Supporting the NOAA National Water Model Through High-Performance, Physics-based Modelingand Data Assimilation
Impacts-based forecasting of hazardous weather and water events requires
- Accurate and timely situational awareness
- Reliable and consistent forecast products
- A comprehensive geospatial intelligence framework to geo-locate risks and hazards
- An effective communications framework to support the deliverable of actionable information to the public and decision makers
To enable these capabilities, NOAA has undertaken a major effort to improve its forecast services through the development of a new National Water Model (National Water Model: http://water.noaa.gov/about/nwm) which runs on NOAA’s centralized Weather and Climate Operational Supercomputing System (WCOSS). Version 1 of the NOAA National Water Model utilizes the community WRF-Hydro modeling system (https://www.ral.ucar.edu/projects/wrf_hydro) and was co-developed for operational implementation by the NOAA Office of Water Prediction, the NOAA National Water Center, NCAR/RAL, the USGS, CUAHSI and other members of the academic community.
The new National Water Model utilizing WRF-Hydro became operational on the NOAA WCOSS supercomputer on Aug. 16, 2016, delivering streamflow forecasts on the 2.7 million river reaches of the USGS NHDPlusv2 hydrography dataset as well as gridded analyses of a host of other hydrologic variables across the Nation. National Water Model version 1.0 operational goals were to:
- Provide consistent, operational forecast streamflow guidance for currently underserved locations
- Generate, regularly cycling, spatially continuous estimates of hydrologic states for the nation (snowpack, soil moisture, channel flow, major reservoir inflows, flood inundation)
- Seamlessly interface real-time hydrologic products into and advanced geospatial intelligence framework
- Implement an Earth system modeling architecture that permits rapid model evolution of new data, science and technology
These goals were accomplished through the rapid development and implementation of the WRF-Hydro modeling framework onto operational NOAA supercomputers. The NWM runs in four configurations: 1) Analysis and assimilation mode, 2) Hourly short-range 15 hour deterministic forecasts, 3) Daily medium-range, ten day deterministic forecasts, 4) 16x daily, long-range 30-day ensemble forecasts. Forecast products from the National Water Model are now being served through a host of official NWS forecast outlets including the National Water Center, River Forecast Centers, Weather Forecast Offices and the National Centers for Environmental Computing. In addition to national streamflow maps and animations, experimental visualization products include:
Supporting the NOAA National Water Model Through High-Performance, Physics-based Modeling and Data Assimilation
A comprehensive, open source model evaluation system, called ‘Rwrfhydro’, was developed to support a wide range of model verification and evaluation activities. As the NWM development effort has proceeded and expanded, the Rwrfhydro toolkit has grown to enable a large variety of analyses of different hydrologic variables. Together these analyses are beginning to provide a more holistic evaluation of the full modeling system. Since operational testing began in the late spring of 2016, real-time verification statistics have been tracked and are now being used as a baseline for future model upgrade and development efforts.
Based on initial verification results from the operational NWM and ongoing research within RAL and its collaborators in the academic and agency communities, a number of model upgrades and improvements are being worked on which will be incorporated in future NWM versions. These research and development activities include:
- Improvements to the land surface physics controlling terrestrial hydrologic behavior
- Enhanced parameter estimation, multi-variate calibration and parameter regionalization
- Improvements in channel hydraulics and in inundation representation
- Development of regional, hyper-resolution nests of the NWM for detailed, extreme event forecasting
- Enhancements to the baseline streamflow data assimilation system
- These and other improvements to the NWM will be transitioned into operations through two upgrades of the model on the NOAA supercomputer during FY2017.
