Past and future hydrogeological risk assessment under climate change conditions over urban settlements and infrastructure systems: the case of a sub-regional area of Piedmont, Italy

Abstract

Hydrogeological impacts due to flood and landslides are among the most significant disasters in the world. In the last decades, these events result in forthcoming losses of life and damages to private and public properties and criticalities increase due to very often increasing losses as consequence of changes in the rainfall regimes as well as the increase in average temperatures. This paper describes an applied hydrogeological risk assessment methodology carried out as part of an interdisciplinary European project between France and Italy (Interreg Alcotra ARTACLIM). We developed a practical framework to assess the risk of urban settlements and infrastructural assets within a sub-regional area of Piedmont (north-west of Italy). Past and future climate drivers are here analyzed using high-resolution data to highlight the temporal and spatial distribution of elements under threat. Using geographic information systems, we elaborate climate and socioeconomic data to develop an effective procedure for a sub-regional integrated research on the intrinsic and external conditions of potential instability. The results of the analyses produce a geo-localized risk score that can be used to rank assets in a screening process that aims to assist urban planner and local policymaker to prioritize the adaptation measures required for taking action to reduce hydrogeological damages to their assets.

Appendix A

In this section, we outline the key procedures used to measure the hydrogeological risk for:

• framework 1—hydrogeological risk due to floods in relation to urban settlements;

• framework 2—hydrogeological risk due to floods in relation to infrastructure systems;

• framework 3—hydrogeological risk due to avalanches and landslides in relation to urban settlements;

• framework 4—hydrogeological risk due to avalanches and landslides in infrastructure systems.

For simplicity, the example examined for the evaluation of the calculation process refers to the framework 1.

Hazard identification

• In a first step, thanks to the collection of proxies on climate hazards, we identified the external climate factors that influence the area in terms of climate change hydrogeological impacts, based on observed (1988–2018) and future (2021–2050) climate, according to RCP4.5 and RCP8.5 scenarios;

• The acquisition of quantitative values related to each climate sub-indicator on a municipal basis is better explained in Sect. 2.4: Local climate trends and hazard indicators identification. In this case, we analyze precipitation frequency and precipitation intensity indicators capable of producing a degree of hazard over the territory. In particular: R10, R20, RR1, PRCPTOT, SDII, RX1day and RX5days (Table 2);

• For each sub-indicator, the values of the observed climate and on the future variations—RCP4.5 and RCP8.5—over the period 2021‐2050 have been considered. There is therefore a dual configuration for each framework. One reflects the dangers on the observed climate and the other two concern the expected changes according to the future scenario. It should also be clarified that while this distinction recognizes the variability of sources of climate danger over time, the frameworks does not consider the variation of exposure, sensitivity and adaptive capacity, which stay constant faithful to the most up-to-date sources;

• At this stage, it is necessary to check the outputs from future projections models. In fact, in case of indicators in which we expect future climate hazard to increase (e.g., increase in rainfall intensity), it is necessary to check whether future anomalies from RCP 4.5 and for RCP 8.5 scenarios presents negative values. If so, the specific values that present negative anomalies have to be imposed to 0;

• After collecting the data for each indicator of climate hazard at the municipal level, the minimum and maximum values useful for the normalization of each indicator were calculated using the “min–max normalization” formula:

$$ X_{i = } \frac{{x_{i} - x_{ \hbox{min} } }}{{x_{\text{max }} - x_{ \hbox{min} } }} $$

(a.1)

where $x_{i}$ represents the individual data point to be transformed, $x_{\text{max }}$ corresponds to the highest value and $x_{ \hbox{min} }$ to the lowest value for each indicator. In this way, every parameter X_i is identified by a numerical value from 0 to 1, where the highest value corresponds to the highest contribution to any factors, considered separately from the others.

This procedure makes possible to transform all the indicators in a range from 0 to 1, facilitating the comparability between them. Furthermore, in some cases we applied the reversal—especially for the indicators related to frost days (e.g., FD), not present in this framework but considered in this studies related to landslides and avalanches—because the minimum values did not correspond to low values of hazard but to high values of hazard;

• We finally proceeded to calculate the climate hazard index (H), through an aggregation that took into account the equal weight of each indicator (w = 1), using the linear aggregation method;

• After that we applied this aggregation for the observed climate and for the future RCP4.5 and RCP8.5 scenarios, the Hazard Index is ready to be represented by GIS software in order to provide an effective representation of the spatial distribution in the ZOP in the last and future 30 years (see Sect. 1b in Box I).

[We performed the mentioned operations for each set of indicators: exposure, sensitivity and adaptive capacity]

Exposure identification

• In the case of framework 1, the exposure indicators are related to the presence of people, livelihoods, services, infrastructures, and economic-social-cultural assets that could be negatively affected by climate variations over the ZOP area. In here, we calculated the surfaces (in hectares) based on the different land uses of the buildings (PAI 1999), then aggregated at the municipal level. Reasonably, the municipalities with a greater surface subjected to flood risks will be more exposed;

• As previously mentioned, after collecting the data for each sub-indicator of exposure, we applied the min–max normalization formula and we aggregated them—with equal weight—to reach the final Exposure Index (E);

• The exposure index is now ready to be represented by GIS software and thus to provide an effective reproduction of the distribution of the factors exposed to the climate hazard, regarding the issue of the urban settlement systems in relation to floods.

Vulnerability identification

• Sensitivity—in this phase the framework envisages the same procedure with a focus on sensitivity, considered as the degree of susceptibility of the place to floods, both in a negative and positive way. In line with the previous index, after the collection of data at the municipal level, the min–max normalization formula was applied, followed by the aggregation process;

• Adaptive capacity—here interpretable as the capacity of the settlement system to deal with flood damages, influenced by the effects of climate change with a longer-term perspective and, inevitably, connected to the capacity of institutions, to the technological capacity of human beings and other organisms to respond to the consequences. Unlike previous indexes, the centrality of the change of sing must be underlined here. In fact, despite the other cases, here lower values reflect high adaptive capacity, while this is the opposite in reality;

• Once the sensitivity and the adaptive capacity are estimated, the process reaches the key point for estimating vulnerability, through the aggregation of the two partial indices. Here, the aggregate Vulnerability Index formula implies a simple average between sensitivity and adaptive capacity indices;

• Once we evaluated the Vulnerability Index, we represented it by GIS software so as to be able to analyze the results obtained according to the classes of vulnerabilities proposed. This exercise makes it possible to classify the territories of the ZOP, subsequently facilitating the identification of priority areas of intervention.

Hydrogeological risk identification

• The calculation of the product among different factors determines a single value of risk propensity at the municipal level for each framework considered. In this case, the final step consists in the calculation of the risk index, derived from the multiplication between hazard, exposure and vulnerability (risk equation in Sect. 2.2), and subsequently normalized on a scale from 0 to 1;

• Considering the need to compare the risk-status with the expected conditions for the future, we decided to adopt a GIS “quantile classification method” in which each class contains an equal number of features. In fact, a quantile classification is well suited to linearly distributed data. The number of classes to adopt for the representation of the risk level is 5: low, medium–low, medium–high, and high. This type of approach allows to investigate through risk what are the possibilities for future territorial transformation in a perspective of adaptation to the effects of climate change.