Klamath TMDL peer reviewer concerns addressed in report 

The North Coast Regional Water Quality Control Board (NCRWQCB), the governing body responsible for developing TMDLs on the Klamath River, recently made available a document containing peer review comments and the NCRWQCB staff responses to the questions and concerns raised, some of which reflect concerns of various commenters at the public workshop.

One issue raised by peer reveiwer Dr. Christopher A. Myrick, of the Colorado State University Department of Fish, Wildlife, and Conservation Biology, concerns the nutrients and organic matter standards proposed for the Iron Gate Hatchery.

Myrick states that he believes “that it may not be realistic to expect the hatchery to be able to achieve a zero net increase of nutrient and organic matter loads between the hatchery intake(s) in the reservoir and the hatchery discharge.

“By their very nature fish hatcheries will produce organic matter (excess feed, fish wastes, etc.) and while the use of settling ponds and careful control of feeding rates can reduce the amount of organic matter produced, they do not wholly eliminate it.”

Responding to Myrick’s concern, staff state that the hatchery must have a National Pollution Discharge Elimination System permit, which must be consistent with the TMDL. The staff response indicates that a “compliance schedule” can be implemented to allow time for infrastructure changes that can bring a discharger into compliance.

“As described in the Implementation Plan, the hatchery may be able to achieve any remaining required load reductions through offset mitigation that would be coordinated through the Klamath tracking and accounting program being developed as part of the Klamath implementation plan. 

“The final point is that the technical TMDL must assign allocations to all sources to the levels needed to meet water quality standards, without consideration of feasibility.  During the implementation phase of the TMDL alternative strategies for achieving load reductions can be evaluated,” staff stated.

Myrick comments on the fish health concerns regarding Ceratomyxa shasta and Columnaris, two of the more common fish diseases in the Klamath Basin, and how that may be affected by the target temperatures set forth in the TMDLs.

“The proposed temperature standards for the Iron Gate Reservoir tail race and the Iron Gate Hatchery (18.8ºC) should provide some protection against severe disease outbreaks, although the temperature does fall within the range categorized as having a high disease risk for juvenile rearing and adult migration.

“Nevertheless, given the natural conditions in the Klamath system above Iron Gate, it is unlikely that a much lower temperature could be achieved,” Myrick states.

Staff responds by saying that the “temperature targets for the Iron Gate tail race and hatchery are not standards, but are interpretations of the conditions that meet the standard, which in this case are natural temperatures. Therefore, the Iron Gate tail race and hatchery targets were developed based on natural conditions, as opposed to conditions that fully support the beneficial use.”

According to the staff, temperature is an important component in the fish disease cycle, but other factors contribute as well. Ultimately, according to the response, the TMDL Monitoring Plan is expected to provide information that will allow researchers to develop a better fish disease model “that will contribute to an improved understanding of the effect of degraded water quality conditions on fish disease in the Klamath River.”

Another peer reviewer, Dr. Gregory W. Characklis, of the University of North Carolina at Chapel Hill Department of Environmental Sciences and Engineering, begins with a concern about the uncertainty of the model used to develop the Klamath TMDLs, a concern also voiced at the workshop.

Characklis states that while he found the water quality models to be based on “sound scientific principles, and that the (largely) deterministic results appear to be reasonable given the data and information available,” he expressed concern over the general ability of models to predict natural conditions.

Characklis states, “my primary concern is that even state-of-the-art water quality models parameterized with extensive datasets are not terribly accurate, and are often unable to predict contaminant concentrations or loadings with what most would consider to be a reasonable level of accuracy.”

Characklis goes on to state that while he understands that it may be impossible to gather enough data “to fully characterize a complex natural system such as the Klamath,” his “primary suggestion would be that a more concerted effort be directed toward the evaluation and communication of the uncertainty inherent in these models.”

Characklis suggests an attempt be made to “corroborate” or validate the data from the model through observations of actual river behavior to try to assess the accuracy of the model predictions.

In the end, Characklis states, “I do believe, however, that the lack of explicit attention to the uncertainty issue could leave the impression that these models are more accurate than they actually are. Consequently, a more concerted effort to evaluate and communicate the uncertainty inherent in these models would seem appropriate.”

NCRWQCB staff respond to Characklis’ concerns by stating, “Due to the size and complexity of the Klamath River, limited resources, and schedules, it was determined that quantitative uncertainty analyses and formal, quantitative sensitivity analysis were not feasible.”

The staff response does state that attempts to minimize uncertainty were made, including peer review of the models; quality assurance project planning, which included data quality assessments; and model corroboration, including “qualitative and/or quantitative evaluation of a model’s accuracy and predictive capabilities.”

The staff go on to state that while a formal, quantitative sensitivity analysis was infeasible due to the “computational complexity of the Klamath River TMDL model,” model parameters and boundary conditions were adjusted based on available data “in order to achieve the best match between prediction and observations.”

Describing the computational limitations on performing multiple scenarios, the staff state that running the Klamath TMDL model requires four days of continuous simulation using a 2.66 gigahertz duo-core computer, generating five gigabytes (five billion bytes) of results.

Regarding the use of observed data to validate the model predictions, the staff state, “Data limitations are largely the reason that a quantitative error analysis was also not performed on the water quality simulation.” It is stated that, “Trends in the observed data and cause-effect relationships between various parameters can be replicated with a model, although precise values at each and every point in time may not be.”

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
NOTE: In accordance with Title 17 U.S.C. section 107, any copyrighted
material  herein is distributed without profit or payment to those who have
expressed  a  prior interest in receiving this information for non-profit
research and  educational purposes only. For more information go to: http://www.law.cornell.edu/uscode/17/107.shtml