[Next Section] [Previous Section] [Table of Contents]

Chapter 2: Hydrogeology

Ground water in the Klamath Basin is considered one of southern Oregon’s most valuable resources . Demands for this resource have recently increased for uses other than irrigation and domestic use. In 1988, the U.S. Fish and Wildlife Service listed two endemic fish species, the Lost River Sucker (Deltistes luxatus) and the Shortnose Sucker (Chasmistes brevirostris) as endangered. To protect these species monthly minimum levels were designated for Upper Klamath Lake. To maintain these levels, upstream diversions of surface water for agricultural use must now be limited. Historical lake level records indicate that these required lake levels would not have been met 45 of the past 73 years without reductions in upstream diversions (Adams and Seong, 1998). Ground water resources will be called upon to compensate for decreased diversion and possibly to augment surface water levels. These issues have prompted a detailed study of the hydrogeology by the U.S. Geological Survey to develop a quantitative conceptual model and a numerical hydrologic model to test proposed use of ground water resources.

Surface and ground water enter Upper Klamath Lake from the Williamson River Basin, Sprague River Basin, and Upper Klamath Lake Basin. The Williamson River basin is bordered to the east and west by large volcanic peaks that are separated by a broad lowland. Klamath Marsh occupies 600 km² of this lowland and receives the majority of its inflow from ground water that originates as precipitation on these volcanic uplands. The level of Klamath Marsh is controlled by the amount of ground water inflow, evapotranspiration, precipitation, surface water diversion, and consumptive uses of ground water. The level of the marsh is also limited by a topographic barrier. This barrier, the Kirk Sill (Figure 4), is a Quaternary basalt flow at the southern edge of the marsh. The lowland that contains Klamath Marsh is known as the Chemult Plateau and its southern boundary is located in the Soloman Butte quadrangle. The Williamson River, which originates from springs on the eastern edge of the basin (Figure 3), flows through the marsh then over Kirk Sill and through a 100-m deep canyon (Williamson River canyon). South of the canyon, the river is joined by Spring Creek, which also originates from springs. The river is then joined by the Sprague River and eventually empties into Upper Klamath Lake. The majority of the area of the Williamson River basin is upstream of Spring Creek, but the surface water contribution from this upper portion of the basin is only 17% of the total flow into Upper Klamath Lake . Approximately 28% of the total flow of the Williamson River into Upper Klamath Lake is from Spring Creek and during low flow months Spring Creek contributes almost the entire flow of the river . Spring Creek has a drainage basin of only 23 km².

This chapter synthesizes data gathered from the mapping of the Wocus Bay and Soloman Butte quadrangles, field investigations, water well logs, and climate records into a hydrogeologic reconnaissance of the southern portion of Klamath Marsh and the Soloman Butte quadrangle. Included in this chapter is a detailed study of the discharge history of the Williamson River at the Kirk Sill and the factors influencing this discharge.

Figure 4: Locations of discharge measurements within the Soloman Butte quadrangle. Coordinates are UTM Zone 10, contour interval 6 m.

The 3,781 km² drainage area of the Williamson River basin has been the subject of several reconnaissance studies. These studies have been limited to assessing the ground water potential of the area and a well-constrained water budget for this area has yet to be calculated.

Illian (1970) could not precisely identify the ground water boundaries of the basin within the larger Klamath basin because of the paucity of well logs in the mountainous areas. The basin was delineated from topographic divides that are thought to form ground water boundaries . Illian (1970) proposed a three-part flow system for the Klamath Basin that is separated into local, intermediate, and regional circulation. Local flow systems are present where discharge occurs close to the recharge source, flow paths are shallow and the residence times are short . The regional flow system has long deep flow paths with long residence times, the waters are enriched in solutes, and have geothermally elevated temperatures, which can reach 93° C in the Klamath Basin . This system receives recharge from the highest elevations in the basin (Cascades, Yamsay Mountain) and discharges in the lowest part of the ground water basin (Lower Klamath Lake) . The intermediate flow system has characteristics of both regional and local flow systems and Illian (1970) describes the flow from the Cascades to Klamath Marsh as intermediate.

