Roberta Padulano (Italy) 1; Guido Rianna (Italy) 1; Alfredo Reder (Italy) 1
1 - CMCC Foundation Euromediterranean Center on Climate Change
Intensity Duration Frequency (IDF) curves describe sub-daily annual maximum precipitation under expected Return Time (RTs). They represent a valuable tool both for hydraulic infrastructures design (e.g. sewage systems) and for the implementation of early warning systems coping with rainfall-induced hazards. Nonetheless, the strong assumption of steady-state conditions could result unreliable due to potential effects of global warming on rainfall patterns. To address such issue, in accordance with several recent tools proposed in literature (e.g. DeGaetano & Castellano, 2017 doi: 10.1016/j.cliser.2017.03.003), a conceptual framework is developed to assess expected variations in IDF curves due to climate change, basing on annual maxima observations collected at the weather station of Naples (Southern Italy), available at sub-daily time scale for several decades up to nowadays. The proposed tool accounts for variabilities in precipitation projections by considering all the 19 simulations available within the EURO-CORDEX initiative at the 12-km horizontal resolution under RCP4.5 and RCP8.5. Current reference period is 1971-2000 while 2071-2100 is considered for estimating future conditions.
Statistical significance of climate signal in maximum yearly precipitations is analysed by means of the two-sample Anderson – Darling (AD) test, which compares two samples (in this case, current and future maximum precipitations) under the null hypothesis that they are extracted from the same population. Furthermore, Quantile Delta Mapping approach (Cannon et al., 2015; doi: 10.1175/JCLI-D-14-00754.1) is adopted to correct biases in modeled/observed quantiles of precipitation growth factors but explicitly preserving modeled/projected relative changes in quantiles. For Naples, statistical variations due to climate change resulted significant (SSCs) for 10 (12) simulations under RCP4.5 (RCP8.5); no clear patterns arise to clarify if this may be related to specific Global or Regional Models.
Ensemble mean results considering all the projections or SSCs only provide for substantial increases in growth factors compared to the current period under both the scenarios; however, they are more evident considering SSCs only and under the more severe RCP8.5, despite a remarkable spread among the simulations. Finally, variations in current rainfall thresholds for early warning systems are assessed for RTs equal to 2, 5 and 10 years (alert, pre-alarm and alarm) and different durations according the hazard. Relevant increases are assessed considering both frequency (under fixed rainfall depth) and magnitude (under fixed RT). This should elicit reflection on timely adoption of adaptation measures aimed to increase the resilience of hydraulic infrastructures and to implement effective protection measures against weather-induced hazards.