Comparative analysis of two rainfall retrieval algorithms during extreme rainfall event: a case study on cloudburst, 2010 over Ladakh (Leh), Jammu and Kashmir

Abstract

Intense rainfall events lead to floods and landslides in the Western Himalayas (WH). These rainfall amounts are considered comparatively moderate over the plains. These events, called ‘cloudbursts,’ are convective triggered followed by orographically locked phenomena producing sudden high-intensity rainfall over a small area. Early warning and prediction of such severe local weather systems is important to mitigate societal impacts arising from the accompanying flash floods, landslides, etc. Due to lack of ground-based observations, cloudbursts over remote and unpopulated hilly areas often go unreported. Present study examines a cloudburst occurred at Ladakh (Leh) in the WH in the early hours of August 5, 2010, using remotely sensed rainfall data from Tropical Rainfall Measuring Mission (TRMM) and Kalpana-1. The storm lasted for 2 days starting from August 3, 2010, followed by flash floods. Rain-band propagation over the region is studied from Kalpana-1 3-hourly rainfall estimates using Indian Satellite (INSAT) multi-spectral rainfall algorithm (IMSRA) and TRMM rainfall estimates using TRMM 3B42 algorithm. Quantitative and qualitative assessment and comparison of these two products is made. It is observed that there is decrease in satellite brightness temperature (BT) during the rainfall event. Initiation of rainfall occurs at about < 255 K. Maximum of 16.75 mm/h rainfall is observed over the Jammu and Kashmir at 21 GMT from TRMM 3B42 estimates. In addition, it is observed that Kalpana-1 IMSRA underestimates the rainfall observations with respect to Indian Meteorological Department (IMD) rainfall estimates. Spatial correlation at 5% significant level is evaluated, and similarities in rainfall estimates based on rainfall retrieval algorithms are made during the cloudburst event. Mean and standard deviations depict that TRMM 3B42 and IMD rainfall estimates are closer in terms of spatial signature, but estimates of rainfall from Kalpana-1 underestimate the mean and standard deviation signature. In view of orographic contribution, it has been observed that linear fit is the better than nonlinear with less rainfall bias at 90% confidence bound over the region.