High economic costs of reduced carbon sinks and declining biome stability in Central American forests

Tropical forests are central to global carbon sequestration and biodiversity conservation but face mounting pressures from deforestation and climate change. Central America, a biodiversity hotspot with complex topography, is already experiencing climate impacts and may become a future hotspot. Prior large-scale vegetation modelling for the region has often used coarse spatial resolution, and the economic implications of climate-driven ecosystem service (ES) changes remain understudied and frequently decoupled from ecological modelling. To inform decision-making, translating ecological outcomes into monetary terms is crucial. This study addresses the following research questions: (1) How may climate regulation and habitat services be influenced by climate change? (2) What are the approximate economic values of these services and the potential costs under climate change? (3) How do results change under national price levels and temporal discounting? (4) Where are hotspots of change that warrant action?

Existing analyses of climate change impacts on forest ecosystem services in Central America have tended to be either local or continental in scope and often separate ecological from economic modelling. High-resolution climate impact modelling is limited by the lack of fine-scale scenarios and computational demands, despite topography being a major driver of tropical forest structure and productivity. Past work indicates potential biome shifts under climate change in the region, though projections vary by model and scenario. There is a recognized need for economic valuation of ES to support policy, yet integrated ecological-economic assessments at regional scales in Central America are scarce. The study builds on prior dynamic vegetation modelling and ecosystem service valuation literature, using social cost of carbon estimates and benefit transfer methods to bridge this gap.

Study design: The authors coupled fine-resolution vegetation modelling with economic valuation to quantify climate change impacts on two ecosystem services: climate regulation and habitat, across Central America from 1985–2100. Vegetation modelling: They used the dynamic global vegetation model LPJ-GUESS (v4.0.1), a process-based DGVM representing assimilation, respiration, growth, mortality, disturbance, and C-N cycling via plant functional types. To capture topographic complexity, they implemented the LPJ-GUESS “landforms” approach, creating subversions per modelling unit from a topographic classification using SRTMGL1 elevation (1 arc-second) while acknowledging that precipitation inputs remained at 0.5° resolution (ISIMIP3b) and were not fully downscaled. Forcing data: Bias-corrected daily climate (ISIMIP3b, 0.5°) from 1850–2100; atmospheric CO2 and nitrogen deposition from literature sources; SRTMGL1 elevation; CCI land cover for masking anthropogenic land use (cropland, urban, mosaics). Simulations spanned historical 1985–2014 and future 2071–2100 periods. Scenarios: Four combinations of two GCMs (GFDL-ESM4, IPSL-CM6A-LR) and two SSPs (SSP1-2.6, SSP3-7.0) to represent low and high forcing futures. Ecosystem service indicators: Climate regulation was proxied by net ecosystem exchange (NEE), converted from kgC/m² to t CO2/ha. Habitat service was proxied by biome stability, defined as the percentage of years in 2071–2100 in which the biome matched the mean biome of 1985–2014. Biomes (eight classes) were assigned mainly following Snell et al.; tropical needle-leaved forests were approximated via temperate broadleaved evergreen and precipitation seasonality due to PFT limitations. Economic valuation: Climate regulation was valued using social cost of carbon (SCC) estimates from Yang et al. (2018), averaging across damage functions and matching years 2015 (historical) and 2085 (future); in no-discounting variants, 2085 values at 0% discount were used and held constant. Habitat services were valued via benefit transfer from the Ecosystem Service Valuation Database (tropics-wide), filtering out provisioning services and climate regulation, removing missing/duplicate/outlier entries (5–95% quantile), grouping by major biomes and ES, summing to biome totals, then averaging across biomes; inflation-adjusted to 2015. Economic calculations were done with two pricing schemes (global uniform prices vs. nationally adjusted by 2020 GDP per capita) and two time preferences (2% discount rate vs. no discounting). Net present values for historical and future periods were computed and their differences interpreted as costs to society. Spatial outputs included maps of stability and CO2 sequestration change, and “economic hotspots” where both services declined, categorized by cost classes.

