The Spring 2014 Mesoscale Ensemble Prediction System "Dust Offensive"

Abstract:

The Mesoscale Ensemble Prediction System (MEPS) at the Air Force Weather Agency (AFWA) is a 10-member ensemble run on a 20-km hemispheric domain and 4-km domains in regions of interest. In the Southwest Asia (SWA) regional domain, dust forecast products from MEPS are of particular interest. Over the past few years, subject matter experts at AFWA have acquired and implemented datasets and developed a soil moisture algorithm that have improved skill scores of dust forecasts in the AFWA GOddard Chemistry Aerosol Radiation and Transport (GOCART) module within the Weather Research and Forecast with Chemistry (WRF-Chem) model. The aforementioned datasets include, but are not limited to, the Desert Research Institute (DRI) dust source region and a dynamic 8-day Leaf Area Index (LAI) vegetation mask. We then tested these individual datasets as ensemble perturbations in all ten MEPS test members during a three-week “dust offensive” test over SWA in spring 2014. Remote sensing specialists at AFWA meticulously documented the locations, duration, and intensity of numerous dust events over the SWA domain during the three weeks. These data were then used for subjective verification of each individual MEPS member and of the individual perturbations within members. Results from the subjective verification showed that individual MEPS members with the DRI dust source region significantly outperformed members that used the standard Ginoux dust source region. The other individual perturbations tested were determined to have a neutral effect (i.e., neither degrading nor improving skill) on individual members during the three-week period. Thus, the DRI dust source region is now used in the majority of MEPS members in both the 4-km SWA domain and the 20-km hemispheric domain. The other individual perturbations are now utilized in a minority of MEPS members. However, additional testing is still needed over other domains to determine if the improved dust forecasting skill has spatial consistency and to ensure that other pertinent forecast products from MEPS are not being degraded.