Illian (1970) estimated that 9 ´ 10⁵ ac-ft yr^-1 is recharged to the ground water system and of this 5 ´ 10⁵ ac-ft yr^-1 resurfaces within the basin where it is lost to evapotranspiration, consumptive uses, and baseflow contribution to the Williamson River. The remaining recharge moves south out of the basin into the Upper Klamath Lake and Lost River sub-basins .

Leonard and Harris (1974) built upon previous work and defined the general geologic units of the Klamath Basin, described their water-bearing properties, and described the general movement and occurrence of ground water as well as the quality of this water. The path of water movement described in this report begins as precipitation and its subsequent infiltration on the slopes of the Cascades to the west and the volcanic centers surrounding Klamath Marsh, and discharge in the lowlands such as Beaver Marsh and Klamath Marsh (Leonard and Harris, 1974). Ground water in the Klamath Marsh area then moves south under a gradient of 0.2 m km^-1 (Leonard and Harris, 1974).

Discharge of the Williamson River at the Kirk Sill (Figure 4) was evaluated through historic stream gauge data and field measurements of discharge at the Kirk Sill and south of the Williamson River canyon at the bridge where Forest Service road 9730 crosses the Williamson River (bridge 9730).

Historic discharge values by stream gauge number are available for water years 1954-1995 and 1999 (U.S. Geological Survey, 1999). These data are daily flow rates recorded in ft³ s^-1 and were converted to m³ s^-1 for this study. To complement these data, climate data are also examined. Temperature, rainfall, and snow water equivalent data for the water years from 1982-1999 are available for the USDA SNOWTEL site Chemult Alternate (USDA, 1999). Precipitation data for the Cascade Range at SNOWTEL site Diamond Lake (USDA, 1999) are also examined. This site lies just to the west of the basin boundary along the Cascade Range (Figure 1), but the precipitation values are thought to be representative for the Cascade Range in the Williamson River basin. The discharges and precipitation values from 1982-1995 and 1999 at the Kirk Sill were plotted separately for each water year to observe seasonal variations (Appendix B). The water year is defined as October 1^st through September 30^th of the following year.

Flow measurements were made with a flow meter on October 10, 1998 and on April 4, 1999. With the flow meter set to average velocity in m s^-1, eight readings were taken at approximately half the channel depth and at a distance interval that was relative to the channel width (Appendix A). These data were then integrated with the following equation to determine the total river discharge at each location:

where Q is the total discharge and q_i is the discharge at each point of measure, which is the product of the flow velocity from the flow meter, the channel depth, and the channel width (distance between measurements). The discharge values and channel dimensions are reported in Appendix A.

Discharges for selected springs in the Wocus Bay and Solomon Butte quadrangles (Figure 5) were measured in June, 1998 and again in October, 1998. Discharges for Recovery, Wocus, and "Bat springs" were measured with the method outlined above for the calculation of the river discharge. Dice Crane spring flows into a reservoir that is drained by a pipe. The time taken to fill a one-liter container by this flow was recorded 10 times and averaged. These data are reported in Table 1.

Three types of springs were identified in the study area: contact, depression, and fault. Contact springs occur where a permeable unit overlies a unit with a hydraulic conductivity that is low enough to prevent transmission of all of the water that is moving through the upper unit . These springs occur along lava flow margins (Bat Spring) and at the interface between the pyroclastic-fall deposits and underlying bedrock units (Cabin Spring). Depression springs occur in topographic lows where the water table intersects the surface (Bryan, 1919; Fetter, 1994). These springs are commonly found within depressions that have been infilled with pyroclastic-fall deposits. It is thought that drainages existed within these depressions prior to the eruptions of Mount Mazama. The presence of phreatophytes in these paleodrainages suggests that water continues to move through these valleys, but not on the surface. The high permeability of the pyroclastic fall and the relatively low permeability of the underlying buried soil promote the lateral down slope movement of water through the pyroclastic-fall deposits. This hypothesis was supported where exploratory augering in several drainages revealed water-saturated pyroclastic-fall deposits near the surface. All of the depression springs in the Wocus Bay quadrangle are ephemeral and were dry or had significantly reduced discharges by mid-summer (Table 1).

