The PAT-CDSJ is an action plan built by the government of Bahia in a participatory process, including decision-makers, scientists, and stakeholders (INEMA et al., 2020). This document is an instrument that helps to guide public policies and put them into practice. The PAT-CDSJ presents a set of objectives and proposed actions that will solve problems and minimize threats to species and their ecosystems (see details of the PAT-CDSJ document and actions in the Supplementary Material 1). The selection of strategic areas for conservation and restoration was originally requested by decision-makers, who are members of the regional government, involved and responsible for the actions of the PAT-CDSJ. Hence, this study is part of the first action of the PAT-CDSJ and will support its subsequent actions.

The study area considers the limits defined for the PAT-CDSJ, extending between the central and eastern regions of the state of Bahia, Brazil (Fig. 1). The limits include portions of 56 municipalities and cover 3,918,743 ha (INEMA et al., 2020). The objective of the PAT-CDSJ is to develop public policies in unprotected areas, as areas protected are already covered by the law. The Chapada Diamantina occurs mainly in the Caatinga Brazilian domain (i.e. semi-arid vegetation), while Serra da Jiboia falls within the Atlantic Forest Brazilian domain. This region comprises a wide range of climatic conditions, geological formations, soil types, vegetation (Fig. 1 and S1), high diversity and degree of endemism (Giulietti et al., 1997; INEMA et al., 2020). Most of the 27 CR-Gap species (24 plants and three invertebrates, Table S1) of the PAT-DCSJ are endemic to mountain regions of Bahia, Brazil, and low density and scattered populations (INEMA et al., 2020; IUCN, 2022), common characteristics of gap species (Rodrigues et al., 2004).

Fig. 1.

The study area of the Territorial Action Plan for the Conservation of Critically Endangered Species and their Ecosystems of the Chapada Diamantina-Serra da Jiboia – PAT-CDSJ, including municipality divisions, elevation, and the Brazilian phytophysiognomies Caatinga (where Chapada Diamantina occurs, orange crosshatched) and Atlantic Forest (where Serra da Jiboia occurs, green crosshatched). The small polygons within the limits of the PAT-CDSJ (white areas) are protected areas, that are not part of the PAT-CDSJ territory.

(0.75MB).

All stages of this study were discussed in several participatory workshops and meetings attended by scientists (from national, state and local universities), decision-makers (i.e. regional government administrators), and stakeholders (e.g. from local NGOs, managers of protected areas close by the PAT-CDSJ limits, and national government members) (Fig. 2, Table S2). In the first meeting, the decision-makers presented the PAT-CDSJ document and their expectations about the mapping action (i.e. SASC and SAER). Subsequently, in the first participatory workshop, the scientist responsible for the mapping suggested a work plan. Using the work plan as a starting point, the participants consensually defined the methodology and the data that should be used, considering local context, data availability, main species threats, project viability, and the relevance and cost of each mapping criterion (see details below). After defining the criteria, the participants assigned weights to each of them.

Fig. 2.

Participatory workflow. Steps and products generated during the participatory mapping process of the Strategic Areas for Species Conservation (SASC) and Strategic Areas for Ecosystem Restoration (SAER) of the Territorial Action Plan for the Conservation of Endangered Species of the Chapada Diamantina-Serra da Jiboia (PAT-CDSJ). Technical products generated (dark orange), a brief explanation about each product (light orange), meetings and monitoring workshops (dark green), the purpose of the meetings and workshops (light green), construction workshops (purple), participants (blue), the final product (pink).

(0.58MB).

The intention for the work plan was to be inclusive, considering several possible environmental and socioeconomic aspects, so the participants could discuss and decide the better aspects and strategies. Among all the possible favorable and unfavorable aspects for the mapping, the objective of the work plan was to reach a reduced number of criteria for the selection of areas. For example, regarding the planning units, participants discussed whether to consider the planning unit as a regular polygon (e.g. hexagons) or microbasins, and which microbasin level would be adopted. Regarding the criteria, the participants discussed several possibilities of landscape metrics (e.g. amount of habitat, connectivity, distance between fragments), human pressures (e.g. proximity to mining areas, fire frequency, type and proximity of agricultural areas), and proximity to Permanently Preserved Area (APP, Portuguese acronym) and protected areas. All participants had to justify why or why not each decision of the methodological choice was made.

