PRMS Model Agreement Statistic (KS Test)

Summary
A study by the U.S. Geological Survey (USGS), in cooperation with the Gulf Coastal Plains and Ozarks Landscape Conservation Cooperative (GCPO LCC) and the Department of Interior Southeast Climate Adaptation Science Center, evaluated the hydrologic response of a daily time step hydrologic model to historical observations and projections of potential climate and land cover change for the period 1952-2099. An application of the Precipitation Runoff Modeling System (PRMS) was used to develop the hydrologic simulations.  The model simulations were used to compute the potential changes in hydrologic response across the southeastern U.S. using historical observations of climate and streamflow, and 13 downscaled general circulation models with four representative concentration pathways representing a range of potential future changes in climate.
 
The PRMS simulated hydrologic response within the entire geographic study area – the model domain.  The model domain was subset into small local watersheds delineating areas expected to have a similar hydrologic response due to changes in the model inputs.  These local watersheds are called “hydrologic response units” or HRUs. 
The PRMS computes flow generated locally on each HRU for each time step. These flow components then are directed to stream segments (SEGs) for flow aggregation.  These segments connect the network of HRUs to simulate accumulated streamflow from the upstream watershed. Each HRU and SEG has a unique ID.  For each HRU and SEG, 52 summary streamflow metrics (Index of Hydrologic Alteration or IHA metrics) were calculated based on the daily flow outputs.  A description of each IHA metric may be found here (streamflow_description_table.xlsx). 
The summary information presented here shows geospatial results from a test (Kolmogorov-Smirnov or KS test) of model agreement. Results are measure of the model skill in accurately reproducing observed historical conditions.  Higher values means higher confidence in model output.  Lower values means lower confidence in model output. 
 
Purpose
The streamflow statistics were selected to describe streamflow conditions that may be most useful in defining the suitability for each river or stream to support sustaining populations of priority aquatic species across the GCPO LCC. The data presented here are intended to provide more easily accessible landscape scale summary information in support of the USGS flow modeling project.