*3.3. Stage 2: Preliminary Rank of Adaptation Measures*

The preliminary assessment phase served as an overall screening of the extensive list of adaptation measures considering cost, effectiveness, and welfare aspects. The process involved the ranking of the weighted output of the application of cost-effectiveness analysis (CEA) and the assessment of co-benefits for each measure within a given strategy.

First, the CEA served as comparative assessment of different options that aim to achieve a given objective not measurable in monetary terms [20]. It did so by assessing alternatives in terms of the cost per unit of benefit delivered. In the RESCCUE context, the objective was to increase a city's resilience through the reduction of the floods and CSO spills impacts. The effectiveness of the consecution of this objective was measured through the variation of the recovery time (VRT) from a modeled climate-related event. The recovery time was based on a 1D/2D hydrodynamic model developed to simulate floods in the assessed cities related to a range of return periods. For both case studies, the 1D/2D model was carried out to obtain the time to recover from a flood episode under different measures. An additional 1D drainage model was developed and applied to simulate CSO spills into water bodies and to estimate the average duration of insufficient water quality for the different scenarios modeled. The detailed methodology of both models can be found in [21,22]. The variation of the recovery time for a measure *i* and event *e* (*VRTi*,*e*) was calculated by subtracting the time obtained in the business as usual (BAU) scenario for the same event (*RTBAU*,*e*) with the one obtained with the measure modeled (*RTi*,*e*) (Equation (1)):

$$VRT\_{i, \varepsilon} = RT\_{BALI, \varepsilon} - RT\_{i, \varepsilon} \tag{1}$$

Hydrodynamic modeling is the preferred option to assess the effectiveness of adaptation actions in urban services. However, if a 1D/2D modeling software is not available, a 1D model would also offer an alternative, as it provides the time during which the drainage network is working under surcharged conditions. This surcharged time could be used alternatively as an effectiveness indicator.

The cost was included as the equivalent annual cost (EAC) (Equation (2)), which is the annual estimated cash flow over the lifespan of the project, considering discounting [23]. This allows harmonization of all costs for comparison of measures through the time horizon of the study, set at 2100. They consider initial investment, annual costs, reinvestment (if necessary), and residual value at the end of the assessment time period:

$$EAC\_j = \frac{NPV\_j}{A\_{(t,i)}},\\ A\_{(t,i)} = \sum\_{t=1}^{T} \frac{1 - \frac{1}{\left(1 + i\right)^{T}}}{i} = \frac{1 - \left(1 + i\right)^{-T}}{i},\tag{2}$$

where *A*(*t*,*i*) is the annuity factor which is the sum of all discount factors for the duration of the project; T is the time horizon, and *i* is the discount rate. The discount rate selected for Barcelona and Bristol case studies was 1.23%. It was based on research on the most suitable long-term rate for both regions (Catalonia, Spain and West Country, England) carried out within the European project EconAdapt [24]. This is aligned with the Stern economic school of thought that considers that climate change impact's increases in the long-term future should be accounted for through low or decreasing discount rates [25,26]. Both sites considered the lower range of the discount rate in the scenario modeled with economic growth.

In the case that costs of measures were not available or not accurate enough, a literature review could be carried out in order to develop a scoring system for expected costs of implementation and maintenance of the measures proposed. In the preliminary assessment, costs accuracy of measures is not essential, but actual relative differences between measures is required in order to obtain a realistic preliminary ranking. However, the measures selected to undergo the detailed assessment (next stage) required more accurate results, as their outputs were expected to be more precise.

The equivalent annual costs (EAC) of each measure *i* divided by its VRT, resulted in the cost-effectiveness assessment (CEA) ratio indicator (Equation (3)). In the framework of increasing urban resilience, the reduction of the city recovery time is an important indicator. Therefore, in the CEA indicator, a "penalty" is levied to those measures that do not reduce it at all. Results were ranked from the most (smaller result) to the least (larger result) preferred option:

$$
\begin{array}{c}
\text{CEA}\_{i} = \left\{ \begin{array}{c}
\frac{\text{EAC}\_{i}}{\text{VRT}\_{i}} \\
\text{EAC}\_{i} \times \text{2}
\end{array} \, \middle| \, \begin{array}{c}
if \, \text{VRT}\_{i} > 0 \\
if \, \text{VRT}\_{i} \le 0
\end{array} \right\}.
\end{array} \tag{3}
$$

In parallel, co-benefits were included in order to assess the indirect effects of measures. In both case studies, there was a lack of accessible quantitative information to assess the co-benefits. Although, there was also a strong interest in including them in the prioritization exercise. The solution was found in assessing co-benefits through a site-specific semi-qualitative scoring system, using local and technical experts from each site. Based on the MCA method, stakeholders from both city councils and utilities and technical experts (engineers, economists, and natural scientists) contributed in participatory processes in order to assess the co-benefits [13,27].

In a first round of workshops, there was a selection of indicators from the co-benefits standardization framework in the C40 context by the London School of Economics [13]. From their extensive framework, a multidisciplinary group of experts selected indicators relevant to resilience and urban services. They were classified by economic, social, and environmental co-benefits (Table 1). In a second workshop, the experts working group were asked to score every measure under the selected indicators using a 0 to 10 scoring system. There was a discussion and voting exercise for each indicator and measure, and consensus was found through a session facilitator. Average values were estimated for each category of co-benefit, in order to include average values per category for each measure in the ranking exercise. One should note the caveat related to the quantification of co-benefits related to their extremely context-driven nature [7]. The sign and size of their impact on welfare depend heavily on local circumstances. Therefore, the co-benefits scores assessed for Barcelona and Bristol are unique for those case studies and measures, and new studies should internally assess their own potential co-benefits impacts.

At the end of the second workshop, the working group agreed to assign a weighted percentage to each variable (i.e., CEA (25%) and economic (25%), social (25%), and environmental (25%) co-benefits) in order to align the importance given to the variables with the overall aim of the decision-makers. To calculate the overall ranking, the results of each variable were normalized in order to use a common scale for the different data types on a common scale. Each data point was given a score between 0 and 1, based on its relative position within all values of the variable. This allowed to calculate the weighted scores of each measure per strategy and produce the rankings.

The aim of this phase was to offer coarse results that facilitate decision-makers to shortlist the measures that deserve further analysis. Thus, this stage is only relevant if there is a large number of measures (e.g., more than ten) to be screened. A low number of measures to assess can be considered affordable in terms of finding the resources to carry out a detailed damage assessment for all measures.
