Protected areas (PAs) are the leading instrument for curbing Amazonian deforestation, but their true benefit depends on where the displaced land-use pressure goes. Fuller and colleagues (2020) showed that for 91 state-governed PAs in Amazônia Legal the local spillover into adjacent unprotected forest is small and more often 'blockage' (reduced clearing) than 'leakage' (displaced clearing), yet they could measure only this local buffer and explicitly left total leakage unresolved: clearing pressure displaced to distant parts of Amazônia Legal, to the Cerrado biome, or to other countries was unquantified, and the counterfactual baselines available for estimating such non-target impacts remain methodologically fragile. A 2026 review of 51 spillover studies confirms that no standardized, robust framework for quantifying protected-area spillovers exists. The open problem is to quantify the total (local plus distant, telecoupled) deforestation leakage attributable to PA establishment in the Brazilian Amazon under a widely applicable, defensible counterfactual, so that the net system-wide deforestation impact of the PA network can be established.

Background

A protected area (PA) is a legally designated region where human land use is restricted, and PAs are the backbone of Brazilian Amazon forest policy. When a PA blocks land clearing inside its boundaries, the economic pressure to clear does not necessarily vanish: it may be displaced to nearby unprotected forest — a spillover called 'leakage' — or, conversely, a PA may dampen clearing in its surroundings, a positive spillover called 'blockage'. Whether leakage occurs, and how large it is, determines whether the avoided deforestation reported inside PAs overstates the real conservation gain; this matters directly for REDD+ (the international mechanism for Reducing Emissions from Deforestation and forest Degradation, which pays for avoided clearing and is undermined if clearing merely moves elsewhere) and for the global 30×30 target of protecting 30% of land by 2030.

Estimating spillover requires a counterfactual: what would deforestation rates have been in the unprotected surroundings without PA establishment? Because PAs are not sited at random (remote areas are preferentially protected), this baseline must be constructed statistically — for example by matching each PA to comparable unprotected land, or by panel regression that compares units before and after intervention (a Before-After-Control-Impact, or BACI, design). Fuller et al. (2020, Biological Conservation 248:108673) assessed 91 state-governed PAs in Amazônia Legal (the Brazilian legal Amazon) established 2005–2016, using PRODES (Brazil's satellite-based annual clear-cut deforestation monitoring system by the National Institute for Space Research) and deforestation-rate baselines. They found blockage outnumbered leakage, with on average 2% or less of the spillover zone affected, but leakage concentrated in the southern and western Amazon. Their study quantified only the local buffer (an adjacent zone of area equal to the PA) and left open whether clearing displaced farther — a possibility made concrete by Dou et al. (2018), who showed that conservation and supply-chain interventions in the Amazon displaced land-use pressure into the Cerrado, the savanna biome south and east of the rainforest.

The wider literature frames this as a measurement problem spanning scales. Locally, Ford et al. (2020) found leakage in 55 of a global sample of 120 tropical and subtropical forest PAs using statistically matched controls and 10 km buffers, while Barros et al. (2022), using a BACI design with matching on 425 Brazilian PAs in the Amazon and Atlantic Forest (2004–2017), found no evidence of local leakage — a disagreement that itself reflects unsettled baseline methodology. At longer distances, displacement is 'telecoupled' (coupled over long distances through markets, commodity supply chains, and migration). Distant leakage has been quantified for supply-chain policies, not for PAs: Heilmayr et al. (2020) found no detectable leakage of the Amazon Soy Moratorium within Amazônia Legal, Moffette and Gibbs (2021) showed the Soy Moratorium displaced soy production onto already-cleared pasture in the less-regulated Cerrado while zero-deforestation cattle agreements increased deforestation elsewhere, and Villoria et al. (2022) estimated 43–50% domestic but under 3% cross-border leakage from soy zero-deforestation policies. Meanwhile, methodological audits by Cumming (2025) and a review of 51 spillover analyses by Gliottone et al. (2026) conclude that protected-area spillover measurement still lacks validated sampling designs, standardized counterfactual frameworks, and treatment of spatial scale, so results cannot yet be generalized or meta-analyzed. In Fuller et al.'s own models, only 14.1–33.6% of the deviance in spillover outcomes was explained (%DE, the fraction of variation accounted for by the fitted statistical model), underscoring how much of the phenomenon remains uncharacterized.

Problem Statement

For protected areas established in the Brazilian Amazon, quantify the total deforestation leakage attributable to PA establishment — that is, the local spillover of clearing into adjacent unprotected land together with any distant, telecoupled displacement of clearing to other parts of Amazônia Legal, to other biomes such as the Cerrado, or to other countries — under a robust and widely applicable counterfactual baseline, so that the net, system-wide deforestation impact of the protected-area network can be established. The existing evidence base is the cohort of 91 state-governed (non-indigenous) PAs established 2005–2016 in Amazônia Legal assessed by Fuller et al. (2020); Indigenous Lands were not assessed in that study, and PAs established before 2005 fall outside the high-resolution PRODES deforestation record. Because these choices determine what total is attributed, the total is defined over the PA population and establishment window considered, the measurement period, the outcome unit, and the spatial system boundary to which the net aggregate refers; that boundary covers the prescribed total — Amazônia Legal together with the Cerrado and other countries. The problem comprises two jointly required, scientifically linked components: (1) an empirical accounting of non-target deforestation impacts at both local and distant scales, in which the local component is quantified against a counterfactual baseline and the distant component is quantified or, where only causal identification is feasible, bounded with the reason full quantification is infeasible stated explicitly; and (2) the counterfactual estimation framework that makes this accounting credible — a baseline construction for non-target impacts that remains feasible in dense reserve networks where locally derived baselines are impossible (as for 18 of the 91 PAs assessed) and that follows clear, widely applicable guidelines. The two components serve one objective — credible measurement of the net deforestation impact of Amazonian protected areas — and an answer that quantifies only the adjacent buffer, or only asserts displacement without a validated counterfactual, does not resolve the question.

