Physical Impacts of Stochastic Perturbed Parameterizations on Microphysical Processes in an Adaptive Habit Model Ensemble: A Case Study of Heavy Rainfall in New York State

Abstract:

A two-dimensional stochastic perturbed parameterization (SPP) is implemented into the Weather Research and Forecasting (WRF) Model version 3.7.1 to individually perturb rates of vapor deposition onto cloud ice and aggregate snow, collection of snow by rain, and riming of droplets by graupel within an adaptive habit bulk microphysics scheme. Identification of processes, if any, that considerably increase forecast spread, is extremely important for high-intensity rainfall forecasts due to the associated potential of societal impacts. Tuning experiments with slightly different spatial, temporal, and amplitude autocorrelation parameters were conducted to elicit those most conducive to the production of physically sound ensemble spread (i.e., standard deviation). These parameters are used, along with random number seeds, to generate a stochastic pattern that perturbs the process rate, thereby producing a microphysical ensemble. The performance of each of the four, 10-member ensembles is verified against the New York State (NYS) Mesonet observations of temperature, wind, and precipitation and also atmospheric soundings launched in Buffalo, Albany, and New York City. This methodology is used to investigate the impact of process rate perturbations on heavy rainfall with a tropical moisture source that impacted NYS from 29-30 October 2017. With the exception of perturbations to riming of droplets by graupel, the greatest spread within each microphysical process rate is the rate that is perturbed throughout the simulation. Additionally, perturbations to riming of droplets by graupel lead to an increased frequency of relatively high QPF spread (> 11 mm) in the quantitative precipitation forecast (QPF). Interestingly, certain regions of NYS are impacted differently by each respective process rate perturbation. For example, QPFs within the Catskill Mountains demonstrate considerable forecast spread when riming of droplets by graupel is perturbed. Further investigation into the reasons driving the unique response of graupel perturbations will be explored in this presentation.