For all of the techniques, only parameters with a significant amount of uncertainty should be included in order to reduce the complexity of the response surface as much as possible.
Parameters that are reasonably well known and parameters whose values are relatively insensitive to the objective function being used should not be included in an automatic optimization run.

The advantage of these optimization techniques is that they are able to test a large number of parameter sets, ultimately producing a set with a lower optimization criterion than the initial set. The great disadvantage of these techniques is that they evaluate the goodness of fit of a simulation based on a single value. The variability of all the hydrograph components are summarized by a single number, the objective function. The following objective functions (optimization criterion options) are available: Daily RMS error (CMSD) Monthly Volume RMS error (mm) Modified Correlation Coefficient Maximum Likelihood Estimates Mean Daily Discharge (Simulated - Observed) Log Mean Daily Discharge (Simulated - Observed)

see NWSRFS Users Manual IV.4.2-OPT3-2