Ensemble Kalman Filter Assimilation of All-Sky Microwave and Infrared Brightness Temperatures for Improved Regional-Scale Tropical Cyclone Analyses and Forecasts

All-sky brightness temperatures (BTs) from microwave imagers/sounders and GOES ABI infrared, as well as other in situ and remotely sensed data, are assimilated into the convection-permitting Weather Research and Forecast (WRF) model. Through use of flow-dependent error covariances in the PSU Ensemble Kalman Filter, infrared BT assimilation well constrains cloudy and clear areas and provides increments to precipitation, while assimilating both low- and high-frequency microwave BT offers better constraint of the location and amount of liquid and ice water contents (independent from each other) as well as other dynamic and thermodynamic state variables.

Adaptive observation error inflation (AOEI) prevents representativeness errors (cloud dislocations) from producing erroneous analysis increments: observations with much larger innovation compared to the ensemble-estimated background error variance have their assumed error inflated to this value. Also, adaptive background error inflation (ABEI) empirically inflates prior covariances to promote cloud and convection formation in regions of high BT (clear-sky) EnKF prior but low BT (cloudy-sky) observations.

The CRTM is the forward operator for both infrared and microwave, but for microwave we use a new cloud scattering lookup table built consistent with the specifications of the chosen model microphysics scheme (WSM6), except snow spherical particles are replaced with non-spherical particles. Replacing microphysics-specified spherical particles is necessary to provide better cross-channel consistency in the simulated BTs relative to observation climatology.

Results from Hurricane Harvey (2017) and other recent case studies show promise in the synergies of assimilating both ever-present geostationary infrared observations and the occasional microwave observations from low-Earth orbit.