Past fire use in deforested land limits secondary forest carbon gains in Amazonia with implications for emissions reduction goals

Abstract Background Understanding the carbon (C) recovery potential of secondary forests (SF) is essential for reducing uncertainties regarding the contribution of Amazon forests to the global carbon budget, and for strengthening national and sub-national climate mitigation policies aligned with Brazil’s commitments under the Paris Agreement. We, hence, focus on Pará state in eastern Amazon to evaluate how regional edaphoclimatic conditions, prior land-use duration, and fire recurrence influenced C removal rates of SFs. Methods We combined Earth observation-based aboveground biomass and SF age data, derived from multitemporal land cover maps, to construct chronosequences representing four past land-use classes: undisturbed SF (regrowth immediately after deforestation), anthropogenic use with fire, anthropogenic use only, and fire only across three edaphoclimatic regions. We characterised spatiotemporal variations in C gain and projected potential removals by 2035 under three protection scenarios (SF age ≥ 1, ≥5, and ≥ 10 years). Results We showed that a history of extensive land use and fire before forest regeneration reduced C gain rates by 27% to 40% compared to undisturbed SFs in Southeast Pará. These impacts persisted across all successional stages. Model projections estimate that fire-affected areas could require up to 226 years to recover mature forest C stocks—three times longer than undisturbed SFs—highlighting long-term recovery lags despite inherent modelling uncertainties. Furthermore, high fire recurrence (≥ 3 events) led to a 34% reduction of the C gain rate. Regarding the mitigation potential, protecting only SFs ