Introduction 


Upper Klamath and Agency Lakes in south-central Oregon constitute a shallow lake system with a large surface area (average depth 2.8 m and surface area 230 km²,fig. 1). As such, the currents in the system are primarily wind-driven. The lakes are situated in the Klamath structural graben, and under prevailing conditions, the westerly to northwesterly winds that blow over the northern part of Upper Klamath Lake are largely constrained by the surrounding topography to a north-northwesterly direction over the lower two-thirds of the lake. The resulting circulation pattern is a clockwise gyre around Upper Klamath Lake (Wood and others, 2006). The largest tributary to the system is the Williamson River, which enters Upper Klamath Lake along the northern shoreline (fig. 1); a smaller tributary is the Wood River, which enters at the northern end of Agency Lake. Lake surface elevations are regulated by the Link River Dam at the southern end of the lake, which is operated by the Bureau of Reclamation. During the summer months, lake elevation declines due to a combination of reduced inflows, evaporation, and downstream diversions to meet irrigation demands and instream flow requirements of the National Marine Fisheries Service for coho salmon in the Klamath River. As a result, in most years lake elevation declines by about 1 m (3.3 ft) between June 1 and September 1. 


Prior to 2008, Agency and Upper Klamath Lakes were two distinct water bodies connected by a narrow channel called Agency Straits. Starting in 2008, the Williamson River restoration project, undertaken by The Nature Conservancy as owner of the property where the Williamson River enters Upper Klamath Lake, transformed the lower 3 mi of the river. Between 2008 and 2009, a series of breaches through levees on both sides of the river channel and through levees surrounding previously drained wetlands on the northwest and southeast side of the channel reconnected the Williamson River with 3,000 ha of land that, prior to the 1940s, had composed the wetlands of the Williamson River Delta (hereafter referred to as the Delta [fig. 1]). Historically, these wetlands had functions that enhanced both the water quality and the ecology of Upper Klamath Lake, notably among them retaining nutrients and providing rearing habitat for larval fish. Sub-populations of the threatened Lost River and shortnose suckers, endemic to Upper Klamath Lake, spawn in the Williamson River system in early spring. After emerging from the gravel at the spawning grounds, larvae drift downstream at night with the river flow (Ellsworth and others, 2008 and 2009). Prior to restoration, all of these larvae drifted directly into Upper Klamath Lake. It is anticipated that the restored wetlands will lengthen the time it takes for most of the larvae to enter Upper Klamath Lake, and provide better conditions in terms of food availability and predator protection for larval growth and survival before they finally enter the lake. 


The goal of the restoration project is to restore the historical functions of the wetlands over time, and because maximizing the connection between the River, the Delta, and Upper Klamath Lake was critical to realizing this goal, the project was designed with this in mind. Modeling scenarios using the two-dimensional MIKE 21 hydrodynamic model (DHI Water and Environment, 2002) were used to select locations for the levee breaches and channel modifications that would maximize the amount of flow leaving the Williamson River channel to flow through either the northwest side of the Delta (Tulana) or the southeast side of the Delta (Goose Bay) (Daraio and others, 2004). The scope of this study, as a matter of practical necessity, was primarily limited to simulating the flow through the Delta at one moderately high Williamson River flow (2,070 ft³/s, representing the 1.5-year flood) and a lake elevation approaching full pool (4,143 ft, compared to a full pool elevation of 4,143.3 ft). Additionally, the scope of the study was limited in two other important ways: wind forcing at the water surface was not considered and, because of this and the location of the downstream boundaries in Agency and Upper Klamath Lakes, the influence of the large wind-driven circulation of those lakes on flow through the Delta could not be considered.


Subsequent to the work by Daraio and others (2004), a three-dimensional hydrodynamic, heat, and transport model of the entire Upper Klamath Lake/Agency Lake system using the hydrodynamic model UnTRIM (Casulli, 1999; Casulli and Cheng, 1992; Casulli and Zanolli, 2002 and 2005) was developed. The boundary of the model grid at the time of the original development reflected the situation with all levees intact, and therefore, flow from the Williamson River entered Upper Klamath Lake entirely at the mouth of the river. The model grid was subsequently modified to include Tulana within the model boundaries, which was opened with explosive breaching of the levees on October 30, 2007, in the first phase of The Nature Conservancy’s restoration project. That modified version of the model grid was used to investigate the probable effect of the construction of a large offstream storage project on the transport of larval fish through the Tulana side of the Delta (Wood, 2009). The grid has since been modified again to incorporate the Goose Bay side of the Delta, which, in a second phase of The Nature Conservancy’s restoration project, was reconnected to the lake with earth-moving equipment on November 18, 2008, along levees to the east of the Williamson River mouth that had previously delineated the model grid boundary. Thus, the model grid as used for the simulations in this report represents the completed changes in the landscape around the lower reach of the Williamson River that resulted from earth removal or movement associated with the Delta restoration.


This study, done in cooperation with the Bureau of Reclamation, is part of a larger modeling study of how the restoration of the Delta has affected the transport of larval suckers between the Williamson River channel upstream of restoration alterations, and the point at which they enter Upper Klamath Lake. Previous and ongoing modeling has demonstrated that lake elevation, Williamson River flow, and wind forcing all affect the movement of water and passively transported constituents through the Delta, but it is often difficult to sort out the magnitude of the dependencies when all three of these variables are changing simultaneously through a simulation. This study attempts to use a series of numerical “experiments” with the hydrodynamic model and tracers to systematically investigate the relative effects of Williamson River flow, lake elevation, and wind on the flow and transport through both sides of the Delta.


Because this study is based on the simulation of numerical tracers, the results strictly apply only to constituents that are carried passively in the flow. Therefore, the results are expected to apply generally to the problem of larval drift, but the implications of non-passive behavior for the movement of larvae through the Delta is an ongoing area of research. The results of this study will also have implications for water quality because the simulated passive transport through the Delta applies to dissolved constituents and suspended material.


First posted March 29, 2012