2.3.2. Benefits

Several benefits can be included into CBA of green infrastructure [5,9,14]. Benefits can be direct and indirect, tangible (i.e., that can be quantified in monetary values) and intangibles [26]. In this study, direct and indirect tangible benefits are taken into consideration. The benefits of the GI scenario were calculated as avoided damages (or added values) compared to the BAU scenario that is considered to be the reference as typically done in similar CBA [3,5]. In this study, the benefits were organized into 3 different categories for a better representation and discussion of the results:


CSO spills to receiving water bodies and for avoided costs of combined sewage treatment. Indirect damages are monetized for coastal economies that are affected by the polluted water.

• Additional benefits. Additional indirect benefits are monetized considering: increased aesthetic value, air quality improvement, reduction of the urban heat island effect and energy consumption, and habitat provision [22,24].

Direct flood damages in both Barcelona and Badalona were quantified using coupled 1D/2D (urban drainage/overland flow) models and damage models based on tailored flood damage curves (developed for indoor flood water levels) and permeability coefficient curves that were developed together with flood insurance experts [27]. The damage model takes as inputs the deterministic and spatially distributed values of maximum flood depth simulated with the 1D/2D urban drainage model. The simulated flood depth from the 1D/2D model (considered as outdoor flood depth) is converted into building indoor water levels using the permeability coefficient curves and then the flood damage curves are applied to indoor water levels. In the case of buildings with basements further model parameters control the indoor flood water exchange from ground floor to lower floors. Both the flood models and the damage models (of Barcelona and Badalona) were calibrated and validated using historical data. The flood models used water level data in the drainage network, rain gauge data and photos of urban floods during different past rain events. The damage models used flood insurance compensation data from different flood events during the last few decades [27]. The most influential model parameters of the 1D/2D model were the roughness coefficients of pipes and urban surfaces and of the damage model the parameters controlling the indoor flood water level exchange from ground floor to lower floors [27].

The 1D/2D model provides the maximum simulated flood depth for different design storms of different return periods between 1 and 500 years: 1, 10, 50, 100 and 500 years for Barcelona and 2, 10, 100 and 500 for Badalona. For each return period, the total flood damages at the urban scale were calculated by multiplying the maximum simulated flood depth at each cadastral parcel by permeability coefficients and flood depth–damage curves that were specifically tailored for Badalona and Barcelona for different land uses (hotels, warehouses, restaurants, dwellings, car parks, etc.) and vehicles [28]. The permeability coefficient curves were used to transform the 2D simulated flood levels on the urban floodable area into indoor water levels. Finally, Expected Annual Damage (EAD) was calculated including both direct and indirect damages as detailed in a previous study of Badalona [27]. Indirect flood damages due to business interruptions were estimated at 29% of the total direct damages using an input–output model [27]. This percentage is in the range of other studies that proposed 19–39% [29,30].

The 1D/2D hydrodynamic models were developed with InfoWorks ICM (www.innovyze.com) and calibrated and validated using local rainfall and water level data. The 1D sewer model of Badalona includes approximately 368 km of pipes, 11,338 manholes, 11,954 sub-catchments, 62 weirs, 4 sluice gates, and 1 detention tank of 30,000 m3. The 2D model has 199,338 cells that form an unstructured mesh generated from a digital terrain model (DTM) of 2 m<sup>2</sup> resolution obtained by a LIDAR with a precision of approximately 15 cm for the altitudes. The size of the 2D cells is in the range of 16–64 m2 in the urban areas where most of the flood damages occur. The 1D sewer model of Barcelona includes approximately 2041 km of pipes, 85,834 manholes, 980 weirs, 44 sluice gates, 75 pumps and 285 storage nodes representing different kinds of chambers and 10 detention tanks with a total volume of more than 400,000 m3. The 2D model has 1,361,324 cells that form an unstructured mesh generated from a digital terrain model (DTM) of 2 m2 resolution obtained from a LIDAR. The size of the 2D cells is in the range of 25–100 m2.

