CRiSP1.6 Theory & Calibration Manual: IV.3 - Summary

IV.3 - Summary

While all the parameters examined have some influence on the outcome of model runs, it is clear that some are substantially more important than others depending on the species in question. The three species (Steelhead, Chinook 0 and Chinook 1) were comparably sensitive to FGE in both the absolute value and scaled runs. Flow increases had less impact than FGE: spring chinook (Chinook 1) were most sensitive to increases in the Snake River flow while fall chinook (Chinook 0) and steelhead were more sensitive to changes in Columbia River flow. Steelhead survival is noticeably higher for cooler temperatures. Chinook 0 are most sensitive to spill. Chinook 1 are very sensitive to all water temperature changes and can best take advantage of increased flows in the Snake River.

There are substantial nonlinearities in some responses. This is due to the underlying theory of the model in these cases, which is itself nonlinear. The relationship between flow and survival is a result of several interacting submodels: because increased flow increases fish velocity, it reduces the mortality suffered due to predation, and thus increasing flow produces increasing survival. As flows increase, however, water is forced into the spillway, which produces elevated dissolved gas levels (TDG). This causes mortality via gas bubble disease; as spill increases, this mortality increases unless this is offset by reduced mortality from other sources.

Analyzed Range and Observed Range

In performing a sensitivity analysis, the range examined for each parameter is to some extent arbitrary. For some parameters, obvious extremes are suggested: for spill fraction, for example, the range from 0% spill to 100% spill is a natural choice to examine. At the same time, managers are interested in the real range of these parameters and how the model responds within reasonable system operations. For this reason, the scaled value sensitivity was performed in order to see how a small percentage alteration in the parameter can affect survival.

For those parameters to which the model is relatively insensitive, this indicates that the real system may be even more insensitive to changes in the modeled process, and that therefore mitigation measures that focus on those areas will be unlikely to produce significant benefit. At the same time, even those parameters which produce moderate to large impacts on survival may in reality be confined to a much narrower band in the actual hydrosystem.

There has been considerable pressure to make improvements in the hydrosystem that lead to improved juvenile salmonid survival; in that sense, the current configuration is, within its constraints, close to optimized for fish survival. The small changes that are allowed within system operation guidelines are unlikely to produce other than equally small changes in fish survival. This is a property of the real world system that is reflected accurately in CRiSP.1.

CRiSP1.6 Theory & Calibration Manual: IV.3 - Summary