Regionalization of precipitation associated with tropical cyclones using satellite‐based precipitation and spatial metrics

Abstract:

Tropical cyclones (TCs) pose a significant flood risk to vast regions along and inland of the coasts because of extreme precipitation. Quantifying spatial characteristics of tropical cyclone precipitation (TCP) and defining homogeneous rainfall regions can benefit forecasts, hydrological modeling, and hazard mitigation of TCs. This study applies an object-based Geographic Information System method to delineate and quantify rainfall fields associated with North Atlantic landfalling TCs from the satellite-based rain rate estimates. The research questions are: first, which spatial metrics effectively characterize spatial patterns of rainfall in different storm intensities? Second, what is the spatial variation in spatial characteristics of rain fields of the storm before and after landfall? Eleven spatial metrics are calculated for both the entire rain field and the largest rainfall polygon within the rain field. Their effectiveness in characterizing rain fields is evaluated using a set of non-parametric tests. Then a multivariate clustering method with hot spot analysis is utilized to investigate spatial variations and regionalize the rainfall patterns using selected spatial metrics. The results suggest that five spatial metrics, namely area, solidity, closure, dispersion, and roundness, can effectively distinguish rainfall patterns among different storm intensity categories. The clustering results highlight the spatial variations of rainfall patterns when storms are over the North Atlantic sub-basins, as well as away from the land, near coasts, and far inland.