3.3 Relationships with climate
To estimate the relationships between dry spells and climatic drivers, namely precipitation and evapotranspiration, the correlation between the annual and seasonal sum of zero flow days and the maximum length of zero-flow days with the Standardized Precipitation-Evapotranspiration Index (SPEI, Vicente-Serrano et al., 2010) was analyzed with the Spearman correlation coefficient (rho). The SPEI uses the monthly difference between precipitation and potential evapotranspiration; thus it represents a simple climatic water balance, which can be calculated at different time scales similarly to the Standardized Precipitation Index (SPI, McKee et al., 1993). The SPEI with 6, 12, 18 and 24 months aggregation time have been downloaded from the CSIC Global SPEI database ( https://spei.csic.es/database.html ). The SPEI values were computed using monthly sum of precipitation and potential evapotranspiration at 0.5 degrees spatial resolution and a monthly time resolution from the Climatic Research Unit (CRU) of the University of East Anglia. Version 3.23 of the CRU dataset has been used to compute the SPEI. The SPEI computation is based on the FAO-56 Penman-Monteith estimation of potential evapotranspiration. The details of the SPEI computation were presented by Beguería et al. (2014). For each station, the value was extracted from the SPEI grid cell covering the station, since the size of the basins considered is small (<2000km²) compared to the CRU mesh (approximately 2500km²) used as a basis for the calculation of the SPEI.
In addition, different climate indices describing large scale atmospheric circulation patterns have been selected: the North Atlantic Oscillation (NAO), the Atlantic Multi-decadal Oscillation (AMO), the Mediterranean Index (MOI), the East Atlantic Western Russia (EAWR), the Pacific Decadal Oscillation (PDO), and the Scandinavian Index (SCAND). The time series for these indices have been retrieved from the CPC database available online at:https://www.cpc.ncep.noaa.gov. The annual, winter and summer mean values were used. For NAO, the seasonal three-month DJF, MAM, JJA, and SON values were also included to account for its within-year variability.