Multiscale drought dataset for the Greater Antilles: a resource for environmental and adaptation studies

Abstract

This study presents a multi-drought metric dataset based on the Standardized Precipitation Index (SPI) for the Greater Antilles, covering Cuba, Jamaica, Puerto Rico and La Española. The SPI was derived from high-resolution (0.1°) Multi-Source Weighted-Ensemble Precipitation (MSWEP) precipitation data for the 1980-2023 period. The MSWEP precipitation and SPI datasets were validated against observational and reanalysis datasets, providing good relationships. Temporal and spatial drought conditions were identified for the SPI temporal scales from 1 to 24 SPI temporal scales, while drought episodes were identified for 1, 3, 6, 12, 18 and 24 temporal scales, and characterised according to their duration, severity, affected area, persistence, onset speed, and recovery. A source-sink moisture analysis was also performed to support the attribution of dry conditions. We demonstrate possible uses of the database, and its consistency in representing past drought events that caused severe damage. Therefore, it proves to be a practical tool for understanding the phenomenon of drought in the Caribbean, managing water resources, agricultural development, and many other sectors.

Fig. 1

Geographical location of the Greater Antilles: Cuba, Jamaica, La Española (Dominican Republic + Haiti) and Puerto Rico, and their elevation. Data obtained from the HydroSHEDS project (hydrological data and maps on shuttle elevation derivatives at multiple scales) are available online at https://www.hydrosheds.org.

Fig. 2

Flow work followed in this study.

Fig. 3

Schematic representation of the onset, duration and end of a drought event.

Fig. 4

Climatological integrated (E−P) >0 from the backwards-in-time experiment for the period 1981–2023, considering the magenta box indicates the domain (a). Schematic representation of the Caribbean-North Atlantic moisture source (NATL), the Greater Antilles, and their location (b), and the moisture contribution to precipitation (in %) over the islands in a forward-in-time experiment from themselves and the NATL (c).

Fig. 5

Schematic representation of the organisation and type of data records available to download for each island.

Fig. 6

Spatial correlation between precipitation data (MSWEP vs ERA5 Land (1980–2023) (a), MSWEP vs CHIRPS (1981–2023) (b), and MSWEP vs. CubaPrec1 (1980–2008) (c)). Temporal correlation between MSWEP vs ERA5 Land, MSWEP vs ERA5 Land (b), and MSWEP vs. CubaPrec1 is shown with bold numbers. All values are statistically significant (p < 0.05).

Fig. 7

Correlation between SPI time series at 1, 3, 6, 12, 18, and 24 months calculated from MSWEP precipitation and SPI calculated from ERA5 Land (a), and CHIRPS (b) precipitation for the 1980–2023 period, and SPI from MSWEP correlated with SPI from CubaPrec1 for the 1980–2008 period (c).

Fig. 8

Time series of the SPI6 and SPI12 for Jamaica. Identification of moderate/severe/extreme drought conditions and comparison with previous registered drought events.

Fig. 9

Temporal evolution of the most severe meteorological drought episode for each island according to the SPI1, and the respective anomalies on the moisture contribution to precipitation (E-P < 0) during the most severe meteorological drought episode over each island.