Tropical cyclone unit-surge surrogate modeling error metrics by training/validation set type and size and by surrogate method

This dataset includes ensemble-averaged surge-aggregated and surge-binned error metrics for a suite of tropical cyclone surge surrogate models. Surrogate models were constructed using training/validation storm track sets for:Four different training/validation sampling schemes: random, systematic, stratified random by grid, and stratified random by storm track.Five different training/validation sample sizes: 275, 486, 1267, 2048, and 4149 and three different surrogate methods.Three different surrogate methods: multilinear interpolation, Kriging, Artificial Neural Network (ANN)Four different, unitless tropical cyclone surge response surfaces defined over a 5-dimensional storm track space (landfall location, central pressure deficit, storm radius, heading, forward speed) were used to specify training/validation set surges prior to building each surrogate model. A 1.3-million storm test set was used to evaluate error metrics:Surge-aggregated mean error, root-mean-square error, median, and interquartile rangeSurge-binned (at 0.1 unit intervals) mean error, root-mean-square error, median, and interquartile rangeResults are presented in Irish et al. (2026):Irish, J. L., Bensi, M. T., Frantzis, C. D., Shao, Y., Carr, M. L., Bender, C. J., Gharagozlou, A. (in review). Storm training set selection for landfalling tropical cyclone surge surrogate modeling. Coastal Engineering. DOI: 10.1016/j.coastaleng.2026.105125.Acknowledgments: This project was funded, in part, by the US Coastal Research Program (USCRP) as administered by the US Army Corps of Engineers® (USACE), Department of Defense. The content of the information provided in this publication does not necessarily reflect the position or the policy of the government, and no official endorsement should be inferred. The authors acknowledge the USACE and USCRP’s support of their effort to strengthen coastal academic programs and address coastal community needs in the United States.