Skip Links

USGS - science for a changing world

Scientific Investigations Report 2013–5014


Evapotranspiration from Wetland and Open-Water Sites at Upper Klamath Lake, Oregon, 2008–2010


Introduction


Upper Klamath Lake is a primary source of irrigation water for the Bureau of Reclamation’s Klamath Reclamation Project, which delivers irrigation water to approximately 210,000 acres of croplands in south-central Oregon and northern California. The lake also provides habitat for two species of fish listed as endangered under the Endangered Species Act, the short-nose sucker (Chasmistes brevirostris) and the Lost river sucker (Deltistes luxatus). The outlet of Upper Klamath Lake forms the headwaters of the Klamath River, the flow of which supports threatened anadromous fish populations and is important to tribal communities, anglers, and recreationalists. Managing the lake to meet all of these demands requires a quantitative understanding of the lake’s hydrologic budget. Evapotranspiration from the lake and surrounding wetlands is a substantial component of the hydrologic budget of the lake.


Although hydrologic budgets have been developed for Upper Klamath Lake (Hubbard, 1970; Kann and Walker, 1999), estimates of evapotranspiration (ET) from the approximately 27,000 hectares (ha) of open water and 7,000 ha of wetlands surrounding the lake were obtained from pan-evaporation measurements and models—and have been a major source of uncertainty. Uncertainty in ET from Upper Klamath Lake and the surrounding wetlands propagates as uncertainty into other components of the budget including groundwater inflow; this uncertainty is problematic for lake management, water-supply forecasting, and management and restoration of drained wetlands.


Year-to-year operation of the Klamath Reclamation Project (which includes lake-level management) varies depending on statistically based water-supply forecasts made by the Natural Resources Conservation Service. One source of forecast uncertainty is that the regression equations use net inflow to Upper Klamath Lake as the dependent variable. Net inflow is a lumped term that includes multiple components of the lake budget such as inflow from streams and springs, ET, and precipitation. Forecasting equations would likely be improved if the dependent variable more directly represented the physical response of the watershed to measured conditions in the basin (precipitation, snow water content, and so forth) as would be the case if ET from the lake and surrounding wetlands could be independently determined.


In addition to the 7,000 ha of wetlands still surrounding the lake, roughly 12,600 ha of wetlands was separated from the lake by levees, drained, and put into cultivation over the past several decades. Large tracts of drained wetland area were purchased by the Federal Government or private conservation organizations between the mid-1990s and late 2000s. Understanding the consequences of various management options for these drained wetlands will be improved with a better understanding of wetland ET and open-water evaporation.


In a review of work to quantify predevelopment water budgets in the upper Klamath Basin (Bureau of Reclamation, 2005), the National Research Council (NRC; 2007) identified ET estimates as a major source of uncertainty in hydrologic calculations, and improving ET estimates was a specific recommendation of the NRC report. The current study was conceived to reduce uncertainty in estimates of ET in and around Upper Klamath Lake, thereby reducing uncertainty in the lake water budget, and improving comparisons between ET from different types of land use in the area.


Purpose and Scope


The purposes of this report are to (1) summarize measurements of year-round evapotranspiration from the dominant vegetative communities in natural wetlands surrounding Upper Klamath Lake; (2) summarize measurements of open-water evaporation that are more comprehensive than measurements made in the past; (3) use the resulting measurements to better calibrate standard ET models for use in water supply planning and water resource management in the basin; and (4) compare ET from natural wetlands, open water, and various types of agricultural lands in the area. This report describes the study area, underlying theory, methods of investigation, and results. Measurements of wetland ET were made at two sites from May 1, 2008, through September 29, 2010, encompassing three growing seasons. Measurements of open-water evaporation were made at two sites during the warmer months of 2008 through 2010.


Previous Studies


Bidlake (2000) measured ET by the eddy covariance method and calculated an energy budget at one location in the wetland northwest of the lake for four periods ranging from 1.2 to 1.9 days in 1997. While the work by Bidlake improved understanding of ET from wetlands around Upper Klamath Lake, significant improvements in understanding could be made by making similar measurements continuously for multiple growing seasons in a variety of vegetative settings.


Bidlake (2002) also measured ET by the Bowen-ratio energy balance method from three fallowed agricultural fields in the Tule Lake National Wildlife Refuge, California, about 45 kilometers (km) southeast of Klamath Falls, Oregon. These sites were selected to typify ET rates that might result from water conservation measures applied to formerly irrigated agricultural land. Alternative management strategies can be evaluated by comparing wetland ET measured in the current study to fallowed cropland ET estimated by Bidlake (2002).