References

  1. Fuller, Carley; Ondei, Stefania; Brook, Barry W.; Buettel, Jessie C. (2020). Protected-area planning in the Brazilian Amazon should prioritize additionality and permanence, not leakage mitigation. Biological Conservation 248:108673. DOI 10.1016/j.biocon.2020.108673. https://doi.org/10.1016/j.biocon.2020.108673
  2. de Assis Barros, Luizmar; Venter, Michelle; Ramírez-Delgado, Juan Pablo; Coelho-Junior, Marcondes G.; Venter, Oscar (2022). No evidence of local deforestation leakage from protected areas establishment in Brazil's Amazon and Atlantic Forest. Biological Conservation 273:109695. DOI 10.1016/j.biocon.2022.109695. https://doi.org/10.1016/j.biocon.2022.109695
  3. Gliottone, Ilaria; Collins, Teigan G.; de Vos, Ennis; Mendez Angarita, Valeria Y.; Kongwarakom, Siwat; Cumming, Graeme S. (2026). A Review and Critique of Methods and Knowledge Gaps in Quantifying Spillover Effects from Protected Areas. Conservation Biology e70361. DOI 10.1111/cobi.70361. https://doi.org/10.1111/cobi.70361
  4. Cumming, Graeme S. (2025). Conceptualizing and measuring ecological spillover effects from protected areas. Frontiers in Ecology and the Environment 23:e70008. DOI 10.1002/fee.70008. https://doi.org/10.1002/fee.70008
  5. Ford, Scott Alan; Jepsen, Martin Rudbeck; Kingston, Naomi; Lewis, Edward; Brooks, Thomas M.; MacSharry, Brian; Mertz, Ole (2020). Deforestation leakage undermines conservation value of tropical and subtropical forest protected areas. Global Ecology and Biogeography 29:2014-2024. DOI 10.1111/geb.13172. https://doi.org/10.1111/geb.13172
  6. Herrera, Diego; Pfaff, Alexander; Robalino, Juan (2019). Impacts of protected areas vary with the level of government: Comparing avoided deforestation across agencies in the Brazilian Amazon. Proceedings of the National Academy of Sciences 116:14916-14925. DOI 10.1073/pnas.1802877116. https://doi.org/10.1073/pnas.1802877116
  7. Dou, Yue; da Silva, Ramon Felipe Bicudo; Yang, Hongbo; Liu, Jianguo (2018). Spillover effect offsets the conservation effort in the Amazon. Journal of Geographical Sciences 28:1715-1732. DOI 10.1007/s11442-018-1539-0. https://doi.org/10.1007/s11442-018-1539-0
  8. Heilmayr, Robert; Rausch, Lisa L.; Munger, Jacob; Gibbs, Holly K. (2020). Brazil's Amazon Soy Moratorium reduced deforestation. Nature Food 1:801-810. DOI 10.1038/s43016-020-00194-5. https://doi.org/10.1038/s43016-020-00194-5
  9. Moffette, Fanny; Gibbs, Holly K. (2021). Agricultural Displacement and Deforestation Leakage in the Brazilian Legal Amazon. Land Economics 97:155-179. DOI 10.3368/le.97.1.040219-0045r. https://doi.org/10.3368/le.97.1.040219-0045r
  10. Villoria, Nelson; Garrett, Rachael; Gollnow, Florian; Carlson, Kimberly (2022). Leakage does not fully offset soy supply-chain efforts to reduce deforestation in Brazil. Nature Communications 13:5476. DOI 10.1038/s41467-022-33213-z. https://doi.org/10.1038/s41467-022-33213-z
  11. Qin, Yuanwei; Xiao, Xiangming; Liu, Fang; de Sa e Silva, Fabio; Shimabukuro, Yosio; Arai, Egidio; Fearnside, Philip Martin (2023). Forest conservation in Indigenous territories and protected areas in the Brazilian Amazon. Nature Sustainability 6:295-305. DOI 10.1038/s41893-022-01018-z. https://doi.org/10.1038/s41893-022-01018-z
  12. Huais, Pablo Yair; Kuemmerle, Tobias; Nori, Javier; Tomba, Ana N.; Cordier, Javier Maximiliano; Baumann, Matthias (2025). Only one-third of protected areas in the Chaco effectively curb woodland loss, but their impact extends beyond their boundaries. Biological Conservation 308:111196. DOI 10.1016/j.biocon.2025.111196. https://doi.org/10.1016/j.biocon.2025.111196