Model Verification Using Gaussian Mixture Models

Verification methods for high-resolution forecasts have been based either

on filtering or on objects created by thresholding the images.

The filtering methods do not easily permit the use of

deformation while threshold-based objects are subject to association errors.

In this paper, we introduce a new approach that breaks down the

observed field into a mixture of Gaussians (the "objects") and

reconstruct the model forecast using scaled and displaced versions

of these Gaussians. We discuss the advantages of this method in

terms of the traditional filtering or object-based methods and

interpret resulting scores on a standard verification dataset.