Implications for Future Monitoring and Management


Through the construction, calibration, and application of the models, a great deal was learned about the characteristics of the Klamath River system, the nature of the organic matter, the relative importance of algae, and the dynamic nature of oxygen demands. Interestingly, years with fewer input data available still produced good model results. For example, the error statistics for dissolved oxygen in 2006 and 2009 (years with fewer data) were comparable to 2007 and 2008 (years with rich datasets). The relatively good fit to the data in 2006 and 2009, however, could not have been achieved without the weekly field sampling and directed experimental work that occurred in 2007 and 2008, all of which was crucial to derive model rates and ratios and improve our understanding of spatial and temporal water-quality patterns in the Link–Keno reach. Future improvements to these and other models will rely on an increased understanding of the complex water-quality processes in this reach of the Klamath River. On the basis of what was learned in this modeling study, suggestions for future monitoring activities include: 


Minimum level of water-quality sampling. Knowledge of interannual and seasonal variability in water quality can be expanded with a minimum level of water-quality sampling in the Link–Keno reach. Both grab samples and continuous water-quality monitors provide valuable and relevant data. A minimum program might include monthly to twice-monthly grab samples and data from a selected set of deployed continuous water-quality monitors. A minimum set of sample locations might include Link River, Miller Island, Keno, and Klamath Straits Drain. Ideally, telemetry would provide real-time data from at least one continuous monitor(s) that could be used to scale up or scale-down the frequency of sampling, depending on the occurrence of unusual events or critical periods of water quality. Useful constituents from these grab samples could include those that provide information on nutrients, organic matter partitioning, and algae. The field data typically collected during each site visit, including vertical profiles, algal density, turbidity, ice cover, and Secchi depth, could aid future analyses and modeling. The inclusion of samples in winter would be valuable because less historical information exists for that period. Measurements of inorganic sediment concentrations during periods of high flow could expand the equation relating turbidity and suspended sediment. The details of any water-quality monitoring plan need to be geared toward providing data to meet specific objectives. A monitoring plan to provide data for modeling purposes, for example, may require more frequent data collection at more sites than might be required otherwise.


Channel bottom sampling and sedimentation rates. The first-order sediment compartment in the model simulates the spatially variable buildup and decay of algal and organic material that settles to the river bottom. Sampling of material at the sediment–water interface in summer and winter could support documentation of this process in the field. Continued work on characterizing inflowing organic matter, sediment deposition rates, and additional studies on the potential for resuspension of sediments, composition of settled materials, decay rates under in-situ conditions, and other factors could greatly aid any future model refinements for sediment-related processes and shed light on dissolved-oxygen conditions. In addition to seasonal variation, sufficient spatial characterization is important when assessing sediment conditions.


Macrophyte survey and algae studies. High densities of rooted aquatic plants (macrophytes) occur only in certain areas of the Link–Keno reach due in part to the low penetration of light through the water column. However, some areas in the downstream half of the reach do have macrophyte populations. A series of macrophyte surveys, including density and species, would be highly beneficial to characterize the extent of these plants, assess their relative importance to the water quality of the river, and allow their inclusion into the model in an effort to improve predictions of near-surface dissolved-oxygen concentrations in the downstream half of the reach. 


Algal communities are immensely complex in their composition, dynamics, and reaction to seasonal changes in flow, light, and nutrient conditions. Model implementation and calibration, although sufficiently representative for many applications and scenarios, indicates that the representation of algae in the model remains a source of substantial uncertainty. A special study focusing on longitudinal and temporal conditions in the Link–Keno reach could lend insight into these important processes, and their impact on dissolved-oxygen conditions and other water-quality effects, particularly in light of future restoration efforts.


Nutrient studies. Experiments focused on understanding the rates and spatial distribution of nitrogen and phosphorus cycling processes will contribute to better models. Potential studies could measure the rates of ammonia production, nitrification, and denitrification, and also examine the forms and availability of phosphorus species in the water column. The relationships of nutrients and primary production play a critical role in the water-quality conditions in this reach.


Link River. The model developed in this study does not include the 1-mi Link River reach. Although the Link River reach is short, questions have arisen about potentially important nutrient transformations there. Before developing a model of the 1-mi Link River reach, additional nutrient and organic matter sampling is needed at several sites—upstream of Link Dam, the outflow of Link Dam, and at the mouth of Link River—to document any such water-quality transformations. This piece of the model could become important if the need exists to connect the Link–Keno Klamath River model to an Upper Klamath Lake model.


Reaeration equations for Keno Dam. The model does not yet simulate the reaeration of dissolved oxygen at Keno Dam. For modeling, equations relating dissolved-oxygen concentrations and spill rates would need to be developed specific to Keno Dam. This addition would become important to connect the Link–Keno model to downstream models to simulate dissolved oxygen in the lower reaches of the Klamath River.


First posted July 14, 2011