currently grazed sites could not be evaluated. If impacts were similar, rested sites can be treated as intermediate between currently grazed and reference sites. If not, however, differences due to succession will be confounded with differences in original level of impact. Reference sites were located on the Blackfoot-Clearwater Game Range, owned by Montana Department of Fish, Wildlife, and Parks. This refuge had not been grazed by livestock for 50 years (M. Thompson, 1997, personal communication). Comparison of grazing impacts between reference and retired and currently grazed sites was complicated by grazing by wild animals, particularly elk, on the Game Range. Sampling and Analysis Hydrologic Measurements A stafFgage was installed at the center of each wetland. Distances from the top of the gage to the sediment, and from the top of the gage to water surface, were recorded in June, July, August, and September. As a supplement to staff gage measurements, stakes were placed at water’s edge in June, July, and August to track water level changes. Shallow open wells (100 cm) were installed in August in six to eight of the unflooded quadrats in each wetland. Wells were placed from within 0.5 meter of water’s edge to about 3 meters horizontally from water’s edge. Water levels were measured in September. Depths to the water table were graphed against ground surface elevation to evaluate use of surface elevation as an index of each quadrat’s position on a hydrologic gradient. Nested shallow (40 cm) and deep (HO cm) piezometers were installed at each wetland in June. July, August, and September water levels were compared to estimate relative pressure heads between the deep and shallow piezometers and to infer vertical direction of flow. Water Chemistry and Physical Measurements Water was collected in late June, early July, and late August. Water samples were collected at the center of each wetland using a bottle attached to a pole to assure that samples were not affected by disturbance caused by walking. Samples were collected 20- 30 cm below the water surface. Samples were unfiltered in accordance with past Montana Department of Environmental Quality (DEQ) research protocol (R. Apfelbeck, 1997, personal communication). Samples for analysis of major ions were preserved with nitric acid in the field. Total concentrations of calcium, magnesium, potassium, sodium, iron, manganese, sulfur, and phosphorus were analyzed by inductively coupled plasma (ICP) emission spectrometry. Water collected for analysis of sulfate and phosphate levels was not preserved chemically but was frozen upon return from the field every night. Sulfate and phosphate were analyzed by ion chromatography. Nitrate and nitrite were preserved with sulfuric acid in the field and analyzed colormetrically. Preservation methods followed Standard Methods for the Examination of Water and Wastewater (American Public. Health Association 1989). Water samples were analyzed by the Soil Analytical Laboratory at Montana State University, Bozeman. Other water quality data were collected in the field using a Horiba U-IO Water Quality Checker. These field data included temperature, pH, and electrical conductivity. Salinity is commonly represented by electrical conductivity measurements (Cowardin et al. 1979). During the first sampling period, bromcresol green-methyl red alkalinity was determined nightly using a 10 ml aliquot and titrating with 0.01 N HCl to a 4.6 pH endpoint. In the second and third sampling period I used a 20 ml aliquot and titrated with 0.02 N HCl to a 4.6 pH endpoint. Physical and chemical characteristics of water samples are reported in Appendix A, Tables 15, 16, and 17. Sediment Chemistry and Physical Measurements Sediment samples, for chemical analysis were collected once in late June. Each sample was collected by wading into the wetland past the emergent vegetation zone and as close to the middle as water depth would allow efficient sampling. Sediment was taken from the upper 20 cm using a shovel, stored in plastic bags and frozen until delivered to the Soil Analytical Laboratory at Montana State University, Bozeman. Samples were analyzed for concentrations of ammonium, nitrate, total Kjeldahl nitrogen (TKN), and Olsen phosphorus. Calcium, sodium, magnesium, and potassium concentrations were determined to estimate Sodium Adsorption Ratio (SAR). Sediment samples for physical analysis and additional chemical analyses were taken in early July. Samples were taken along a transect that extended from the