Human forecaster discussions improve tropical cyclone intensity predictions

Quantifying the added value of human forecaster discussions for tropical cyclone intensity prediction

Machine LearningComputation and Language

Summary

Predicting how strong a hurricane will get is very hard. This study found that the written discussions by expert forecasters add useful information to computer models that predict hurricane strength. Using both the forecasters’ discussion text and computer data improved 24-hour wind speed forecasts for tropical cyclones. Words like "intensify" and "eyewall" in the discussions were especially helpful for predictions. The research shows expert judgment can complement numerical data to better forecast hurricane intensity, though these results are retrospective and not for immediate use.

What this means in practice

  • For weather forecasting teams: Enhance 24-hour tropical cyclone intensity forecasts by incorporating forecaster discussion text alongside numerical predictors to reduce wind speed prediction errors.
  • For emergency management planners: Use improved hurricane intensity predictions that combine expert discussions and model data to better anticipate tropical cyclone impacts for preparedness planning.

Authors

Ilia Azizi, Frederick Iat-Hin Tam, Marc-Olivier Boldi, Valérie Chavez-Demoulin, Tom Beucler

Abstract

The National Hurricane Center routinely releases Tropical Cyclone Discussions that offer insights into its tropical cyclone forecasts. This study examines the added value of these discussions for intensity forecasting beyond a set of numerical predictors. For 637 Atlantic and eastern North Pacific tropical cyclones from the 2002-2023 hurricane seasons, we analyze the discussion text and its combination with 13 hurricane environmental predictors from SHIPS and ERA5 10-meter wind fields, for forecasting intensity changes at lead times of 24 to 120 hours using three modeling methods. In one approach, an end-to-end deep learning framework for the text shows that the discussions alone offer comparable value for 24-hour intensity forecasts to the SHIPS predictors. Adding discussion information to the SHIPS predictors in an ensemble of five such models reduces the mean absolute error of 24-hour maximum sustained wind forecasts for 2022-2023 cyclones from 9.89 to 8.90 kt, with a confidence interval that excludes zero when whole storms are resampled. The ERA5 wind fields have a minor effect on mean absolute error in wind predictions, and model rankings differ across predicted variables, lead times and model types. We provide prediction intervals for these forecasts, whose coverage falls below the 90% target during rapid intensification. Analysis of the text shows that words related to intensity and intensity change (such as intensify, weaken, eyewall and category) are important for the model predictions. This retrospective study shows the added value of the forecasters' discussions and suggests that multi-modal learning can help understand the predictive contribution of expert human judgment. The reanalysis data differ from real-time analyses, and the discussions are issued at least three hours after the valid time of the other inputs, so these results should not be interpreted operationally.