Figure 5: Location of springs in the Wocus Bay quadrangle. Coordinates are UTM Zone 10, contour interval 6 m.

Fault springs occur where interflow zones are offset or where fractures and porous zones are clogged with fault gauge thus creating a barrier to flow and forcing ground water to the surface. Dice Crane Springs is the only recognized fault spring in the Wocus Bay quadrangle with a measurable discharge. It occurs along a distinct northwest striking lineament and the topography upstream of the spring is insufficient to produce a depression type spring (Figure 5). The discharge of this spring only slightly decreased throughout the summer (Table 1). The relatively steady flow throughout the summer and the temperature of the water (7° C (June 98) and 8.5° C (October 1998)) suggest that this stream is being fed by intermediate flow described by Illian (1970).

Table 1: Locaton (UTM Zone 10), discharge, and spring type for selected springs in the Wocus Bay and Soloman Butte quadrangles.

Water samples collected in October, 1999 from Kirk Sill, Bridge 9730, Cabin Spring, and Dice Crane Spring were analyzed for deuterium (²H) by the Department of Earth Science at Dartmouth College. These samples were collected in 100-ml containers and completely filled to evacuate air from the container. The samples were then analyzed using a stable isotope mass spectrometer. The results are reported in the usual d notation as per mil deviations from SMOW (standard mean ocean water) and presented in Table 2. Analytical precision for these values is estimated at ± 5 ⁰/_00.

Table 2: dD values for water samples from the Wocus Bay and Soloman Butte quadrangles.

Water-well logs for Klamath County were obtained from the Oregon Water Resources Department (158 12^th Street NE, Salem, OR, 97310). These logs provide limited information on the lithologies encountered during the drilling of water wells for domestic, stock, and irrigation uses. Limited information exists on the few wells drilled in the Wocus Bay quadrangle so wells drilled north of this area in the Military Crossing and Wildhorse Ridge quadrangles were included in this study. Well logs for the Soloman Butte quadrangle are plentiful, but are concentrated in the southwest. Representative well logs from these areas have been summarized in Appendix C.

The information available from geologic mapping, water-well logs, and work by Leonard and Harris (1974) were used to define the aquifers in the southern Klamath Marsh area (Table 3) and in the Soloman Butte quadrangle (Table 4).

Analysis of historic discharge data provide information on basin characteristics and climate. A plot of the annual mean discharge of the Williamson River from 1955-1995 displays a high variability in discharge (Figure 6). A statistical measurement of variation is the coefficient of variation, which is defined simply as the ratio of the standard deviation of a variable, in this case annual mean discharge, to its mean. A value of zero describes no variation and values greater than one describe extreme variation. The value for the Williamson River is 0.59. When compared to reported values for North America by Riggs and Harvey (1990) this value is similar to those from the desert southwest of the United States. The variability of flow is related to the variability of precipitation, which is greatest in semiarid and arid regions (Riggs and Harvey, 1990). High variability is usually associated with low precipitation, but in areas with low precipitation and large ground water bodies or surface water storage, flow variability is reduced (Riggs and Harvey, 1990). High infiltration rates and large aquifer storage in a basin result in runoff with little variability (Riggs and Harvey, 1990). The high infiltration rate of this basin and the high variability of the Williamson River discharge suggest low storage capability of the aquifers.

Table 3: Hydrologic units of the southern Klamath Marsh and their water-bearing properties.

Table 4: Hydrologic units of the Soloman Butte quadrangle and their water-bearing properties.