deepest part of the wetland to the outer edge of the wetland. The transect was divided into three equal sections and three evenly spaced samples were taken in each of the three sections. Cores were taken from the upper 15 cm of the substrate. Samples from each section were composited and analyzed for particle size distribution, pH, and electrical conductivity: The hydrometer method was used for particle size analysis. Electrical conductivity and pH were measured in saturated paste extracts. Organic matter was estimated by loss on ignition. During data analysis I concluded that it would be more effective to analyze data. for the entire transect rather than three arbitrary sections, so values for the three sections were averaged. Physical and chemical characteristics of sediment are reported in Appendix A, Table 18. Vegetation Sampling In July a map of each wetland was drawn approximately to scale. In August, a random systematic sampling design was used to quantify vegetation (Thompson 1992). A grid was overlaid on the map of each wetland, and a subgrid was overlaid on one section of the grid. A sampling location was randomly selected within this smaller grid and the equivalent subgrid sample was systematically repeated throughout the rest of the major grid. Because the transect lines forming the grid were spaced approximately the same for all wetlands sampled, sampling effort was proportional to each wetland’s size. The vegetation was patchy, and this method efficiently provided a representative and unbiased sample. Sampling by zones was rejected because zonation was not found to be truly concentric or consistent within or among wetlands. A 0.5 x 1.0-meter quadrat frame was placed at each sampling location. Species were identified according to Dorn (1984) and Hitchcock and Cronquist (1973). I estimated percent cover visually for each species as well as for litter, rock, bare earth, and water. Cover estimates were maintained as percentages throughout data analysis. Voucher specimens were deposited at the Montana State University, Bozeman, herbarium. Vegetation data were organized for analysis by first dividing the data into two strata, canopy species and ground cover. Submergent species were included with ground cover. Species that were submerged due to water levels in 1997 but that ordinarily would be found in moist but not flooded soil, such as Mentha arvense or Cirsium arvense, were treated as canopy species. This study primarily examined wetland vegetation at the community level by comparing relative abundance of species and species groups among land uses. Raw relative abundance data for each species are reported in Appendix B, Table 19. Groups of vascular plants were defined prior to the study, including life history groups (annual/perennial), lifeform groups (grasses, forbs, grasslike plants, submerged species), native and introduced species, weedy species, and dominant species. Some groups were defined after initial data analysis was completed: native perennials, native grasslike plants, and weedy annual and perennial species. These generally were refinements of groups defined before analyzing data. Individual species were selected for statistical analysis based on graphical data exploration. Several species groups and species were selected for investigation based on metrics under development in other states, including species with persistent litter, non-vascular ground cover (in this study, only moss), the Carex genus, and Utricularia vulgaris. Relative abundance of dominant species was calculated in three ways: relative abundance of the single most dominant species in each wetland, the combined relative abundance of the first and second most dominant species, and the combined relative abundance of the first, second, and third most dominant species. Group designations for each species are summarized in Appendix B, Table 20. Time of and reasons for inclusion of each species and species group and expected direction of response to disturbance are summarized in Appendix B, Table 21. Vegetation data were expressed in terms of relative abundance, frequency, and species richness for each wetland. These variables are all used commonly in vegetation analysis (BLM 1996). I evaluated these alternative measures because there is limited information on using vegetation for multimetric indices to guide analyses. I expected that relative abundance would be most informative, based on other researcher’s experience with macroinvertebrate indices. However, frequency was examined concurrently because it is a less time consuming observation than cover estimation and thus would be efficient . to use for management. When data were explored graphically, the relative abundance data showed land use effects more clearly than frequency or richness. Therefore, remaining statistical analyses examined relative abundance only. \ Topographic Measurements Elevations of important points at each wetland site were surveyed in August: open water surface, stakes marking the water’s edge in June, July, and August, staff gages, unflooded quadrats, piezometers, and wells. A benchmark was installed and surveyed for future hydrologic research on these sites. Water surface elevation in September was designated as baseline elevation (0 cm) for each wetland and all other elevations were expressed relative to this elevation. In this manner elevation of vegetation quadrats and well measurements were standardized relative to water surface elevation for all wetlands. Because vegetation was sampled in August and well measurements were made in September, quadrat elevations were corrected for the drop in water level between August and September. Distribution of species in relation to elevation was compared between land uses. Elevations of all quadrats in which a particular species was found were compiled. Comparisons were made for all species that had two or more observations in both the reference subset and either the retired or current subset within the range of elevations shared between subsets. Elevation distributions of retired, and current subsets were compared separately to elevation distributions at reference sites. ; . Statistical Analysis Differences among land use groups were tested using analysis of variance (ANOVA) or median tests. All statistical analyses were completed in SAS (SAS Institute, Inc., Cary, NC). Ninepipe and Ovando data sets were analyzed separately. Untransformed data were analyzed by one-way ANOVA unless data failed tests for normality and equal variance. If data failed these tests, they were log-transformed and retested. Iftransformed data failed tests for normality and equal variance, a test comparing medians of land use groups was used. After ANOVA, retired and currently impacted sites were compared to reference sites using Dunnetf s test, which compares treatments to a single control. There is no equivalent test for pairwise comparisons following a median test. Data analyzed with a median test, generally included one or more land use groups with many zero values. Differences between land use group means were considered significant at p<0.10. This liberal p value was used because my main goal was to detect changes. From a land management perspective, awareness of possible changes in wetlands and thq ability to act on indications of land use impacts is more important than being certain that change is occurring (Mapstone. 1995). Overall trends in changes in vertical distributions of rooted, vascular species were also compared among subsets using a sign (+/-) test. Only species that were present in reference sites and one or both disturbed land use groups were evaluated. The mean elevation was calculated for each species excluding quadrats at elevations not present in both land use groups. If the mean of the retired or currently impacted sites was higher than that of the reference sites, a “+” sign was assigned; if lower, a “-’’ was assigned. Where means differed between land use groups by one cm or less, the means were considered equal and not used in the statistical comparison. The proportion of species with lower mean elevations (- score) in the retired or current sites than reference sites was tested at significance levels of p < 0.10. 40 RESULTS Hydrology Surface water levels dropped an average of 35 cm during the growing season but all sites remained flooded (Table 3). Piezometers indicated that 23 sites were recharging groundwater by September (Table 4). Three wetlands in the Ninepipe area and four wetlands in the Ovando area received groundwater discharge initially, but as water levels fell over the summer, sites were no longer supplied by groundwater and instead acted to recharge groundwater. I was unable to discern flow direction at one wetland in the Ovando reference area because of data collection errors in July and dry piezometers in August and September. This site probably was also a seasonal discharge or a recharge site because water levels dropped at rates similar to those in other Ovando reference recharge wetlands. There was a clear difference in the pattern of groundwater depth relative to ground elevation at the Ninepipe and Ovando areas (Figure 4). Water levels in the Ninepipe area indicated that beneath exposed areas of wetlands, groundwater dropped away from the wetland water surface. Depth to water increased at nearly twice the rate of increase in ground surface elevation around the wetland. In contrast, increasing water depth in the Ovando wells corresponded almost directly to increasing ground surface elevation, indicating that groundwater was nearly level with water surface in the wetlands. 