High spatial resolution hydrological modeling on the Hydrologic Rainfall Analysis Project (HRAP) grid

As a contributor to the National Oceanic and Atmospheric Administration's (NOAA) National Weather Service (NWS) Office of Hydrologic Development (OHD) Hydrology Test Bed (HTB), we developed the capability to run hydrological and land surface models (e.g., Sacramento-Heat Transfer/SNOW17 and Noah models) within the NASA LIS modeling system on the high-resolution HRAP (Hydrologic Rainfall Analysis Project) grid used by the NWS. The HRAP grid is notable in that it covers the entire CONUS at a high spatial resolution (4km), and in that it serves as the main hydrological modeling grid for the NWS. It is the implementation of these modeling and supporting interpolation systems and data files within LIS which enables CONUS-wide execution at this resolution. In support of this effort, we made improvements to model physics and data assimilators, and conducted a hydrologically-oriented evaluation of the forcing data.

Because we are targeting high spatial resolution modeling, one thrust of this research effort is to implement an elevation adjustment scheme for downscaling the 30-year, 1/8th degree NLDAS forcing to the HRAP grid. As part of this, we investigated the temporal and spatial characteristics of lapse rate versus temperature using a total of 989 in-situ stations in the western United States. The lapse rate can be quantitatively predicted from air temperature via a linear regression relationship. In combination with high-resolution digital elevation map (DEM) data, this regression equation and the resulting lapse rates will support the elevation adjustment of NLDAS temperature forcing data. Validation against gridded PRISM and in-situ temperature data is planned, as is a comparison against existing 1/8th degree NLDAS temperature data.

The 30-year (1979-2008) retrospective run on the HRAP grid was performed with temperatures adjusted using both a standard (constant) lapse rate, as well as lapse rates derived with the aforementioned regression equation. After taking into account spin-up processes, a 30-year model climatology was generated to support hydrological applications. Drought and flood monitoring is further supported through execution of real-time runs.