Wind gust refers to a sudden increase in wind speed. It is measured as the forecasted maximum hourly wind gust. Due to their highly variable spatial and temporal characteristics, wind gusts are particularly difficult to predict.
The surface wind gust metric predicts wind gust magnitudes through a machine learning-based parameterization scheme. Parameterization is a technique that oversimplifies complex processes to expose the physical effects of a given process.
Wind gust percentile expresses a forecasted wind gust value in relation to the historical distribution of wind gust observations or modeled values at a given location. A percentile rank indicates what proportion of historical wind gust values fall at or below the forecasted value. For example, a wind gust at the 90th percentile means the forecast exceeds 90% of historically observed gusts at that location and time of year.
Percentile-based values provide context that raw wind speed alone cannot. A gust of 45 mph may be unremarkable in a coastal corridor with a history of strong onshore flow, but highly anomalous in a sheltered inland valley. By framing a forecast against local historical norms, wind gust percentiles help distinguish genuinely unusual events from typical variability.
In Technosylva's platform, wind gust percentiles support identification of future weather events that fall outside the normal range for a given area, which is relevant to PSPS threshold evaluation and operational risk assessment.