Loading...
05/02/02 DeGrandpre/Development in the Evergreen Alluvial Aquifer AreaFlathead Basin Commission 33 Second Street East Kalispell, Montana 59901 406.752.0081 406.752.0095 Fax fbc@digisys.net May 2, 2002 Tri-City Planning Office 17 Second Street East — Ste. 211 Kalispell, MT 59901 Re: Development in the Evergreen Alluvial Aquifer area Administratively attached to Office of the Governor Capitol Station Helena, Montana 59620 406.4443111 Fax 406.444.5529 The Flathead Basin Commission is a non -regulatory state commission established by the Montana Legislature to monitor, protect and preserve the water quality of the greater Flathead watershed. Among the FBC's primary objectives are to use voluntary, education -based methods to involve the public in pro -active efforts to recognize and address non -point source pollution issues. The Evergreen area east of Kalispell, given its high water table, ,highly permeable soils, proximity to the main stem of the Flathead River, dense settlement pattern, and rapid growth, has long been a source of concern as these factors relate to the protection of water quality. The FBC believes the public interest is best served when the widest possible range of objective, science -based information is taken into account when land management decisions are made. In the case of the Evergreen area, there is ample credible information available that provides detailed analysis of the function of the local aquifer and how water moves between surface, groundwater, and river environments. It is the documented existence of this so-called "interchange zone" that raises concerns about additional potential sources of non -point source pollutants in the Evergreen area. The FBC strongly urges the Kalispell City Planning Board to give protection of water quality its highest priority when considering development proposals. The FBC does not comment on individual projects, but believes the approval or rejection of such proposals or stipulation of appropriate mitigation design features should follow a full examination of the best scientific information available. Fortunately, top experts in this field of scientific endeavor have conducted internationally recognized research on the issue in the Flathead. We have taken the liberty of attaching copies of two publications that provide definitive information on the specifics of alluvial aquifer in the Evergreen area. Sincerely, David QeGdpre, Chair /G Flathead Basin Commission Attachments: "The hyporheic habitat of river ecosystems" (FLBS and CSU) "Ecology of the Alluvial Aquifers of the Flathead River, Montana" (FLBS) Ecology of the Alluvial Aquifers of the Flathead River, Mon tana J. A. Stanford,* J. V. Ward,f and B. K. Ellis* * Flathead Lake Biological Station The University of Montana Polson, Montana 59860 t Department of Biology Colorado_ State University Fort Collins, Colorado 80S23 I. Introduction II. Biophysical Setting of the Catchment Ecosystem III. Methods IV. Hydraulic Connectivity of the Alluvial Aquifers with the River V. Temperature and Solute Dynamics VI. Distribution and Abundance of Groundwater Biota A. Microbial Ecology B. Invertebrate Ecology C. Food Web Relationships D. Cave Streams VII. Conclusion References I. INTRODUCTION Stanford and Gaufin (1974) reported the presence of abundant insect larvae deep in the alluvia of the Tobacco River, a small gravel bed river in northwestern Montana. The insects were entrained in an infiltration gallery that provided potable water from the alluvial aquifer to the town of Eureka. - Montana. The collection system was located ca. 1SO m from the river and 3 m below the floodplain surface. The most abundant forms were stone flies (Plecoptera) of the genera Isocapnia and Paraperla. However, many other taxa, traditionally thought of as river macrobenthos, were also present, 368 Stanford et al. either actively or passively entering the collection system from the interstitial aquifer of the river. The discovery, coupled with the wide range in western North America of adult Isocapnia and Paraperla, led the authors to propose that the deep penetration of benthos into the saturated alluvia of gravel bed rivers was likely a widespread phenomenon that had significant ramifications for river ecology. A decade later Stanford and Ward (1988) confirmed the occurrence of the same Plecoptera, along with over 70 other invertebrate taxa, in the shallow alluvial aquifer of the much larger Flathead River in the Kalispell Malley (Fig. 1); Plecoptera larvae were pumped from drilled wells on the valley bottom up to 3 km from the river channel. Others working on intersti- tial fauna of gravel bed streams and rivers in Europe (Bretschko and Klemens, 1986; Danielopol, 1976; see also Chapters 12 and 13), Canada (Coleman and Hynes, 1970; Pugsley and Hynes, 1986; Williams and Hynes, 1974), and elsewhere in the United States (Boulton et, aL, 1991, 1992; Palmer, 1990; Pennak and Ward, 1986; Ward and Voelz, 1990; see also Chapter IS) also demonstrated speciose interstitial assemblages, but at the scale of meters or Iess. Based on our work summarized below and research on the Rhone River in France (Marmonier et al., _ 1992; Dole -Olivier et al., this volume), it is clear that a diverse and abundant interstitial fauna may inhabit the alluvium of entire floodplains kilometers from the river channel. Our long-term studies in the Flathead catchment now emphasize the four- dimensional nature of the riverine and floodplain landscapes (Ward, 1989): and the processes and implications related to water and materials exchange between floodplain alluvial aquifers and the river (Stanford and Ward,1993). The purpose of this chapter is to summarize the data and interpretations linking the groundwaters and surface waters in the Flathead catchment. Much of the work remains unpublished, and some original data are included herein. 11. BIOPHYSICAL SETTING OF THE CATCHMENT ECOSYSTEM The Flathead River —Lake Catchment encompasses 22,241 km2 in west- ern Montana and British Columbia (Canada) (Fig.1). The attitudinal gradi- ent ranges from 750 in above sea level (a.s.l.) at confluence with the Clark Fork of the Columbia River to 882 m a.s.l. at Flathead Lake and over 4000 in a:s.l. on the Continental Divide. The geology is complex. Rock ages and composition vary from the very old and nutrient -poor Belt Series of the Precambrian age in Glacier National Park (drainage area between the Continental Divide and the North and Middle Forks; Fig.1) to much younger limestones and other Mesozoic strata in the headwater reaches of the North, Middle, and South Forks. All of these formations were modified and molded into the present landscape by the cordilleran orogeny and subsequent erosion 14. Alluvial Aquifers of the Flathead River, Montana 369 BRrnsH COLUMBIA MONTANA N 20 Km to COLUNMIA RIVER NORTH ALLUVIAL FISSURED MONTANA ALBFRYA CONTINENTAL FORK �~ DIVIDE Y ' N " K MIDDLE FORK FLATHEAD LAKE FLATHEAD RIVER M SOUTH JOCKO CLARK FORK FIGURE 1 Flathead River catchment in Montana and British Columbia, Canada, showing alluvial floodplains (stippled), including the primary study sites at Nyack (N, a montane floodplain) and Kalispell Valley (K, a piedrount valley floodplain). Well transects at the Kalispell Valley site are shown as bold lines. Lakeshore alluvia containing groundwater biota occur at site S (Flathead Lake Biological Station). Cave streams in fissured bedrock substrata were studied at site T (Trail Creek) and were reported at sites M (McDonald Creek) and B (Sported Bear River) [after Stanford and Ward (1993)], by Pleistocene glaciation and fluvial outwash processes (Locke, 1990; Ross, 1963). The land mass is drained by a network of mostly pristine streams and lakes. Streams generally begin as spring brooks emerging from fissured aquifers in the bedrock or from alluvia usually associated with the glacial history (see Fig. 6 in Chapter 1). The annual hydrologic pattern of surface flow is distinctively maximized in May —July by spring snowmelt in the 370 Stanford et aL mountain headwaters of each subcatchinent. Streams, lakes, and near - surface aquifers fill to capacity or flood, depending on the intensity and