- Spatial ES changes: High biome stability persisted along the Caribbean-facing half of Central America, southern Yucatán, and parts of northwest Central America (aligned with tropical rainforest extent). Areas of consistent decline included dry forests along the Pacific coast and northern Yucatán, with hotspots of concurrent declines in biome stability and CO2 sequestration covering 4.0–8.7% of the study area. - Biome stability: Share of area with stability >50% was 84.8% (GFDL-SSP126), 84.9% (IPSL-SSP126), 79.4% (GFDL-SSP370), and 84.5% (IPSL-SSP370). - CO2 sequestration: Area projected to store less CO2 than the reference period was 64.1% (GFDL-SSP126), 55.9% (IPSL-SSP126), 24.1% (GFDL-SSP370), and 33.4% (IPSL-SSP370). SSP126 showed larger areas of reduced sequestration, while SSP370 showed stronger declines in biome stability and higher extremes. - Regional patterns over time: Some changes were nonlinear, with losses accelerating late in the century. - Economic costs (region-wide): Total ES decline costs were $51–314 billion/year until 2100. Disaggregated, habitat losses were $29–313 billion/year and climate regulation losses $1–65 billion/year across scenarios and valuation variants. - Discounting and pricing effects: For habitat, applying a 2% discount rate produced costs approximately six times higher than without discounting due to valuation dynamics; costs arose in 100% of the area under discounting irrespective of stability change. For climate regulation, no-discounting costs exceeded discounted results by up to a factor of 30. Global uniform price estimates were ~1.6× higher than nationally adjusted prices. - Country-level impacts relative to GDP: Losses reached up to 335% of GDP (Belize), 189% (Nicaragua), and 115% (Honduras). Lower-middle income economies (Belize, Honduras, Nicaragua, El Salvador) were disproportionately affected, while higher-income countries (Panama, Costa Rica) showed lower projected losses. - Comparative ES costs: Habitat decline costs exceeded climate regulation costs in all but one scenario (GFDL-SSP126, no discounting). Under some scenarios, average CO2 costs were higher than habitat in Colombia, Panama, Costa Rica, Honduras, and Guatemala; the opposite held for Nicaragua, El Salvador, and Mexico. - Economic hotspots: Under global pricing and no discounting, highest-cost hotspots appeared in Yucatán and montane regions of Honduras, Costa Rica, and Colombia (GFDL-SSP126); Colombian Andes intensified under SSP370; IPSL scenarios highlighted the American cordillera, with SSP370 focusing costs in Honduras, Nicaragua, and the Colombian Andes. National pricing shifted hotspots toward higher-income areas, reducing shares in upper cost classes.

The results demonstrate that climate change threatens both climate regulation and habitat services in Central American forests, with substantial economic implications. The modelling suggests that tropical rainforest areas may be relatively stable, while montane and dry forests face heightened risks of biome instability, consistent with topographically driven climate sensitivities and potential altitudinal shifts. CO2 fertilization contributes to greater net uptake under higher-forcing scenarios, though its magnitude remains uncertain; meanwhile, stability declines intensify with forcing, implying significant biodiversity risks and fragmentation challenges for species migration. Economically, valuation outcomes are strongly shaped by methodological choices: discounting approach and price levels have greater influence on total costs than scenario forcing, though spatial cost distributions are mainly driven by climate scenarios and, to a lesser extent, by GCM choice. Habitat-related costs generally exceed climate regulation costs, emphasizing that biodiversity-supporting services represent substantial economic stakes beyond carbon alone. The distribution of costs disproportionately affects lower-middle income countries, potentially exacerbating regional inequalities. These findings underscore the need for integrated policies and payment for ecosystem services schemes that consider multiple services and avoid perverse incentives from carbon-only approaches, aligning ecological realities with equitable economic instruments.

This study links fine-resolution dynamic vegetation modelling with economic valuation to estimate climate change impacts on climate regulation and habitat services in Central America. It provides spatially explicit projections of biome stability and CO2 sequestration, translates them into monetary terms under varying pricing and discounting assumptions, and identifies economic hotspots for targeted mitigation. Key contributions include quantifying that 24–62% of the region may experience ES declines, with associated costs of $51–314 billion per year, and highlighting the outsized economic risks for montane and dry forests and lower-middle income countries. The results argue for broadening policy focus beyond carbon to include habitat conservation and for designing PES and market mechanisms that account for multiple ES. Future research should improve precipitation downscaling and incorporate land-use and anthropogenic pressures, refine SCC estimates and discounting choices, standardize and contextualize habitat valuation methods, advance monitoring of ES for PES schemes, and explore integrated strategies that capture co-benefits for carbon and biodiversity.

- Climate inputs: Precipitation remained at 0.5° resolution without full downscaling to account for wind fields, leaving artefacts and limiting fine-scale accuracy. - Scope: The analysis focused on climate-driven ecosystem changes; anthropogenic land-use changes and other human pressures were excluded but are likely significant. - Model and processes: Uncertainties persist in CO2 fertilization magnitude despite accounting for nitrogen limitation; biome classification adjustments were needed due to PFT limitations (tropical needle-leaved forests). - Valuation uncertainties: Social cost of carbon estimates are sensitive to discount rates and socio-economic/damage function assumptions; habitat benefit transfer carries biases from source studies, ecosystem coverage, and selection filters. Placement of “habitat” within ES frameworks is debated, and valuation methods for cultural/existence/option values lack standardization. - Discounting asymmetries: Different temporal dynamics between SCC and habitat valuation complicate joint interpretation of discounted results. - Monitoring and implementation: Practical PES applications depend on accurate, cost-efficient ES monitoring, which remains challenging.