The annual hydrographs (1982-1995, and 1999) of discharge plotted separately with precipitation at Chemult (Appendix B) display the dependence of discharge on precipitation and the storage capabilities of aquifers in southern Klamath Marsh. On average, there is a large annual precipitation event usually as snowfall, between December and January. This does not produce an immediate response in discharge; there is a lag of approximately one-month. The usual crest in discharge is in either March or April. This lag between precipitation and discharge is attributed to snowmelt. The input from snowmelt is gradual due to variations in topography and slope aspect. The local snowpack is usually completely melted by mid-April to early May, while snow along the Cascade crest to the west is often present throughout the year. The snowpack provides temporary storage that sustains the system through periods of little or no precipitation in late spring and early summer. Variations in snow pack and the rate at which the snow pack melts have direct affects on the discharge of the river. Discharge in 1993 (Figure 7) is above average for the values observed from 1990-1995. Precipitation is also high for this year relative to other years, but the high discharge is partly the product of a large snow pack. The snow-water equivalent for this year was 16 cm above the average of 33 cm. The precipitation for 1995 was 63.5 cm, which is the

Figure 6: Annual mean discharge from 1955-1995 for the Williamson River with the three year moving average. Data are from U.S. Geological Survey stream gauge 11493500 (U.S. Geological Survey, 1999).

third largest in the fourteen-year span, but the total discharge for 1995 was the fourth lowest (Figure 8). This low discharge is partly the result of a snow-water equivalent of only 34 cm. The correlation coefficient between annual discharges and annual precipitations at Chemult is only 0.56 while the correlation between discharge and snow-water equivalent is 0.67. Another factor affecting discharge is the rate at which the snow melts. In 1993, it took 28 days for the snowpack at Chemult to decrease from 25 cm to 0 cm, while in 1995 it took 60 days for the same amount of snowmelt. This suggests that the faster the snow melts the more likely surface runoff will be achieved in the permeable pyroclastic deposits, which would lead to an increase in river discharge.

Figure 7: Discharge of the Williamson River at the Kirk Sill (U.S. Geological Survey, 1999) with precipitation and cumulative snow water (USDA, 1999) for 1993.

Figure 8: Discharge of the Williamson River at the Kirk Sill (U.S. Geological Survey, 1999) with cumulative snow water and precipitation (USDA, 1999) for 1995.

The low discharge in 1995 could also be the result of the preceding year’s low precipitation (Figure 9). The precipitation value in 1994 was the lowest in the data set for both Chemult (34 cm) and Diamond Lake (68 cm). This indicates a dependence of discharge on preceding precipitation values.

Figure 9: Annual mean discharge for the Williamson River at Kirk Sill (U.S. Geological Survey stream gauge 11493500) and annual precipitation from SNOWTEL sites Diamond Lake and Chemult Alternate.

The plot of annual mean discharge with precipitation from 1982-1995 (Figure 9) displays the response of the aquifer to low precipitation years. Precipitation was below the annual average for Chemult and Diamond Lake in 1981, 1985, 1987, 1988, 1989 (Diamond Lake only), 1990, 1991, 1992, and 1994. Discharge was below the average annual mean (5 m³s^-1) in 1981, and from 1988-1995. The low precipitation lead to less recharge and to a depletion of water that was in storage in the southern Klamath Marsh aquifers. Less water in the aquifers resulted in decreased contribution to the Williamson River. Precipitation was above average for several years that the discharge was below average, but sufficient recharge to the aquifer had not taken place to increase the discharge of the Williamson River above Kirk Sill. The hydrographs from the years prior to 1990 are sustained throughout most of the season indicating sufficient ground water contribution to the river for discharge to occur at Kirk Sill (Appendix B). The depletion of storage was not immediate, it required three nonconsecutive years of precipitation below the annual average until the discharge of the river at Kirk Sill was below detectable limits for portions of the water year. Data are not available to determine how many years of average and above average precipitation are required for the replenishment of storage to the point that flow over the Kirk Sill was sustained throughout the water year. Precipitation was however above average in 1993, and 1995-1999. The discharge data available for a portion of 1999 and field observation of the discharge in August and October, 1999 indicate flow over the Kirk Sill for the entire water year. Unfortunately, the stream gauge was not operational from 1995-1998 so the time required for the aquifer to recover is unknown.