duration of the-snowmelt spate. Lower volume floods may occur at other times of the year, resulting from short-term, heavy precipitation created by the collision of Pacific maritime and continental air masses within the catchment. Average annual flow of the sixth -order (Strahler, 1965) Flat -head River at confluence with 480-km' Flathead Lake is 340 m/sec. Three dams regulate the flow of the mainstern river above and below Flathead Lake and the top 3 m of the lake (Fig. 1). Glaciation undoubtedly was responsible for the initial formation of the larger alluvial reaches (Fig. 1), although local Quartenary alluvial structures in the catchment have not been studied in detail. Our primary study sites are the Kalispell Valley alluvial aquifer (K; Fig. 1), between the Whitefish (mean Q = 20M3 /sec) and Flathead River channels upstream from Flathead Lake and the Nyack Valley aquifer on the Middle Fork (N; Fig. 1). The Nyack Valley is an aggraded segment that constricts into steep canyons at the upstream and downstream ends. In both cases the valleys apparently were shaped by glaciation, and postglacial fluvial outwash left an expansive deposit of cobble and gravel alluvia about 10 m thick, overlying an imperme- able layer of Tertiary clay Subsequent flooding and migration of the active river channel across the valley in response to centuries of cut and fill alluvi- ation and, perhaps, block fault tectonics deposited 1-3 m of fine sediments on the ancient floodplain surface. Hence, the extant landscape of the Kalispell Valley is composed of the modem floodplain and channel, low terraces of the ancient floodplain, higher, lateral terraces created by glaciation (see Fig. 3), and a shallow, unconfined alluvial aquifer. Ancient river channels (paleochannels) are visible as depressions in the ancient floodplain surface, and the buried channels are remarkably evident in traces done with ground, penetrating radar (G. Poole, Flathead Lake Biological Station, unpublished). The montane valley (N) differs from the piedmont valley (K) only in being smaller, and it lacks the high lateral terraces. Rubble bottom streams also occur in karstic caves in Mesozoic lime- stones (B and T in Fig. 1) (J. A. Stanford, et al., unpublished) and in the Belt Series substrata (M; Fig. 1) (G. Gregory, Glacier National Park, West Glacier, Montana, personal communication). The drainage patterns of the cave streams have not been documented. The waters of the Flathead are cold, and concentrations of dissolved nutrients, such as nitrogen, phosphorus, and organic matter (DOM), in surface waters are naturally near or below detection limits; so the lakes and streams naturally do not produce much plant and animal material compared with most other waters in the United States (Ellis and Stanford, 1988; Stanford et al., 1992). An important feature of the Flathead is that many aquatic species are present, owing in part to the pristine, oligotrophic nature of the waters and 14. Alluvial Aquifers of the Flathead River, Montana 371 the wide variety of aquatic habitats- along the altitudinal gradient (Hall et al., 1992). Moreover, the Flathead exists midway in the north —south continuum of the Rocky Mountains, and three major river systems of the North America continent share adjacent headwaters in Glacier National Park (i.e., the area between the North and Middle Forks and the Continental Divide; Fig. 1). these include the Saskatchewan, Missouri, and Columbia Rivers, which drain into the Arctic, Atlantic, and Pacific Oceans, respectively. Hence, many species on the fringe of their continental distribution are pres- ent, and the area is a "melting pot" of biodiversity. For example, 67% (101 species) of Plecoptera reported for the Rocky Mountains from Alaska to New Mexico have been collected in the Flathead catchment [updated from Stanford and Ward (1983)]. Native biota include species [e.g., bull charr (Fraley and Shepard, 1989) and west -slope cutthroat trout (arnell, 1988)] that are relatively rare in North America, having been eliminated from less pristine areas by humans or existing as relict populations since PIeistocene glaciation (Stanford and Prescott, 1988). Biophysical features of the catchment are naturally interrelated; pro- cesses in one place or time may be influenced or controlled by adjacent processes. For example, the riparian plant assemblages of the alluvial flood - plains appear to be closely linked to. piezometric gradients associated with groundwater down- and upwelling (Stanford and Ward, 1993). Although the river —lake ecosystem is reported to be relatively pristine for similarly sized catchments in the human -dominated latitudes of the world, stream - and lake -level regulation by dams (Stanford and Hauer, 1992). and other human sources of disturbance have exerted measured influences on natural ecosystem integrity (Spencer et al., 1991; Stanford and Ward, 1992). It is within this complex biophysical setting that we wish to summarize what currently is known about the groundwater ecology of the alluvial` aquifers of the Flathead River. III. METHODS Research wells, installed with a hollow auger drilling rig (Fig. 2), con- sisted of slotted PVC pipe. We used an electric saw to cut S min x S cm alternating slots in the pipe before installation. The wells were cappped on the top and bottom, although bottom caps were also slotted to allow fauna that collected in the wells between sampling periods to escape. Wells were arrayed across the modern and ancient floodplains of the Kalispell Valley, including two transects (Fig. 3) that were intensively studied. Most wells penetrated at least S m into the saturated alluvia. Over 200 wells were installed to. determine aquifer characteristics, such as flow direction and basic ion chemistry, in comparison- with those of the river. Several large wells were installed by the Montana Bureau of Mines and r f 14. Alluvial Aquifers of the Flathead River, Montana 373 000 0 0 0 0 ova o M o C of A r r A !7 Cu {m IONS to Flathead Lake FIGURE 3 Water table elevations (1.S-m isopleths, bold contours) obtained from time series measures of depth to water in 271 monitoring wells installed in the Kalispell Valley alluvial aquifer. Thin contours are isopleths of specific conductance (µS/cm) of the groundwater. Arrows indicate the direction of groundwater flow (declining elevation of water table). Solid circles are wells (including two transects; see Fig. S) used to sample groundwater biota; the stippled area indicates the portion of the aquifer in which amphibites were routinely collected Ica. corresponding to the 3S0 µS/cm contour; after Stanford and Ward (1988)]. FIGURE 2 (top) Hollow auger drilling rig used to install sampling wells (slotted PVC pipe) in the alluvial floodplains of the Flathead River. A finished well is shown in the foreground. (bottom) Janine Gibert (left) and Bonnie Ellis (right) remove small rocks that have been left in the well to allow colonization of epilithic biofilm. 