[Manual] [Contents] [Prev] [Next]
CRiSP.1 version 5 (CRiSP.1.5) is a complex model, with hundreds of parameters. It is impossible to examine the potential interactions of all of these parameters. Consequently, sensitivity of fish survival for a number of single parameters is evaluated independently, as well as for several potentially interesting pairs of parameters.
In general, sensitivity is determined by first obtaining some baseline output - the survival of some arbitrary species under arbitrary model conditions. Individual parameters are then changed to some reasonable limit values, while all other parameters are held constant, and the resulting impact on model output is recorded. For pairs of parameters, several values for each parameter are chosen and the output for all possible combinations is recorded.
Analysis of the individual parameters are presented as graphs of survival as a function of the parameter, and pairs of parameters are shown as similar response surfaces in three-dimensional space.
Time-dependent and time-independent parameters
The parameters in CRiSP.1 fall into two categories: those that vary over the season of migration (e.g. flow, temperature, migration characteristics) and those that do not vary over the season (e.g. dam mortality components, predator density). Some of the parameters will not produce any change in model output. For example, if flow and temperature are held constant; it makes no difference when fish are released if there is no seasonal variation.
Most parameters were analyzed under constant conditions, that is, without seasonal variation. Variation was added when considering variables that would be meaningless without it.
Analysis of single parameters
The model was initiated and run using 1993 parameter values, except for dam survival-related parameters (turbine, bypass, spillway, and transport mortality) and FGE values, both of which were assigned based on calibrated values (see FGE Calibration section on II.9.2 and Dam Passage Survival section II.9.5). A single release of yearling (spring) chinook was input into the model at the head of Lower Granite Pool, and survival to Bonneville Dam, or to the estuary, was recorded. The effects of variation in:
- reach predator density
- forebay predator density
- tailrace predator density
- predator activity exponent
- Velocity variance (Vvar)
- Migration rate parameters
- Turbine mortality
- Bypass mortality
- Spillway mortality
- Transport mortality
- Fish Guidance Efficiency
- Total Flow (no seasonal variation)
- Temperature (no seasonal variation)
- Spill (same fraction at all dams)
have been examined. Note that because predation is determined by a multiplicative combination of predator density and predator activity coefficient, varying one of these two parameters gives identical results to varying the other; here predator density was varied rather than the activity coefficient.
[Manual] [Contents] [Prev] [Next]
Columbia River Salmon Passage Model CRiSP.1.5 Theory, Calibration & Validation Manual
Copyright © 1996, Columbia Basin Research. All rights reserved.
web@cbr.washington.edu