Discharge measurements made at the Kirk Sill (0.013 m³s^-1) and downstream at Bridge 9730 (0.87 m³s^-1) on 10 October, 1998 indicate that the discharge increases 66 times between the two points (Appendix A). There are no tributaries to the river along this stretch so the increase is the result of ground water contribution that is occurring between the two points of measure. This indicates that the upper Williamson River was disconnected from the water table at and above the Kirk Sill, but then intersects the ground water table at a lower elevation in the canyon. The measurements taken on 04 April, 1999 show only a 19% increase between Kirk Sill (21.98 m³s^-1) and Bridge 9730 (26.65 m³s^-1). At this time of year, the percentage of ground water contribution to the discharge is less because of the high volume of water flowing over the Kirk Sill. The actual volumetric contribution from ground water was three times larger than the October, 1998 value. The high discharge at the Kirk Sill is attributed to contribution from snowmelt and ground water discharging in Klamath Marsh.

The isotopic data from the Williamson River and springs in the Wocus Bay quadrangle support the information gathered from the October, 1998 discharge data. The Williamson River is enriched with deuterium at the end of summer as it leaves Klamath Marsh (Kirk Sill sample, Table 2). This enrichment occurs because the river and marsh are subject to large amounts of evaporation with little precipitation during the summer. These factors lead to high concentrations of deuterium because the ¹H is preferentially lost to evaporation and not replenished by precipitation. The sample from bridge 9730 has a value that is very similar to the ground water discharge at Cabin and Dice Crane Springs (Table 2). This indicates that when the sample was collected the Williamson River was intersecting the ground water table between the Kirk Sill and bridge 9730 and the discharge is ground water dominated.

Wells on the west side of Klamath Marsh have high yields and often have artesian flow. Well 32S/8E-17ab (Appendix C) is drilled into a basalt aquifer at 100 m depth and has a yield of 252 L s^-1 (liters per second) (4000 gpm (gallons per minute)) with 24 m (79 ft) of drawdown. This basalt aquifer is the principle aquifer defined by Leonard and Harris (1974) for the Williamson River Basin and supplies most irrigation wells for Klamath Marsh. Recent geologic mapping in the Klamath Marsh area (Conaway and Cummings, in preparation) has shown that basalt, basaltic andesite, and andesite units are not laterally extensive. Therefore, the basalt aquifer of Leonard and Harris (1974) is likely comprised of several Pliocene and older basalt units throughout the basin. Lower yield wells around Klamath Marsh remove water from unconfined sands and gravels from the ancestral Williamson River channels. These deposits are over 100 m thick in well 32S/8E-30ab (Appendix C). Low yield wells drilled in areas with thick pyroclastic deposits are primarily for stock use because of their susceptibility to short term fluctuations from inflow. The primary confining layers in the eastern portion of southern Klamath Marsh are poorly welded Tertiary tuffs (Tt of Sherrod and Pickthorn, 1992, and Conaway and Cummings, in preparation) and cemented continental sediment. The confining layers in the western portion of southern Klamath Marsh are assorted basalt, basaltic andesite, and andesite with little jointing or vesiculation.

Wells in the southwest corner of the Soloman Butte quadrangle and the southeast corner of the Fort Klamath quadrangle are drilled into sands, diatomite, and clay and some have artesian yields. These deposits are over 200 m thick (well no. 34/7-4G). The well at Spring Creek Ranch Motel (well No. 34s/7e-4a) penetrates this aquifer of continental sediments to a depth of 79 m and has an artesian yield of 12 L s^-1 (200 gpm). These high artesian flow rates are the result of a ground water mound that occurs at the base of the Chemult Plateau and the confining layers of clays within the older continental sediments, and interfingering hydrovolcanics and overlying younger continental sediments. This ground water mound is topographically induced and is likely supplied by the high recharge from the east slopes of the Cascades. Several low yield domestic wells (33s/7e-35dc, 33/7-35) are located at the mouth of the Williamson River Canyon and are drawing water from ancestral Williamson River deposits.