374 Stanford et al. Geology to allow pump tests for the estimation of hydraulic conductivity, transmissivlty, and flow rates; others we fitted with temperature and pressure sensors and data loggers to document thermal patterns and water table dynamics. Wells plotted in Fig. 3, in particular the upstream and downstream transects, were sampled to determine the distribution of groundwater fauna. Data loggers for temperature and flow also were installed at river sites, and additional data were available from the U.S. Geological Survey. On each sampling date, the depth to the water table was recorded, and then wells were purged using a gas -powered - diaphragm pump connected to a 5-cm-diameter tube inserted into the bottom of the wells. The pump rate for most wells was ca. 40 1 per minute. All water, fauna, and detritus pumped out of a specific well during the first 10 min were filtered with a 100-ILm-mesh plankton net. Pumping was continued until a clear water flow was attained, and then water samples for chemical constituents (nitrogen and phosphorus forms and DOM) and microbial determinations [acetone - extractable chlorophyll and epiflouresence (DAPI) cell counts] were taken from the pump stream. Temperatures, specific conductance values, and dis- solved oxygen concentrations were recorded in each well before and after pumping. Samples for dissolved oxygen, specific conductance .determina- tions, chemical analyses, and benthos also were collected in the river on specific sampling dates for comparison to well data. Biophysical data summa- rized herein were obtained from quarterly (spring,_ summer, fall, and winter) samplings of the wells in the two transects (Fig. 3) during the period January 1988—April 1989'. Two of the wells, C and E, were sampled less frequently due to access problems. The transects were installed ih an attempt to docu- ment the flux of water and materials through the aquifer and to avoid the urbanized area of the aquifer farther downstream, where groundwater effects were evident (Noble and Stanford, 1986). Wells were drilled in a grid on the Nyack- Valley and sampled in a similar -fashion. IV. HYDRAULIC CONNECTIVITY OF THE ALLUVIAL AQUIFERS WITH THE RIVER The boundaries of the Kalispell Valley alluvial aquifer are not clearly delineated, but the system is ca. 5-6 km wide at the upstream end and ca. 13 km long. The aquifer apparently is fed by the Flathead River and perhaps to a lesser extent by the much smaller Whitefish River and other small tributaries that flow into the alluvial domain of the valley from the Whitefish Range at the upstream end (Konizeski et al., 1968). Additional geohydrolog- ical investigations are needed to precisely determine the extent of the aquifer and the relative importance of different sources of water. However, we demonstrated that the aquifer is hydraulically connected to the Flathead River; seasonal and daily dynamics of the water table are clearly correlated 14. Alluvial Aquifers of the Flathead River, Montana 375 with river flow (Fig. 4). Water flows through the aquifer from north to south and is highly interactive with the .river (Fig. 3). Upwelling groundwaters are evident as spring brooks at many locations on the active and ancient floodplain; most erupt in paleochannels (Fig. S). The aquifer has an average slope of 2°. Hydraulic conductivity and, by inference, interstitial porosity are variable because some wells and some layers within wells yielded water at higher rates than others. However, in general, hydraulic conductivity estimates (0.1-10 cm/sec) from pump tests were among the highest recorded in unconsolidated alluvia (Freeze and Cherry,1979). Wells in or near paleo- channels always produced the highest pump rates, which inferred that the paleochannels were zones of the highest porosity. The aquifer discharges into the Flathead River at or before the anastomosed zone downstream of the confluence of the Whitefish River (Fig. 3), where the aquifer terminates in contact with fine-grained paleodeltaic sediments of postglacial Flathead Lake. More detailed geohydraulic investigations were performed at the Nyack Valley site (J. A. Stanford et al., unpublished data). The valley is 3 km across at the widest point by 8 km long, and the alluvial aquifer occupies 3 well - 0.65 km i---------- well - 0.02 km r 0 J F M A M river tt���i#lam A s O N 1989 FIGURE 4 Hydrographs obtained in the Flathead River channel and in two wells located 0.02 and 0.6S km laterally (west) from a channel in the Kalispell Valley (sires Ru, A, and D in Fig. S). Nonseasonal fluctuations were caused by discharges from a Iarge hydroelectric dam on an upstream tributary on the South Fork (Fig. 1). Data are depths to water surface in the wells and the river channel relative to surface benchmarks as measured by hydrostatic sensors reporting on the hour to data loggers during 1989 [from Stanford and Ward (1993)]. 376 Stanford et al. North ANCIENT FLOODPLAIN MODERN FLOODPLAIN TERRACE 1 km FIGURE 5 Mean abundance of amphibites (upper numbers) and srygobites (lower numbers) in wells (letters) in two transects (see Fig. 3) across the alluvial aquifer in the Kalispell Valley (upstream transect, wells A—G from river sampling site Ru; downstream transecr, wells H—O from river sampling site Rd). Paleochannels on the ancient floodplain (see text) are indicated by pathways in the stippling. Spring Creek and Gooderich Bayou are perennial spring brooks that emerge from paleochannels. No biological data were collected in well C. the entire valley bottom. Hydrographs in wells near the valley walls were directly correlated with the river flow pattern. Maps of piezometric gradients and accretion studies showed that 30% of the river flow enters the alluvium at the upstream end of the valley. Interstitial water upwells from the middle to the end of the aquifer in relation to ponding caused by the constriction 14. Alluvial Aquifers of the Flathead River, Montana 377 of the .valley at the downstream end- All of- the interstitial flow enters the river channel directly or by spring brooks. Upwelling groundwaters create • network of spring brooks on the modern and ancient floodplain and are • primary determinant of riparian plant assemblages (Stanford and Ward, 1993). V. TEMPERATURE AND SOLUTE DYNAMICS Interaction between the Kalispell Valley aquifer and the Flathead River was inferred by mapping isolines of specific conductance (Fig. 3) relative to the river (190 AS/cm). Values in the aquifer were similar near the river, indicating a large hyporheic or transition zone between the river and 'the aquifer; values were over two times higher in wells most distant from the river, indicating a longer residence time of phreatic water. The annual temperature amplitude in the river was 2-20°C; the river did not freeze in spite of very cold winter air temperatures (— 30'C), owing to regulated flows from Hungry Horse Dam and - the discharge of warm water from the aquifer into the river. The annual temperature amplitude in the aquifer was 6-1 VC, and the amplitude was most pronounced in wells nearest the river. Dissolved oxygen concentrations in the wells were consis- tently >50%. saturation (4-10 mg/1). In the river dissolved oxygen was always near saturation or supersaturated A. Stanford et al., unpublished data). We observed a strong tendency for dissolved constituents, especially nitrate to concentrate in the center of the aquifer (Fig. 6). Mass balance calculations suggested that 12-17% of the base flow nitrate load of the river downstream from the aquifer was derived from nitrate generation, probably due to nitrification, within the aquifer. Soluble .reactive phosphorus was as an order of magnitude higher in well samples than in the river, and we estimated that 4-25% of the river load during base flow was due to aquifer discharge (Stanford and Ward, 1988). High solute loads in groundwaters discharging into the oligotrophic surface waters produced mats of green algae, Ulothrix zonata, and diatoms that otherwise would not be present, and thereby they influenced the distribution of benthic consumers (Hauer and Stanford, 1981, 1982, 1986) and may strongly influence bloproduction in surface waters (Stanford and Ward, 1993). On the other hand, DOM values were consistently low (<2 mg/1) in the aquifer, suggesting the possible carbon limitation of microbial metabolism relative to other dissolved constit- uents within the aquifer (Stanford and Ward, 1988). We initially established the alternate study site at Nyack because we were concerned about contamination of the aquifer in the Kalispell Valley by human activities, especially fertilizers from farms, and the possible inter- active effects of the regulation of river flow by Hungry Horse Dam. In 378 Stanford et a1. t Z Soo 0 Ru A S C D E F G SITE 2000 0 z 500 0 Rd H I 7 K L M N O SITE FIGURE 6 Average concentration of nitrate + nitrite (ILgll) observed in quarterly (see text) samples obtained during the period January 1988—April 1989 in the Flathead River and wells located in the transects shown in Fig. S (bars indicate ranges). comparison, the Nyack Valley is virtually pristine. However, we observed no significant differences in solute or dissolved oxygen patterns. Thermal patterns were also similar, except that the river was near 0°C in midwinter and froze over at the upstream (downwelling) end for long