CRiSP.1 In Context with Other Models

I.2.1 - CRiSP.1 In Context with Other Models

CRiSP.1 is one of several models describing the passage of juvenile salmon through the river system. The major mainstem passage models and their time steps are outlined below. All of these models describe fish survival through the river system but differ in the details included. The simplest model is PAM and the most complex is CRiSP.1.5. With increasing complexity, the other models fall in between. PAM and CRiSP.0 are close to SPM, FLUSH lies in the middle and CREM and RESPRED are in turn more complex.

CREM/RESPRED - Smolt passage and survival through a reservoir interactions of temperature, smolt run timing, size and number of both predator and prey. Dam survival is modeled with passage efficiencies and mortalities. Time step is a fraction of an hour.
PAM - Smolt passage and survival is described through a flow relationship, and dam survival through passage efficiencies and mortalities. Time step is one year.
FLUSH - Reservoir survival in terms of fish residence time depending on flow and fish size and a predation mortality rate that depends on temperature. Dam survival through passage efficiencies and mortalities. Time step is daily for fish passage and monthly for flow.
CRISP.0 - Smolt reservoir survival is through a flow relationship. Dam survival through passage efficiencies and mortalities. Time step is daily.
CRiSP.1 version 3 - Smolt reservoir survival depends on travel time, which depends on flow and mortality rates which in turn depends on temperature, predator abundance and nitrogen supersaturation. Dam survival depends on diel passage mortalities and passage efficiencies. Time step is a fraction of a day.
CRiSP.1 version 4 - Smolt reservoir survival depends on travel time, which depends on flow, reservoir elevation, smolt age and mortality rate which in turn depends on fish age, temperature, predator abundance and nitrogen supersaturation. Mortality in dam passage is identified separately for the forebay, dam passage routes and the tailrace. These mortalities depend on diel factors and passage efficiencies, which may depend on smolt age. Time step is a fraction of a day.
CRiSP.1 version 5 - Smolt reservoir survival depends on travel time, which depends on flow, reservoir elevation, smolt age and release date, temperature, predator abundance and nitrogen supersaturation and fish depth. Mortality in dam passage is identified separately for the forebay, dam passage routes and the tailrace. These mortalities depend on diel factors and passage efficiencies, which may depend on smolt age. Time step is a fraction of a day. Several options are available to define mortality rates in transportation.

Simple and complex models each have advantages and disadvantages and the choice depends in part on the types of questions being addressed. Simple models are easy to understand and do not require large amounts of data since they are generally developed to the degree of resolution of the available data. Unfortunately, since they are designed to fix existing data, the variables are identified from past studies and there is no formal process to assess if chosen variables are insignificant or the driving variables. Simple models, being empirical and based on existing data, typically fit the available data but it is not clear that they would still represent the system if the system were to change.

In general, complex models are based on underlying mechanisms which are inferred from analogous systems. As such, they often can be extended beyond the data of the system they are modeling. Mechanistic models are particularly useful for developing hypotheses that are to be tested through experimentation. A disadvantage of such models is that they require a considerable effort to understand and calibrate, and there is a chance that the underlying mechanisms may be misapplied or inappropriate, as can also be the case with simple empirical models.

The tags of "simple and empirical" vs. "complex and mechanistic" are, in fact, only points along a continuum. Even the simplest models have some underlying mechanism and complex models, at some level of detail, contain empirical descriptions in which the parameters have no mechanistic foundation and are applied because they fit data or a desired mathematical form. The level of detail at which a model switches to empirical formulations distinguishes its degree of mechanism.

All of the existing mainstem fish passage models leave out some variables which may be important. The simplest models, such as PAM and CRiSP0, rely solely on flow to describe survival and are of limited value for investigating other factors. Although FLUSH and CRiSP.1 version 5 consider the effects of fish age on survival, none of the models considers differences in fish conditions, fish bioenergetics or the river and tributary carrying capacities.

[Manual] [Contents] [Prev] [Next]

web@cbr.washington.edu