USGS Home
Contact USGS
Search USGS

U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2008–5076

Back to Table of Contents

Introduction

Upper Klamath Lake (UKL) is a large (surface area 232 km²) and shallow (mean depth 2.8 m at full pool) lake located in southern Oregon, in the semiarid landscape in the rain shadow east of the Cascade Range, at a full-pool elevation of 1,262.9 m above the National Geodetic Vertical Datum of 1929 (fig. 1). Agency Lake is connected to the northern end of Upper Klamath Lake through a narrow channel (Agency Straits), and adds about 38 km² of surface area. Paleolimnological evidence indicates that the lake has been highly productive for at least the last 1,000 years (Eilers and others 2004). This can be attributed to, among other things, the fact that much of the basin drains volcanic soils with high phosphorus content, and, because most of the lake is shallow, photosynthetically active radiation penetrates most of its volume on a daily basis, providing energy that is converted to biomass through the photosynthesis performed by algae and cyanobacteria.

Starting about 150 years ago, changes in land use in the UKL basin led to increased sedimentation rates and nutrient loads, and to a decrease in the nitrogen to phosphorus ratio in the nutrient loads to the lake. Simultaneous with these changes, the diverse taxa that characterized the assemblage of phytoplankton prior to 150 years ago gradually came to be dominated by a single species of buoyant cyanobacterium, Aphanizomenon flos aquae (AFA) (Phinney and Peek, 1960; Miller and Tash, 1967; Kann, 1998; Eilers and others, 2004; Eilers and others, 2001; Bradbury, Colman, and Dean, 2004; Bradbury, Colman, and Reynolds, 2004; Colman and others, 2004; Colman, Bradbury, and Rosenbaum, 2004). The shallowness of the lake also makes the transfer of nutrients from the sediments to the water column and back again to the sediments particularly efficient, although the specific mechanism for the transfer is still debated (Kann, 1998; Walker, 2001; Fisher and Wood, 2004). Annual blooms of AFA are perpetuated primarily by this internal recycling of nutrients stored in the sediments, rather than by external sources (Kann and Walker, 1999; Walker, 2001; Kann, 1998).

The annual cyanobacterial blooms have important implications for water quality. Highly supersaturated dissolved oxygen concentrations with large diel swings, and high pH values caused by the photosynthetic removal of carbon dioxide from the water column, are associated with the rapidly expanding phase of the bloom. A rapid bloom decline is accompanied by undersaturated dissolved oxygen concentrations as photosynthetic production of oxygen slows dramatically and ongoing respiratory demands continue, and decay processes associated with cell senescence consume oxygen rapidly. Severe low dissolved oxygen events (LDOEs) associated with bloom declines are detrimental to the survival of the shortnose and Lost River suckers, which are listed as endangered under the Endangered Species Act by the U.S. Fish and Wildlife Service (2001). As determined from an analysis of three large fish die-offs during 1995–97 (Perkins and others, 2000), severe LDOEs in which dissolved oxygen concentrations of less than 4 mg/L occur throughout the water column for a large part of the day and persist for several days at a time may cause fish death. These conditions may also cause fish death indirectly by forcing fish into crowded conditions and facilitating the spread of disease among animals already weakened by exposure to hypoxia, high pH, and the high un-ionized ammonia concentrations that sometimes accompany conditions of high pH.

Continuous water-quality monitors with well-documented data quality were first installed by the U.S. Geological Survey (USGS) for long-term deployment in 2002. Thus, reconstructing the details of the LDOEs of 1995–97 on time scales of a few days to a week is not possible, but information gathered since 2002 indicates that there is a rapid decline in the AFA bloom in most years around the end of July or beginning of August (Wood and others, 2006; Hoilman and others, 2008). The longest dataset of water-quality measurements in UKL is a 17-year record of biweekly profiles of conventional water-quality variables, as well as depth-averaged chlorophyll a and nutrient concentrations (Kris Fisher, Klamath Tribes, oral commun., 2007). This dataset does not provide enough temporal resolution to determine precise dates of bloom peaks and troughs, but the general pattern appears consistent over this long record. The effect of the late July to early August decline on dissolved oxygen concentrations over a large area in the northern part of the lake, which is the preferred habitat for adult suckers, varies in severity from year to year. In 2003, for example, dissolved oxygen concentrations at a centrally located site in the northern part of the lake were less than 3 mg/L continuously for 8 days, and the spatial extent of the LDOE was nearly 40 km² (Wood and others, 2006). This event culminated in a smaller fish die-off than those of the mid-1990s (Adams and others, 2003). A die-off of the severity of the mid-1990s has not occurred since the water-quality monitors were installed in 2002; therefore, it is not an annual occurrence. The impact of large fish die-offs on the endangered sucker populations is, nonetheless, devastating; thus, it is important to understand what combination of conditions results in the most severe events before a management strategy can be devised to modulate the algal bloom-decline cycle. Understanding the causes of the precipitous bloom decline is critical in this regard because the loss of photosynthetic production in addition to the added oxygen demand generated by senescing cells is the root cause of LDOEs. It also has been observed that the events are more severe in the northern sucker habitat area of the lake than in areas of similar depth in the rest of the lake, and observations of currents have indicated that circulation patterns play an important role.

The surface area to volume ratio of a shallow lake contributes to greater primary production per volume and enhanced nutrient cycling between the water column and sediments in comparison to a deep lake (Scheffer, 1998). Thus, the hypereutrophication of UKL has similarities to processes observed in other shallow lakes that experience massive cyanobacterial blooms. At the same time, each lake is unique in some way, and UKL is no exception. A particularly important feature of UKL that has implications for both the hydrodynamics and water quality of the lake is its bathymetry. Although most of the lake is shallow, a relatively deep trench runs along the western shoreline (fig. 1). This trench is apparent in the bathymetry as far south as Buck Island. It runs across the mouth of Howard Bay, along the western shoreline and to the west of Bare Island, and then turns west around Eagle Point, runs across the entrance to Shoalwater Bay and along Ball Point, before turning north at the entrance to Ball Bay and fading away. The currents in the trench are the strongest in the lake and are aligned with the bathymetry; under prevailing wind conditions, water flows northward through the trench (Gartner and others, 2007).

Recognizing that a detailed understanding of the movement of water around the lake was essential to a complete understanding of the water quality of the lake, the Bureau of Reclamation and the U.S. Geological Survey (USGS) entered into a cooperative agreement in 2005 to develop a hydrodynamic model of the lake. This report is the first comprehensive documentation of the UKL model and its application, although singular aspects of the model have appeared in proceedings papers (Cheng and others, 2005; Wood and Cheng, 2006).

Purpose and Scope

The purpose of this report is to provide a comprehensive accounting of the boundary, forcing, and calibration data available during the summer field season in 2005 (approximately mid-June through mid-September) and 2006 (approximately mid-May through mid-October), to provide a detailed description of the source and boundary terms that have been added to the model in order to accurately simulate heat transport, and to document the calibration and validation of the numerical model. The report also discusses the implications of lake circulation for water quality, with particular emphasis on how circulation might affect dissolved oxygen concentrations in the northern part of the lake. The most important numerical features of the model are briefly summarized below, but the reader is referred to the references cited in this report for the details of the numerical methods used in the computational core of the UnTRIM model.

Back to Table of Contents