Assessing environmental conditions associated with spatially varying rainfall structure of North Atlantic tropical cyclones: An object-based climatological analysis

Keywords: tropical cyclones; rain fields; object-based analysis; Geographic Information System; spatial regression; North Atlantic
Abstract Type: Paper Abstract

Abstract

This study utilized Geographic Information System-based shape analyses with spatial regressions to examine the spatial variations in the relationships of North Atlantic TC rain field patterns to the environmental conditions. We measured the area, solidity, dispersion, and closure of rain fields associated with North Atlantic TCs during 1998-2014 from satellite-based rain rate estimates. The potential contributing factors examined included storm intensity, translational speed, total precipitable water, upper-tropospheric divergence, deep-layer vertical wind shear, and distance to land. Spatial metrics of rainfall, TC attributes, and environmental conditions were aggregated onto a hexagon grid, and the average values were used in regression models. Spatial autoregressive regression and Geographically Weighted Regression were applied to investigate the relationships between contributing factors and rainfall patterns. Storm intensity was positively correlated with area, solidity, and closure and negatively correlated with dispersion. These relationships were stronger in subtropical regions. Higher moisture and larger upper-level divergence contributed to a larger rainfall field with a more dispersed and less solid pattern, especially over the western Gulf and the western Caribbean. Wind shear and storm motion affect rain field patterns differently depending on the juxtaposition and relative magnitude of storm motion and wind shear vectors and thus showed more regional variations. Last, upon approaching land, the storms gain a less solid but more enclosed pattern. This research provides an improved understanding of basin-wide and sub-basin-wide relationships between TC rain fields and contributing factors,which would benefit rainfall forecasting and hazard mitigation.