Hydrologic Sampler Results
It is important for the user to be able to verify that the Hydrologic Sampler outputs reproduce the probability distribution inputs the user defined in a Hydrologic Sampling alternative used for a Flood Risk Analysis (FRA) simulation. Therefore, in HEC-WAT, certain Hydrologic Sampler results are available within the HEC-WAT Framework for comparison to those probability distributions, and the user is encouraged to examine these after a simulation has successfully computed. Results can be accessed using two methods: viewing Hydrologic Sampler Results menu items (accessed from the HEC-WAT menu bar), review Hydrologic Sampler Results; and using the more generalized Output Variable Editor (accessed from the HEC-WAT toolbar), review the HEC-WAT User Manual available here: HEC-WAT Documentation.
A generated flood event includes a maximum flow at each location, with specified correlation between primary locations, a shape set to transform these maximum flows into hydrographs, and a date to place the event in time. Hydrologic Sampler results (accessed from the Results menu, from the HEC-WAT menu bar) include plots and tables that allow the user to verify each aspect of this flood event generation as follows:
- A histogram of sampled event dates, for comparison to the input flood season distribution.
- Sampled annual maximum flows, for comparison to input frequency curves.
- Correlations between maximum flows at different primary flow locations, for comparison to the input correlation matrix.
- Shape set frequencies, for comparison to the combination of input weights and exceedance probability filters.
The user may also choose to view individual event hydrographs, having chosen a realization, lifecycle and event number. Finally, the Hydrologic Sampler output file is available for the lifecycle of that chosen event if the prerequisite steps have been completed (review Prerequisites).
Due to the nature of Monte Carlo simulations, the results will match the input distribution more closely as the total number of simulated events increases. A poor match may occur when few events are computed (resulting either from few events per realization, or few realizations). Any obvious failures to match input distributions, given adequate simulated events, should be explored more closely.