Pessamit is an indigenous community located on the North Shore of the St. Lawrence maritime estuary. Pessamit’s coast, which extends over a 12 km span, is mainly composed of unconsolidated cliffs (53.7%), littoral spit (21.0%), salt marsh (20.7%), and beach terrace (4.6%). The tidal range is mesotidal and the offshore significant wave height (95^th percentile) is 0.85 m (depth of 132 m). The study site is located on a sandy littoral spit. A wide unvegetated sandy foreshore is present in front of the coast. The study site is a part residential, and part public sector, highly frequented by the local community for a variety of uses and activities: off-road vehicles, boat launching, gatherings, walking, beach activities, archeological site, waterfowl concentration area, etc.

Sept-Îles is located on the North Shore of the Gulf of St. Lawrence. The study site, Gallix, is located in the Sainte-Marguerite Bay, which extends along 26.90 km, and is mainly composed of littoral spit (10.4 %), unconsolidated cliff (47.0 %), beach terrace (41.8 %), and rocky shore (0.8 %). The tidal range is mesotidal and the offshore significant wave height (95^th percentile) is 0.75 m (depth of 90 m) (Corriveau et al., 2021). The study site is located on a sandy beach terrace, and is characterised by the formation and disappearance of a sandy triangular salient ( Figure 3 ). This dynamic is generated by estuarian currents and littoral drift. Storm events, between 2009 and 2017, have caused high shoreline retreat (7.96 m/yr), leading to a shoreline enlargement on the west side of the coastal sector ( Figure 3 ). A new salient slightly to the east was formed between 2013 and 2016, and was still present in 2017. That new salient modifies the longitudinal sediment transport processes, leading to an offshore sediment deviation. Sediments are redirected towards the coast, further to the West, resulting in a sediment progradation in the coastal sectors (Corriveau et al., 2019). A rock armour structure is present between the salient formed in 2013-2016, and the high shoreline retreat sector ( Figure 3 ). The sandy lower foreshore is partially covered (0-25 %) by a narrow band of macroalgua (mainly laminaria longicruris). A bar system is also present in front of the study site. The study site is a residential sector, with high scenic and socio-cultural values, and is mainly used by the local community for fishing, nautical and beach activities, gatherings, walking, etc. It is also a waterfowl concentration area, and spawning ground for capelin.

Cap-Chat is located on the Gaspé Peninsula on the south shore of the St. Lawrence river. Baie-des-Capucins, a bay located east of Cap-Chat, extends over approximately 3.2 km, and is mainly composed of beach terrace (54.4 %), unconsolidated cliff (15.0 %), rocky cliff (10.3 %), and salt marsh (8.1 %). The study site is located on a beach terrace composed of a mixture of coarse sand, cobbles and pebbles, and by a wide salt marsh vegetated with spartina alterniflora. The bay’s entrance is partially (1-25%) vegetated with Zostera marina and Fucus sp. The tidal range is mesotidal and the wave energy is low. The littoral drift is directed towards the inside of the bay. The site is a biodiversity hotspot, and the main activities are nautical activities, walking and fishing. The national road 132 is, in some areas, less than 5 metres from the shoreline.

Gaspé is located on the Gaspé Peninsula on the south shore of the Gulf of St. Lawrence. Cap-des-Rosiers, a former village annexed to the town of Gaspé, is located at the far north-east end of the Gaspé peninsula. The study site is located on a rocky cliff stretch of coast with an unconsolidated lower (cobbles) foreshore and a rocky infralittoral zone. The lower foreshore is partially (0-25%) covered by macroalgua (mainly fucaceae). The tidal range is mesotidal and the offshore significant wave height from ESE-SE reached more than 3 m between November 2017 and 2019 (Savoie-Ferron et al., 2020). The study site is a residential sector with low density activities. Still, tourism infrastructures, such as motels, are present in Cap-des-Rosiers due to the proximity of the Forillon National Park, which brings a high volume of tourism during the summer season. The main activities are gatherings and relaxation. Cap-des-Rosiers is a biodiversity hotspot, and has a high socio-cultural value with, among others, the Cap-des-Rosiers lighthouse, the tallest in Canada. The national road 132 is, in some areas, less than 5 metres from the cliff.

