Evaluation of Independent Stochastic Perturbed Parameterization Tendency (iSPPT) Scheme on Ensemble TC Intensity Forecasts Using HWRF

Despite significant progress in numerical modeling, tropical cyclone (TC) intensity prediction remains a challenge. Ensemble forecasting provides a methodology to better understand the predictability of TC intensity, which can be used to represent both initial condition uncertainties and uncertainties in model formulation. Previous studies have considered TC initial condition uncertainties by perturbing the initial conditions or by employing ensemble data assimilation; however, initial-condition uncertainty alone is not sufficient to produce reliable ensemble forecasts; uncertainty in model formulation is also necessary.

Stochastic parameterization schemes (e.g., stochastically perturbed parameterization tendency (SPPT)) have become a widely-used method of representing uncertainties in model formulation. This work employs the independent SPPT (iSPPT) scheme to the hurricane WRF (HWRF) ensemble prediction system to improve TC ensemble forecasts and to better understand the predictability of TC intensity. In contrast to SPPT, which perturbs the total parameterization tendency, iSPPT individually perturbs each parameterization (i.e., PBL, radiation, microphysics and cumulus) tendency with a unique stochastic pattern. As a consequence, iSPPT can evaluate the impact of uncertainties associated with a particular physical process. Preliminary results indicate that applying stochastic perturbations to the PBL parameterization results in larger 72-h TC intensity ensemble standard deviation compared with control ensemble forecasts.