Verification of tropical cyclone motion and rainfall forecast over North Indian Ocean

The accurate prediction of tropical cyclone (TC) track, intensity, and rainfall is necessary from a disaster management perspective. The simulations of 24 TCs that occurred from 2016 to 2020 over the North Indian Ocean (NIO) were carried out to examine the efficacy of 5 days forecast. The analysis reveals that the direct position error (DPE) values over NIO for 12-, 24-, 48-, and 72-hr lead time is 68.12, 91.79, 149.8, and 232.36 km, respectively. The forecast track is southeastward of the observed track till 72 hours and northwestward at later lead times. The landfall position error is less over the Bay of Bengal (BoB) as compared to the Arabian Sea (AS), and the model indicated a delayed landfall response. The intensity error for TC forecast over NIO magnifies with forecast lead time from 4 to 12 ms −1 . Quantitative verification of rainfall indicated overestimation of model rainfall with respect to GPM-IMERG rainfall. Verification of rainfall forecasts during the landfall of the TCs is carried out using contiguous rain area (CRA) method. It is seen that pattern error dominates for light-moderate rains, and displacement and volume error contribution dominate, especially for heavy rain. CRA adjustment has greatly improved longer lead times, especially heavy rainfall thresholds. Research highlights The ATE and CTE evaluation shows that the forecast track is southeastward of the observed track till 72 hours and the northwestward at later lead times. The intensity is overestimated for BoB cyclones and is undermined for AS cyclones. The category rainfall verification shows the highest prediction skill for very heavy rainfall events. The CRA method for spatial rainfall verification reveals that pattern error is dominant for moderate rainfall while volume and displacement error contribute more for very heavy rainfall rain. The landfall error for TCs generated over BoB is less than those generated over AS.