Effects of Aerosol PSD Data Assimilation on Precipitation Prediction in Western Puerto Rico using a Cloud-Resolving Atmospheric Model

It has been shown that the assimilation of observed aerosol particle size distribution (PSD) in cloud resolving models can improve precipitation estimates in many regions, including the Caribbean. The main goal of the research presented here is to determine the effects of increasing aerosol PSD data assimilation complexity on precipitation for the purpose of improving precipitation prediction. A localized event that produced over 80 mm of precipitation near the west coast of Puerto Rico (18.25N, 67.05W) on 16 June 2013 was simulated under varying PSD scenarios using the Regional Atmospheric Modeling System (RAMS) with a two-moment explicit microphysics scheme. PSD scenarios included: (a) idealized unimodal PSD, (b) observed bimodal PSD from the Aerosol Robotic Network (AERONET), and (c) observed bimodal PSD from two different AERONET sites coupled with vertical Ceilometer data. For case (c), a three dimensional PSD distribution resulted from the combination of AERONET and Ceilometer data which was assimilated into RAMS model. Results for precipitation were compared against National Weather Service (San Juan office) NEXRAD and local weather station data to determine which method yielded the most accurate results. Scenarios (a) and (b) produced 26 and 34 mm of rain, respectively. Scenario (c), which produced 44.5 mm of rain, provided the best accumulated precipitation results when compared to NEXRAD. These results show that PSD assimilation in cloud-resolving models is necessary to improve precipitation prediction efforts.