Fig 1.
(A) Large-scale map of Madagascar and the African continent, (B) Inset of the 25 villages of the Velondriake Locally Managed Marine Area in southwestern Madagascar. Vertical box extent is ~75 km. (C) Representative example of a periodic octopus fishery closure. Indicated in the map are two villages, Andavadoaka and Ampasilava, with their respective octopus fishing sites mapped in orange and yellow. In green, you can see the sites Amagnahitse and Nosinkara, in which these two villages have repeatedly co-implemented a periodic octopus fishery closure.
Fig 2.
Experimental design and samples sizes used to investigate effects of periodic fishery closures.
(A) Site fishery catches, (B) Village fishery income, and (C) Site net economic benefts. Colors highlight the distinctions among “no-ban” and “ban” closures, and between closure sites/villages and either open-access controls (A&B) or simulated landings (C).
Fig 3.
Closure effects on Catch-Per-Unit-Effort.
Site-level catch-per-unit-effort (CPUE, kg/fisher-day), 30 days before closure and after reopening at closed sites and paired control sites. Data are separated by season, thus separating those closures that occurred independently of a regional fishery shutdown (“no ban”), and those that extended the shutdown (“ban”). Significance indicators show distinctions between a particular group and its “before” group comparison, independent contrasts from linear mixed models. NS = Not Significant; * = p < 0.05; *** = p < 0.001. For components of CPUE and data aggregated across seasons, please see S1 & S2 Figs.
Fig 4.
Closure effects on Village Fishery Income.
Total village-level octopus fishing income ($PPP) 30 days before, during, and after closures, at villages both with and without closures. The data depicted are from 28 closure periods showing closure-implementing villages and their control villages from 2004–2011. Data are separated by season, thus separating closures that occurred independently of a regional fishery shutdown (“no ban”), and those that extended the shutdown (“ban”). As “during” periods are not exactly 30 days, “during” values are scaled to a per-30-day measure. Significance indicators show distinctions between a particular group and its “before” group comparison, from linear mixed-effect model. NS = Not Significant; * = p < 0.05; ** = p < 0.01; *** = p < 0.001. For effort, value per unit effort, and data aggregated across seasons, please see S5 & S6 Figs.
Table 1.
Mean Village Level Octopus Income Before and After Temporary Closures.
Fig 5.
Net Economic Benefits of Closures.
Site-specific Net Earnings (NE) and Internal Rate of Return (IRR) of 36 local closures, 2004–2011, using site-based cost model. Point coding represents rates of illegal fishing during the closures: Low (< = 5% of baseline ‘before’ catches, blue circles), Moderate (< = 50% of baseline catches, green squares), and High (>50% of baseline catches, red diamonds). Data points represent median values across 1000 model runs.
Table 2.
Closure Site Net Economic Benefits.
Table 3.
Effect of Illegal Fishing on Mean Closure Site Net Economic Benefits.