A comparative analysis of a global and a regional model on the tropical cyclone track forecast

The skill of model estimations is examined in reproducing the trajectory of tropical cyclones (TCs) Thane (2011) and Nilam (2012) by comparing the results of a global model (Global Forecast System; GFS) to a regional model (Weather Research and Forecasting—Advanced Research WRF; WRF-ARW). The spatial resolution of the global model (25 km) is chosen similar to the horizontal grid spacing (27 km) of the regional model for a fair comparison over the Bay of Bengal (BOB) region. The objective is to set the basis for future studies that analyze the physical reasons why a particular simulation performs better in the BOB than others in terms of reproducing trajectory. Collectively, 408 h and 288 h of the run of each model are used in this paper for analyzing the TCs Thane and Nilam, respectively. The wind vector plots at 850 hPa, pressure at mean sea level (PMSL), and rainfall are also used for enhancing the analysis. The results show that GFS simulations significantly improve the trajectory of both TCs as compared to WRF with a crucial 13.9%, 24%, and 31.6% (Thane: 6.3%, 21.4%, 31%, and Nilam: 26.6%, 29.2%, 32.9%) reduction in forecast error for a 24-h, 48-h, and 72-h forecast, respectively. By and large, the GFS simulations improve the TCs’ trajectory forecast results by reducing more than one-fourth (26.5%; 30.3% in TC Nilam and 24% in TC Thane) of the forecast error as compared to the WRF data. Moreover, the maximum one-third (31.6%) to a minimum of 13.9% of the trajectory forecast error is reduced in the overall 72-h and 24-h GFS simulations, respectively, in comparison to the WRF simulations. Typically, both models provide less forecast error in the case of TC Nilam in comparison to TC Thane, which is also partially contributed by the anti-clockwise recurvature taken by the TC Thane, again GFS shows better agreement relative to the observations. 795.4 km and 40.2 km are the maximum and minimum forecast errors, which are provided by the 72-h and 24-h WRF simulation on 27 and 30 December 2011, respectively, making the forecast error range 22.9% broader (755.2 km) for WRF simulations over GFS (614.3 km). The paper creates a significant base for future work for assessing model simulation data against topography, in-land water bodies, ocean, and upper-air observations. These future analyses can enhance the understanding of how these models represent the distribution of those variables. Also, it will be utilized to fathom the physical reasons behind the enhanced performance of GFS over WRF in the BOB in terms of reproducing trajectory, although the current situation indicates the possible dominance of the global environmental processes and the BOB ocean water over the regional topographical effects. Historically, the regions around BOB have been known long for huge human and economic losses during the TCs. The extremely heavy population in these regions is considered the central cause behind it. Through a 49-year data analysis in the North Indian Ocean (NIO), this paper also shows a significant enhancement in the strength of TCs, especially during the recent 14-year (2007–2020), which further enhances the crucial need to examine the performance of models to capture the location of TCs for mitigating their impact on human lives in these TC-prone regions.