*2.2. The Climate Change Adaptation Scenarios*

Green infrastructure was one out of the several climate change adaptation options (do nothing, pipe enlargement, new pipes and detention storages and early-warning systems) proposed and analyzed in agreement with the different local stakeholders. In this study, two different adaptation scenarios with future rainfalls are presented: the business as usual (BAU) scenario where no adaptation

is considered, and the GI scenario. The BAU scenario is used as a reference scenario when calculating benefits as part of the cost-benefit analysis. Both BAU and GI scenarios were based on future simulated rainfalls. Two kinds of future rainfalls were estimated for each of the two case studies: a future design storm event relevant for single event flood simulations and a future continuous rainfall time series relevant for continuous urban water simulations that aimed at stimulating combined sewer overflows, water quality impacts on the Mediterranean Sea and annual combined sewer water fluxes at waste water treatment plants. The future design storm events were calculated by applying climate factors (CF) to current design storm events according to Arnbjerg-Nielsen et al. [12]. It is noted that significantly different approaches were used in Barcelona and Badalona in order to derive CF. In Barcelona the 50th percentiles of all Representative Concentration Pathways (RCP) 8.5 and 4.5 scenarios were used to compute CF as a function of both different return periods and rainfall durations. Instead, in Badalona the average values of RCP 8.5 were used to compute CF as a function of different return periods. Nevertheless, the obtained climate values are in both cases within the range proposed in other local studies [13]. Details on the derivation of CF and future rainfall time series are provided in the following. Further future climate variables like temperature, sea level rise, wind, solar radiation, etc., were not considered in the current climate change adaptation scenarios even though they likely impact the future urban drainage systems and GI performances [14].

#### 2.2.1. Green Infrastructure in Barcelona

The proposed GI in Barcelona was agreed with local project stakeholders and it was mostly derived from a study of the Municipality of Barcelona [15] that aimed at increasing stormwater exploitation in the city. Three different types of GI were proposed: green roofs, bioretention cells and retention and detention basins. Figure 1 shows the location of GI throughout Barcelona. Extensive green roofs are assumed to be retrofitted to approximately 5% (143 ha) of all the roof area of Barcelona. This percentage was derived from a study for the Municipality of Barcelona [16] that analyzed the roof areas suitable for green roof retrofitting. Bioretention cells with a total area of approximately 181 ha are supposed to be implemented in almost all the streets of Barcelona as shown in Figure 1. The location and preliminary design of these bioretention systems were proposed in a study for the Municipality of Barcelona [15] that suggested five different spatial distribution and capacity of bioretention systems depending on street slope and width (the five street types classified were presented in Figure 1). The proposed systems are made of a top soil and vegetation layer and a deeper layer of more porous material for water detention and infiltration into the underlying soil. The bioretention cells are devised for managing stormwater runoff from part or the whole streets where they are built. Finally, ten retention and detention basins with a total volume of 128,700 m<sup>3</sup> are supposed to be located at the upstream parts of the urban area in order to collect stormwater runoff from the upstream rural areas for a 10-year return period design storm. Approximately half of the basin volume is allocated to retention with infiltration into the ground and the rest to detention and reduction of peak stormwater runoff. Other examples of the combination of retention and detention volumes can be found in the literature [17].

Overall, the GI implementation in Barcelona would reduce the total impervious area by approximately 14% for all the modelled area. Nevertheless, this reduction is higher in the city center reaching approximately 29%. It is noted that bioretention cells and retention and detention basins do manage stormwater runoff from their associated catchment areas (larger areas compared to their physical construction areas).

#### 2.2.2. The Future Rainfalls in Barcelona

The future rainfalls in Barcelona were computed based on the results of CMIP5 climate models considering the RCP scenarios 8.5 and 4.5. Downscaling methods were then applied and verified using both the ERA-Interim re-analysis as a reference for reproducing the past climate variables and other statistical indicators. Future rainfalls were finally derived using both rainfall observations from local rain gauges and different atmospheric circulation models: ACCESS1, BCC-CSM1, CanESM2,