41 Table 3. Monthly water depths in depressional wetlands in the Ninepipe and OvandO areas. Reference, retired, and current refer to land use categories that are explained in the text. Water depths were determined by measurement of a staff gage centrally located in wetlands and surveying stakes placed at water's edge each month. Measurements are in cm. The absence of sites 2 and 6 reflects their elimination during final site selection. __________________________________________ month net site , June July August September decrease Nineoine Reference I 97 93 82 77 19 3 105 102 87 86 20 4 80 72 54 33 47 5 73 68 45 36 37 Retired 7 96 89 69 61 35 8 124 108 90 84 40 9 73 67 46 33 39 10 84 78 60 50 34 Current 11 72 71 ' 53 46 26 12 65 63 40 34 32 13 81 ' 75 52 45 36 14 50 47 26 19 32 Ovando Reference 15 213 210 195 186 .27 16 86 . 84 77 59 ' 27 17 86 76 45 33 54 18 114 . 94 49 29 85 Retired 19 114 112 101 89 24 ' 20 47 58 39 34 . 13 21 59 47 42 37 6 21 22 104 ' 99 78 66 38 Current 23 114 108 ' 92 83 31 24 187 176 152 130 56 25 216 211 194 185 31 , 26 265 261 247 240 25 42 Table 4. Direction of groundwater flow in depressional wetlands in the Ninepipe and Ovando areas. Reference,, retired, and current refer to land use categories that are explained in the text. Duection of flow was mfered from piezometers open at 40 cm and HO cm below ground surface. ND = no data. Asterisks, * and **, signify that the shallow or deep piezometer, respectively, was dry, preventing calculation of hydrologic gradient The absence of sites 2 and 6 reflects their elimination during final site selection. month site July August September I ND Ninenine Reference discharge recharge 3 recharge recharge . * 4 recharge recharge * 5. discharge recharge recharge 7 recharge Retired recharge . recharge 8 discharge recharge recharge 9 recharge recharge recharge 10 recharge recharge . recharge 11 recharge Current recharge recharge 12 recharge recharge recharge 13 recharge ND recharge 14 recharge ND * 15 discharge Ovando Reference recharge recharge 16 ' * recharge recharge 17 ND * ** 18 recharge recharge recharge 19 discharge Retired recharge i recharge 20 discharge recharge recharge 21 discharge recharge recharge 22 recharge recharge recharge 23 * Current recharge recharge 24 */** recharge recharge 25 */** recharge recharge 26 */** recharge recharge 43 y = 1.84x R2= 0.51 a) Ninepipc area ground elevation (cm) y = I 06x b) Ovando area ground elevation (cm) Figure 4. Depth to groundwater relative to ground surface elevation in depressional wetlands in the (a) Nincpipe and (b) Ovando areas. Well water depths and wetland water surface levels were measured in September and converted to relative elevations based on surveying wells, staff gages, and ground surface elevation. Ground surface elevations are reported relative to water surface elevation (O) in each wetland. Dashed lines represent groundwater depth relative to surface elevation if expected groundwater was level with the surface water in each wetland; that is, if depth to groundwater equaled surface elevation. Solid lines represent best-fit linear regressions of depth and surface elevation. Water and Sediment Characteristics Field measurements related to background environmental characteristics of water are reported for the Nine Pipe area (Table 5) and the Ovando area (Table 6). In the Ninepipe area there were no significant differences among land uses in any sampling period. In the. Ovando area, the only significant difference among land uses was for water temperature in June. Significant differences in water temperature were not seen in July or August. Electrical conductivity (EC) did not differ significantly among subsets in the Ninepipe or Ovando areas. When selected, all sites had EC below 800 pS/cm - (pmhos/cm), the upper limit for freshwater (Cowardin et al. 1979). Some sites rose above 800 pS/cm by August, but the highest value was only 879 pS/cm, less than a 10% deviation from the