Evaluating the Impact of Improvement in the Horizontal Diffusion Parameterization on Hurricane Prediction in the Operational Hurricane Weather Research and Forecast (HWRF) Model

Abstract

Abstract Improving physical parameterizations in forecast models is essential for hurricane prediction. This study documents the upgrade of horizontal diffusion parameterization in the Hurricane Weather Research and Forecasting (HWRF) Model and evaluates the impact of this upgrade on hurricane forecasts. The horizontal mixing length Lh was modified based on aircraft observations and extensive idealized and real-case numerical experiments. Following an earlier work by the first two authors, who focused on understanding how the horizontal diffusion parameterization worked in HWRF and its dynamical influence on hurricane intensification using idealized simulations, a series of sensitivity experiments was conducted to simulate Hurricane Earl (2010) in which only Lh was varied. Results from the Earl forecasts confirmed the findings from previous theoretical and idealized numerical studies, in that both the simulated maximum intensity and intensity change rate are dependent on Lh. Comparisons between the modeled and observed structure of Hurricane Earl, such as storm size, boundary layer heights, warm-core height and temperature anomaly, and eyewall slope, suggested that the Lh used in the HWRF Model should be decreased. Lowering Lh in HWRF has a positive impact on hurricane prediction based on over 200 retrospective forecasts of 10 Atlantic storms. Biases in both storm intensity and storm size are significantly reduced with the modified Lh.