Numerical Forecasts of Deep Convection in a Tropical Mountainous Environment in Support of the Colombian Civil Aviation Authority (Aerocivil)

A 3-year collaborative effort between NCAR, Meteostar and Aerocivil is in progress, with the goal of establishing a regional forecast system aimed towards skillful forecasts, with an emphasis on deep convection in a high-altitude tropical environment. Regional nested domains of 27-9-3 km with 63 vertical levels are centered over Columbia and designed to include all airports within the high-resolution inner nest. Of particular interest is highly resolved and timely forecasts of precipitation and fog at airport locations in Colombia. Efforts have been directed at mitigating the effects of lateral boundary conditions on the outer domain, having adequate vertical resolution to handle complex terrain as well as meeting the expectations of available computational capabilities. Two physics schemes (ThompsonMP and WSM6) suitable for the tropical coastal and mountain climates of the region were selected to perform rigorous testing of the initial forecast system. A testing period of 1 week duration was chosen during the rainy season for a comprehensive statistical analysis of surface point observations and gridded precipitation observations using the Model Evaluation Tools (MET). Baseline results were established using WRF-ARW version 3.6.1 during the first year of the project, and have since been compared to and upgraded to the latest stable releases of the WRF model. In addition to maintaining the forecast system, a 10 member multi-physics ensemble is also being developed to examine sources of uncertainty in the forecast. Diagnostic and objective verification for both the forecast system and ensemble will be presented.