A Quantitative Method for Validating Tropical Cyclone Model Forecasts Using Satellite Microwave Observations

Mesoscale numerical weather prediction of tropical cyclones (TCs) has advanced to the point where detailed graphical presentations of eye development, vortex structure and convective banding features are now common. While these modeled structures often appear “reasonable”, they are difficult to verify. In this presentation we demonstrate a new application, in which synthetic microwave retrievals from numerical models can be quantitatively compared and evaluated against actual microwave imagery. We employ the Automated Rotational Center Hurricane Eye Retrieval (ARCHER) algorithm that was originally designed to objectively fix TC centers from multispectral satellite imagery. ARCHER objectively characterizes TC organizational features such as eyewall size/completeness and vortex spiral banding into scores, and from these also yields estimated storm intensity categories. Thus, the algorithm offers a unique tool for evaluating several critical aspects of model performance, without significant interference from brightness temperature biases in the model's associated radiative transfer scheme. We will present some preliminary results of applying ARCHER to several simulated TCs to demonstrate the prospects of this novel approach to assessing TC model forecasts.