Abstract
The Standardized Precipitation Index (SPI) has been widely used to detect and classify drought events based exclusively on precipitation. However, its application in arid and semi-arid regions, characterized by a high frequency of zero-precipitation months, presents limitations that may compromise drought detection and duration assessment. This study evaluated the limitations of the SPI in the state of Paraíba, northeastern Brazil, a region with strong spatial contrasts in rainfall regimes. Monthly precipitation series from the Paraíba State Water Management Agency (AESA) and the BR-DWGD gridded dataset were used to calculate the SPI, while the Standardized Precipitation Evapotranspiration Index (SPEI) was computed using AESA precipitation and BR-DWGD reference evapotranspiration. SPI normality was assessed using the Shapiro–Wilk test. Results showed that SPI values became asymmetric whenever the probability of zero precipitation, P(0), exceeded zero and failed to satisfy the normality assumption when P(0) exceeded 0.30. When P(0) > 0.30, the SPI cannot detect drought events of any magnitude, compromising duration estimates. The BR-DWGD dataset underestimated the frequency of zero-precipitation months, although agreement with AESA improved when monthly precipitation of ≤2 mm was treated as monthly zero rainfall. The SPEI reduced artificial wet classifications during the dry season, providing a more realistic representation of regional water balance. Nevertheless, its application at the monthly timescale also requires caution in regions with frequent zero-precipitation months.