Multi-case study on VIL Nowcast enhanced with conditional bifurcation using vertical wind

Abstract

Localized heavy rainfall events have caused an increasing number of water-related disasters in recent years, particularly in urban areas where drainage capacity is limited. Accurate prediction of such events requires high-resolution, high-frequency three-dimensional (3D) observation data to capture rapidly evolving atmospheric structures. The Multi-Parameter Phased Array Weather Radar (MP-PAWR) provides rapid, high-density observations, offering significant advantages over conventional parabolic radars. However, rainfall prediction methods such as Vertically Integrated Liquid water content (VIL) Nowcast (VIL Nowcast; VILNC), which traditionally rely on limited-resolution radar data, often overestimate rainfall intensities, especially during the dissipating phase of cumulonimbus clouds. In this study, we propose an enhanced VILNC model that incorporates 3D wind field estimations derived from MP-PAWR to better represent the development and decay of rainfall areas. By integrating vertical wind information, the modified VILNC aims to mitigate overestimations associated with downdrafts. Comparative analysis of several localized heavy rainfall events demonstrates that the proposed method improves prediction accuracy, particularly at shorter lead times. These results suggest that the utilization of high-resolution 3D wind fields from MP-PAWR contributes to more reliable short-term rainfall predictions.