Very short-term QPFs for summer convection over complex terrain areas using the NCAR RTFDDA system: a comparison with NCEP NAM and RUC operational forecasts in New Mexico and Arizona

NCAR and the Army Test and Evaluation Command (ATEC) have jointly developed a real-time rapid-cycling four-dimensional data assimilation (FDDA) and forecasting (RTFDDA) system. This system has been deployed, and is running operationally at five Army testing ranges. The RTFDDA system, built upon the MM5 and WRF models, differs from the NCEP operational national models (NAM and RUC) in terms of its ability to perform continuous FDDA and produce forecasts from dynamically and diabatically "spun-up" analyses. This ability is particularly important for short-term forecasts of weather and precipitation over complex terrain, where models with cold-start can suffer greatly from dynamic and physical adjustment. This paper compares the verification of precipitation analyses and forecasts (6 - 12h) of RTFDDA with the NAM and RUC models for summer convection over New Mexico and Arizona, where complex terrain dominates and frequent afternoon convection occurs (26 rain days out of 31 in August 2005). The precipitation forecasts from the three models during August 2005 are verified against the NCEP STAGE IV precipitation analyses using grid-point statistics and an object-based approach. The verification demonstrates the significant advantage of the RTFDDA system in forecasting the orographically induced convection, relative to the cold-start NAM and RUC models. Compared to the national models, the RTFDDA system better forecasted the detailed climatography and precipitation intensities of the convective rain areas associated with mountain ranges. Also, the timing of the diurnal cycle of convective initiation (in the early afternoon), the growth and migration of cells (later on), and their dissipation (a few hours after sunset) in the region was better handled. Furthermore, the results of object-based precipitation verification show quantitative improvements of the RTFDDA system over the two national operational models for this complex terrain area.