Ensemble methods for neural network-based weather forecasts
This work addresses uncertainty estimation in weather forecasting for meteorologists, but it is incremental as it applies existing ensemble methods to neural networks.
The study transformed a deterministic neural network weather forecasting system into an ensemble system using four methods, with retraining yielding the highest improvement in ensemble mean forecasts, though neural network skill remained lower than state-of-the-art numerical weather prediction models.
Ensemble weather forecasts enable a measure of uncertainty to be attached to each forecast, by computing the ensemble's spread. However, generating an ensemble with a good spread-error relationship is far from trivial, and a wide range of approaches to achieve this have been explored -- chiefly in the context of numerical weather prediction models. Here, we aim to transform a deterministic neural network weather forecasting system into an ensemble forecasting system. We test four methods to generate the ensemble: random initial perturbations, retraining of the neural network, use of random dropout in the network, and the creation of initial perturbations with singular vector decomposition. The latter method is widely used in numerical weather prediction models, but is yet to be tested on neural networks. The ensemble mean forecasts obtained from these four approaches all beat the unperturbed neural network forecasts, with the retraining method yielding the highest improvement. However, the skill of the neural network forecasts is systematically lower than that of state-of-the-art numerical weather prediction models.