Spatial forecast dependence has limited value in single-period dispatch
The Operational Value of Spatial Dependence in Renewable Forecast Scenarios for Single-Period Economic Dispatch: A Controlled Ablation Study
Machine Learning
Summary
Power grid operators rely on forecasts of renewable energy to decide how much electricity to produce. This study shows that knowing how different forecast locations relate to each other (spatial dependence) adds very little benefit when planning electricity dispatch for a single period. The authors used real data from European grids and tested various scenarios, finding that the extra accuracy gained from modeling these spatial relationships barely reduces costs. Instead, techniques focused on improving forecasts for decision-making directly provide much more value.
What this means in practice
- •For grid operators: Evaluate spatial correlation models in forecast scenarios by their impact on dispatch costs, prioritizing decision-focused training over complex dependence modeling.
- •For renewable forecast vendors: Focus model improvements on decision-value gains rather than on modeling spatial correlations among forecast sites to better meet grid operator needs.
Authors
Jayakumar Manoharan
Abstract
Renewable forecasts are evaluated by statistical skill (e.g., CRPS), but grid operators pay for realized dispatch cost. We diagnose what drives dispatch value in a single-period newsvendor-style economic dispatch using real public data from two European transmission systems (CWE, DE-4TSO). Spatial coherence across forecast sites falls below the pre-specified 1% practical-significance threshold: a controlled ablation holding per-zone marginal forecasts bit-identical and varying only cross-zone dependence (10 configurations, 3 seeds, paired-bootstrap confidence intervals) shows a coherence gain of at most 0.64% of dispatch cost, indistinguishable from zero in 3 of 10 configurations, reached only under an unrealistic 8-fold forecast-error stress test. Decision-focused training, an established paradigm in this venue, delivers a robust 2.82-5.19% gain. A parametric Gaussian-copula approximation matches the empirical copula at realistic error magnitudes but performs worse than no dependence under extreme stress. A single-seed sweep shows that a 12% energy-score gain changes cost by less than 0.1%. Results characterize this single-period dispatch class; a lightweight four-period extension supports the same conclusion. For this dispatch class, spatially-correlated scenario generation provides limited operational value on its own; grid operators and forecast vendors should instead evaluate dependence models by downstream decision value and prioritize decision-focused training.