Abstract One of the major uncertainties in atmospheric modeling is parametric uncertainty. It is important to infer appropriate parameters in various parameterizations from observation. Despite previous efforts on the calibration of parameters in atmospheric models, there is no existing work that comprehensively calibrates parameters using geostationary satellite observations although they are of paramount importance to monitor tropical cyclones. In this study, we estimate the posterior distribution of parameters of a meso-scale atmospheric model using brightness temperature observations from a geostationary satellite toward the improvement of the simulation of tropical cyclones. With the aid of an image-processing inspired evaluation index and machine-learning-based surrogate models, we developed a method to calibrate model parameters of a meso-scale atmospheric model by geostationary satellite observations. We found that parameters in cloud microphysics and boundary layer schemes could be efficiently estimated by geostationary infrared satellite observation. The estimated posterior distribution of parameters not only improves the accuracy of the prediction of satellite images but also partly reduces errors in the prediction of tropical cyclone intensity. We demonstrate the potential of adjusting multiple parameters based on satellite data and its implications of model development to improve the accuracy of the simulation of tropical cyclones. Although we did not consider uncertainty in initial conditions, this work would be extended to meso-scale ensemble forecasting systems which explicitly consider uncertainty in both initial conditions and model parameters.
Hirose et al. (Wed,) studied this question.