The PROMETHEE method is based on a pairwise comparison between different alternatives, following their assessment A = {a₁, a₂, …, a_n} against each criterion c_k (Δ_k(a_i, a_j)) from a defined set of criteria C = {c₁, c₂, …, c_m} (equation 1). The variations in the results of the assessments Δ_k(a_i, a_j) associated with criterion c_k, are translated into a preference index P[Δ_k(a_i, a_j)] through a preference function, which lies between 0 and 1, 1 being a strong preference, and 0 meaning no preference. In this study, the usual preference function was used; it corresponds to the optimization of values without threshold, that is, larger values are better than lower ones. The multicriteria index π(a_i, a_j) is the weighted sum of the preference index P, and is calculated by dividing the preference index by the weight w_k, which is a measure of the importance of the criterion c_k determined by workshop participants (equation 2). The leaving Φ⁺(a_i) and entering Φ^-(a_i) flows are then calculated in relation to the multicriteria index. The leaving flow expresses the extent to which a_i outranks all other alternatives (equation 3), and the entering flow expresses how much a_i is outranked by all other alternatives (equation 4). Thus, the best alternative has the highest leaving flow and the lowest entering flow. The net flow is the sum of the leaving and entering flows and represents an overall ranking. All criteria were set as maximum with the exception of the criteria whose evaluation was based on the 5 points impact qualitative scale (geomorphological effects, ecological effects and aesthetics). The specificities of the PROMETHEE method are described in the works of Brans et al. (1986) and in the Visual PROMETHEE software’s manual (VPSolutions, 2013).

Coastal zone stakeholders were consulted in the course of two action research projects, with the aim of developing tools to improve coastal planning and protection, and to facilitate the choice of solutions adapted to climate change, in the short, medium and long terms. First, four workshops were organized between April and December 2019, in Eastern Quebec, in the context of the Coastal Resilience project. Stakeholders who were invited to the workshops included administrative personnel and professionals from local municipalities and coastal MRCs (regional county municipalities), relevant ministries, local and regional organizations, and members of the First Nations ( Table 2 ). Second, in June 2019, in the context of a project entitled Identification de solutions d’adaptation aux aléas côtiers pour augmenter la résilience des communautés des Premières Nations dans un contexte de changements climatiques, a workshop was organized in Pessamit. Participants to this workshop comprised administrative personnel and professionals from the community. Through various activities, one of the aims was to integrate stakeholders into the decision-making process leading to the implementation of coastal defence measures, by asking them to identify and weigh the CDMs selection criteria ( Figure 4 ). The locations of the four study sites and five workshops are shown in Figure 2 .

For Pessamit, all participants were employees of the municipality, many of them, members of the First Nations; they were all recorded under Municipality rather than First Nations to avoid duplication.

Using an adaptation of the World Cafe methodology (Brown et al., 2005), the participants were separated into rotating discussions groups of between five and fifteen people. The aim was to allow participants to express their views on different subjects related to coastal management, each within a 25-minute time frame. With two facilitators in charge of the discussion, one table was dedicated to discussing relevant criteria to be integrated in the decision-making process for the selection of coastal defence measures.

At the beginning of the discussion, participants were presented with a brief description of factors affecting the CDMs decision-making process, in order for them to understand the context before identifying the CDMs selection criteria. The question “In your opinion, which are the criteria to be considered when a coastal defence measure must be selected?” was asked of participants to start the discussion. Five criteria categories written on cards were presented to participants as a guide ( Table 3 ). Specific criteria identified by participants were written on post-it notes and affixed to their corresponding category. The selection criteria identified by participants of one table served as a basis for discussion in the following rotating groups.

A weighting method was established to allow participants to assess the importance of the criteria they previously identified. A three steps classification system was created: first, a list of criteria, as identified by the participants, was compiled; second, from that list, criteria that were similar in nature were amalgamated by the facilitators under a maximum of five general criteria per criteria category; third, the general criteria were grouped under five criteria categories.

Each criteria category was limited to a maximum of five general criteria to limit the total number of criteria, and because it was deemed sufficient to capture and adequately represent all of the criteria identified by the participants. Each general criterion was written on a card with the list of similar criteria originally identified by the participants. The weighting of the general criteria was carried out in two steps. The voting table was divided into six sections: one for each criteria category with its list of general criterion, and one which listed the five criteria categories ( Figure 5 ). As a first step, each general criterion card was placed on the voting table under its respective category, with a corresponding voting box. Each participant was given ten tokens per criteria category to weight each general criterion according to the question “For each criteria category, how important is each criterion when selecting a coastal defence measure?”. As a second step, the five criteria categories’ cards were placed in the sixth section, each with a corresponding voting box. As for the previous exercise, participants were given ten tokens to weight the criteria categories according to the question “How important is each criteria category for the selection of a coastal defence measure?”.

The evaluation of CDMs in relation to each criterion was based on experts’ judgement, and on different sources of data such as literature reviews, reports on CDMs, etc. In the PROMETHEE method, different rating scales can be defined, depending on the nature of the criterion. Here, two scales were used: a 9 points qualitative scale for the criteria that lead to the comparison of CDMs’ performance relative to each other [very good (9), very good – good, good, good – average, average, average - bad, bad, bad – very bad, and very bad (1)]; and a 5 points impact qualitative scale for the criteria that lead to the evaluation of the direct effects of CDMs on the coastal system [very low (5), low, moderate, high, very high (1)]. Two qualitative scales were used for the evaluation of criteria due to a lack of quantitative local data. For example, the evaluation for the cost related criteria was based on the cost of previous projects at a national scale.

The CDMs evaluation is dependent on each criterion ( Table 5 ), and on the socio-ecological features associated with each site. The evaluation scale was adapted to the nature of each criterion.

The evaluation of the CDMs in relation to each criterion and to each study site’s socio-ecological features is presented in Table 6 , with rationale where relevant.

The CDMs hierarchization is thereafter presented in diamond shape figures for each study site ( Figure 7 ). This shape shows the overall ranking (Φ net) along the vertical axis, as well as the leaving Φ⁺ and entering Φ^- flows along the left edges of the diamond. Each CDM is represented by a grey dot. For one CDM to be ranked higher than another, it must outperform it on both the leaving and entering flow axes.

For the Pessamit site, the analysis of the results shows that, according to the workshop participants’ preferences, vegetation is ranked first followed by beach nourishment. Despite the large outperformance of vegetation over beach nourishment on the net and leaving flows, their respective entering flows are quite close together, thus weakening the general outperformance of the vegetation option. In third and fourth ranks, while permeable groin outperforms foreshore nourishment on the leaving flow axis, foreshore nourishment outperforms permeable groin on the entering flow. This results in a quasi-tie on the net flow axis. Meanwhile, impermeable groin and emerged breakwater come last, in fifth and sixth ranks, respectively.

For the Gallix site, the vegetation option is ranked first as shown by its significant outperformance over all other CDMs. In second and third ranks, a quasi-tie between beach nourishment and permeable groin on the net flow axis is explained by an outperformance of foreshore nourishment on the leaving flow, and an outperformance of permeable groin on the entering flow. Impermeable groin and foreshore nourishment are also on a quasi-tie in fourth and fifth ranks, respectively, while submerged breakwater and emerged breakwater are ranked sixth and seventh.

The top five CDMs for the Pessamit and Gallix sites are in the same order. Two reasons can explain the similarity. First, their socio-ecological systems are comparable. The environmental characteristics of both sites are low-lying sandy coast with a sandy lower foreshore, mesotidal shoreline, and a low energy environment (<1 m). Also, both sites are part residential part public localities, mostly frequented by the local community for a variety of uses and activities. Due to these similarities, the evaluation of the CDMs in relation to each criterion is equivalent for all criteria with the exception of social perception ( Table 6 ). Second, the difference in average weight is below 3 % for seven of the 13 and 15 criteria identified, respectively, in Pessamit and Gallix ( Figure 7 ). The local stakeholders’ preferences were quite similar in nearly half of the identified criteria, though variances explain differences in the net, leaving, and entering flows between the two sites ( Figure 7 ). While the final ranking for the top 5 options is equivalent in both sites, inner differences shown by the above-mentioned three indicators, provide information that is relevant to the decision-makers.