Janssen (2005) calculated open-water evaporation from Upper Klamath Lake for part of summer 2003 using the energy-budget method and compared the results with rates determined using the modified Penman method (Penman, 1956). Janssen’s work improved understanding of open-water evaporation from the lake and showed that the modified Penman method overestimated evaporation, but longer-term measurements are needed.


Description of Study Area


The upper Klamath Basin is semiarid but spans a steep precipitation gradient extending from the Cascade Range, where mean annual precipitation is 1.7 meters (m; Crater Lake National Park, Oregon, 1971–2000), to the interior parts of the basin where mean annual precipitation can be as low as 0.291 m (Tulelake, California, 1971–2000); all climate data in this section are from Western Regional Climate Center (2012). Mean annual precipitation (1971–2000) at Klamath Falls at the southern end of Upper Klamath Lake is 0.365 m. Of the total annual precipitation at Klamath Falls, 54 percent falls (mostly as snow) during the winter season (November through February), while only 15 percent falls during summer months (June through September). The remaining 31 percent falls during the transition months of October and March through May. Klamath Falls experiences cold winters with average December minimum and maximum temperatures of -5.6 degrees Celsius (°C) and 4.2°C, respectively, and warm summers with average July minimum and maximum temperatures of 10.9°C and 29.7°C, respectively (1971–2000).


Upper Klamath Lake (fig. 1) is the source of the Klamath River, and the lake encompasses approximately 270 square kilometers (km2) with an average depth of roughly 2.8 m at full pool. The lake occupies a fault-bounded structural basin immediately east of the Cascade Range volcanic arc. Principal tributaries to the lake include the Williamson River, which accounts for about half the inflow, the Wood River, Sevenmile Canal, and several other smaller canals and streams originating from the Cascade Range (Hubbard, 1970). Most of the lake is just a few meters deep except for a trench that is as much as 15 m deep that runs along the western margin. Although Upper Klamath Lake is a natural water body, the stage of the lake has been increased and controlled artificially since 1921 by Link River Dam, a control structure built to facilitate management of the lake for irrigation. Total surface inflow to the lake is not routinely measured, but the mean annual flow of the Williamson River from 1971 to 2000 (measured below its confluence with the Sprague River, USGS station 11502500), which accounts for roughly half the total inflow, averaged about 31 cubic meters per second (m3/s; 1,100 cubic feet per second (ft3/s)). Outflow from the lake is principally to the Link River (which becomes the Klamath River downstream) and to the A Canal which delivers irrigation water to the east and south. Link River discharge averaged 38.57 m3/s (1,362 ft3/s) from 1971 to 2000 (USGS station 11507500), and annual diversions to the A Canal averaged 7.45 m3/s (263 ft3/s) during the same period (Jason Cameron, Bureau of Reclamation, written commun.). A third component of surface outflow is the Keno Canal. Reliable measurements are no longer available for this canal, but discharge averaged 5.32 m3/s (188 ft3/s) during the period from 1967 through 1983.


Upper Klamath Lake was originally surrounded by roughly 196 km2 of wetlands. Between 1889 and 1971 about 126 km2 was diked off and drained for agricultural use, whereas about 70 km2 of natural wetlands remain (Snyder and Morace, 1997; Dan Snyder, USGS, written commun., 2011). The Upper Klamath Lake National Wildlife Refuge (NWR) is the largest natural wetland (58 km2) adjacent to the lake. The Nature Conservancy is in the process of restoring about 28 km2 of previously drained wetlands on the Williamson River delta by removing or breaching dikes. Draining wetlands around the lake, however, has caused considerable oxidation and compaction of the thick peat soils, resulting in subsidence of the land surface by as much as 4 m (Snyder and Morace, 1997). As a consequence, water depths are too deep in most areas for reestablishment of wetland vegetation and the areas become open water when previously drained wetlands are reflooded. It is not known how much time will be required for reestablishment of wetland vegetation.


Upper Klamath Lake has been historically eutrophic but is now considered hypereutrophic. For the past several decades, the lake has experienced annual blooms of a near‑monoculture of the blue-green algae Aphanizomenon flos‑aquae which results in toxic conditions for fish (Snyder and Morace, 1997; Wood and others, 2006). During blooms, the algae congregate in the photic zone within a few centimeters of the lake surface, possibly affecting the albedo (ratio of reflected to incident light) and hence energy budget of the lake.