periods (J. A. Stanford et al., unpublished data). We concluded that- contamination problems in the Kalispell Valley were minimal, at least on the two well 14. Alluvial Aquifers of the Flathead River, Montana 379 transects (Figs. 3 and S). But, contamination of wells was highly probable (i.e., highest specific conductance, lowest dissolved oxygen readings, highest DOM and nitrate concentrations, high coliform bacteria counts, and fewer groundwater biota) in a limited area of intensive housing development and individual sewage drain fields on the ancient floodplain near the town of Kalispell at the lower end of the aquifer (Noble and Stanford, 1986). VI. DISTRIBUTION AND ABUNDANCE OF GROUNDWATER BIOTA A. Microbial Ecology The main objective was to determine the relative importance of free living and attached bacteria, fungi, and Protozoa in the groundwater food web of the alluvial aquifers. We also examined the entrainment of riverine algae and. bacteria in the aquifers. This section is based. primarily on the work of Ellis et al. (199S). In the Kalispell Valley, diatoms common in the Flathead River, such as Synedra, Acbnantbes, Navicula, Gompbonema, and Cymbella were present in well samples up to 0.3S km from the river, with autofluorescing chromato- phores very " visible. Green algae were found primarily in wells far from the river channel (0.6S-2.55 km), whereas blue—green algae were rather ubiquitous. This strongly suggested that riverine particulate organic carbon was entrained in the aquifer. Indeed, measurable concentrations of chloro- phyll a were detected in samples from many wells; values were consistently highest in the upstream transect (where the river is influent to the aquifer) in wells nearest the river and reached a maximum of 14 ng/l. However, in spite of the strong hydraulic connectivity between the aquifer and the river in the Kalispell Valley (Fig. 4), the total microbial density, in wells near the river was about 80% lower than that in' the river, and only 2-3 % of the riverine microbial density occurred in the most distant wells. In the remote wells densities averaged ca. 9.0. X 10' cells/ml, which is at the low end of reported values for groundwater. The immediate decline of bacteria relative to the river may be due in part to a similar decline in low -molecular -weight utilizable carbon. Multiple regression analysis indi- cated that SRP, ammonium, nitrate + nitrate, and DOM concentrations (in that order) explained about half the variance in the free living bacteria assemblages in the aquifer. Protozoans were detected at all river sites and in all the wells. Flagellated and nonflagellated protozoans, including amoebae and a single ciliate, were observed. Densities of all forms ranged from 0 to 213/ml. Replicate samples indicated high variances in the density estimates (i.e., standard deviations averaged 26% of the mean). In both the river and the aquifer, solitary cocci dominated densities of free living bacteria. Bacilli dominated the biomass in the river and many of 380 Stanford et al. the wells. However, actinomycetes, fungi, and cocci dominated the biomass in some wells. Fungi have been found in low numbers in a few aquifers (Madsen and Ghlorse, 1993). The presence of substantial populations of fungi suggests a strong connection to the surface. Infiltration of precipitation carrying soil detritus and microbiota (perhaps including algae) may be an important -process in spite of the river being by far the dominant water source for the aquifer. Moreover, most of the attached bacteria obser"ved in water samples from the aquifer were assoicated with organic detrital particles. The fine mineral sediment particles flushed from the aquifer during pumping appeared to be free of any bacterial colonization. This suggested that epilithic biofilins may be important in the aquifer. In environments characterized by low nutrient concentrations, organ- isms . attached to solid surfaces have an advantage in nutrient uptake in relation to free living forms, owing to continuous delivery. of nutrients by water movement and because of nutrient release and feedback (looping) between bacteria and primary consumers (Protozoa and very small metazo- ans) that compose biofilms. In an effort to understand the epilithic microbial community of the Kalispell Valley aquifer, artifical ceramic tiles and rocks obtained from the drilling of the wells were suspended in wells with high flow rates and high oxygen concentrations for 38 days (Fig. 2). They were then removed, and rH]thymidine incorporation into DNA was determined. Results from all incubations (n = 11) indicated that production was signifi- cantly greater on the natural substrata than on artificial tiles (p < 0.01, ANOVA). Protozoan densities were also higher on rocks than on tiles, with values ranging from 99 to 6399/CM2 on rock substrates versus 2.6 to 79/cin 2 on tiles. Although it is difficult to compare bacterial numbers per unit surface area with those per unit volume of water, it is evident that extensive growths occurred on rock substratum in comparison with densities in the interstitial *Waters (Table 1). The same pattern also appeared .. to be. true for protozoan densities: 99 to 6399/;cm 2 on the rock substratum versus 0 to 213/ml in interstitial waters. The incubation time selected was arbitrary, and we surmise that the populations were early stages in the colonization of the experimental substra- tum. Very small size classes of bacteria dominated the rock substratum, as well as goundwater and the river. But large bacteria were more common on rocks than in the interstitial waters, suggesting that biofilins provide more optimal environmental conditions. High numbers of fungi and Protozoa on the rocks also have important implications. Fungi and actinomycetes may be very active in the aerobic heterotrophic metabolism of the groundwater microbial assemblage. Even when they are present in low numbers, their contribution to total subsurface metabolic activity may be significant (Mad- sen and Ghiorse, 1993). The predominant form of nutrition for free living Protozoa is the ingestion of bacterial prey (Fenchel, 1987), and predation by Protozoa has been shown to accelerate the cycling of carbon and other nutrients (Stout, 1980; Fenchel, 1987). Z Vf• O O O ram. O N w C \c 'ON �? 6n W ON Gn W cc O 1+ a x 0 SwJ �J x K N a ? a� C\ Q1 �D npt: 1+ I+ I+ 1+ 1+ 0 0000 -� N w Gn CZ x x x x x r' O to o 0 0 0 N to M Uf 382 Stanford et al. Although the database is not as extensive, microbial dynamics appeared to be similar at the Nyack site, in that the densities and biomass of bacteria were higher and more variable in the river than in the wells and that the numbers and biomass declined with distance from the river channel. Proto- zoa were detected in wells 0.28 and 0.3 km from the river. Abundance ranged from 0 to 2.0/ml in the river and from 0 to 2.4/ml in the aqui- fer. Although flagellates were observed, the majority had no flagella and were probably cysts. No ciliates or amoebae were detected. Chlorophyll - containing algal cells were present in a well 0.28 kin from the river. We concluded that a rich, but perhaps carbon -limited, microbial biofilm is the primary base of the food web in both aquifers. The -existence of a microbial loop, which tightens the efficiency of nutrient utilization, seems highly likely and probably explains how a diverse metazoan biocenosis can exist in these aquifers. B. Invertebrate Ecology To date we have found over 80 taxa of groundwater animals (i.e., stygophiles and stygobites; Table II) in the two alluvial aquifers of the Flathead River. Most of these were members of the very speciose benthos (200 + species) that occasionally showed up in samples from wells near the river or other groundwater collections within a few meters of water flowing in the river channel (i.e., occasional hyporheos). However, 10 crustacean (Ward et al., 1994), 14 insect, and several other species (J. A. Stanford et al., unpublished data) were routinely collected in wells far from the river channel (stygophiles and stygobites; Table II). Of these, a copepod species new to science has been described (Reid et al., 