freshwater criteria. Hydrogeochemistry was similar within the Ninepipe and Ovando areas but concentrations of dominant cations and anions differed between the two areas (Figure 5). The Ninepipe sites were dominated by bicarbonate and sodium. In the Ovando area, bicarbonate was the dominant anion and magnesium and calcium were equally dominant cations. Nutrient concentrations in water generally did not differ significantly among land uses for either the Ninepipe areas (Table 7) or Ovando sites (Table 8). In the Ninepipe area, nitrate differed significantly between reference and retired sites in August. This difference was not found in June or July. In the Ovando area, total phosphorus varied Table 5. Physical and chemical characteristics of water in depressional wetlands in the Ninepipe area. Reference, retired, and current refer to land use categories that are explained in the text. Data were collected in late June (period I), early July (period 2) and late August (period 3). Electrical conductivity (EC), pH, and temperature were measured on site using a Horiba U-IO Water Quality Checker near the center of each wetland. Alkalinity was measured by titration, which was performed within twelve hours after collecting water samples. Electrical conductivity is reported in uS/cm, which is directly equivalent to umhos/cm. Means were compared by ANOVA (A) or a median test (M) using untransformed (U) or Iog10 transformed (L) values. The Minimum Significant Difference (MSD) is the critical value for pairwise comparisons using Dunnett's test. "Retired-Reference" and "Current-Reference" refer to differences between land use group means. Asterisks (*) designate significant differences (p < 0.10). See Appendix A, Tables 15, 16, and 17 for raw data.________ ____________________________________________________________ variable period Reference mean s.d. Retired mean . s.d. Current mean s.d. test trans. F P MSD Retired- Reference Current- Reference EC (uS/cm) I - 450 215 274 39 414 182 A U 1.29 0.32 253 -178 -40 pH I 6.9 0.5 6.7 0.2 7.1 0.4 A U 1.23 0.34 0.6 0.3 0.1 temperature (0C) I 20.8 4.3 18.8 3.3 20.0 2.5 A U 0.34 0.72 5.3 -2.0 -0.8 alkalinity (mg CaCOg/l) I 218 111 117 12 189 88 A L 1.61 0.25 . 126 -101 -28 EC (uS/cm) 2 469 ' 237 273 47 454 194 . A U 1.49 0.28 276 -198 -18 pH 2 7.1 0.2 6.6 0.2 7.1 0.8 A U 1.35 0.31 0.8 -0.5 0.0 temperature (9C) 2 21.1 3.1 22.3 3.2 18.7 • 1.7 A U 1.81 0.22 4.3 1.3 -2.4 alkalinity (mg CaCO3A) 2 214 .117 116 17 191 93 A U 1.40 0.30 134 -98 -23 EC (uS/cm) 3 574 291 323 76 607 265 A U 1.82 0.22 356 -253 33 PH 3 7.0 0.6 7.0 0.6 6.9 0.3 A U 0.03 0.97 0.8 -0.05 -0.1 temperature (9C) 3 19.8 1.9 20.3 2.8 21.2 2.4 A • U 0.34 0.72 3,7 0.5 1.4 alkalinity (mg CaCO3A) 3 272 148 139 29 263 120 A U 1.79 0.22 171 -133 -9 Table 6. Physical and chemical characteristics of water in depressional wetlands in the Ovando area. Reference, retired, and current refer to land use categories that are explained in the text. Electrical conductivity (EC), pH, and temperature were measured on site using a Horiba U-IO Water Quality Checker near the center of each wetland. Data were collected in late June (period I), early July (period 2) and late August (period 3). Alkalinity was estimated by titration, which was performed within twelve hours after collecting water samples. Electrical conductivity is reported in uS/cm, which is directly equivalent to umhos/cm.. Means were compared by ANOVA (A) or a median test (M) using untransformed (U) or Iog10 transformed (L) values. The Minimum Significant Difference (MSD) is the critical value for pairwise comparisons using Dunnett's test. "Retired-Reference" and "Current-Reference" refer to differences between land use group means. Asterisks (*) designate significant differences (p < 0.10). See Appendix A, Tables 15, 16, and 17 for raw data.______________________________ ;____________;___________________ _______ variable Reference Retired period mean s.d. mean Current Retired- Current- s.d. mean test trans. EC (uS/cm) I 528 176 657 54 450 189 A U 1.90 0.21 235 133 -75 . PH I 7.4 0.2 7.3 0.2 7.1 0.4 A U 0.79 0.48 0.4 -0.1 -0.2 temperature (0C) I 21.4 4.3 15.3 1.8 16.9 1.6 A U 4.88 0.04 * 4.4 -6.1 * -4.5 * alkalinity (mg CaCO3/!) I 256 68 313 23 213 94 A U 2.13 0.18 105 57 -43 EC (uS/cm) 2 525 200 651 46 441 183 A L 2.18 0.17 499 366 -84 PH 2 7.5 0.5 7.2 0.2 7.7 0.3 A U 2.34 .0.15 0.5 -0.2 0.3 temperature (0C) 2 21.3 2.3 19.3 ' 4.3 24.3 3.0 A U 2.36 0.15 5.0 -2.0 3.0 alkalinity (mg CaCO3/!) 2 243 76 305 19. 