For the Baie-des-Capucins site, the vegetation alternative also outperformed all other CDMs. In second and third ranks, while low-crested breakwater outperformed beach nourishment on the net, and entering flow axes, both are on a quasi-tie on the leaving flows axis. It is to be noted here that, because of the presence of salt marsh, the low-crested breakwater scenario was actually based on a living shoreline rock sill concept (Bilkovic et al., 2017). Finally, permeable groin was ranked fourth.

As for Cap-des-Rosiers, rock armour outperformed all other CDMs, but seawall is a close second. Beach nourishment, low-crested breakwater and emerged breakwater were ranked third, fourth and fifth, respectively. The evaluated CDMs and the final ranking is quite different from the other three sites, which is explained by their significantly dissimilar local conditions.

An analysis was performed to evaluate the robustness of the CDMs hierarchization method in relation to each criterion ( Figure 8 ). The circles and triangles show the highest and lowest thresholds at which there is a change in ranking of the top 3 CDMs for a given criterion. For example, in Pessamit under the criterion Economic repercussions, the CDM ranked first is outperformed by the CDM ranked second above 21.06%, but it is never outperformed by the CDM ranked third. Also, in Pessamit with regard to the three criteria Ecological effects, Regulatory liability, and Technical feasibility, the circles representing CDMs 1, 2, and 3 are all at 100%. This means that no change occurs in the top 3 CDMs’ ranking, no matter the weight attributed to those criteria. An empty criterion cell indicates that the criterion was not identified at the site’s workshop.

The grey plus and minus signs show the weighted average plus or minus the standard deviation. For example, if there is no circle or triangle within the range of the standard deviation (between the plus and minus signs), any change to the weight of the criteria would not affect the ranking, which is an indication that the results are robust. In Pessamit and Gallix, the CDMs ranked 1 and 2 are outside the standard deviation range for all criteria, which indicates that the results are robust for the first two CDMs. On the other hand, because CDM ranked 3 sometimes falls within the standard deviation range, it may be outperformed by the CDM ranked 4 if a change occurs in some of the criteria weight values. This shows that the results for CDM ranked 3 are not as robust as the ones for the first two CDMs. In Cap-des-Rosiers, all of the CDMs fall outside of the standard deviation range. This indicates that, in the case of the first three CDMs, the results would not be affected by any change in the criteria weights. In Baie-des-Capucins, the CDM ranked 3 falls within the standard deviation range for the criteria Construction costs, and Social repercussions. In this case, two out of sixteen criteria would not be enough to change the ranking.

Considering Figures 6 and 7 , even though the results are not equally robust in all cases, vegetation, beach nourishment, and permeable groin were ranked first, second, and third, respectively, for the three study sites with a sandy low coast (Pessamit, and Gallix). As for Cap-des-Rosiers, with a cliff coast that is naturally reflective, rock armour and seawall were solidly ranked first and second. These results are in line with those obtained through the CDMIA developed by Sauvé et al., (2022), but the MCDA adds a layer of refinement to the assessment.

In Pessamit, the quasi-tie between permeable groin and foreshore nourishment in third and fourth ranks (a) is validated by the robustness analysis, which shows that the lowest and highest thresholds of the CDM ranked 3 (represented by a green triangle and a green dot, respectively), each fall within the standard deviation range in 5 of the criteria (a). In Gallix, the quasi-tie between beach nourishment and permeable groin (b, ranks 2 and 3, respectively) is confirmed by the robustness analysis, which shows the permeable groin’s lowest threshold falling within the standard deviation in 3 of the criteria, and the highest threshold, in 6 of the criteria (b). In Baie-des-Capucins, there is no quasi-tie in the CDMs hierachization (c). Indeed, the robustness analysis shows that the thresholds for CDMs ranked 1 and 2 all fall outside of the standard deviation, and even though, in one criterion, the highest threshold falls within the standard deviation for the CDMs ranked three (permeable groin) (c), any change in the criteria weighting would not be enough to affect the final hierarchization. In Cap-des-Rosiers, the CDMs hierarchization is unambiguous (d), and is validated by the robustness analysis which shows that, for all criteria, the top 3 CDMs fall outside the standard deviation (d).