1991), and perhaps as many as 10 other crustacean and 2 insect species remain to be described. The Acarina and Oligochaeta also require further taxonomic resolution. Nema- todes, segmented worms, and other unknown organisms with DAPI-stained' DNA were observed in epilithic biofilms in wells and in the river and were classified as permanent hyporheos; nematode densities ranged from 0 to 7.9/ml in wells and from 0 to 2.0/ml in river samples (Ellis et al., 1995). A unique aspect of the grouhdwater fauna of the Flathead River is the presence of amphibiric Plecoptera (Table II). At least six species, some with multiple morphs, are involved. These insects are large bodied (2-3.5 cm in -length as mature nymphs) and reside deep in the alluvia as larvae, moving to the river or spring brooks during the spring to emerge as flying adults. They mate and produce eggs, which are deposited into the river, where presumably they are entrained in the aquifer before hatching or from which the first instar larvae immediately migrate into the interstitial environment (Stanford and Gaufin, 1974; Stanford and Ward, 1988).We now think that the larvae follow thermal gradients to sites of epigean ecdysis during the emergence process, because the wells across the valley floor 14. Alluvial Aquifers of the Flathead River, Montana 383 TABLE II Types, Common Species, and Total Numbers of Species of Epigean and Hypogean Invertebrates Found to Date in the Flathead River, Based on the Typology of Cibert et al. (Chapter 1) Life history type Numeric dominants Total number of species Stygoxene Trichoptera Archtopsyche grandis Plecoptera Taenionema pacificum Suwallia pallidula Isoperla fulva Ephemeroptera Ephemerella inermis Rhithrogena spp. Baetis tricaudatus Diptera Simulium spp. Procladius sp. Stygophile Occasional hyporheos Stygophile Amphibite Stygophile Permanent hyporheos Ephemeroptera Ameletus 2 spp. Paraleptophlebia sp. Plecoptera Capnia confusa Suwallia pallidula Saveltsa coloradensis Hesperoperla pacifca Diptera Cricotopus sp. Plecoptera Isocapnia crinita L grandis L vedderensis L missouri Paraperla frontalis Kathroperla perdita Copepoda Acanthocyclops montana Diacyclops crassicaudus brachycercus D. languidoides Ostracoda Cavernocypris wardi Oligochaeta Lumbricillus sp. Lumbriculidae indet sp. 137 137 m 0 9C (continues) 384 Stanford et al. TABLE II (continued) Acarina Oribarei index sp. Halacaridae spp. Stygobire Amphipoda 8 Ubiquitous' Stygobromus 2 spp.' lsopoda Salmasellus steganotbrix Microcbaron sp. Bathynellacea batbynella 2 spp. Copepoda Parastenocaris sp. Axchiannelida Troglochaetus beranecki Stygobite ? ? Phreatobite Only Stygobromus sp. has been found in cave streams and alluvial aquifers. sequentially become depauperate of larvae when river water penetrates the aquifer in association with the spring spate, when river temperatures are up to 4°C warmer than groundwater. Other explanations exist, including the possibility that the larvae can find surface water by following ion gradients. However, large numbers climb up the well pipes to undergo ecdysis above the water table, and water temperatures are significantly higher at the top of the wells during warm weather in the spring, The -distribution of stygobites and amphibites within the aquifer is clear. Stygobites are ubiquitous; amphibites are most common near the, river (Figs. 3 and S);; although they were collected in all except three wells in the two transects (Fig. 5). By examining the distribution of the crustacean fauna using gradient and correspondence analyses, Ward et al. (1994) showed that crustacean stygobites attained maximum relative abundance near the center of the ancient floodplain, whereas permanent hyporhe'os (e.g., cope- pods and ostracods; Table H) dominated closer to the river or in wells on the modern floodplain. Areas in the center of the aquifer (i.e., about midway in the two transects across the ancient and modern floodplains) are associated with higher ion content (Fig. 3). We think that higher ion content is indicative of the longer residence time of water in the aquifer and the presence of zones of lower hydraulic conductivity and, by inference, lower porosity. Hence, faunal patterns to some extent may be attributed to variations in porosity. However, well G (Fig. S) penetrated very porous substrata based on high pump rates (>40 I/min), yet it was characterized by high ion content, fairly abundant styobites, and no amphibites. It is likely the high -ion waters 14_ Alluvial Aquifers of the Flathead River, Montana 385 are phreatic and not very interactive with the river for reasons that relate to the complexity of the interstitial milieu of this aquifer. As did Marmonier et al. -(1992) for an alluvial floodplain aquifer on the Rhone River, Ward et al. (1994) concluded that the distribution and abundance of the groundwater fauna on the floodplain scale are structured by hydrogeologic and geomor- phic processes that are not necessarily related in any meaningful way to distance from the river owing to spatial discontinuities in the lattice -like alluvium. An analysis of the relation of paleochannels, which are known to be areas of higher hydraulic conductivity, to faunal patterns (e.g., installation of wells in high -porosity sites determined by ground penetrating radar or seismic techniques) may be more revealing. The groundwater fauna of the Nyack Valley aquifer is very similar to that of the Kalispell Valley, except that some of the occasional hyporheos are different and are found in wells near the valley wall. Phreatic waters may be more limited. Spring -brook fauna also were studied to determine affinities with the groundwater biocenosis. Although permanent hyporheos (mainly copepods) were abundant in the spring brooks, stygobites were never found and amphibites were rare. The spring -brook fauna was primarily composed of stygoxenes with temperature and substratum as primary deter- minants (Case, 1994). C. Food Web Relationships The energy base of the food web of the Flathead River alluvial aquifers appears to be the speciose microbial bioftlm (bacteria and fungi). and associ- ated microconsumers (Protozoa, nematodes, and other microscopic fauna). The productivity of this microbial loop is likely limited by the availability of short -chain DOM, because other nutrients appear to increase in concen- tration relative to those in the river. Although POM is entrained in the aquifer, based on the presence of algal cells with viable chromatophores and measurable concentrations of chlorophyll, the importance of detritus relative to DOM is not clear. Although others have shown that microbial metabolism in groundwaters is controlled by entrainment of POM (Chapter 7), most of the work has been done in situations in which some or most of the organic matter is derived from pollution. The pristine Flathead systems are extremely oligotrophic, and carbon limitation relative to the supply of other nutrients (N, P, and trace metals) remains at issue. Because nitrate accumulates in aquifers and oxic conditions persist in the wells, nitrification is an important process within the groundwater microbial loop. Our work is now emphasizing the assessment of N flux and the importance of the groundwater N subsidy to biofilms in surface waters. Higher level consumers in the food web are composed of a wide variety of sizes and taxa of. stygobite and stygophile consumers. Amphibitic stone flies seem to be the top consumers in the groundwater food web, at least 386 Stanford et al. in the hyporheic waters.. However, in phreatic waters, where large stone flies are rare, Stygobromus spp. likely is the top consumer. . With R. Wissmar (University of Washington, Seattle, unpublished), we analyzed the stable isotopic enrichment of 15N to trace trophic relations of the groundwater food web in comparison with that of the river (Peterson and Fry, 1987). The isotopic enrichment values for river and aquifer biofilms were significantly higher in the aquifer, supporting the notion- of different conversion pathways for organic compounds. Nitrogen compounds incorpo- rated into the food web by biofilm uptake were different in the river and in the