210 91 A U 1.92 0.20 107 63 -33 EC (uS/cm) 3 542 239 695 54 468 204 A U 1.59 0.26 284 155 -73 pH 3 7.5 0.7 7.2 0.2 7.1 0.2 A U 1.21 0.34 0.7 -0.4 -0.5 temperature (0C) 3 . 18.6 4.0 17.2 2.6 17.6 2.6 A U 0.21 0.81 4.8 -1.4 -1.0 alkalinity (mg CaCO3/!) 3 . 259 106 346 20 233 103 A U 1.87 0.21 132 86 -26 47 >L X X X X1 A / H C 034-C03v X X X Z V \\ );<■ V \ A--X--X- f t ' s ' ' z «• —J C - ! % / I / I ' *> I ~ i N I \ ' \m mm 2̂1 e . m m m £ m m . .J. m m m ym m m m f m m m . ' S ' I ' I I S • N • \ / \ I f s ' t ' • * I ' s I N I s ' N 1^X - - - ^ - xXy'y Z % Z I / I ' I » ' • \ Z ' I ' V Z Figure 5. Chemical composition of water in dcprcssional wetlands in the Nmepipe and Ovando areas. Ninepipe area wetlands are represented by diamonds and Ovando area wetlands are represented by circles. Water samples were collected in late June. See Appendix A, Table 15 for raw data. Table 7. Nutrient concentrations in water in depressional wetlands in the Ninepipe area. Reference, retired, and current refer to land use categories that are explained in the text. Data was collected in late June (period I), early July (period 2) and late August (period 3). All concentrations are mg/1. Phosphate (PO4) values below 0.01, and nitrate (NO3) and ammonium (NH4) values below 0.1 have a lower analytical accuracy than data above these values. Means were compared by ANOVA (A) or a median test (M) using untransformed (U) or Iog10-Iransformed (L) values. The Minimum Significant Difference (MSD) is the critical value for pairwise comparisons using Dunnett's test. "Retired-Reference" and "Current-Reference" refer to differences between land use group means. Asterisks (*) designate significant differences (p < 0.10). See Appendix A, Tables 15, 16, and 17 for raw data. ______________________________________________ variable period Reference mean s.d. Retired mean s.d. Current mean s.d. test trans. F P MSD Retired- Reference Current- Reference NO3 I 0.04 0.03 0.03 0.04 0.02 0.01 A U 0.66 0.54 0.04 -0.02 -0.01 NH4 I • 0.09 0.04 0.14 0.05 0.11 0.06 A U 0.78 0.49 0.07 0.02 0.04 P I 1.06 0.91 0.55 0.17 0.84 .0.52 A U 0.71 0.52 0.94 0.52 -0.23 PO4** I 2.29 2.33. 0.34 0.10 1.13 . 0.98 A L 2.0 0.19 0.67 -0.61 -0.24 NO3 ' 2 0.05 0.02 0.04 0.02 0.04 0.01 A U 0.30 0.75 0.03 0.00 -0.01 NH4 2 0.09 0.05 0.10 0.02 0.08 0.05 A U 0.24 0.79 0.06 0.01 -0.01 P 2 1.02 0.96 0.56 0.15 0.96 0.47 A U 0.66 0.54 0.96 -0.47 -0.06 NO3 3 0.06 0.03 0.03 0.01 0.04 0.01 A U 2.9 0.104 0.031 -0.033 * -0.025 NH4 3 0.12 0.05 0.09 0.06 0.13 0.06 A U 0.62 0.56 0.09 -0.04 0.01 P 3 1.11 111 0.55 0.16 1.09 0.35 A U 0.90 0.44 1.04 -0.57 -0.02 PO4 3 2.14 2.76 0.32 0.15 0.78 . 0.42 A L 0.93 0.43 0.78 -0.45 -0.04 **P04 values are reported as total molecular mass of phosphate per liter. To convert to phosphorus as phosphate, multiply by 0.33. Other values are reported as mass of nitrogen or phosphorus per liter. Table 8. Nutrient concentrations in water in depressional wetlands in the Ovando area. Reference, retired, and current refer to land use categories that are explained in the text. Data was collected in late June (period I), early July (period 2) and late August (period 3). All concentrations are mg/1. Phosphate (PO4) values below 0.01, and nitrate (NO3) and ammonium (NH4) values below 0.1 have a lower analytical accuracy than data above these values. Means were compared by ANOVA (A) or a median test (M) using untransformed (U) or Iog10-Iransformed (L) values. The Minimum Significant Difference (MSD) is the critical value for pairwise comparisons using Dunnett's test. "Retired-Reference" and "Current-Reference" refer to differences between land use group means. No significant differences were found (p < 0.10). See Appendix A, Tables 15, 16, and 17 for raw data.____________________________ variable period Reference mean s.d. Retired mean s.d. Current mean s.d. test trans. F P MSD Retired- Reference Current- Reference NO3 I 0.01 0.01 0.02 0.01 0.02 0.01 A U 0.91 0.44 0.01 0.01 0.01 NH4 I 0.04 0.03 0.02 0.01 0.03 0.01 ■ A U 1.73 0.23 0.03 -0.02 0.00 P I • 0.33 0.45 0.15 0.01 0.09 0.02 M U — 0.03 "— — — PO4*.* I 0.40 0.80 0.00 0.00 0.00 0.00 M U — 0.37 — - — — - NO3 2 . 0.01 0.00 0.01 0.01 001 0.00 M U — 0.37 — — ——— NH4 2 0.05 0.03 0.03 0.02 0.05 0.02 A U 0.71 0.52 0.03 -0.02 0.00 P 2 0.29 0.35 0.11 0.02 0.11 0.03 ■ A L 0.59 0.58 0.46 -0.21 -0.19 NO3 3 0.07 0.08 0.02 0.01 0.02 0.01 A L 2.37 0.15 0.45 -0.36 -0.41 NH4 • 3 • 0.16 0.19 0.23 0.22 0.12 0.11 A U 0.37 0.70 0.27 0.07 -0.04 P . 3 0.18 0.18 0.06 0.01 0.08 0.02 M U — 0.12 — — — 5 3 0.26 . 