Comparison of parametric and non-parametric methods for non-stationary hydrological modelling: Rainfall Puglia region case study (Italy)

Study region: Puglia Region, Southern Italy. Study focus: The intensifying signal of climate change necessitates a transition from stationary to non-stationary hydrological modeling. This study addresses a critical research gap in the Mediterranean context by evaluating the non-stationarity of extreme rainfall through a comparative analysis of non-parametric (Mann-Kendall, Sen’s slope, Pettitt) and parametric methodologies. Notably, this research introduces the Two-Stage parametric method for the first time in Italy, allowing for a nuanced characterization of climatic shifts by accounting for temporal variability in both the mean and standard deviation of rainfall extremes. New hydrological insights for the region: Analysis of long-term time series (1921–2019) reveals a significant intensity dichotomy: a prevailing decreasing trend in daily annual maxima contrasted by a marked intensification in sub-daily (1 h and 3 h) extremes. The Two-Stage method specifically demonstrates that non-stationarity affects not only the mean intensity but also the climatic variance, indicating that rainfall is becoming increasingly volatile. The high consistency between non-parametric results and Two-Stage modeling reinforces the reliability of these detected trends. These findings provide a robust scientific basis for updating regional Rainfall Depth-Duration-Frequency (DDF) curves and enhancing the resilience of hydraulic infrastructure in the Mediterranean area.