UPSurgeML Model Data for "Cross-scale Probabilistic Storm Surge and Tide Forecasts: Impacts of Tropical Cyclone Wind Field Structure"

UPSurgeML model output data and perturbation information is provided to reproduce the analysis in the article: "Cross-scale Probabilistic Storm Surge and Tide Forecasts: Impacts of Tropical Cyclone Wind Field Structure". There are a total of 14 historical events. Five of these have three landfall lead times (24 h, 48 h and 72 h) and the rest have one lead time (48 h). Each of these contains three model types (regression_gahm, regression_holland, and persistent_holland). Note that Hurricanes Irene and Matthew have two landfall locations, which are run independently, as indicated by the state acronym contained within the tarball filename. Each event tarball contains the following files for each leadtime and model run: Hurricane forecast track files in Automated Tropical Cyclone Forecasting (ATCF) system format (*.22 files) with associated error variable parameters (*.json files) Error variable parameters for ensemble member in each forecast (perturbations.nc) The maximum water elevation output for every model run subsetted within the 34-kt wind swath and < 25 m ocean depth (subset.nc)