0.30 0.00 0.00 0.00 0.00 M U — 0.03 — — — **P04 values are reported as total molecular mass of phosphate per liter. To convert to phosphorus as phosphate, multiply by 0.33. Other values are reported as mass of nitrogen or phosphorus per liter. significantly among land uses in June. Because the median test was used, pairwise comparisons were not made. Mean phosphorus concentrations were 50% lower in retired sites than reference sites and 70% lower in currently grazed than reference sites. Total phosphorus values did not differ significantly during other months. Phosphate varied significantly among the Ovando land use groups in August. Again, because the median test was used, pairwise comparisons were not made; however, means indicated that phosphate was present at low but detectable levels in reference sites but was below detection limits in both retired and current sites. This tendency.toward higher mean total phosphorus and phosphate values in reference sites than retired or. current sites was observed during all months, although differences were not significant except as reported above. There were no significant differences in sediment geochemistry in the Ninepipe area (Table 9). SARvaried significantly among land use groups in the Ovando area (Table 9), but because a median test was used, specific land use groups were not compared. However, mean SAR for the retired subset was an order of magnitude higher than for reference or currently grazed subsets. Sediment pH and EC did not differ significantly between the reference sites and other, land use groups in the Ovando area. Differences in nutrient concentrations in sediments were found among land use groups in the Ninepipe area (Table 10). Ammonium was significantly higher at retired than reference sites. Total phosphorus was significantly greater on retired and currently cultivated sites than reference sites. There were no significant differences in sediment chemistry among land uses in the Ovando area (Table 10). Table 9. Geochemical characteristics of sediment in depressional wetlands in the Ninepipe and Ovando areas. Reference, retired, and current refer to land use categories that are explained in the text. Samples used for SAR determination were collected in late June. Samples used for pH and electrical conductivity (EC) determination were collected in early July. Means, were compared by ANOVA (A) or a median test (M) using untransformed (U) or Iog10-Iransformed (L) values. The Minimum Significant Difference (MSD) is the critical value for pairwise comparisons using Dunnett's test. "Retired-Reference" and "Current-Reference" refer to differences between land use group means. Asterisks (*) designate significant differences (p < 0.10). See Appendix A, Table 18 for raw data.________________________________________________________________ Reference Retired Current Retired- Current- variable mean s.d. mean s.d. mean s.d. test trans. F P MSP Reference Reference Ninepipe SAR . 6 5 3 I 3 2 A pH 8.1 0.8 7.5 0.2 7.7 0.3 A EC (uS/cm) 1024 875 471 136 581 200 A Ovando SAR 0.4 0.1 3 5 0.4 0.1 M U ---- 0.03 * —™ — pH 8.0 0.4 8.3 0.2 7.9 0.3 A U 1.5 0.27 0.5 0.3 0.0 EC (uS/cm) 468 231 541 128 401 93 A U 0.75 0.50 249 73 -68 U 0.67 0.54 5 -2 -2 U 1.7 0.24 0.8 -0.6 -0.4 U . 1.2 0.33 807 -553 -440 Table 10. Nutrient concentrations in sediment in depressional wetlands in the Ninepipe and Ovando areas. Reference, retired, and current refer to land use categories that are explained in the text. Samples were collected in late June. Means were compared by ANOVA (A) or a median test (M) using untransforined (U) or Iog10-Iransformed (L) values. The Minimum Significant Difference (MSD) is the critical value for pairwise comparisons using Dunnett's test. "Retired-Reference" and "Current-Reference" refer to differences between land use group means. Asterisks (*) designate significant differences (p < 0.10). See Appendix A, Table 18 for raw data. __________________________________________________ Reference Retired Current Retired- Current- variable mean s.d. mean s.d. mean s.d. test trans. F P MSD Reference Reference Nineoine TKN (%N) 0.55 0.15 0.45 0.06 0.46 0.14 A U 0.78 0.49 0.19 -0.10 -0.09 NH4 (mg/kg) 37 19 87 28 62 16 A U 5 0.03 * 33 50 * 25 NO3 (mg/kg) 9 7 5 3 7 6 A U I I 9 -5 -2 . P (mg/kg) 30 11 51 10 57 14 A U 6 0.02 * 18 21 * 27 * Ovando TKN (%N) 0.73 0.15 0.95 0.31 1.09 0.39 A U 1.4 0.29 0.46 0.22 0.36 NH4 (mg/kg) 4 2 23 22 .18 33 13 A U I 0.4 28 -19 -9 NO3 (mg/kg) 14 5 20 12 23 15 A U I I 18 6 9 P (mg/kg) 42 32 30 15 20 . 