Local Analysis and Prediction System (LAPS) Technology Transfer

1. Local Analysis and Prediction System (LAPS) Technology Transfer Steve Albers 1,2 Brent Shaw 3 Ed Szoke 1,2 1 – ESRL, 2 – CIRA, 3 – Weather Decision Technologies • Introduction • Used for Data Assimilation, Nowcasting, and Model Initialization • Wide variety of (~150) group & individual users • Federal and state agencies (e.g. NWS, USAF) • Private Sector • Academia (e.g. Univ of HI) • International (e.g. Taiwan CWB) • Meets NOAA Weather and Water Goals National Weather Service NOAA & CIRA personnel at ESRL have been involved in the first quasi-operational implementation of the new Weather Research and Forecast (WRF) mesoscale model within a NWS Weather Forecast Office. As one portion of the NWS Office of Science and Technology’s Coastal Storms Initiative, WRF is being run on a Linux cluster at the Jacksonville, FL, office and is being made available to forecasters directly on their operational AWIPS workstations. This project seeks to (1) demonstrate whether or not local modeling is feasible, (2) whether they add value to the forecast products, and (3) how much impact local data assimilation may have. To date, all three goals have been met with positive results. • Environmental Support for Space Launch Operations • Collaborators: Lockheed-Martin, USAF • LAPS / MM5 system Installed at both ranges (Cape Canaveral AFS, FL and Vandenberg AFB, CA) supporting space launch operations. • First operational modeling system in U.S. government running on an affordable Linux cluster and fully integrated with the AWIPS system... • Prototype for the future Weather Forecast Office! • Fire Weather Analysis and Prediction • Used for National and regional wildfire managers. • Demonstration real-time fire weather products as one of the five regional centers of the Fire Consortia for Advanced Modeling of Meteorology and Smoke (FCAMMS). High-resolution (4km grid spacing) forecast of Ventilation Index and PBL winds. Convective Available Potential Energy (CAPE) helps thunderstorm forecasts on Cape Canaveral. LAPS Analyses on AWIPS/D3D for 10km domain centered in Colorado Modular Local Data Assimilation and Forecast System Integrated within the JAX WFO 1 • Real-Time Fire Products • Ventilation Index and PBL Winds • PBL Height • Haines Index (Low, Mid, High)‏ • Fosberg Fire Weather Index • LAPS Fire Weather Index • Tabular User-Defined Point Forecasts • Standard Surface and Upper-Air Fields • Local observations include ASOS, wideband WSR-88D reflectivity and radial velocity, meteorological towers, RTAMPS balloons, SODAR, wind and RASS profilers. • Combined with national data sets such as GOES imagery, narrowband WSR-88D, METARs, and ACARS to make this an extremely robust data assimilation system. LAPS Attributes: • LAPS blends a wide variety of global, national data sets and local data sets. For example, it can combine both national surface data and local mesoscale networks. • Utilizes large-scale numerical weather prediction models as backgrounds for both analysis and prediction. LAPS analyses are used to initialize mesoscale forecast models. • Creates forecasts at any horizontal or vertical resolution needed, with output at the times and frequencies you require with many mesoscale models (MM5, RAMS, WRF). • Highly portable and runs on inexpensive hardware: desk top to laptop, using the UNIX operating system. Surface winds in the JAX area of responsibility are influenced by small scale estuaries that require mesoscale models to be effectively resolved. Shown above is verification of wind speed RMSE (left) and bias (right) from FSL’s Real Time Verification System from the LAPS-initialized WRF (blue), Eta-initialized WRF (green), and the Eta model for all 0600Z runs from 1-20 July 2003. The mesoscale models demonstrate a clear advantage, which is greater when local data assimilation is included. Triple nested MM5 domains (10, 3.3, and 1.1 km) resolve terrain-driven flow missing from national-scale models (NCEP Eta streamlines in tan are shown for comparison). BlueSky smoke dispersion model prediction using LAPS-initialized MM5 forecast for the SW US domain. The local modeling capability, combined with the LAPS diabatic initialization, shows promise for improving short range forecasts of precipitation by eliminating the model “spin up” problem with affordable computational solutions. • Collaborators • NOAA / Earth System Research Laboratory • USDA Forest Service • Rocky Mountain Research Station • Fire and Environmental Research Applications Team • Northwest Regional Modeling Consortium Reliable, inexpensive, high-performance computing solutions for local forecast offices using Linux clusters A measure of success for technology transfer is whether or not the end product is useful. This Area Forecast Discussion issued by the JAX NWS office where the LAPS hot-start initialized WRF is an indicator! Collaborators: NOAA/GSD/AB (Forecast Verification Section), NOAA/NWS/OST, NOAA/NWS/JAX WFO, Florida State University, NOAA/NOS