The scientist responsible for the mapping, then, gathered all data to develop the mapping – partly compiled by online databases (e.g. occurrence records and environmental variables), partly provided by the regional government members (e.g. limits of the PAT-CDSJ and APP), and partly provided by the other scientists (e.g. occurrence records and phytophysiognomy classification). The scientist responsible also developed all technical procedures for mapping, however, most steps were assisted by other scientists and decision-makers in brief meetings. For example: (i) the scientist responsible compiled occurrence records from databases and the taxonomists validated them, and (ii) the scientist responsible proposed an APP design and a geography technician from INEMA validated the proposal and provided additional data (see all details in the subsections below).

The preliminary mapping was presented in a second participatory workshop, where all participants gave suggestions to the final mapping. The SASC and SAER mapping were also presented and discussed in two monitoring workshops of the general PAT-CDSJ project, where the progress of the other actions was also presented by other scientists or decision-makers responsible. Finally, the new mapping proposal was presented to the decision-makers, then, sent to all participants for suggestions and final approval.

We considered four and five criteria to define the areas of conservation and restoration, respectively (see details in Supplementary Material 1). In this order, the most relevant criteria were: ‘Habitat amount’, ‘Environmental suitability’, ‘Fire frequency’, ‘Permanently Preserved Area amount’ (this one was used only for SAER), and ‘Diversity of phytophysiognomy’. As the planning units (PUs) used to evaluate each criterion, we chose the microbasin limits established by National Hydrographic Division, level 6 (ANA, 2015), dividing the PAT-CDSJ into 314 Pus – planning unit definition often used in other established public policies and because it can also be a natural ecological barrier for some species or populations of animals and plants (e.g. Loyola et al., 2018; Monteiro et al., 2018).

The criteria listed above do not necessarily have the same relevance for the SASC and SAER mappings. To define the relevance, we used the method Analytic Hierarchy Process (AHP) – commonly used in Brazilian action plans (e.g. ConsórcioGeoBahia et al., 2018; Rocha et al., 2019) – which compares pairs of criteria according to the fundamental scale of Saaty (1977) (see Table S4).

To achieve the consensus among the participants during the workshop, the AHP was applied by panel cards (Fig. S7). First, we discussed the relevance of criteria for both mappings (SASC and SAER), separately. Then, each criterion was exposed in a panel and the participants ranked their relevance, based on theoretical grounds and application feasibility. Subsequently, all participants consensually assigned weights to each criterion. To estimate the consistency among the weights, we calculated the Consistency Ratio (see details in the Supplementary Material 2), where values near 1 indicate low consistency and near 0 indicate high consistency (Saaty, 1980). Then, we considered ratios below 0.1 as consistent to validate the weights assigned. Finally, we multiplied each standardized result (i.e. normalized criteria from 0 to 1) by its weight. All weighted criteria were then summed to generate a single result for each species, according to the equation:

We defined the order of criterion relevance for SASC (except for item 4) and SAER mapping, as follows: 1) Habitat amount, 2) Environmental suitability, 3) Fire frequency, 4) APP amount, and 5) Diversity of phytophysiognomy. The criterion 1 was the most relevant because the presence of habitat is determinant for species occurrence. Criterion 2 came in sequence, as the environmental suitability is determinant for species persistence, followed by criterion 3, as it is one of the main threats to local species (INEMA et al., 2020). For restoration, criterion 4 was chosen considering the importance of maintaining water resources and encouraging compliance with the law. Finally, criterion 5 was the least relevant, as the heterogeneity is important but not determinant for species occurrence and persistence. Once the priority order was raised, each criterion received a weight linked to its relevance.

We generate two synthesis maps (for conservation and restoration) summing the 27 altimg="si3.svg"Rf -maps, in which the greater the PU value, the greater its relevance. We selected from the synthesis map only the 30% most relevant PUs. According to the decision-makers, this is a viable percentage for the implementation of subsequent public policies, considering the budget available to the PAT-CDSJ.

To select the most relevant PUs, we first considered all PUs with CR-Gap species occurrence records, to ensure that all sites where the species have been found were considered, helping to reduce the gap species in the current protected areas in Brazil (Rodrigues et al., 2004). Next, we selected the remaining PUs with the highest relevance according to all criteria together (i.e. those that no longer had occurrence records). This selection was proportional to Chapada Diamantina and Serra da Jiboia regions. We distinguished the strategic areas based on a set of PUs that were close to each other and within political units. The participants delimited and named the strategic areas according to their knowledge about the socio-environmental characteristic of the region (see details for delimiting and name-giving procedures for strategic areas in the Supplementary Material 1).

