Fig 1.
A map of the tropical Western Pacific Ocean and Micronesia with study islands indicated by stars.
Fig 2.
Ecosystem condition evaluation process depicting the contribution of individual biological metrics on their respective latent variables, and the weighting of latent variables on overall condition scores.
Latent variables were generated from independent datasets. Biological metrics were established by a working group of regional scientists, and linked with key processes required for ecosystem function and maintenance (S1 Table).
Fig 3.
Normalized ecosystem-condition scores across Micronesia.
Monitoring sites were stratified by reef habitat, management, geography, and/or wave exposure, as appropriate (S1 Fig). Dashed lines depict the ‘effective-conservation’ threshold used to assess progress towards the Micronesia-Challenge conservation goals. Percentages indicate the proportion of sites currently above the threshold. Site-circle sizes on the maps adjacent to the bar graphs were scaled by their normalized ecosystem-condition scores. Marine protected areas are designated on the bar graphs with an asterisk (*).
Fig 4.
Sensitivity of ecosystem condition scores depicted by correlation coefficients between individual metrics and their respective latent variables (circles-fish, squares-benthic, and crosses-corals), or between latent variables and overall ecosystem condition (bottom triangles).
Error bars (thin lines) represent standard deviations associated with the differing jurisdiction and reef-habitat strata.
Fig 5.
Regression models depicting the relationships between localized stressors and ecosystem condition.
Fishing pressure predicted ecosystem condition on reefs > 5 km from land across Micronesia, with notable variation among islands and reef habitats (a). Independent slopes, in turn, predicted the variance among the top four non-normalized ecosystem condition scores (b, S2 Table), and also the relative influence of wave exposure—coral assemblages—ecosystem condition (blue), or the fishing pressure proxy-fish assemblages—ecosystem condition (red), or a direct relationship between the two integrated variables, fishing pressure proxy—ecosystem condition (black). Conversely, interactions between fishing and pollution proxies, termed localized stressors, best predicted condition for lagoon reefs <5 km from land (c), with stressors acting equally (blue) or unequally (red) across benthic, coral, and fish latent variables. Excluding marine protected areas (MPA, dashed circles around site symbols) did not improve the regional model fit, but removing two successful MPAs on Pohnpei (circles on furthest right) did improve the localized association between stressors and ecosystem condition.