Direct and indirect water quality benefits were computed using continuous simulation of a 1D urban drainage model to estimate annual volumes of CSO and combined sewage water sent at the treatment plant. The urban drainage models used were the 1D/2D models presented earlier but without the 2D overland flow model. The urban drainage models were then coupled to a sea water quality model [31] to simulate the sea water contamination from CSOs and to estimate the average duration of insufficient bathing water quality. The duration of insufficient bathing water quality was used as

an input to a coastal economy model that estimates indirect damages to coastal economies caused by pollution of bathing waters and the consequent reduction of sea related leisure, sport and restoration activities. The coastal economy model includes different contributions. First, the daily direct added value of the coastal economy was calculated by selecting the expected business sectors affected by a beach closure (restaurants, small retails and maritime sector). This selection was based on the results of a field study based on surveys to beach goers and personal interviews to coastal business owners carried out in Barcelona and Badalona (see both H2020 BINGO and RESCCUE projects). Second, based on data from Barcelona's economic annual report [32] a 50% share of the annual coastal economic added value was assumed to come from the bathing season [33], which lasts approximately 3–4 months in Badalona and Barcelona. Only the direct added value of coastal districts (identified by comparing the CSO spill points with the districts maps) affected by CSO spills were included. Furthermore, assumptions of the magnitude of the impact per sector were made based on the results of the local surveys: 50% impact to restaurants, 25% to retails and 25% to maritime sector (water sport and private fishing). The daily economic impact obtained by dividing the value added by the number of days of the bathing season, was finally multiplied by the average number of sea water pollution days (where the beaches could potentially be closed to bathing) to estimate the potential annual indirect damages to the coastal economy.

The different GI systems were simulated in both the 1D and the 1D/2D drainage models by converting the planned GI areas from impervious areas into pervious areas with hydrological losses. This simplified approach was also used by Velasco et al. [4]. However, to the knowledge of the authors, this method was not validated with hydrological data and can be a limitation.

The direct damages produced by CSO spills to receiving water bodies were calculated using a reparation cost method, which assumes that the value of the damage is equal to the cost of repairing it [34]. The direct damage produced by CSO spills was obtained multiplying the average annual CSO volume by the unit CSO damage cost of 0.7 €/m<sup>3</sup> in Badalona. Instead, in the case of Barcelona different values were used: 2.69 €/m<sup>3</sup> for CSOs to the sea and 1.50 €/m<sup>3</sup> to the river and the harbor according to a Spanish regional normative devised for industrial spills [35]. Another benefit considered was the reduction of the sewage water to be treated by the wastewater treatment plants (WWTPs). The monetization of this benefit was calculated as the avoided costs of combined sewer water treatment that were estimated by multiplying the average treated annual volumes from the urban drainage model with a selected unit treatment cost of 0.12 €/m3 that is considered reasonable for local WWTPs based on local expertise. The tangible indirect damages (and the consequent benefits calculated as avoided damages) to coastal economies were estimated using the pollution time from the sea water quality model and the coastal economy model explained before.

The additional indirect benefits considered are based on four contributions. The first is aesthetic value which is monetized as the willingness to pay for properties nearby, or that include green infrastructure, is measured through the increase of the value of these properties. This value could also include the increased property taxes acquired by the taxation authorities [8]. In this case, the benefits were estimated with a benefit transfer method to be the 3% of the CAPEX of GI [22]. The benefits derived from the reduction of energetic consumption (for indoor heating and cooling) and heat island effect are quantified using 0.049 €/m2/y per green roof unit surface [24]. Benefits derived from urban heat island reduction obtained with bioretention cells (that in Barcelona are planned) were not included and this can be a limitation. The air quality benefits are derived from both emission reduction (of CO2 and Nox) capacity of GI that was estimated to be 0.072 ton/ha and the cost of emissions of 3051 €/ton [24,36]. The habitat provision was based on the potential increase of urban ecosystems that support wildlife and it was estimated to be 2.8 €/m2 for both case studies. This was estimated using a benefit transfer method from a study that assumed the value of habitat creation could be estimated at 15% of the value of natural land [22].