The movement of ground water in the Williamson River basin is controlled by the areas of recharge and the geometry of the units through which the water is moving. Faults in this basin are thought to exert only local control over the movement and occurrence of ground water . The area of greatest recharge is along the slopes of the Cascades to the west of Klamath Marsh with lesser amounts of recharge occurring on the volcanic centers bordering the marsh to the east and south. The high recharge along the slopes of the Cascades results from a combination of heavy precipitation and high infiltration through the pyroclastic deposits. Ground water then enters volcanic and sediment units and moves towards Klamath Marsh under a topographic gradient.

Two conceptual models for the movement of ground water in the Williamson River basin are presented. The preliminary analysis of climate data, water-well logs, historic hydrographs, measured discharges, isotope data, and data from geologic mapping must be considered in any conceptual models of ground water flow in the Williamson River basin.

1. Diamond Lake in the Cascade Range, receives approximately twice as much average annual precipitation as Chemult Alternate (128 cm verses 68 cm). Figure 9 illustrates that there is little difference between the annual precipitation patterns at the two sites.
2. Despite the high precipitation values on the western edge of the basin, the Williamson River has periods of no flow at the Kirk Sill, and the variance in its discharge is similar to rivers in arid regions of the southwestern United States. Variations in discharge at the Kirk Sill illustrated by the hydrographs, high infiltration rates in the basin, and precipitation data suggest low storage capabilities in the aquifers.
3. High yield wells and large discharge springs are located on the west side of Klamath Marsh.
4. During periods of low flow, the Williamson River is ground water dominated and has an isotopic composition that is similar to perennial springs in the Wocus Bay quadrangle.
5. Spring Creek, which is spring fed, contributes approximately 28% of the total flow to the Williamson River into Upper Klamath Lake and contributes almost the entire discharge during low flow months.
6. The high infiltration rates of the pyroclastic deposits prohibit any surface water contribution to the river from the slopes to the west of the marsh. Therefore, any contribution to discharge of the Williamson River above Kirk Sill from the west must be made by ground water.

The flow paths that originate in the Cascade Range and that are being supported by the high precipitation do not surface in Klamath Marsh to contribute to the river discharge (Figure 10). These flow paths occur at depths (intermediate flow path of Illian, 1970) that do not extend up into the unconfined aquifers of the marsh. The Pliocene basalt aquifer (Table 3), which supplies the high-yield agricultural wells on the west side of the lower marsh, could be recharged by these intermediate flow paths. The deuterium data and measured discharge on the river suggest contribution from a ground water source. The isotopic composition of these waters is similar to those of the perennial springs to the east of the river. These springs occur along faults that may be forming barriers to the intermediate flow paths. Another possibility is that the perennial springs are fed from the volcanic centers surrounding the marsh. Isotope work on waters from the Cascade Range would help to identify the source of the ground water feeding the springs and the Williamson River in the canyon.

Ground water flow from the Cascade Range encounters a structural barrier before it reaches Klamath Marsh and the river is supported only by inflow from the volcanic centers and lowlands surrounding the marsh (Figure 10). The springs and artesian wells on the western side of the marsh could indicate ground water being forced to the surface by such a structure. More significant are the voluminous cold springs that feed Spring Creek and the high-yield artesian wells near the headwaters of Spring Creek. They are a likely outlet of the ground water that originates in the Cascade Range. No large-scale faults have been mapped on the western edge of Klamath Marsh. However, Sherrod and Pickthorn (1992) and Lee and Cummings (in preparation) show several small volcanic centers along the eastern front of the Cascade Range that could indicate the presence of such a structure.

Figure 10: Proposed conceptual models for ground water flow in the southern Williamson River basin. Arrows indicate proposed ground water flow paths. Figure is not to scale.

