|
nif:isString
|
-
LMPAs were selected based on four criteria: 1) biodiversity conservation as a primary goal; 2) large: defined as >10,000km2 (several magnitudes larger than the median size of MPAs (3.3km2; [6])); 3) five years of active management: defined as having a management plan and some implementation for at least five years; and 4) sufficient data on outcomes. We identified LMPAs that met our first three criteria from MPAtlas.org [27], and then conducted a preliminary literature search to determine whether there was evidence of management actions (i.e. environmental monitoring, enforcement). We considered there to be sufficient data for coding outcomes when there were published peer-reviewed or grey literature sources that assessed ecological and/or social outcomes. Globally, 16 MPAs met the first two criteria, four of which were later excluded because they either lacked active management or adequate data on outcomes (Greenland National Park, Dominican Republic Marine Mammal Sanctuary, Franz Josef Land, Pelagos Sanctuary). Our final sample of 12 MPAs range in size from 11,859 km2 (Raja Ampat MPA Network) to 362,073 km2 (Papahānaumokuākea Marine National Monument), and in age from 10 years (Raja Ampat MPA Network) to more than 40 years (Svalbard Eastern Nature Reserves and Great Barrier Reef; S1 Fig).
We used the Social-Ecological Systems Meta-Analysis Database (SESMAD) [28] to provide a consistent approach for coding outcomes across the 12 LMPAs. SESMAD is a relational database based upon the social-ecological systems framework [29] that uses mostly categorical and ordinal variables to describe components of a social-ecological system and enable comparisons across cases where different metrics might be used. For each LMPA, we focused on five outcomes (Table 1): three outcomes associated with the ecological system (changes in an ecosystem health, a target fishery, and a key migratory species), and two outcomes associated with resource users (changes in the well-being of a user associated with ecosystem health, changes in the well-being of a user associated with the target fishery). Additional information concerning methods and coding are found in [13], and coded cases can be viewed at: https://sesmad.dartmouth.edu/ses_cases.
Table data removed from full text. Table identifier and caption: 10.1371/journal.pone.0195760.t001 The three main components of the social-ecological system and the five outcomes measured in this study. We conducted a detailed literature review of peer-reviewed and grey literature for each LMPA to identify potentially relevant components across the social-ecological system. Natural components (i.e. fish, migratory species, and indicators for ecosystem health) were selected for coding based upon: 1) their influence at the scale of the LMPA; 2) explicit mention of the natural component in the LMPA management plan or governance guidance; 3) data availability (i.e., changes in the natural component have been documented), and; 4) where multiple options existed, we selected components that would be expected to respond to governance. For instance, in Macquarie Island Marine Reserve, we selected Royal Penguins as an indicator of ecosystem health because they are a higher trophic level species and breed exclusively on islands within the Reserve. User groups, meanwhile, were determined by considering whether there was a group of actors that derived a non-trivial fraction of their livelihood benefits, whether directly or indirectly from the selected components. For instance, fishers in the Great Barrier Reef clearly derive livelihood benefits from reef fish, but also depend indirectly upon coral cover to maintain the supply of reef fish. In contrast, the livelihoods of fishers in the Heard Island and McDonald Island Marine Reserve are not substantially related to the health and abundance of King Penguins. The specific components coded for each case are identified in Table 2.
Table data removed from full text. Table identifier and caption: 10.1371/journal.pone.0195760.t002 Details on the components coded for each large marine protected area. Changes in resource conditions and well-being were coded using SESMAD protocols [28] to explore the implications of a wide range of social, ecological and institutional factors for sustainability (11), and generate insights about potential trade-offs. Each LMPA was assessed for a specific time-period (or ‘snapshot’) in which the governance structure remained relatively stable (i.e., no major re-zoning), while outcomes were coded as changes in resource conditions or well-being over this time period. All outcomes were ordinal with three possible values, and were recorded as missing (NA) in the absence of a user group (Table 1).
Potential trade-offs were identified using radar plots in R (version 3.2.2 [30]) and the fmsb package [31]. Radar plots provide visualisations of multivariate data in a simple two-dimensional chart. First, radar plots were analysed visually to identify potential trade-offs where one outcome was stable or increasing and another was declining, as depicted in Fig 1. Potential supply trade-offs are indicated by different outcomes in ecosystem health, fisheries, and/or migratory species. Potential supply-demand trade-offs are indicated by differences between outcomes for ecosystem health or fisheries, and the well-being of associated user groups. Potential demand trade-offs are indicated by differences in the well-being of different user groups. Although variation in outcomes is indicative of a potential trade-off, these may be coincidental rather than causal. As a result we complement our analysis of outcomes with a qualitative analysis of the plausibility of a causal mechanism linking the two outcomes (Table 3) to understand if different outcomes were potentially causal (trade-off) or merely coincidental (divergent outcomes).
Table data removed from full text. Table identifier and caption: 10.1371/journal.pone.0195760.t003 Mechanisms that may give rise to trade-offs, including description and examples from the literature. Figure data removed from full text. Figure identifier and caption: 10.1371/journal.pone.0195760.g001 Visual representations of how the three conceptual trade-offs (as identified by Mouchet et al. (23)) may appear across the seven outcomes assessed in our study.Each example radar plot (A,B,C) shows all five focal outcomes (ecosystem health, migratory species, fishery resources, well-being of user groups (e.g., fishers), and well-being of users of the ecosystem (e.g., coastal residents, tourists), with the inner-most band representing a decline and the outside line representing an increase (indicated with ‘worst’ to ‘best’ on the radar plot). Key outcome trade-offs have been circled to aid understanding of the trade-off typology and how it applies to our data. Outcome abbreviations used in radar plot: Eco = ecosystem health change; WB_Eco = well-being change of the user of the ecosystem health indicator; WB_Fish = well-being change of the user of the fisheries indicator; Mig = migratory species change; Fish = fisheries change. A: Supply trade-off: ecosystem health improving, but fisheries declining (or vice versa; conservation versus use). B: Supply-demand trade-off: fisheries improving, but well-being of a user (fisher) declining (or vice versa). C: Demand trade-off: differentiated impacts in the well-being of different users, with a well-being decline of a user dependent on fisheries, and a well-being improvement of a user dependent on ecosystem health (e.g. tourism) (or vice versa).
We categorised four types of causal mechanisms that can lead to trade-offs (Table 3): 1) deliberate a priori management decisions to prioritize some outcomes over others (32), or the allocation of finite resources to some activities over others (33); 2) everyday resource use decisions by resource users that influence well-being and resource conditions [39, 40]; 3) unintended consequences of resource use where the exploitation of one resource has a direct impact on others (e.g., by-catch) [41]; 4) indirect consequences that occur when two or more resources are connected via biophysical relationships or ecosystem processes (e.g., food webs) [42]. This last type of trade-off mechanism is less visible than others and can take longer to manifest.
|
![[This page as RDF]](http://data.gesis.org/somesci/static/rdf-icon.gif){kind=icon}