Introduction


The Klamath River flows about 255 mi (410 km) from the outlet of Upper Klamath Lake through southern Oregon and northern California to the Pacific Ocean. The first 20 mi of the river, just downstream of Upper Klamath Lake, are bounded by Link River and Keno Dam (fig. 1). Water quality in this reach has been classified as “very poor” by the State of Oregon (Mrazik, 2007) and was designated as “water quality limited” for exceeding ammonia toxicity and dissolved-oxygen criteria year-round, and pH and the chlorophyll a criterion in summer (Oregon Department of Environmental Quality, 2007). A Total Maximum Daily Load (TMDL) for the Klamath River including this reach (Oregon Department of Environmental Quality, 2010) was approved by the Environmental Protection Agency in May 2012. The TMDL specifies load decreases of total nitrogen, total phosphorus, and 5-day biochemical oxygen demand (BOD5) for the nonpoint sources Lost River Diversion Channel and Klamath Straits Drain and for point sources including the Klamath Falls and South Suburban wastewater treatment plants. At the time of this study, the temperature section of the upper Klamath River TMDL was not approved and was undergoing additional analysis. 


Most nutrient loads in the Link River to Keno Dam (Link-Keno) reach came from Upper Klamath Lake through Link River with additional input by nonpoint and point sources in the Link-Keno reach (table 1). Total phosphorus loads in Link River consisted of phosphorus in algae, particulate organic matter, dissolved organic matter, and orthophosphorus. Total nitrogen loads in Link River were comprised of algae, particulate organic matter, dissolved organic matter, nitrate, and ammonia. The relative contribution of those various nutrient sources varied through the year (fig. 2).


A TMDL for Upper Klamath Lake, just upstream of the study reach was approved by the U.S. Environmental Protection Agency in 2002. That TMDL specifies decreases in inflow loads and provides an in-lake phosphorus target (Oregon Department of Environmental Quality, 2002). Changes of water quality in Upper Klamath Lake would affect water quality in Link River, the upstream boundary of the Link-Keno model.


In addition to TMDL actions, other management options are being considered that may improve water quality in the Link-Keno reach. Running scenarios based on a calibrated water-quality model of the reach allows the effects of such options to be predicted and fully considered so that management and restoration efforts can focus on strategies with the highest likelihood of success. Only the water-quality effects of potential management actions are discussed here; other aspects of water-quality improvement options, such as cost or implementation timeframes, are not included in this report.


Model Background


River water quality can be affected by hydrology, weather and climate atmospheric conditions, inputs and withdrawals, chemical reactions, and biota. Mechanistic computer models such as CE-QUAL-W2 (Cole and Wells, 2008) include many of these processes and are regularly used to make predictions about the potential water quality response to system changes. Models commonly are used in constructing TMDLs, and the upper Klamath River TMDL was based in part on results from a CE-QUAL-W2 model that used conditions from years 2000 and 2002 (Rounds and Sullivan, 2009; Tetra Tech, Inc., 2009; Rounds and Sullivan, 2013). That model from Tetra Tech, Inc. and the Oregon Department of Environmental Quality (ODEQ) is referred to as the “TMDL model” in this report.


In an effort to improve the understanding of instream processes in this river reach and create a more accurate predictive model, the U.S. Geological Survey (USGS), Watercourse Engineering Inc. (Watercourse), and Bureau of Reclamation (Reclamation) collaborated in a research, monitoring, and modeling study that produced a calibrated CE-QUAL-W2 version 3.6 model of the Link-Keno reach for conditions during 2006-09 (Sullivan and others, 2011). This new model was based on extensive field data, with additional field research on issues of flow, suspended matter settling, and dissolved oxygen and organic matter dynamics (Sullivan and others, 2008, 2009, 2010; Poulson and Sullivan, 2010; Deas and Vaughn, 2011) to better define model parameters and rates. Subsequently, this calibrated model was updated to include macrophytes and improvements to the simulation of pH (Sullivan and others, 2013). This USGS-Watercourse-Reclamation model (referred to as the “USGS model”) can simulate stage, flow, water velocity, ice cover, water temperature, specific conductance, inorganic suspended sediment, total nitrogen, particulate nitrogen, nitrate, ammonia, total phosphorus, orthophosphorus, particulate carbon, dissolved organic carbon, organic matter in the sediment, three algal groups, three macrophyte groups, dissolved oxygen, and pH.


CE-QUAL-W2 is a two-dimensional model, simulating variability from upstream to downstream and from the river surface to the channel bottom. The third dimension, from bank to bank, is laterally averaged. As such, the model is well suited for the simulation of conditions in long, narrow waterbodies such as rivers and reservoirs that tend to stratify thermally; in such waterbodies, the vertical variability of water quality tends to be more distinct than any lateral variability. The main branch of the Link-Keno model grid consists of 102 segments that connect together in the direction of flow (fig. 3); segments average 1,009 ft (308 m) in length. Each segment represents a cross-sectional shape of the river channel, with stacked layers of varying width from the river surface to the channel bottom. A side view of the model grid is available in Sullivan and others (2011). Vertical layers in the USGS model grid were 0.61 m in height. The model keeps track of all simulated constituents in all layers of every segment, and can output results at selected locations and time intervals, often hourly.


Although the Link-Keno model was constructed and calibrated for conditions in the years 2006-09, the mechanistic nature of the models allows for useful predictions of hydrodynamic, thermal, and water-quality changes resulting from altered conditions. However, all model predictions have some uncertainty. Results from model scenario runs are most useful in providing insights regarding changes to the system through comparative analysis, rather than in providing high certainty regarding the values of predicted concentrations. For example, scenario results can be used to evaluate decisions about which treatment or restoration processes might be most effective at improving water quality by assessing the predicted changes in key constituent concentrations.


These scenarios were developed by the USGS and Watercourse in cooperation with Reclamation. Scenario results will inform local and regional managers who need information about potential approaches to improve water-quality conditions while efficiently managing the system for multiple uses.


Purpose and Scope


The purpose of this study was to predict the potential water-quality effects of management strategies and other system changes through the application of the USGS model of the upper Klamath River from the mouth of Link River to Keno Dam. Most model scenarios were superimposed on the wide range of conditions that occurred for the years 2006-09, thus allowing simulation of a range of climatic, hydrologic, and water-quality conditions. These model scenarios were formulated and run to:


1. Assess the effect of TMDL total phosphorus, total nitrogen, and BOD5 targets and allocations on upper Klamath River water quality (scenarios 1-2), and compare those results to the relevant Oregon dissolved oxygen, pH, and ammonia toxicity criteria (scenario 3);

2. Evaluate the importance of differences in the formulations of the USGS and TMDL models (scenario 4);

3. Assess various water quality improvement options related to particulate material, wetland treatment, flow management, shading, and oxygen injection (scenarios 5, 6, 7, 8, 9);

4. Examine the possible effects of a warmer climate on river water quality (scenario 10).


Preliminary results from model scenarios 1 through 5 were published previously (Sullivan and others, 2011, 2012). The results presented in this report expand upon or augment those results; the previous results are still valid, although the results in this report are more current.


First posted July 24, 2013