Wind energy resource assessment typically requires numerical modelling, but this is too computationally intensive for multi-year timescales. Increasingly, unsupervised machine learning techniques are used to identify small numbers of representative weather patterns that can help simulate long-term behaviour. Here we develop a novel wind energy workflow that for the first time combines the weather patterns from unsupervised clustering with a numerical weather prediction model (WRF) to obtain efficient and accurate long-term predictions of wind farm power and downstream wakes. We use ERA5 reanalysis data and, for the first time, compare clustering on low altitude pressure and wind velocity, a more relevant variable for wind resource assessment. We also compare varying domain sizes for the clustering. A WRF simulation is run at each cluster centre and the results aggregated into a long-term prediction using a novel post-processing technique. We consider two case study regions and show that our long-term predictions achieve excellent agreement with a year of WRF simulations in 2% of the computational time. The most accurate results are obtained when clustering on wind velocity. Clustering over a Europe-wide domain is sufficient for predicting wind farm power output, but clustering over smaller domains is required for downstream wake predictions. Our approach facilitates multi-year predictions of power output and downstream farm wakes, by providing a fast, accurate and flexible methodology applicable to any global region. Moreover, this constitutes the first tool to help mitigate effects of wind energy loss downstream of wind farms, via the identification of optimum wind farm locations.

The trustworthiness of neural networks is often challenged because they lack the ability to express uncertainty and explain their skill. This can be problematic given the increasing use of neural networks in high stakes decision-making such as in climate change applications. We address both issues by successfully implementing a Bayesian Neural Network (BNN), where parameters are distributions rather than deterministic, and applying novel implementations of explainable AI (XAI) techniques. The uncertainty analysis from the BNN provides a comprehensive overview of the prediction more suited to practitioners’ needs than predictions from a classical neural network. Using a BNN means we can calculate the entropy (i.e. uncertainty) of the predictions and determine if the probability of an outcome is statistically significant. To enhance trustworthiness, we also spatially apply the two XAI techniques of Layer-wise Relevance Propagation (LRP) and SHapley Additive exPlanation (SHAP) values. These XAI methods reveal the extent to which the BNN is suitable and/or trustworthy. Using two techniques gives a more holistic view of BNN skill and its uncertainty, as LRP considers neural network parameters, whereas SHAP considers changes to outputs. We verify these techniques using comparison with intuition from physical theory. The differences in explanation identify potential areas where new physical theory guided studies are needed.