aquifer, which supports the inference of nitrification as an -important biotic pathway in groundwater. As a consequence, high-level consumer groups in the aquifer showed high enrichment values; Paraperla and Stygo- bromus had the highest enrichment values (+8-11 %o), supporting the notion of these genera as high or- top -end consumers in the food web. However, they may not be the top consumers in the interstitial food web, because a small (S cm) salamander, tentatively identified in the genus Ambys- toma, was collected from well D (Fig. S). In addition to the importance of nutrient conversions in groundwater food webs, transport of groundwater macrofauna to surface waters may also provide an important food source for riverine consumers. The amphibi- tic stone flies are very abundant during emergence in the river and on the shoreline, relative to most of the benthic species (Stanford, 197S). The large volume of the groundwater system apparently allows greater production of the amphibitic stone flies, and, when they move from the groundwater into the river during the emergence process, they dominate the drift and are a major food source for fish and other top consumers in the river (Perry and Perry, 1986). Amphibitic assemblages similar to those in the Flathead likely exist in many well -oxygenated interstitial aquifers fed. by river interflow in North America and elsewhere, depending on the origin and attributes of the porous milieu (Stanford and Ward, 1993). We have found the same interstitial fauna in the Methow River in Washington, and others in western North America will likely be reported in the future. D. Cave Streams We know only that the stygobite Stygobormus spp. is present in cave streams in the Flathead catchment. Several specimens were collected with a kick net at the Trail Creek site. The stream intersects the cave about 300 in from the dry portal. Presumably the stream flows into the alluvium of Trail Creek, but additional work is needed. Hence, we conclude that Stygobromus is a ubiquitous genus (Table II), but we do not know if phrea- tobites are present in the alluvial aquifers, owing to a lack of species -specific associations with habitat types. However, this is the first report of groundwa- ter fauna in cave systems in the Belt Series substrata of the Rocky Mountains. 14. Alluvial Aquifers of the Flathead River, Montana 387 VII. CONCLUSION Shallow alluvial aquifers biophysically connected to the main channel are fundamental units of the geohydraulic continuum of the Flathead River. They are not delineated by the active river floodplain; they also include adjacent low terraces composed of ancient glacial outwash floodplains and modified by recent cut and fill alluviation mediated by flooding. Paleochan- nels are evident on the ancient floodplain surface and provide clues to the complicated lattice -like structure of the alluvial aquifer. Interactions between river inflow and aquifer discharge produce biotic patterns that manifest as complex groundwater food webs and riparian plant assemblages, al- though detailed study of the latter remains to be performed. Over 80 taxi of stygophiles and stygobtes are present in the groundwater food webs. Considerable taxonomic evaluation of the groundwater fauna also remains to be performed, and the list of stygobites will likely increase. Clearly, riverine biodiversity is in large measure influenced by the groundwater bio- tope. Bioproduction may be controlled by the availability of low -molecular - weight DOM, although nitrification and other microbially mediated conver- sions increase the solute load of the water as it fluxes from the river, through the aquifer, and back into the river. These and other process —response mechanisms also require greater elucidation. However, it is clear that the interaction zone between the river and the aquifer is expansive (i.e., involving amphibitic and stygobitic 'forms; Figs. 3 and 5), and the flux of materials in this zone probably controls the biophysiology of the river and also may influence advective processes in Flathead Lake (e.g., pelagic autotrophy in response to nutrient loads from the river). These findings have profound ramifications for the management of river corridors in the Flathead and other gravel bed rivers, especially those that. remain relatively pristine. The expansive interaction zone and the, flux of water and materials between surface water and groundwater require a broader conception of the floodplain (riparian).. corridor than usually is applied in river management plans. In the Kalispell Valley human activities on the ancient floodplain are beginning to subsidize the nutrient load of the aquifer and alter faunal distribution patterns. The sensitive nature of these groundwater environments requires greater appreciation in the context of pollution control. In the Flathead and elsewhere these shallow alluvial aqui- fers are strategic sources of potable supplies. Moreover, we are beginning to recognize important interactions between wetlands created by groundwater upwelling and wildlife. For example, for the Nyack Valley, we have observed that spring brooks are used as spawning and rearing habitats for fishes and are modified by dam building activities of the beaver (Castor canadensis). A variety of ungulates and their predators rely on the riparian plant commu- nities produced by the interactions between interstitial and overland flow processes. We conclude that conservation of groundwater aquifers is im- portant not only from the standpoint of hypogean flora and fauna that are 388 Stanford et al_ not well known, but also from the standpoint of influences on terrestrial flora and fauna on scales and in time frames that are not understood at all. ACKNOWLEDGMENTS This study was supported by a National Science Foundation grant (BSR-8705269) and .by funding from the National Park Service. We give special thanks to John Dahmata and Neil and Frances Graham for access and Iogistical support on the Nyack and Kalispell Valley floodplains. Thanks also to Jim Craft and Roger Noble for technical help. REFERENCES Boulton, A. J., Stibbe, S. E., Grimm, N. B., and Fisher, S. G. (1991). Invertebrate recolonization of small patches of defaunated hyporheic sediments in a Sonoran Desertstream. Freshwater Biol. 26, 267-277. Boulton, A. J., Valetr, H. M., and Fisher, S. G. (1992). Spatial distribution and taxonomic composition of the hyporheos of several Sonoran Desert streams. Arch. Hydrobiol. 125, 37. Brerschko, G., and Klemens, W. E. (1986). Quantitative methods and aspects in the study of the interstitial fauna of running waters. Stygologia 2, 297-316. Case, G. L. (1994). Benthic and interstitial faunal patterns on a floodplain of an alluvial river. Thesis, University of Montana, Missoula. Coleman, M. J., and Hynes, H. B. N. (1970). The vertical distribution of the invertebrate fauna in the bed of a stream. Limnol. Oceanogr. 15, 31-70. DanieIopol, D. L. (1976). The distribution of the fauna in the interstitial habitats of riverine sediments of the Danube and the Piesting (Austria). Int. J. Speleol. 8, 23-51. Ellis, B. K., and Stanford, J. A. (1988). Nutrient subsidy in montane lakes: Fluvial sediments versus volcanic ash. Verb. Int. Ver. Limnol. 23, 327-340. Ellis, B. K" Stanford, J. A., and Ward, J. V. (199S). Microbial ecology of the alluvial aquifers of the Flathead River, Montana (USA). J. North Am. Benthol. Soc. (in review). FencheI, T. (1987). "Ecology of Protozoa." Science Tech. Publisher, Madison, WI. Fraley, J.J., and Shepard, B. B. (1989). Life history, ecology and population status of migratory bull trout (Salvelinus confluentus) in the Flathead Lake and River system, Montana. Northwest Sci. 63, 133-143. Freeze, R. A., and Cherry, J. A. (1979). "Groundwater." Prentice -Hall, Englewood Cliffs, NJ. Hall, C. A. S., Stanford, J. A., and Hauer, F. R. (1992). The distribution and abundance of organisms as a consequence of energy balances along multiple environmental gradients. Oikos 65, 377-390. Hauer, F. R., and Stanford, J. A. (1981). Larval specialization and phenotypic variation in Arctopsyche grandis (Trichoptera: Hydropsychidae). Ecology 62, 645-6S3. Hauer, F. R., and Stanford, J. A. (1982). Bionomics of Dicosmoecus gilvipes (Trichoptera: Limnephilidae) in a large western montane river. Am. Midl. Nat. 108, 81-87. Hauer, F. R., and Stanford, J. A. (1986). Ecology and coexistence of two species of Brachycen- trus (Trichoptera) in a Rocky Mountain river. Can. J. Zool. 64, 1469-1474. Konizeski, R. L., Brietkrietz, A., and McMurtrey, R. G. (1968). Geology and groundwater resources of the Kalispell Valley, northwestern Montana. Mont., Bur. Mines Geol., Bull. 68, 1-42. Locke, W. W. (1990). Late Pleistocene glaciers and climate of western Montana, U.S.A. Arct. Alp. Res. 22, 1-13. 