AI weather models help time power outage crew deployments better
Forecast lead time and the timing of crew deployment for weather-driven power outages using AI weather models
Artificial Intelligence
Summary
Power utilities send crews to fix outages after storms, but calling crews early relies on less accurate weather forecasts. The authors studied two AI weather models to predict outages up to five days ahead across the US. They found forecasting accuracy drops with longer lead times, especially beyond three days, but deploying crews early with all requested workers generally helped more than waiting to deploy fewer crews later. Some prediction methods can overestimate skill if not computed properly, and the authors suggest careful calibration to improve reliability, especially for tropical cyclones.
What this means in practice
- •For utility operations teams: Optimize timing for dispatching mutual aid crews using AI weather forecast data to improve outage response.
- •For emergency management planners: Incorporate calibrated AI forecast uncertainty to better plan resource staging before severe weather hits.
Authors
Mohammad Kian Golkar, Luciano Alves de Oliveira, Mohamad Alipour
Abstract
Utilities request mutual aid crews to restore power after severe storms, and requesting early secures more crews but relies on less accurate forecasts. We examined how forecast lead time affects the value of outage predictions for crew deployment, using operational forecasts from two AI weather models, GraphCast and FourCastNet, to predict outage occurrence for counties across the contiguous United States 24 to 120 hours ahead. Forecasts were evaluated by the share of outage burden captured by the counties where crews would be staged. Captured burden declined with lead time: the loss between three and five days was significant in all analyses, whereas the loss within three days was significant only for outages that had not yet started. Early deployment with all requested crews outperformed late deployment with fewer, and we report the availability ratio at which the two break even. Antecedent land surface predictors computed at forecast valid time include weather from the forecast period, which inflates their skill by 42 to 53% and reverses their apparent advantage over forecast meteorology at long lead times. Computed correctly, they improved occurrence skill by 17.5% at 24 hours but did not significantly change captured burden, a gap that attribution analysis linked to their static component. Prediction intervals covered only 0.476 of tropical cyclone events against a nominal 0.90. Calibration by hazard class corrected most of this shortfall at an 11% cost in interval width, as a diagnostic only, because hazard class is known only after the event.