Study Sites


Two ET monitoring sites were chosen within the Upper Klamath National Wildlife Refuge (NWR) wetland area and two sites were chosen to monitor open-water evaporation (fig. 1). At full pool, the ratio of open water to wetland area is about 4:1. Despite this partitioning, two wetland sites were selected to address questions relating to differences in ET from different wetland communities and how these rates compare to ET from land historically drained to support irrigated agricultural crops, formerly irrigated land allowed to go fallow, and land recently returned from crop production to wetland status through breaching of dikes.


Bulrush and Mixed Vegetation Sites


The 5,800-ha Upper Klamath NWR is dominated by extensive wetlands and forms the northwest margin of Upper Klamath Lake, north and south of Pelican Bay (fig. 1), at an altitude of about 1,262 m. Lake level is controlled by the Link River Dam at the outlet and fluctuates about 1.3 m annually, causing large areas of the refuge (including the two study sites) to be flooded from about midwinter to midsummer of each year. The wetland areas of the NWR are covered in dense vegetation, heavily dominated by hard-stem bulrush (formerly Scirpusacutus, now Schoenoplectus acutus), with smaller amounts of cattail (Typhalatifolia) and wocus (Nuphar polysepalum), and trace amounts of other vegetation. To typify ET from the most common species, one site was selected in a stand of almost exclusively bulrush, and another site was chosen within a patchwork of about 70 percent bulrush, 15 percent cattail, and 15 percent wocus. The sites were mapped on a visible light satellite image of the Upper Klamath NWR taken on August 28, 2011, and color patterns associated with each site were described. The bulrush site is located in an extensive, uniform, sepia-colored area, and the mixed vegetation site (hereafter called the mixed site) is located in a patchwork of sepia, light green, and dark green areas. Based on viewing the image of the whole NWR, we estimate that the bulrush site is typical of about 70 percent of the wetland, and the mixed site is representative of about 30 percent. Other objectives for site selection were to provide adequate fetch (extent of uniform vegetative cover around the sensors) for the ET measurements and to minimize the effort needed to access the sites during maintenance visits. These two objectives were best balanced by navigating along existing streams through the refuge far into the wetland, then along lateral, open, narrow, channels away from the streams, and finally out 30 to 50 m through the vegetation perpendicular to the channels. This strategy exploited the relatively easy travel along the streams and channels to penetrate deep into the wetlands and still kept the effects of the streams and channels on the ET measurements negligible. The two sites are located near the centers of the two largest lobes of wetland in the NWR (fig. 1). The bulrush site location is 42°30′48.88″ N., 122°2′4.89″ W., and the mixed site location is 42°28′36.80″ N., 122°4′ 6.05″ W. The bulrush and mixed site altitudes are 1,261.9 m (4,140.0 feet (ft)) and 1,262.0 m (4,140.5 ft) above mean sea level, respectively.


Bulrush, cattail, and wocus are phreatophytes, meaning their roots can remain flooded for extended periods without stress or damage caused by anoxia. In contrast, non-phreatophytic vegetation requires some degree of air penetration in the root zone to supply oxygen for root respiration. Consequently, the species at the study sites can thrive and transpire during the roughly 6-month period each year when lake water floods the local land surface, sometimes by as much as about 1 m. Bulrush and cattail are emergent vegetation, typically sending stalks and leaves to 2 to 3 m above the water surface. Their growth forms are similar, both having a central stalk and narrow, elongate leaves, all arranged nearly vertically when alive and vigorous. During winter dormancy, both plants cease photosynthesis, turn brown and brittle, and usually lodge over in response to snow and wind loading. At the study sites, the current year’s fallen plants merge with the existing dead layers from past years, forming a loose network of branching debris, which persists year-round, averaging about 0.8 m in depth. Unusually high winds also caused patchy areas of live-vegetation blow-down, primarily at the mixed site during the 2009 growing season.


Wocus is a pond lily, classified as floating rather than emergent vegetation. Its roughly circular leaves are about 0.2 to 0.3 m in diameter and often rest on the water surface, distributed evenly, nearly covering the surface. At the study site, the plant sprouts and the leaves surface in the spring during the rising water level, remain at the water surface through midsummer, and become slightly aerial when water level recedes sufficiently in late summer. The leaves senesce and become part of the substrate as the water level approaches land surface. In early spring, areas occupied by wocus appear to be open water until the leaves surface.