14. Alluvial Aquifers of the Flathead River, Montana 389 Madsen, E. L., and Ghiorse, W. C. (1993)- Groundwater microbiology: Subsurface ecosystem processes. In "Aquatic Microbiology: An Ecological Approach" (T. E. Ford, ed.), pp. 167-213. BIackwell, Boston. Marmonier, P., Dole -Olivier, M. J., and Creuz6 des ChatelIiers, M. (1992). Spatial distribution of interstitial assemblages in the floodplain of the Rhone River. Regulated Rivers 7, 75-82. MarnelI, L. F. (1988). Status ofthewestslope cutthroat trout in Glacier National Park, Montana. Am. Fish. Soc. Symp. 4, 61-70. Noble, R. A., and Stanford, J. A. (1986). Groundwater resources and water quality of the unconfined aquifers in the Kalispell Valley, Montana. Open File Report 093-86. Flathead Lake Biological Station, The University of Montana, Polson, Montana. Palmer, M. A. (1990). Temporal and spatial dynamics of meiofauna within the hyporheic zone of Goose Creek, Virginia. J. North Am. Benthol. Soc. 9, 17-25. Pennak, R. W., and Ward, J. V. (1986). Interstitial faunal communities of the hyporheic and adjacent groundwater biotopes of a Colorado mountain stream. Arch. Hydrobiol., Suppl. 74, 356-396. Perry, S. A., and Perry, W. B. (1986). Effects of experimental flow regulation on invertebrate drift and stranding in the Flathead and Kootenai Rivers, Montana, USA. Hydrobiologia 134, 171-182. Peterson, B. J., and Fry, B. (1987). Stable isotopes in ecosystems analysis. Annu. Rev. Ecol. Syst. 18, 293-320. Pugsley, C. W., and Hynes, H. B. N. (1986). Three-dimensional distribution of winter stonefly nymphs, Allocapnia pygmaea, within the substrate of a southern Ontario river. Can. J. Fish. Aquat. Sci. 43, 1812-1817. Reid, J. W., Reed, E. B., Ward, J. V., Voelz, N. J., and Stanford, J. A. (1991). Diacyclops languidoides (Lilljeborg,1901) s.l. and Acanthocyclops montana, new species (Copepoda, Cyclopoida), from groundwater in Montana, USA. Hydrobiologia 218, 133-149. Ross, C. P. (1963). The belt series in Montana. Geol. Surv. Prof. Pap. (U.S.) 346. Spencer, C. N., McClelland, B. R., and Stanford, J. A. (1991). Shrimp stocking, salmon collapse, and eagle displacement: Cascading interactions in the food web of a Iarge aquatic ecosystem. BioScience 41, 14-21. Stanford, J.. A. (1975). Ecological studies of PIecoptera in the Upper Flathead and Tobacco Rivers, Montana. Ph.D. Dissertation, University of Utah, Salt Lake City. Stanford, J. A., and Gaufin, A. R. (1974). Hyporheic communities of two Montana Rivers. Science 185, 700-702. Stanford, J. A., and Hauer, F. R. (1992). Mitigating the impacts of stream and lake regulation in the Flathead River Catchment, Montana, USA: An ecosystem perspective. Aquat. Conserv. 2, 35-63. Stanford, J. A., and Prescott, G. W. (1988). Limnological features of a remote alpine lake in Montana, including a new species of Cladophora (Chlorophyta). J. North Am. Benthol. Soc. 7(2), 140-151. Stanford, J. A., and Ward, J. V. (1983). Insect species diversity as a function of environmental variability and disturbance in stream systems. In "Stream Ecology: Application and Testing of General Ecological Theory" (J. R. Barnes and G. W. Minshall, eds.), pp. 265-278. Plenum, New* York. Stanford, J. A., and Ward, J. V. (1988). The hyporheic habitat of river ecosystems. Nature (London) 335, 64-66. Stanford, J. A., and Ward, J. V. (1992). Management of aquatic resources in large catchments: Recognizing interactions between ecosystem connectivity and environmental disturbance. In "Watershed Management" (R. J. Naiman, ed.), pp. 91-124. Springer-Verlag, New York. Stanford, J. A., and Ward, J. V. (1993). An ecosystem perspective of alluvial rivers: Connectivity and the hyporheic corridor. J. North Am. Benthol. Soc. 12, 48-60. 390 Stanford et al. Stanford, J. A., Ellis, B. K., Chess, D. W., Craft, J. A., and Poole, G. C. (1992). "Monitoring Water Quality in Flathead Lake, Montana. 1992 Progress Report." Open File Rep. 128- 92. Flathead Lake Biological Station, University of Montana, Poison. Stout, J. D. (1980). The role of protozoa in nutrient cycling and energy flow. Adv. Microb. Ecol. 4, 1-49. Strahler, A. N. (1965). "Introduction to Physical Geography." Wiley, New York. Ward, J. V. (1989). The four-dimensional nature of lodc ecosystems. J. North Am. Benthol. Soc. 8, 2-8. Ward, J. V., and Voeiz, N J. (1990). Gradient analysis of interstitial meiofauna along a longitudinal stream profile. Stygologia 5, 93-99. Ward, J. V., Stanford, J. A., and Voelz, N. J. (1994). Spatial distribution patterns of Crustacea in the floodplain aquifer of an alluvial river. Hydrobiologia (in press). Williams, D. D., and Hynes, H. B. N. (1974). The occurrence of benthos deep in the substratum of a stream. Freshwater Biol. 4, 233-256. The hyporheic habitat of river ecosystems J. A. Stanford* & J. V. Wards' * Flathead Lake Biological Station, University of Montana, Poison, Montana 59860, USA t Department of Biology, Colorado State University, Fort Collins, Colorado 80523, USA Contemporary river ecology is based primarily on biogeochemical studies of the river channel and interactions with shoreline vegeta- tion, even though most rivers have extensive floodplain aquifers that are hydraulically connected to the channel. The hyporheic zone, the interstitial habitat penetrated by riverine animals, is characterized as being spatially limited to no more than a few metres, in most cases centimetres, away from the river channel'. However, riverine invertebrates were collected in hundreds per sample within a grid of shallow (10 m) wells located on the flood - plain up to 2 km from the channel of the Flathead River, Montana, USA. Preliminary mass transport calculations indicate that nutrients . discharged from the hyporheic zone may be crucial to biotic productivity in the river channel. The strength and spatial magnitude of these interactions demonstrate an unexplored dimension in the ecology of gravel -bed rivers. The Flathead River, a major tributary of the Clark Fork of the Columbia, drains a catchment area of 22,241 kmz (average flow=340 m3 s-t) in northwestern Montana and southeastern British Columbia The morphology and surficial geology of the Kalispell Valley (Fig. 1) were largely determined by Pleistocene glaciation. Fine-grained lacustrine sediments and fault traces delineate the northern geological boundary of the palaeodeltas of proglacial Flathead Lake. The river is heavily braided in the area delineating this boundary or ectone (E in Fig. 1). An alluvium of cobbles, gravel and sand covers an. impermeable clay formation of Tertiary age upstream of the heavily -braided reach. Groundwaters in this alluvium interact hydraulically with the Flathead River and a tributary system, the Whitefish .River (Fig. 1). Wells near the river channel produced hydrographs that closely tracked daily and seasonal changes in river flow (Fig. 2). Even wells in the centre of the valley, 2 km or more from the rivers, were influenced by river -flow patterns (Fig. 2). The aquifer lies on a 2° slope, bounded on both sides by the river channels; water flows through the system in a north to south direction at an average rate of 0.7 m3 s-t. The primary area of aquifer discharge is near the confluence of the two rivers, where faulting and finer sediments on the proglacial Flathead Lake palaeodeltas begin to limit groundwater flow rates (E in Fig. 1). Water also moves to and from the channels. During spring freshet, water moves into the aquifer from the river channel until a hydrological equilibrium occurs or river flow begins to decrease. As flows