Evaluation of
High‐Resolution Rapid Refresh model
forecasts of extreme rainfall associated with concurrent multiscale atmospheric phenomena during the 2024
Indian
Greeshma M. Mohan, Ashish Routray, K. B. R. R. Hari Prasad, H. Haripriya Abstract
This study evaluates the heavy to very heavy rainfall categories associated with the simultaneous occurrence of multiple atmospheric phenomena during selected rainfall episodes of the southwest monsoon 2024, as forecast by the High‐Resolution Rapid Refresh (HRRR) model. The Multi‐Object Analysis of Atmospheric Phenomena (MOAAP) tool is used alongside the India Meteorological Department's (IMD's) reports to ensure the concurrent atmospheric phenomena that occurred during these rainfall events. A total of 20 cases were identified, comprising 87 subregions demarcated with a rainfall intensity in the range of heavy or above. The evolution of concurrent atmospheric phenomena from these 87 subregions reveals that 4.6%, 37.9%, 24.1%, 27.6%, and 5.7% of instances involved the simultaneous occurrences of six, five, four, three and two different phenomena, respectively, with upper‐level anticyclones being a consistent monsoon feature. Mesoscale convective systems (MCSs; 65.5%) and atmospheric rivers (ARs; 59.8%) are the most frequent atmospheric phenomena associated with these events, while fronts and moisture streams are less frequent. The capability of the HRRR system to capture the rainfall that occurred during such concurrent atmospheric phenomena is evaluated against Global Precipitation Measurement (GPM)‐observed rainfall. It showed generally positive correlations with observed rainfall, though some cases exhibited weaker relationships and lead time biases, with mostly early occurrences of the peak rainfall activity in the model. Fractions skill score (FSS) analysis indicated that HRRR rainfall predictions remain reliable up to a 20‐mm threshold, with reduced skill at higher thresholds. Overall, HRRR forecasts captured the spatial and diurnal rainfall distributions well, despite underestimating rainfall intensity and, therefore, showing limited skill at higher thresholds.