Soils at the wetland study sites consist of a mat of roots and decaying plant matter from previous years, roughly 0.3 m thick, underlain by a saturated clay-silt-peat mixture sometimes called gyttja (Hansen, 1959; Snyder and Morace, 1997). The gyttja is basically a high-viscosity fluid just beneath the root mat, and it gradually grades into an elastic solid with depth. A small (about 5-millimeter (mm)) diameter rod pushed into the land surface encounters moderate resistance through the root mat and then penetrates easily below that point. Sizable impacts delivered to the land surface propagate laterally through the root layer as a wave, indicating the root layer is a somewhat tough skin, floating on the denser, more fluid gyttja below. This lithology posed unique challenges for working in the wetland and deploying the sensor stations, as discussed in Eddy-Covariance Sensors and Data Collection. The soil/root mat surface is hummocky, with a relief of about ± 5 centimeters (cm) over small horizontal scales (0.5 m), but the mat surface is extremely flat and level at horizontal scales larger than about 5 m.


Open-Water Sites


Open-water evaporation was measured using the Bowen‑ratio energy balance (BREB) method (Anderson, 1954), with specialized sensors deployed at two open-water locations near the middle of Upper Klamath Lake, three open-water locations near the deep trench along the western margin of the lake, at the mouths of the Williamson and Wood Rivers (the two major inlets to the lake), the mouth of Sevenmile Canal, and at the beginning of the Link River and A Canal (fig. 1). In addition, lake stage was determined using existing sensors operated by the USGS located at Rocky Point, at Rattlesnake Point, and near the Link River Dam. Surface-to-air temperature and vapor-pressure differences, and water temperatures at various depths were measured at roughly the middle of the two largest lobes of the lake, at sites labeled MDL and MDN (fig. 1). These are previously established buoy locations used in other earlier and ongoing studies (Wood and others, 2006), where water depth is about 3 m. Minimum fetch to shore at the buoy sites is about 2.4 km, ensuring that the temperature and vapor‑pressure differences were fully equilibrated to the lake surface (Stannard, 1997; Stannard and others, 2004).


The four components of net radiation were measured at a shallow open-water location near the shore at the mouth of the Williamson River (fig. 1). A four-component net radiometer was deployed at this site to investigate the effects of the high concentration of suspended algae on reflected solar (short‑wave) radiation from the lake. In many lake evaporation studies, incoming radiation is measured on shore, and reflected solar radiation is modeled, assuming optical properties of clear water (Koberg, 1964; Sturrock and others, 1992). The known proliferation of algal blooms in Upper Klamath Lake suggested that a separate measurement of reflected solar radiation was warranted. The net radiometer site was chosen to typify the optical properties of water in the deeper parts of the lake but still provide water depths shallow enough to deploy a tripod on the lake bottom to support the sensor. Deployment from a raft was undesirable due to the inability to level the sensor reliably. Water depth at the site generally ranged from several centimeters to about 1 m, which was sufficient to behave optically like deeper midlake water during the bulk of the algal bloom periods.


Computation of open-water evaporation using the BREB method requires measurement or estimation of heat advected to or from the water body caused by inflows and outflows of surface water and groundwater. Therefore, water temperature and stage were measured at the two major surface-water inflows (the mouths of the Williamson and Wood Rivers) and at the Link River and A Canal outflows (fig. 1). Other surface water inputs were much smaller and were, therefore, not instrumented. The net heat exchange with groundwater also was considered to be minor (Janssen, 2005), and no attempt was made to measure it.


First posted March 4, 2013

For additional information contact:
Director, Oregon Water Science Center
U.S. Geological Survey
2130 SW 5th Avenue
Portland, Oregon 97201
http://or.water.usgs.gov

Part or all of this report is presented in Portable Document Format (PDF); the latest version of Adobe Reader or similar software is required to view it. Download the latest version of Adobe Reader, free of charge.

Accessibility FOIA Privacy Policies and Notices

Take Pride in America logo USA.gov logo U.S. Department of the Interior | U.S. Geological Survey
URL: http://pubsdata.usgs.gov/pubs/sir/2013/5014/section2.html
Page Contact Information: GS Pubs Web Contact
Page Last Modified: Tuesday, 26-Feb-2013 13:09:55 EST