ERA5-land–based drought assessment using SPI and SPEI to support smart agriculture in Prince Edward Island, Canada

Recognizing regional drought patterns is essential for managing agricultural risks and ensuring stable crop production for policymakers and farmers. However, relying only on ground-based data limits historical drought analysis due to inconsistent ground meteorological observations. This study assessed historical drought (1980–2023) using ERA5-Land (ERA5L) reanalysis data and the 1-, 3-, 6-, and 12-month Standardized Precipitation Index (SPI) and Standardized Precipitation Evapotranspiration Index (SPEI) at three weather stations in Prince Edward Island (PEI), Canada's largest potato-producing province. A significant portion of the island’s farmland depends on rainfall for irrigation, making it highly vulnerable to precipitation and variations in drought occurrence. In this study, the performance of ERA5L data as a substitute for ground-based measurements needed to calculate the SPI and SPEI was evaluated within an integrated framework linking climate-variable validation, evapotranspiration assessment, and drought index analysis. Raw ERA5L reanalysis data proved effective in providing climate variables, including max/min temperatures, dew point temperature, and wind speed, based on statistical comparisons with ground data from 1993 to 2023 at the Summerside, Charlottetown, and East Point weather stations, representing the island’s west, center, and east. The performance of data in estimating FAO-Penman-Monteith evapotranspiration (ET o ) was also confirmed. However, for monthly precipitation, the ERA5L performance was reasonable after applying bias correction. By integrating validated reanalysis data with multi-timescale drought indices and trend analysis, a long-term, spatially consistent drought characterization for PEI is provided, along with a transferable methodological framework for data-scarce coastal agricultural regions. Results indicate simultaneous drought occurrence across the island, with mild droughts being the most frequent (33% of events) and severe and extremely severe droughts being rare (