Using Ensemble of Neural Networks to Learn Stochastic Convection Parameterizations for Climate and Numerical Weather Prediction Models from Data Simulated by a Cloud Resolving Model
Figure 11
Decadal boreal winter mean cloudiness (CLD, in fractions) distribution for the CAM (upper panel) and CAM-NN (middle panel) runs over tropical Pacific region (with the area size of and the following coordinates: 150° E < lon < 90° W shown with the 10° interval; 15° S < lat < 15° N shown with the 3° interval). The lower panel shows the corresponding NCEP reanalysis decadal mean distribution. The TOGA-COARE location, for which the pseudo-observations were generated and the NN ensemble was trained, is shown by a star in the middle panel. The contour interval is 0.05°C.
