The Hydrologic Sampler is a plug-in built for the HEC-WAT (Hydrologic Engineering Center's (HEC) Watershed Analysis Tool) software.  For further details on HEC-WAT review the HEC-WAT User Manual available here: HEC-WAT Documentation. The purpose of the Hydrologic Sampler is to allow users to generate the hydrologic time series necessary for a Flood Risk Analysis (FRA) compute in HEC-WAT. The FRA compute is a Monte Carlo simulation that evaluates thousands of randomly sampled flood events within a watershed. Each flood event is initiated with time series of either flow or precipitation that are used as input boundary conditions in subsequent watershed models such as HEC-RAS (River Analysis System), HEC-HMS (Hydrologic Modeling Systems) or HEC-ResSim (Reservoir System Simulation).

The Hydrologic Sampler uses pseudo random number generation to create flood events defined by flow or precipitation input time series required for an FRA simulation. For flow sampling, two methods for generating hydrographs are available: Correlated Flow Frequency Curves and Bootstrapping Historical/Synthetic Basin-wide Events. For precipitation sampling, one method of generating hyetographs is available: Basin Average Frequency Curve. In each case, the Hydrologic Sampler randomly samples the hydrology, generating as many events (sets of hydrologic time series such as hydrographs or hyetographs) as necessary for the FRA compute. Each set of generated time series is available to the other software applications in the HEC-WAT program.

HEC-WAT distinguishes between and separates random sampling of natural variability (aleatory uncertainty) from knowledge uncertainty (epistemic uncertainty) using a nested Monte Carlo sampling structure, with inner and outer loops. Natural variability is sampled at an event level (inner loop), while knowledge uncertainty is sampled at a realization level (outer loop); where a realization contains enough flood events across the feasible magnitude range to fully characterize the natural variability. A realization contains many HEC-WAT "analysis periods" or lifecycles, with an event placed in each year of a lifecycle. At the event level, the Hydrologic Sampler generates each flow or precipitation event using a random instance of natural variability, such as peak flow or total precipitation magnitude, and date. The primary knowledge uncertainty captured in the Hydrologic Sampler is in the annual maximum flow or precipitation frequency curves. This uncertainty is defined by an equivalent record length, N, to capture sampling error, and at the realization level, is randomized by re-estimating each frequency curve from an N-year bootstrap. Since each realization is based on one random instance of knowledge uncertainty, many realizations are required to capture the full distribution of possible frequency curve estimates.

The introduction to the hydrologic sampler includes the following generalized discussion topics: