Climate change is projected to severely impact global food production, particularly in tropical regions, affecting both fisheries and agriculture. By 2100, tropical areas could lose up to 200 suitable plant growing days annually due to climate change, and fishable biomass in the ocean could decrease by up to 40% in some areas. However, understanding the magnitude of these losses is insufficient; the social dimensions of vulnerability determine the extent to which societies are affected. Vulnerability encompasses exposure (stress from environmental/social conditions), sensitivity (susceptibility to harm), and adaptive capacity (ability to respond to change). This study focuses on potential impacts, combining exposure and sensitivity. Many coastal communities rely on both agriculture and fisheries, yet impact assessments often consider these sectors separately. National-level analyses obscure the variability in how communities experience combined impacts. This study addresses the need for localized analyses by integrating model projections of losses to fisheries catch potential and agriculture with survey data on socioeconomic factors from over 3,000 households across 72 tropical coastal communities in five Indo-Pacific countries. The research investigates the potential impacts of projected changes on coastal communities, the effect of mitigation measures, and whether socioeconomic status influences vulnerability.

Previous research has highlighted the significant potential losses in both fisheries and agriculture due to climate change, particularly in tropical regions. Studies have also emphasized the importance of incorporating social vulnerability factors, including exposure, sensitivity, and adaptive capacity, to understand the differential impacts on communities. However, most previous assessments have treated fisheries and agriculture in isolation, often at national scales, masking the significant variations in impact at the community level. Several studies have begun to examine simultaneous impacts at the national level, but these coarse-scale analyses fail to capture the complex interplay of these sectors within individual communities. Existing literature points towards higher vulnerability among lower socioeconomic communities due to their greater dependence on natural resources and limited access to alternative livelihood options. This study directly addresses the gap in understanding community-level, simultaneous impacts on fisheries and agriculture.

This study combined projected losses from the Inter-Sectoral Impact Model Intercomparison Project (ISIMIP) Fast Track phase 3 dataset with socioeconomic survey data from 3,008 households across 72 coastal communities in Indonesia, Madagascar, Papua New Guinea, the Philippines, and Tanzania. The ISIMIP dataset provided projections of changes in fisheries catch potential (exploitable marine biomass) and agricultural productivity (rice, maize, and cassava) under two Shared Socioeconomic Pathways (SSPs): SSP1-2.6 (low emissions) and SSP5-8.5 (high emissions). The socioeconomic surveys measured material wealth (based on 16 material items) and occupational diversity. Sensitivity was calculated based on household dependence on fisheries and agriculture, considering the relative importance of these sectors compared to other income sources. Exposure was assessed using the projected changes in fisheries catch potential and agricultural yields from the ISIMIP models. Potential impact was quantified as the Euclidean distance of exposure and sensitivity from the origin. Linear mixed-effects models (with country as a random effect) were used to analyze the data, accounting for within-country variations. A sensitivity test compared the exposure of the study sites to that of a random selection of coastal locations in the study countries. The study also explored the observed changes in sensitivity and material wealth over time in two Papua New Guinean communities to evaluate the stability of these metrics over decadal time frames.

The study yielded three key findings. First, projected impacts on fisheries catch potential were generally greater than those on agriculture, although significant within-country variability existed. Under the high-emission scenario (SSP5-8.5), the mean projected loss in fisheries catch potential was 14.7% (± 4.3% SE), approximately double the potential increase from strategically protecting 28% of the ocean. Agriculture showed small average gains (1.2% ± 1.5% SE), not significantly different from zero, largely due to rice; excluding rice revealed average declines. This difference in projected losses was evident both in the study sites and a random selection of coastal locations. Sensitivity was also higher for fisheries than agriculture. Second, under the high-emission scenario, 64% of study sites were projected to experience simultaneous losses in both sectors. Mitigation (SSP1-2.6) reduced this to 37%, consistent with results from the random sample of locations. Many sites with high projected losses also had high sensitivity, indicating significant potential impacts. Third, communities with lower socioeconomic status were more likely to experience greater potential impacts, irrespective of the emission scenario. This was because low socioeconomic communities exhibit greater dependence on natural resources (higher sensitivity) and were more exposed to climate change impacts. A time-series analysis of two Papua New Guinean communities showed that sensitivity and material wealth were reasonably stable over a 15-16-year period, despite absolute changes in material wealth.

This study offers crucial insights into the simultaneous impacts of climate change on fisheries and agriculture in tropical coastal communities. The findings highlight the higher vulnerability of fisheries compared to agriculture, though substantial within-country variability necessitates localized mitigation strategies. The substantial overlap in projected losses for both sectors emphasizes the need to consider intersectoral dependencies in climate change adaptation planning. The disproportionate impact on lower socioeconomic communities underscores the critical need to address existing inequalities alongside climate change mitigation. The results emphasize the importance of community-level assessments, given the spatial variability in impacts that are masked in national-level studies. While alternative livelihood programs might reduce sensitivity in some communities, their frequent failure highlights the need for more holistic and context-specific approaches.

This study provides quantitative evidence of the potential combined impacts of climate change on fisheries and agriculture in tropical coastal communities. The higher exposure and sensitivity to fisheries compared to agriculture, coupled with the significant proportion of communities experiencing double burdens, underscore the urgency for comprehensive, integrated management strategies that account for intersectoral dependencies. The clear relationship between socioeconomic status and vulnerability highlights the critical need to address social inequalities in climate change adaptation. Future research should focus on improving model resolution and incorporating adaptive capacity and other stressors such as overfishing and soil degradation. Expanding the scope to include other climate change impacts and exploring additional socioeconomic dimensions is also warranted.

The study has several limitations. The exposure measure was dynamic (projected into the future), while sensitivity and material wealth were static (from a single point in time). The models used had limitations: tropical small-scale fisheries targeting seagrass and coral reef habitats were not fully represented, and the spatial resolution of the models was relatively low. The sensitivity metric focused on economic dependence on natural resources and excluded other dimensions of sensitivity. Adaptive capacity was not explicitly incorporated due to data limitations. The study focused solely on the impacts on fisheries and agriculture, neglecting other potential climate change impacts on community well-being. The study also used only two available SSP scenarios from the ISIMIP dataset.