The Specified Values method allows users to sample from provided values instead of using a parameter distribution. While this method is not an uncertainty analysis in the traditional sense—where values are sampled from a parameter distribution—it facilitates variations in parameter values based on user specifications. This enables users to specify multiple parameter sets intended to be treated as a unit, defining relationships between parameters based on specified values for a particular index. When using this method, it can only be performed sequentially if the number of values per parameter is the same (e.g., 10 values for parameter 1 and 10 values for parameter 2). Alternatively, users can opt for random sampling with the same index to maintain consistency across the parameters.

As outlined, for each iteration, the values can be selected sequentially, randomly with the same index, or randomly with independent indices. If the sequential method is chosen, the model will start the first iteration from Index 1, the second iteration will use parameters from Index 2, and so on. Once the iteration reaches the final value, it will loop back to the first index and repeat the cycle.

If one parameter has more values than the other parameter, the iteration will continue to use each value before returning back to Index 1.  In the table below, at iteration 5, Parameter 1 will use value 0.1 while Parameter 2 will use value 0.58.  

When the randomly with the same index method is selected, the parameter selection will use the same index value for all parameters.  When using this method, the number of values for each Parameter must be the same length.  

The last selection method is a random selection independent of the index value.  This allows each parameter to be selected without any consideration to the index value.  

One strategy for determining parameter values are to extract the values from calibration events.  For example, if the model has been calibrated to 6 storms events, each Parameter-Index can be set to a calibration event.