Assimilating HDOB Data in COAMPS® NAVDAS for Hurricane Earl

HDOB data are high-density, high-accuracy flight-level observations from NOAA and USAF hurricane reconnaissance aircraft. These observations include 30-sec averages of pressure, geopotential height, temperature, dewpoint temperature, wind speed, and wind direction reported every 30 sec, as well as extrapolated surface pressure, maximum 10-sec average wind speed, and maximum 10-sec average surface wind speed and rain rate from the Stepped Frequency Microwave Radiometer (SFMR).

These data are provided at a much higher density than conventional automated aircraft data from commercial airlines, so further averaging is applied to level-flight data to yield 2.5 min or 25 km spacing. The HDOB pre-processing also included checking for egregious errors (e.g., constant values, spikes, missing data) prior to averaging. An algorithm was also devised to exclude HDOB data in the center of tropical systems, based on the magnitude of the along-track gradient in estimated surface pressure. Details of this pre-processing will be presented at the conference.

Experiments were run with the U.S. Navy's mesoscale 3DVAR data assimilation (NAVDAS—NRL Atmospheric Variational Data Assimilation System) and numerical weather prediction model (COAMPS®--Coupled Ocean/Atmosphere Mesoscale Prediction System) for Hurricane Earl in late August-early September 2010. The control experiment was configured to mimic the operational run, with additional experiments adding HDOB data with and without the tropical cyclone synthetic observations. Observation impact experiments were also run with a one-nest version of COAMPS®; these showed a dramatic beneficial per-ob impact for the HDOB data. Details of the experiments and observation impacts will be shown.