Future work should first be directed towards the development and testing of a conceptual model for ground water flow. The proposed preliminary conceptual models in this study could be tested easily with further stable isotope analysis. Once an isotopic signature for waters originating in the Cascade Range was established it could be compared to samples from Spring Creek, waters from the high yield wells in Klamath Marsh, and the data from this study. Age determinations of these waters would help identify the length of flow paths. Older age determinations from Spring Creek and the high yield wells would suggest a contribution from a more regional source.

Evaluation of historic discharge of the Williamson River from the Kirk Sill shows large variations in both annual and seasonal discharges. Seasonal variability is controlled by the magnitude and form of annual precipitation and the magnitude of historic precipitation. Greater discharge will occur at the Kirk Sill if the precipitation occurs locally as snow. The rate at which the snow melts also affects the discharge because the faster the snow melts the more likely that surface flow will be achieved in the permeable pyroclastic deposits. The below average annual precipitation years of 1985, 1987, 1988, and 1990 led to depletion of ground water stored in the shallow unconfined aquifers of southern Klamath Marsh and noncontinuous discharge over the Kirk Sill for the 1990-1995 water years. Precipitation was above average in 1995, and 1997-1999, but the stream gauge was deactivated from 1996-1998, so the time required to recharge the aquifer to the point of continuous flow over Kirk Sill is unknown. The high infiltration rate of this basin and the high variability of the Williamson River discharge suggest low storage capability of the aquifers.

The measured discharge data quantify the ground water contribution to the Williamson River within the canyon. The discharge at Kirk Sill was only 1.5% of the flow south of the Williamson River canyon at bridge 9730 in October, 1998. The percentage of ground water contribution to the overall flow decreased in April, 1999 to only 18%, but the volumetric contribution was three times larger than the October, 1998 measurement.

Three types of springs were identified in the Wocus Bay quadrangle: contact, fault and depression. Depression springs are the most common and occur in topographic lows that are thought to be pre-Mazama drainages that were filled with pyroclastic deposits. The discharge from these springs is ephemeral and is the discharge point for local flow systems. The fault springs were the only springs that maintained consistent flow throughout the year. These springs are being recharged by intermediate flow systems.

Values of deuterium from the Williamson River at Kirk Sill and bridge 9730, and two springs in the Wocus Bay quadrangle support the conclusions from the measured discharge data. The deuterium enriched value from Kirk Sill and the similarity of the values from bridge 9730 and the springs indicate that the Williamson River discharge south of the canyon is ground water dominated.

Review of water-well logs and information gathered from geologic mapping defined the aquifers and their water-bearing properties for the southern Klamath Marsh and the Soloman Butte quadrangles (Tables 3 & 4). The primary aquifer with the greatest yields in the southern Klamath Marsh is a basaltic unit that has only been identified from the well logs. The alluvial facies is also an important aquifer, but its extent is not as great as the basalt. Recharge to this aquifer is thought to occur on the east slopes of the Cascades and then flows through pyroclastic and pre-Mazama fluvial deposits to the Klamath Marsh. The aquifer with the greatest yield in the Soloman Butte quadrangle is an older continental sediment (Table 4) package that is over 200 m thick in some locations. Recharge to this aquifer and the basalt aquifer of southern Klamath Marsh is thought to occur in the volcanic eruptive center facies of the Cascades and the surrounding volcanic centers, and the alluvial facies.

Ground water in the Williamson River basin originates primarily as precipitation on the volcanic centers that surround Klamath Marsh. Ground water that originates in the Cascade Range to the west is thought to have little influence on the inflow to Klamath Marsh and the Williamson River. The ground water that originates in the Cascade Range is thought to either be prohibited from discharging in Klamath Marsh by a fault zone along the eastern slope of the Cascade Range or the flow paths are deep enough that discharge does not occur and the ground water moves south off of the Chemult plateau. The ground water from the Cascade Range is likely supplying the voluminous cold-spring-fed Spring Creek and many high yield artesian wells near the Spring Creek headwaters.

[Next Section] [Previous Section] [Table of Contents]