Correspondence Rebecca Weeks, School of Marine and Tropical Biology and Australian Research Council Centre of Excellence for Coral Reef Studies, James Cook University, Townsville, Queensland 4811, Australia. Tel: +61-7-47814853; fax: +61-47816722. E-mail: Rebecca.Weeks@jcu.edu.au
Although the importance of socioeconomic factors in conservation planning is increasingly recognized, there are few examples demonstrating how such factors can be practically incorporated into the design of protected area networks. Here, we illustrate how spatial zoning software can be used to consider local marine tenure explicitly in the design of marine protected area (MPA) networks, using a case study from the Philippines. By stipulating the minimum area of fishing grounds that must remain open to each local fishing community, we were able to design MPA networks that impacted local resource users more equitably and were therefore more likely to be socioeconomically viable. MPA networks that considered local tenure boundaries had a greater overall area and cost than those that sought to minimize costs to small-scale fishers as a single stakeholder group. However, in this context, established concepts of “efficiency” in conservation planning are likely to be less important than minimizing costs to each fishing community individually.
Systematic conservation planning is the process of designing, implementing, and maintaining protected areas to achieve explicit objectives for biodiversity conservation (Margules & Pressey 2000; Pressey & Bottrill 2009). In the last two decades, conservation planning has evolved from a largely academic discipline to influence conservation action on the ground and in the sea (e.g., Fernandes et al. 2005; Klein et al. 2008; case studies in Pressey & Bottrill 2009). Nevertheless, in many regions of the world where conservation action is urgently needed, systematic planning has had relatively little influence on local-scale management initiatives. One reason for this planning–implementation gap (Knight et al. 2008) is the mismatch between the regional scales at which planning has typically been undertaken and the local scale at which implementation occurs. Another is that the social, economic, and political context in which systematic conservation planning methods and tools have been developed is different to that in which we now need to apply them (Christie et al. 2007; Cinner 2007). The fine spatial scale of governance, existence of formal or informal local tenure, and limited spatial mobility of small-scale fishers (who are often the primary stakeholders for conservation in coastal waters) all present new challenges to the application of conservation planning.
How resource tenure systems shape conservation outcomes in different social and ecological contexts has been identified as one of 100 questions of conservation importance (Sutherland et al. 2009). Local or customary marine tenure has been documented worldwide (Johannes 2002; Cinner & Aswani 2007) and often coincides with forms of customary management (see Cinner & Aswani 2007). While there has been much discussion in the literature as to whether no-take marine protected areas (MPA) are an effective management tool in this context (Foale & Manele 2004; Cinner & Aswani 2007), less attention has been paid to the effect that constraints on spatial resource use brought about by local tenure may have on spatial planning and the design of regional-scale MPA networks (but see Aswani & Lauer 2006; Green et al. 2009).
In regions with local marine tenure, fishing rights are managed by individual communities, often at spatial scales of hundreds of meters to a few kilometers of coastline. This is in contrast to most developed countries, where coastal and marine resources are governed at national or state level. Under local tenure systems, fishers’ rights are not as strong, or frequently exercised beyond the boundaries of their own community (Foale & Manele 2004). Thus, local tenure acts to restrict fishers’ spatial mobility, as they may not be able to redistribute effort to areas outside of their community following MPA implementation. Whereas commercial fishers might incur increased operating costs associated with traveling to more distant fishing grounds, and opportunity costs if their catch per unit effort in those areas is lower (Ban & Klein 2009), small-scale fishers in regions with local tenure might lose access to fishing grounds entirely if large contiguous areas are protected (Aswani & Hamilton 2004b). Consequently, if they are to be socioeconomically viable, the design of MPA networks needs to consider the spatial scale and location of marine tenure boundaries (Aswani & Hamilton 2004a; Foale & Manele 2004; Cinner 2007).
Previous attempts to reduce the impact of MPA network implementation on small-scale fishers have used population pressure (Ban et al. 2009a), the density of small boats (Sala et al. 2002), or numbers of fishers (Weeks et al. 2010b) as a proxy for opportunity costs. These examples do not account for constraints on the spatial distribution of fishing effort, which can occur either as a result of local tenure or of the limited spatial mobility of small, nonmotorized fishing vessels. Approaches that seek to minimize impacts on small-scale fishers as a single stakeholder group are likely to prioritize sites for protection that will disproportionately impact some resource users, leaving others unaffected. Such inequitable distribution of the costs and benefits of conservation among stakeholders may result in social or political conflict, failure during implementation or poor compliance (Cinner 2007; Klein et al. 2009).
There are few examples where local marine tenure has been explicitly incorporated into regional-scale conservation planning. In the development of an MPA network for the Roviana and Vonavona Lagoons, Solomon Islands, knowledge of customary tenure boundaries was used to identify sites at which conflict over natural resources was likely to be minimal, and MPA implementation more successful (Aswani & Lauer 2006). In Kimbe Bay, Papua New Guinea, information on customary marine tenure boundaries was taken into account when refining an MPA network initially designed using reserve selection software (Green et al. 2009). While adjustments to the design of MPA networks will always be required prior to implementation, plans that incorporate the constraints of local tenure from the outset are less likely to be subject to extensive alterations that may compromise biodiversity objectives.
Here, we demonstrate how spatial zoning software can be used to incorporate local marine tenure into the design of MPA networks in the Philippines. The resulting MPA networks distribute the costs of conservation equitably among local fishing communities, producing solutions that are socioeconomically viable without compromising biodiversity objectives. Our approach turns the traditional reserve selection problem on its head, treating the areal extent of fisheries as targets to be achieved, rather than costs to be minimized (see also Ban & Vincent 2009; Klein et al. 2009). We show, somewhat counter intuitively, that larger MPA networks might have a greater likelihood of acceptance by local communities.
We used the island Province of Siquijor, Philippines, as our case study (Figure 1). More than two-thirds of the population (c. 88,000) live in coastal barangays, (analogous to a village or city ward) and small-scale fisheries (defined as those using vessels of three gross tons or less) contribute considerably to income and food security. Coral reefs in the Philippines are increasingly threatened by overexploitation, destructive fishing techniques, and coastal development. Networks of comprehensive, ecologically representative, connected, and resilient MPAs are considered necessary both to safeguard the livelihoods of coastal communities (Coral Triangle Initiative 2008) and to conserve the Philippines’ rich marine biodiversity.
Prior to colonization, the Philippines had a long history of traditional marine tenure at the barangay level. During the Spanish colonial period, the traditional property rights of barangays over their fishing grounds were steadily eroded and superseded by national government control (Pomeroy & Carlos 1997). More recently, this trend has been reversed, with decentralization of management of coastal resources (to 15 km offshore) to municipal level following the Philippine Local Government Code of 1991 (Republic Act 7160). While it has been stated that traditional fishing rights and barangay-based management systems have disappeared (Pomeroy & Carlos 1997), we found that in our study region the spatial distribution of small-scale fishing effort still conforms to boundaries between barangays (Weeks et al. 2010b). Although MPAs are legislated at the municipal level, barangay governments play a central role in planning and implementation. Furthermore, fisherfolk associations, which often play a key role in MPA management, also operate at this scale.
Biodiversity features targeted for inclusion in MPA networks were coral reef-associated habitats (fringing reefs, sunken shoals, seagrass, lagoons, and mangroves) and bioregions (Figure 1). Habitat types were identified from satellite imagery and verified using hand-held global positioning system (GPS) as part of the Siquijor Coastal Resource Enhancement Project (2003). We subdivided these five habitat types into six bioregions identified on the basis of reef fish community composition (Figure 1). This resulted in 22 targeted biodiversity features (not all habitat types were present in all bioregions).
Data on the spatial distribution of fishing effort were collected through semi-structured interviews with fishers (see Supporting Information). Interviewees were asked to identify fishing grounds used by members of their barangay and to estimate the number of fishers using each site. We found that fishers operating inshore only identified fishing grounds in areas adjacent to their own community, adhering to informal boundaries between barangays (fishers using hook and line in open water traveled beyond their barangay, but rarely identified fishing grounds outside their municipality). We thus identified 67 stakeholder groups of small-scale fishers around Siquijor: one for each coastal barangay. We mapped marine tenure units by applying guidelines for delineating municipal waters (Department of Environment and Natural Resources 2001) to boundaries between barangays. Thus, the tenure boundaries used are conceptual and should not be considered definitive. We did not consider commercial fishers as stakeholders, as following the Local Government Code of 1991 and Philippine Fisheries Code of 1998 (Republic Act 8550) commercial fishers are excluded from operating within 15 km of the coastline.
MPA network design
We used the conservation planning software Marxan with Zones (Watts et al. 2009; available online at http://www.uq.edu.au/marxan/) to identify MPA networks that achieved specified biodiversity objectives while minimizing impacts on small-scale fishers. In contrast to other conservation planning software (including earlier versions of Marxan), Marxan with Zones allows users to allocate sites to a range of different zones that offer different levels of protection (e.g., no-take, habitat protection, open access). This functionality allows users to address multiple objectives simultaneously (Watts et al. 2009). For example, conservation planners typically require the presence of biodiversity features (e.g., habitat types, species) within protected zones. Using Marxan with Zones, it is possible to target simultaneously socioeconomic activities (e.g., fishing, recreation) for inclusion in zones in which that activity is permissible. Klein et al. (2009) demonstrate how this functionality can be used to ensure that impacts on different commercial fishery sectors are equitable. Here, we demonstrate the flexibility of this approach to incorporate the constraints of local marine tenure in the design of MPA networks.
We divided the planning region into regular hexagonal planning units of 0.05 km2. This planning unit size was selected to be at a scale relevant to management: the median size of no-take MPAs in the Philippines is 0.12 km2, and the minimum for Siquijor is 0.04 km2 (Weeks et al. 2010a). Thus, a single planning unit, or two contiguous units, would be a typical size range for a no-take MPA in the region. Each planning unit could be assigned to one of two zones: no-take or open to fishing.
To assess the effect of local marine tenure on MPA network design, we compared two different scenarios. In line with the Philippine Marine Sanctuary Strategy and the Coral Triangle Initiative (Arceo et al. 2004; Coral Triangle Initiative 2008), the biodiversity objective for both scenarios was to include 10% of the area of each biodiversity feature (reef-associated habitat types in each bioregion) within MPAs. In scenario one, our objective was to represent 10% of all biodiversity features within MPAs while minimizing the cost of implementation to small-scale fishers as a single group. We assumed that minimizing opportunity costs to fishers would increase the likelihood that they would support and comply with MPA implementation, resulting in more effective conservation. Thus, costs were assigned to each potential MPA site as the number of fishers that would be displaced if that site were protected, from interview data (Weeks et al. 2010b). This scenario corresponds to a provincial-scale planning process, with no consideration of local tenure.
In scenario two, in addition to requiring representation of biodiversity features in no-take zones (as above), we required that a minimum percentage of the area of inshore fishing grounds used by each community remain within the “fished” zone. To examine tradeoffs between achieving biodiversity and fishery targets, we incrementally increased the fishery targets until it was not possible to achieve all biodiversity and fishery targets simultaneously. We varied Marxan's “feature penalty factor” to express a preference for meeting either biodiversity or fishery targets in order to determine where tradeoffs occurred. For each scenario and target level, we performed 100 replicate Marxan runs, and report results as the mean values of these replicates.
In scenario one, which did not include fishery targets, the percentage of fishing grounds lost was highly variable among barangays. While many communities retained all of their fishing area, others lost as much as 60% (Figure 2A). In contrast, the fishery targets specified in scenario two ensured that the costs of MPA network implementation were distributed more equitably across local fishing communities (Figure 2B).
Biodiversity and fishery targets could be achieved simultaneously with fishery targets up to 87% (i.e., at least 87% of the fishing grounds in each barangay remained open to fishing). When fishery targets were increased above this amount, not all biodiversity targets could be met (Figure 3). Some biodiversity targets were achieved more easily than others: while the percentage area of seagrass, fringing reef, and lagoons within MPA networks declined steadily as fishery targets increased above 87%, representation targets for mangroves and shoals were still achievable with up to 95% of the fished area in each barangay open to fishing (Figure 3).
MPA networks designed to achieve barangay-specific fishery targets (scenario two) had greater total area (Figure 4A) and cost (Figure 4B) than those that sought to minimize costs to small-scale fishers as a single stakeholder group (scenario one). MPA networks with fishery targets of 87% were 40% larger and almost twice as costly overall than those with no barangay-specific targets (Figure 4).
To examine which barangays were causing tradeoffs between biodiversity and fishery objectives, we changed Marxan's feature penalty factor to give preference to meeting biodiversity targets. When fishery targets were set at 90%, on average 10 barangays (out of 67) did not achieve this target (i.e., less than 90% of their inshore fishing grounds remained open to fishing). The identity of these barangays was relatively consistent, with 11 barangays accounting for two-thirds of all missed targets. Of these, only two retained less than 87% of their fishing grounds (mean values from 100 Marxan solutions).
In regions where resource use patterns are defined by local tenure, it might be more important to minimize costs to each local community individually than to minimize the overall cost or area of an MPA network. For example, Ban et al. (2009b) found that when presented with a choice of MPA networks, indigenous communities in Canada did not necessarily prefer the most efficient solution; other factors, such as the locations of individual MPAs were more important. This is in contrast to previously stated concepts of “efficiency” in conservation planning that emphasize minimizing the overall area or cost of a protected area network, under the assumption that costs are borne by a single group (e.g., a nongovernmental organization purchasing land) (Stewart & Possingham 2005; Naidoo et al. 2006).
We found that by specifying zone-specific fishery targets, we were able to distribute the cost of MPA network implementation more equitably across coastal communities. Klein et al. (2009) achieved similar results for commercial fisheries. This approach has the additional benefit of presenting stakeholder interests as targets to be achieved, rather than obstacles to be overcome: a concept likely to be well received in participatory planning.
Our results indicate that to achieve biodiversity conservation objectives for Siquijor, each barangay would have to close up to a maximum of 13% of their inshore fishing grounds. To put this figure into context, the no-take MPA at nearby Apo Island covers approximately 10% of the coral reef area (Alcala & Russ 2006). This MPA is widely considered to be both an ecological and socioeconomic success (Russ et al. 2004; Alcala & Russ 2006). Another MPA at Sumilon Island covers 25% of coral reef area; however, this island does not have a resident community (Alcala & Russ 2006). The inshore fishing grounds considered here represent a surprisingly small proportion of the total fishing grounds used by coastal communities: in many barangays the majority of fishers operate in open water or on offshore shoals, beyond the boundaries of our planning region. Thus, a 13% reduction in inshore fishing grounds equates to a much smaller loss of fished area overall. Nevertheless, inshore habitats are used by gleaners and fishers with nonmotorized boats; these resource users have the least spatial mobility and are therefore most vulnerable to spatial closures (Johannes 2002).
MPA networks that incorporated local tenure boundaries through the use of zone-specific socioeconomic targets had greater total cost and area than those that did not. Mathematically, it is not surprising that a more constrained optimization algorithm is less efficient (McDonald 2009). Several studies have demonstrated that planning across larger spatial extents results in protected area networks that are more cost and area efficient than if planning processes are conducted for subregions separately (Erasmus et al. 1999; Vazquez et al. 2008; Kark et al. 2009). Greater efficiencies have led some to advocate “scaling-up” conservation planning efforts through multinational coordination (Kark et al. 2009). However, McDonald (2009) points to the increased transaction costs (time, money, politics) in coordinating conservation efforts across borders. Such transaction costs can be observed even at the village scale (Cinner & Aswani 2007) and might be especially acute if they require changing traditional patterns of resource use, for example, negotiating access for fishers from a different community to compensate for the loss of their traditional fishing grounds to establish an MPA.