decrease, aquifer water is discharged from the hyporheic zone into the river. Spatial plots of specific conductance data (Fig. 1) also demon- strated lateral dilution. The 35 mS per metre isopleth (Fig. 1) appeared to delineate true interstitial groundwaters that are inhabited by subterranean fauna1114 and are less interactive with surface waters. This was confirmed by the distribution and abundance pat- terns of the interstitial fauna within the aquifer. Biota collected from wells in the hyporheic zone (as defined by the 35 mS per metre isopleth, Fig. 1), consisted almost exclusively of stonefly (Plecoptera) larvae and other typically riverine taxa (Table 1). In contrast, subterranean forms were abundant in most of the wells located on the high concentration side of the 35 mS per metre isopleth. There was some overlap between riverine and subterreanean taxa in several wells located near the 35 isopleth, but either subterranean amphipods (Stygobromus spp.) or Fig.l Map ofhyporheichabitat (stippled areas) within the Kalis- peIl Valley of the Flathead River. Isopleths of specific conductance (mS per metre) were determined from 271 wells. Closed circles locate only those wells from which biological samples were obtained. Inset map shows the three major tributaries upstream of the study site. NF, North Fork; MF, Middle Fork; SF, South Fork; the Whitefish River (WF) and Flathead Lake (FL). The area designated E is the interface between porous and less transmissive substrata (see text). stonefly larvae were very abundant (hundreds per sample), never b oth. Levels of dissolved oxygen within the wells were always greater than 50% saturation. Temperatures remained at 7-9'C throughout the year in all wells, except those Iocated near the river channel where greater exchange of water elevated or reduced temperatures in a seasonal way. It may be that the biota navigate through the interstitial groundwaters by following ther- mal cues when near the river channel, or ion concentration gradients when far from the river. Table 1 Relative abundance of biota Hyporheic zone Paraperla frontalir (Iasecta: Plecoptera) 100 Isocapnia 4 spp. (Insects: Plecoptera) 100 Chironomidae (Insecta: Diptera) 10 Capniidae (Lnsecta: Plecoptera) 20 Early instar zoobenthos 20 Incidental zoobenthos 20 Phreatic zone* Stygobromus 2 spp. (Crustacea: Amphipoda) 100 Cyclopoid copepods (Crustacea: Eucopepoda) 100 Asellus sp. (Crustacea: Isopoda) 10 Bathynellidae (Crustacea: Bathynellacea) 10 Biota were collected from 17 wells located within the alluvial aquifer adjacent to the Flathead River and on either side of the 35 mS per metre specific conductance isopleth, which differentiated phreatic (>35 mS per metre) and hyporheic (<35 mS per metre) habitats. Data are the percentage of total wells located in either the phreatic or hyporheic zones in which a particular taxon was present in numbers greater than 10 per well on every sampling date (n = 8) in 1984-6. * All taxa listed under phreatic zone are crustaceans of subterranean facies. Bathyneifids and Stygobromus are exclusively subterranean groups. Asellus (Caecidotea) sp. is a blind, depigmented isopod of attenuated morphology. Although most Cyclopoid copepods are plank- tonic or littoral, several species are widespread in interstitial bio- topes10-14 f 3 coo E Flath®ad RN., Yl .. n Fig. 2 Hydrographs from wells X and Y (see Fig. 1) compared with the Flathead River in 1984-85. Nonseasonal fluctuations were caused by discharges from a large hydroelectric dam on an upstream tributary (5F in' Fig. 1). Well X is 2.4 km and Well Y 50 m from the Flathead River channel. On the basis of the relative distribution of the biota and the 35 mS per metre isopleth, we were able to measure the hyporheic habitat within the 10 km (approximately) study area (Fig. 1). The hyporheic zone is on average about 3 km wide with an average depth of about 10 in; whereas, the Flathead River chan- nel (median flow) is about 50 in wide with most zoobenthos in the upper 025 in of channel substrata. There is therefore about 0.3 km3 of hyporheic habitat, compared to about 125,000 m3 of channel habitat. Standing crop biomass in the hyporheic zone could easily exceed benthic biomass in this river. Stouieflies are apparently the top consumers within an as yet undescribed and probably detritus -based food chain. Others415 have suggested that the hyporheic zone is a functional sink for fine _(<500 µm) particulate organic detritus from the channel. Particulates may also be recruited by infiltration of precipitation through the soil profile. Decomposition of organic detritus, coupled with ionic leaching, desorption and other biophysical processes, Iike nitrification, may sequester labile (bioavailablel6) nutrients within groundwaters. Nutrient concentrations were significantly higher in the hyporheic (for example well X, Table 2) than in the river. Preliminary calculations of mass transport Table 2 Average concentrations and standard deviations of nutrients at well sites Soluble Soluble Nitrite+ Soluble Sampling reactive total nitrate organic site phosphorus phosphorus nitrogen carbon Well X 1.5 0.5 4.9=�1.0 916.0i11.0 1,810=138 Well Y 1.2±0.5 2.2±0.8 145.0=!:5.0 1,810�_112 River BDL* 2.3 =�: 0.4 38.0:±-1.0 1,730=t= 162 Concentrations are in µg per litre. Well sites were located 2.51cm (well X) and 50 m (well Y) from the Flathead River channel compared to values in the river. These data summarize six bimonthly sampling dates during 1984-85. * Below detection limit =1.0 µg I-t. of bioavailable phosphorus16 and nitrate -nitrogen indicate that baseflow loads in the Flathead River increased by 25% and 12% respectively. During extended baseflow periods, certain near -shore areas of the relatively unproductive Flathead River channel are matted with algae. We believe that these areas of enhanced phytobenthos production occur in direct response to the inflow of nutrient -rich hyporheic waters. Gravel -bed rivers are common worldwide17. The practice of screening groundwater monitoring wells accounts for the absence of faunal records in samples collected by hydrogeolo- gists. The spatial extent and strength of hyporheic-channel inter- actions undoubtedly vary from river to river. Nonetheless, hyporheic-channel interactions as reported here are probably common features of gravel -bed river segments, and should'be included in holistic constructs"' of riverine ecosystems. We thank R_ Noble (Montana Bureau of Mines and Geology) for his help. This research was supported by a National Science Foundation (USA) grant and the Montana Department of Natural Resources and Conservation. Received 19 January; accepted 11 July 1988. 1, Stanford, J. A. & Gattfin, A. R. Science 185, 700-702 (1974). 2. Sehwoerbel, J. Arch HydrobicL SuppL 33, 1-62 (1967). 3. Husmann, S. Smithson. Contr. ZoaL 76, 161-169 (1971). 4. Williams, D. D. & Hynes, H. B. N. Freshwater BioL 4, 233-256 (1974). S. Danielopol, D. L Int. I. SpeleoL & 23-51 (1976). 6. Danielopol, D. L Inc Rea ges Hydrabiol 65, 777-791 (1980). 7. Bretsrhko, G. Verb. int. Verein. Limnol. = 2049-2052 (1985). S. Williams, D. D. in The Ecology of Aquatic Insects (eds Rah, V. H. do Rosenberg, D. M.) 430-455 (Praeger, New York, 1994). 9. Pennak, R. W. & Ward, J. V. Arch HydrobiaL SuppL: 74, 356-396 (1986). I0. Holsinger, J. R. Am. SeierrL 74, 146-153 (1988). 11. Ward, J. V. Trans. Am. microsc Sac 96, 452-466 (1977). 12. Holsinger, J. R. Crustaceana SuppL 4, 244-281 (1977). 13, Ward, J. V; & Holsinger, 1. R. Int. I. SpeleoL 11, 63-70 (1981). 14. Pennak, R. W. & Ward, J. V. Trans Am microse. Sac 104, 209-222 (1985). 15. Hynm,•H. B. N. Hvdrobiologia 100, 93-99 (1983). 16. Ellis, B. K. & Stanford, J. A. Bioawilability of Phosphorus Fractions in .Flathead Lake and its Tributary Waters (Open File Rpt. 91, Flathead Lake Biology Station, University of Montana, Poison, 1986). 17. Sediment Transport in Gravel Bed Riaers (eds Thorne, C. R., Bathhurst, S. C. & Hey, R. D.) (Wiley, Chichester, 1987). ' 18. Vannote, R. A., Minshall, G. W., Cummins, K. W., Sedell, J. R. & Cushing, C. E. Can. I. Fish. aquas ScL 37, 130-137 (1980). Printed in Great Britain by Turnergraphic Limited. Basingstoke Hampshire