Bridging the data gap: A new approach for reconstructing historic typhoon intensity using single station observations

This study addresses the challenge of assessing the minimum central pressure of typhoons in the northwest Pacific before the mid-20th century, crucial for understanding typhoon disaster risk in the long-run. It makes a methodological contribution by developing an approach to estimate historical typhoon intensities using limited observational data. Our method integrates two models: the Holland pressure field model to infer central pressure of typhoons and the minimum cumulative error model to adjust for observational inconsistencies across different periods. This combined approach achieves low Mean Absolute Errors (MAE) of 10 hPa, 9 hPa, and 12 hPa across training, testing, and re-testing datasets. Sensitivity analysis reveals the model’s robustness, maintaining an MAE of 11 hPa and 12 hPa for various typhoon locations and sizes in the test set. The model’s reconstructions align well with the historical records of the Hong Kong Observatory (HKO), especially the 1906 typhoon data, confirming its reliability. By accurately reconstructing typhoon intensities from scarce data, this method offers a valuable framework for distinguishing natural variability from human-induced climate change effects on typhoon intensity, enhancing the analysis of historic catastrophic typhoons.