4 A U I 0.4 31 -12 -21 Ninepipe area substrates had higher clay contents than the Ovando substrates (Figure 6); sediments in nine out of 12 sites were clays or silty clays and three were silty clay loams. In contrast, sediment in 10 out of 12 Ovando sites classified as loams, silty loams, or clay loams; two fell low in the clay category. Vegetation In the Ninepipe area, life history groups showed the strongest differences between reference and currently or previously cultivated sites (Table 11). Perennial species differed significantly between reference sites and currently cultivated or retired sites. Because a median test was used, pairwise comparisons among land uses were not made. However, mean values for cultivated sites were 15% less than for reference sites, while means for retired sites were slightly higher than for reference sites. Native perennials showed greater differences between reference and cultivated sites than all perennials combined. Relative abundance of native perennials was significantly lower on cultivated than reference sites, reflecting a 40% decrease in mean values. Retired sites were not significantly different from reference sites. Relative abundance of all native species did not differ significantly among land use groups, although values tended to be lower on cultivated than reference sites. Annuals were significantly less abundant in retired sites than in reference sites; in fact, they were virtually absent. Although the difference was not statistically significant, the mean relative abundance of annuals was four times greater at reference sites than 54 Q Ovando reference sites B Ovando retired sites A Ovando current sites • Ninepipe refererence sites ■ Ninepipe retired sites A Ninepipe current sites Figure 6. Sediment particle-size distributions in depressional wetlands in the Ninepipe and Ovando areas. Ninepipe land use groups are represented by solid symbols and Ovando land use groups are represented by open symbols. Samples were collected in early July. See Appendix A, Table 18 for raw data. Table 11. Composition of vegetation in depressional wetlands in the Ninepipe area: a) relative abundance of species groups, b) relative abundance of selected species, and c) species richness. Reference, retired, and current refer to land use categories that are explained in the text. Vegetation was sampled in late August. Means were compared by ANOVA (A) or a median test (M) using untransformed (U) or Iog10-Irahsformed (L) values. The Minimum Significant Difference (MSD) is the critical value for pairwise comparisons using Dunnett's test. "Retired-Reference" and "Current-Reference" refer to differences between land use group means. Asterisks (*) designate significant differences (p < 0.10). See Appendix B, Table 19 for relative abnnHanee data for individual species. See Appendix B, Table 20 for summary of group designations for each species.__________ variable Reference mean s.d. Retired mean s.d. Current mean s.d. test trans. F P MSD Retired- Reference Current- Reference a) Species groups Annuals 0.05 0.05 0.001 0.001 0.19 0.14 A L 11.5 0.003 * 1.14. -1.42 * 1.08 Perennials 0.95 0.06 0.999 0.001 0.80 0.13 M U 0.01 * —-- Natives 0.71 0.17 0.67 0.30 0.50 0.19 A U 0.94 0.43 . 0.35 -0.04 -0.21 Non-natives 0.29 0.17 0.29 0.30 0.43 0.21 A U 0.51 . 0.62 0.36 0.01 0.15 Native perennials 0.69 0.17 0.67 0.30 0.31 0.16 A U 3.9 0.06 * 0.34 -0.03 -0.38 * Weedy 0.51 0.26 0.64 .0.23 0.48 0.23 A U 0.49 0.63 0.37 0.12 -0.04 Weedy annuals 0.02 0.04 0.00 0.00 0.00 0.00 M U. 0.11 Weedy perennials 0.48 0.28 0.64 0.23 0.48 0.23 A U 0.53 0.61 0.39 0.16 -0.01 Grasses 0.23 0.17 0.33 0.30 0.47 0.17 A U 1.2 0.36 0.34 0,10 0.24 Grasslikes 0.39 0.17 0.29 0.19 ' 0.39 0.25 A U 0.32 0.74 0.32 -0.10 -0.01 Grasslike perennials • 0.39 0.17 0.29 0.19 0.23 0.19 A U 0.78 0.49 0.28 -0 .10 -0.16 Forbs . 0.13 0.10 0.03 0.03 0.08 0.07 . A U 1.9 0.21 0 11 -0 .10 -0.05 ' Moss 0.11 0.11 0.000 0.001 0.001 0.003 A L 12.0 0.00 * 1.36 -2.62 * -2.67 * Submergents .0.07 0.14 0.002 0.004 0.06 0.13 M U 1.00 Spp. with persistent litter 0.51 0.12 0.52 0.21 0.09 0.06 A U 12.3 0.00 * 0.22 0.02 -0.42 * 1st most dominant spp. 0.38 0.08 0.48 0.18 0.44 0.13 A U 0.52 0.61 0.21 0.10 0.06 1st & 2nd most dominant spp. 0.61 0.16 0,71 0.16 0.66 . 0.12 A U 0.48 0.64 6.23 0.10 0.05 1st, 2nd, & 3rd most dominant spp. 0.74 0.19 0.89 0.10 0.78 0.05 A U 1.5 0.28 0.20 0.15 0.04 Tillage tolerant spp. 0.18 0.03 0.17 0.12 0.23 0.22 A U 0.17 0.85 0.23 -0.01 0.05 Table 11. (continued) Reference______ Retired_______ Current Retired- Current- variable ________