Use of a Genetic Algorithm to Optimize a Numerical Weather Prediction System

Abstract An important step in an ensemble Kalman filter (EnKF) algorithm is the integration of an ensemble of short-range forecasts with a numerical weather prediction (NWP) model. A multiphysics approach is used in the Canadian global EnKF system. This paper explores whether the many integrations with different versions of the model physics can be used to obtain more accurate and more reliable probability distributions for the model parameters. Some model parameters have a continuous range of possible values. Other parameters are categorical and act as switches between different parameterizations. In an evolutionary algorithm, the member configurations that contribute most to the quality of the ensemble are duplicated, while adding a small perturbation, at the expense of configurations that perform poorly. The evolutionary algorithm is being used in the migration of the EnKF to a new version of the Canadian NWP model with upgraded physics. The quality of configurations is measured with both a deterministic and an ensemble score, using the observations assimilated in the EnKF system. When using the ensemble score in the evaluation, the algorithm is shown to be able to converge to non-Gaussian distributions. However, for several model parameters, there is not enough information to arrive at improved distributions. The optimized system features slight reductions in biases for radiance measurements that are sensitive to humidity. Modest improvements are also seen in medium-range ensemble forecasts.