The effects of non‐stationarity on <scp>SPI</scp> for operational drought monitoring in Europe

It is a good practice to follow common guidelines in the computation of Standardized Precipitation Index (SPI) data sets as part operational drought monitoring systems. In European Drought Observatory (https://edo.jrc.ec.europa.eu/), reference statistics are computed following World Meteorological Organization Guidelines on Calculation Climate Normals, where definition period for applications introduced most recent 30-year finishing year ending with 0. this study, temporal consistency SPI time series using fifth-generation Centre Medium-range Weather Forecast reanalyses model precipitation set tested over Europe quantify effect transition from one baseline (1981–2010) another (1991–2020) and evaluate capability these static baselines reproduce behaviour non-stationary SPIs (nSPIs). The results comparison suggest that threshold commonly used identify droughts (SPI = −1) only marginally affected by change (mean absolute deviation, MAD 0.15 ± 0.1) short-term SPI, whereas larger differences (MAD up 0.6) can be observed certain areas (i.e., Southern Italy Eastern Europe) longer accumulation periods SPI-9 or SPI-12). Examples show changes classification extremely dry < −2) moderately not uncommon, which may lead misinterpretation users. Finally, analyses against nSPI highlight an overall correspondence between stationary nSPIs, even if both displayed difficulties reliably capturing magnitude entire 10-year they should used. regard, it has been demonstrated more spatially uniform achieved 5-year updates, matching <0.25) SPI-1 acceptable 0.50) SPI-12 than 80% Europe.