Answers are grounded in this paper and cite their sources.
Atlantic hurricanes are among the most significant meteorological disasters affecting the United States, often causing considerable damage and loss of life and property. To prepare for the Atlantic hurricane season, the National Oceanic and Atmospheric Administration (NOAA) issues outlooks each year for the expected seasonal tropical cyclone activity. A suite of dynamical, statistical, and hybrid models supports the outlooks. Given the need to enhance predictive skill, efforts to build models of varying complexity for seasonal hurricane outlooks are necessary. This paper introduces a newly developed multiple linear regression (MLR) model for Atlantic seasonal hurricane forecasting and compares it with two hybrid models developed and operationally used at NOAA. The MLR model is constructed using observed antecedent vertical zonal wind shear anomalies between 200 hPa and 850 hPa over the tropical North Atlantic and sea surface temperature anomalies in the hurricane Main Development Region of the North Atlantic as two predictors. Overall, the MLR model outperforms the two operational hybrid models in the NOAA seasonal hurricane forecast tool suite for forecasts/hindcasts from 1982 to 2024 using June observations, although the validation period is relatively short and this may impact the robustness of skill comparisons to some extent. The MLR model predictions, incorporating the predictors from March and June 2025, contributed to the NOAA official hurricane season outlooks issued in May 2025 and the update issued in August 2025, respectively, and were consistent with the observed storm counts. The encouraging results of the MLR model suggest the importance of continued efforts to explore optimal predictors to improve the skill of the Atlantic hurricane season outlook.
No takes yet. Share an insight, caveat, or question.
Hu et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: