Flood damage assessment in two western mediterranean regions. Present conditions and future scenarios

Resumen

Flooding is one of the main natural hazard in the world causing huge economic and human impacts. Assessing the flood damage in the Mediterranean region is of great importance, mainly due to its pronounced sensitivity to climate change. A large number of floods affecting the western Mediterranean region of study are surface water floods that can cause catastrophic damage. These floods are caused by intense precipitation events, thus, in order to understand properly these type of events, the analysis of the relationship between precipitation and flood damage is crucial. The overall objective of this thesis is to analyse flood damages in two Mediterranean regions, namely Catalonia and the Valencian Community, frequently affected by intense precipitation events, as well as to estimate their changes when future climate change projections and different socioeconomic scenarios are considered. To do this, the relationship between heavy precipitation and flood damage estimates from insurance datasets, provided by the Spanish Insurance Compensation Consortium (Consorcio de Compensación de Seguros, CCS), have been analysed. Other than for the above mentioned two areas of study, this analysis was performed for the Metropolitan Area of Barcelona (MAB), a highly vulnerable urban area. The study period covers 1996-2015. Several regression models have been tested in order to gauge the probability of large damaging events occurring given a certain precipitation amount and taking into account other variables related to the exposure of the territory. Results have shown that generalized linear mixed models are the most appropriate tool for studying the relationship between precipitation and flood damage. The probability of a damaging flood event increases with precipitation and population of the basin. Moreover, 30-min precipitation data proved to be a better predictor of the probability of large damages than daily precipitation, however, this type of information is not always available. On the other hand, we have analysed the projected changes in precipitation extremes in the Iberian Peninsula considering the Representative Concentration Pathway (RCP) 8.5 scenario and an ensemble of seven EURO-CORDEX simulations spanning the period 1976-2100. In order to do this, different climate indices were calculated to estimate the changes in precipitation assuming global warming scenarios of 1.5, 2 and 3 °C above preindustrial levels. Results show a general decrease of the total annual precipitation and an increase in the length of dry spell in most of the Peninsula. This increase accentuates with higher levels of global warming and during summer months. In terms of heavy precipitation, we have found increases with global warming in the maximum 1-day and consecutive 5-day precipitation indices as well as in the number of days with precipitation exceeding 40 mm, especially during the months of autumn and winter and in the north and north-east of the Iberian Peninsula. Finally, changes in the probability of occurrence of damaging flood events have been assessed for both regions when considering a global warming of 1.5, 2 and 3 °C and taking into account different climate projections and socioeconomic scenarios. To do this, the previously developed statistical climate model that links precipitation, population and flood damage estimates, has been used to assess future climate conditions. The daily precipitation data from the seven climate models used previously and population projections based on five different socioeconomic scenarios (Shared Socioeconomic Pathways, SSPs) were incorporated into the model. Results have shown a general increase in the probability of a damaging event for most of the cases and in both regions of study, being higher in the case of Catalonia. This change is usually larger when greater warming is considered and for higher percentiles of damage. Moreover, the increase in probability is larger when both climate and population changes are included. Our findings highlight that, when it comes to flood damage analysis, it is crucial not only to account for climate change but also consider socioeconomic conditions. Furthermore, results show that limiting global warming is a must in order to minimise the consequences of flood events in the study area.