Automated Thunderstorm Forecasting Validation
Abstract
Increased use of long-duration unmanned aerial vehicles (UAVs), as well as other factors, has led to a growing demand on the naval meteorology and oceanography (METOC) community for mission-tailored meteorological support. To creatively meet this need, the Naval Research Laboratory (NRL) Marine Meteorology Division has developed methodologies to process output from the Navy Global Environmental Model (NAVGEM). Automated products replicate human weather forecasters written and unwritten rules to produce guidance for icing, turbulence, and thunderstorms. This is a force multiplier, as it takes the skilled, but few in number, human forecasters from being in the loop to being over the loop. Much like the National Weather Service Aviation Weather Centers World Area Forecast System, utilizing Global Forecast System (GFS) output, NRLs existing automated thunderstorm algorithm uses convective precipitation and cloud thickness variable output to compute areas of expected thunderstorms. This thesis examines the current NRL method, and a few alternatives, in the Geostationary Satellite East coverage area and in two operationally relevant locations. Partial validation of forecasts is conducted using statistical analyses and comparison with satellite lightning observations. Results indicate that NRLs current methodology performs the best overall; new methodologies slightly lag behind.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 2020
- Accession Number
- AD1126790
Entities
People
- Joshua D. Carter
Organizations
- Naval Postgraduate School