Sensitivity of tropical cyclone predictions in the coupled atmosphere–ocean model WRF‐3DPWP to surface roughness schemes

Abstract

The sensitivity of tropical cyclone simulations in a coupled ocean–atmospheric model WRF‐3DPWP to surface roughness parameterizations is investigated. Six cyclones, Hudhud, Phailin, Helen, Jal, Nilam and Khaimuk, which occurred in the North Indian Ocean (NIO) during 2008–2014, are simulated. Three schemes for drag and enthalpy co‐efficients (control (CNTL): Garrett C d and C k, Opt1: Donelan C d + Constant C k, Opt2: Donelan C d + Garret C k) are used. India Meteorological Department (IMD) best track parameters are used for comparison. The results indicate that the simulated cyclone intensity is highly sensitive to the formulation of drag and enthalpy co‐efficients in all cyclone categories, whereas the simulated tracks for strong cyclones only are affected by these schemes. The results showed that CNTL produced better predictions for track and Opt2 for intensity for the three cyclones. An improvement of 12.3, 35.2 and 34.9% in central sea level pressure (CSLP) and 2.9, 31.5 and 45.9% in maximum surface winds (MSW) at 24, 48 and 72 hr respectively are found with Opt2 over Opt1. Simulations showed the model enthalpy co‐efficients mainly affected the intensity of the cyclones by producing a wind‐induced surface heat exchange (WISHE)‐type of feedback through heat and moisture fluxes and warm‐core structure. Opt1 produced highly intensified cyclones due to simulating large enthalpy fluxes associated with a high C k/C d ratio. The stronger wind‐induced moisture fluxes, higher vorticity and strong vertical motion associated with large enthalpy fluxes all support stronger simulated cyclones in Opt1 compared with CNTL and Opt2. The study shows that the decreasing C k for higher winds in Opt2 produced realistic intensity predictions.