MPA networks that incorporate local tenure at fine spatial scales are necessarily more fragmented than those that do not, as large, contiguous closures will only be viable if fishers have sufficient spatial mobility to retain access to fishing grounds (Aswani & Hamilton 2004b). However, MPA networks that comprise many small no-take areas may not be ecologically desirable: guidelines for MPA network design recommend “bigger rather than smaller” no-take areas, to ensure that they are adequate to protect biodiversity features and ecological processes (McCook et al. 2009; McLeod et al. 2009). Thus, there may be a tradeoff between the socioeconomic acceptability and ecological viability of an MPA network. Nevertheless, a network of small MPAs that are supported by local communities (and thus have good levels of compliance) is likely to be more effective than a system of large MPAs that exist on paper only.
It has also been suggested that the cost of implementing and managing many small reserves will be greater than that for fewer, larger closures (Roberts et al. 2003). This argument is most relevant to MPA systems in developed countries, with centralized management. In regions such as the Philippines, where local communities take responsibility for MPA management, the cost of managing many small reserves may not be significantly greater than that of a few large no-take areas, especially if the smaller closures are supported by local fishers, reducing the need for enforcement.
An alternative approach to establishing many small MPAs within tenure units would be to compensate communities that sacrificed a large proportion of their fishing grounds to establish an MPA. The results for our case study suggest that to achieve biodiversity objectives, relatively few communities would require compensation. However, attempts to develop sustainable alternative livelihoods are often unsuccessful (Johannes 2002) and might not reduce dependence on coral reef ecosystems (Cruz-Trinidad et al. 2009). In regions with highly exclusive marine tenure, it may be possible to designate larger no-take areas by placing them across boundaries, as fishers will only be impacted by the loss of fishing area that falls within their tenure unit.
Conservation planning software such as Marxan with Zones provides a tool for evaluating tradeoffs between competing objectives, which is critical for informed decision making. Setting zone-specific socioeconomic targets allows conservation planners to be explicit about tradeoffs not only between biodiversity and socioeconomic objectives, but also between different stakeholder groups. For simplicity, in our analysis, we applied the same fishery target for all barangays. An alternative approach would be to set variable targets, for example, to reflect the number of fishers in each community, indices of occupational mobility, or dependence on marine resources. This could be achieved either by specifying different percentage area targets for each fishing community, or by assigning variable penalty factors (which inform Marxan how important it is that a target is met).
By setting the minimum area of fishing grounds to be retained for each community, we were able to design MPA networks that impact resource users more equitably and are therefore more likely to achieve support from local communities. This support is essential in regions where effective management is reliant upon voluntary compliance (e.g., Alcala & Russ 2006). Although it appears counter intuitive that larger MPA networks may be more socially acceptable, in this context, minimizing costs to each stakeholder group individually is likely to be more important than overall “efficiency.”
Conservation planning has been criticized for creating “grand designs” (Sayer et al. 2008) that fail to consider adequately the socioeconomic context in which they are to be applied (Knight et al. 2008; Polasky 2008). While it is true that many early applications focused exclusively on biological aspects of conservation, more recent studies have demonstrated consideration for the socioeconomic viability of plans, for example, by incorporating opportunity costs to resource users (e.g., Cameron et al. 2008; Klein et al. 2008). We contribute to these efforts by demonstrating that spatial zoning software can be used to design MPA networks that achieve biodiversity conservation targets within the constraints of local marine tenure. The general approach that we describe here could be applied in any region in which spatial resource use is subject to local-scale constraints. It has particular relevance to the Coral Triangle Initiative, a key objective of which is to develop MPA networks throughout the Philippines, Indonesia, Malaysia, Papua New Guinea, Solomon Islands, and Timor Leste (Coral Triangle Initiative 2008), all countries that are characterized by local marine tenure.
We thank the fishers of Siquijor for their cooperation, without which this research would not have been possible. We also thank S. Foale and three reviewers for helpful comments on an earlier version of the manuscript. R.W. was supported by a Northcote Graduate Scholarship.