For both conservation and restoration maps, criteria showed a more dispersed pattern throughout the territory, except for habitat amount criterion in the conservation map, that showed higher relevance mainly around the protected areas (Fig. 3). The synthesis map for conservation showed the same pattern of the habitat amount criterion – as this criterion received the highest weight compared to the others – while for restoration, it showed a more dispersed pattern (Fig. 3). By restricting the result to the 30% most relevant PUs, we selected 86 PUs in the Chapada Diamantina (in which 32 PUs had species occurrence records, Fig. S8) and 8 PUs in the Serra da Jiboia (in which 1 PUs had species occurrence records, Fig. S8). Finally, the strategic area delimitation provided 11 SASC, more concentrated around the protected areas (Fig. 4A, Table S5), and 12 SAER, better distributed across the PAT-DCSJ (Fig. 4B, Table S6).

Fig. 3.

Relevance of UPs by criterion and the synthesis result (i.e. all criteria and species) for ecosystem conservation (left) and restoration (right) mapping. The relevance of UPs is associated with the cost, presented along the gradient from ‘relevant’ (highest cost) to ‘highly relevant’ (lowest cost). The upper right numbers refer to the weight of each criterion.

(1.29MB).

Fig. 4.

Strategic Areas for Conservation (SASC, 11 in total) (A) and Strategic Areas for Ecosystem Restoration (SAER, 12 in total) (B), including municipality divisions in the Territorial Action Plan for the Conservation of Critically Endangered Species and their Ecosystems of the Chapada Diamantina-Serra da Jiboia – PAT-CDSJ.

(0.85MB).

Approximately 29 participants attended the two participatory workshops and the two monitoring workshops, while assistance meetings (approximately seven) had fewer participants, with decision-makers always present. The four workshops lasted approximately 4-hs, and the assistance meetings were shorter (∼2-hs). The group of participants was heterogeneous, consequently, during the meetings, some divergent suggestions were exposed, however, we always reached a consensus before moving on to the next step.

In general, while scientists focused on the ideal methods and biological aspects (as most scientists were biologists), the decision-makers and stakeholders focused on the usefulness and feasibility of the map implementation and political aspects. For example, considering that the main conservation target is the CR-gap species, at first, scientists suggested the use of basins as planning units due to natural limits, which is more relevant for biological aspects. Some decision-makers, although, suggested political units (i.e. municipality limits) due to the ease of carrying out public policy actions in these units. Considering both aspects, they decided to use basin limits for the smaller units (i.e. PUs), and aggregate these units considering political limits into strategic areas.

Another example was related to the presentation of the strategic areas. Some scientists suggested presenting all PUs with gradual colors, from the most to the least relevant areas. They argued that even inside of the strategic areas, the gradual design provides the idea of a priority ranking that could guide the order of the implementation of the action plans. However, the decision-makers defended the delimitation of the areas without ranking. The reasons were: (i) a gradual map is not informative, as they already know that the whole region is important – instead of deciding later which are the best areas to implement new actions, they wanted to delimit it to this exact action, (ii) a delimited result is more intuitive to present to other stakeholders and decision-makers, avoiding the misinterpretation that “some of these selected areas are not so important”, and (iii) the beginning of the next actions will be almost simultaneous in all the defined areas, and the order of implementation of the actions will depend especially on the articulation with the local decision-makers. The solution was to keep both results, the gradient maps as a supplement and the delimited SASC and SAER as the final design of the map.

Other important discussions occurred regardless of which group the participants belonged to. For example, they evaluated that most criteria related to human pressures (e.g. mining, agricultural, cattle, and urban areas) coincided with the areas reported by ‘Habitat amount’ and ‘Fire frequency’ criteria. Therefore, they decided not to include these human pressures to avoid redundancy and highly correlated criteria.

Regarding the fire frequency, some participants adverted to the high cost of considering PUs with high fire frequency to conservation and restoration, as: (i) for conservation, these PUs would not be a priority due to the high level of degradation, and (ii) for restoration, as there is a need for high-cost restoration efforts and the susceptibility of these areas to recurrent degradation is prominent. On the other hand, other participants warned of the urgent need to conserve and restore the most threatened areas with higher fire frequency. They decided to prioritize low fire frequency for strategic areas, based on the items (i) and (ii), and because fire control is one of the proposed actions of the PAT-CDSJ, but not the principal, and the resources are limited to each action of the PAT-CDSJ (see Supplementary Material 1).

According to the preliminary results of the synthesis maps (i.e. all criteria together), Serra da Jiboia presented lower priority than Chapada Diamantina. The participants argued that these results implied a possible non-prioritization of the Serra da Jiboia region, going against the objective of the PAT-CDSJ. Thus, it was suggested to separate Chapada Diamantina and Serra da Jiboia in the analyses and proportionally prioritize both regions (as described in the Methods).