Improving the Hurricane Outage Prediction Model by including tree species

Abstract

Hurricanes can be a major threat to electrical power systems, often resulting in costly repairs and lengthy restoration times. The Hurricane Outage Prediction Model (HOPM) predicts the location and number of outages for a major coastal Gulf of Mexico utility company up to 5 days before a hurricane makes landfall. This model can help electrical utilities to improve their resource allocations and potentially shorten restoration times. The goal of this study is to evaluate whether the accuracy of the HOPM can be improved by including information on tree species. Our results demonstrate that the model accuracy increased by ∼3% when tree species were included. The most important tree species in our study region were Sweetgum and Loblolly Pine. As the relative abundance of these tree species increased, the number of outages tended to increase. In contrast, increases in the relative abundance of Water Oak and Chestnut Oak did not result in more outages. This suggests that certain oaks may be more resistant to uprooting or snapping when there are strong winds. Therefore, the inclusion of tree species variables in the model provides a means for capturing the spatially varying vulnerability of the trees to strong winds. Even though including tree species improved the accuracy of the model, information about the frequency of tree trimming is more important. Accurate prediction of the location and number of power outages requires information on both the meteorological hazard and the vulnerability of the vegetation.