Projections of tropical cyclone rainfall over land with an Eulerian approach: Case study of three islands in the West Indies

Abstract

The latest version of the atmospheric general circulation model ARPEGE-Climat was used to perform 5-member ensemble simulations for both present and RCP8.5 scenario climates (mid-21st century). The rotated/stretched configuration enables a local horizontal resolution of less than 15 km over the tropical North Atlantic basin. Moreover, a tracking algorithm was used to extract tropical cyclones (TCs) simulated by the model. Through an Eulerian approach, this paper focuses on the relationships between TCs and rainfall over three French islands in the West Indies. Although the model underestimates the occurrence of TCs over this latitude band, especially in September, precipitation rates during TC days are realistic. Indeed, the model shows a good capacity to reproduce different relationships between island rainfall and TC characteristics such as proximity and intensity. In addition to rainfall distribution, the TC contribution to annual cumulative rainfall is also well captured by the model. We used three different series characterizing precipitation at the island scale to underline that the model overestimates the area that is impacted by TC rainfall. According to the simulations, the number of minor TCs tends to decrease in the future (about −15%) over the study domain (50-70 W × 10-25 N) despite a +1.6 C over ocean warming. In contrast, no trend was detected in the number of major hurricanes. Except for annual precipitation that decreases significantly in the future (about −15%), no significant change was detected in the relationships between TC properties and island rainfall.

Designing scenarios for upscaling climate-smart agriculture on a small tropical island

“…Climate projections were obtained using the ARPEGE-Climat model (Chauvin et al, 2020;Cantet et al, 2021) with radiative forcing parameters based on the Representative Concentration Pathway (RCP) 8.5 (Cubasch et al 2013;Moss et al, 2010). This is the Roehrig et al, 2020;Voldoire et al, 2019).…”

“…A specific configuration allowed a local horizontal grid spacing of < 15 km over the western tropical northern Atlantic. This enabled representation of Guadeloupe's climate explicitly despite the island's reduced size, unlike CMIP coarse-resolution CGCMs (Cantet et al, 2021), which are critical for island-scale climate projections (Cantet et al, 2014).…”