Sensitivity of intensity and track forecasting of Hurricane Earl (2010) to single- and double-moment microphysics parameterization schemes

The impact of various single and double-moment microphysics parameterization schemes on the track and intensity forecast of hurricane Earl (2010) is examined using the Advanced Research Weather Research Forecasting model (ARW-WRF) at a cloud-resolving 4-km resolution in a very large domain. Different microphysical schemes, including Lin-Purdue, Goddard-Lin, WRF Single-Moment 6-class (WSM6), Thompson double moment, Milbrandt-Yau double-moment (MYDM), Morrison double-moment, and WRF Double-Moment 6-class (WDM6), are tested for Hurricane Earl of 2010. Among the double moment schemes, only the Milbrandt-Yau scheme predicts two moments for all hydrometeor species while others only predict part of the species with two moments. Forecasts up to 60 hours are expected, and the forecastsare initialized using either the operational Global Forecast System analyses or experimental global ensemble Kalman filter analyses.

The track forecast of Earl is found to have little sensitivity to the microphysical schemes used during the storm intensification with the track error around 60-70 km for 48-h forecasts and 170-200 km for 60-h forecasts. Larger sensitivity is found with the intensity forecast. The MYDM scheme produces the strongest hurricane with the smallest intensity error among all schemes during the intensification phase of Earl. The hydrometeors produced by the microphysical schemes differ markedly in both quantities and spatial distributions. Much more snow is generated in the forecast with Thompson and Goddard schemes. The forecast using MYDM produces the most of cloud ice while Thompson produces the least amount. Latent heating and microphysical budgets are also examined to gain more understanding of the roles of different microphysical processes in hurricane intensification.