Retrieving latent heating profiles from GOES for precipitation forecasts (Invited)

Accurate short-term precipitation forecasts require both a good understanding of precipitation processes and the initial state observed by precipitation-related measurements. Even when forecast models simulate precipitation realistically, often the location or intensity of the precipitation needs to be adjusted using observational data for the best results. A common approach to initializing convection in forecast models is to apply observed latent heating during a spin-up period. Recently, methods to derive latent heating from Geostationary Operational Environmental Satellite (GOES) Advanced Baseline Imager (ABI) and Geostationary Lightning Mapper (GLM) were developed using artificial intelligence techniques, and applying these to convective initialization showed improvements in short-term precipitation forecasts. This study advances the approach by understanding and quantifying uncertainties that can arise from GOES retrievals of latent heating.

Latent heating, which is a temperature change resulting from phase changes of the water substance in clouds, is sensitive to the microphysical scheme used in the forecast model. Different microphysical schemes can lead to different magnitudes and vertical structures of latent heating. It is important to understand these differences because cloud-resolving model simulations provide our best estimate of latent heating, and are used to derive the retrievals from GOES. Therefore, this study examines vertical structures of latent heating and their relationship to the evolution of different hydrometeors in different microphysical schemes. The potential impacts on precipitation forecasts will be discussed in the context of model sensitivity to latent heating. Furthermore, the GOES retrieval results will be compared with other latent heating products from the Global Precipitation Measurement Mission (GPM) satellite.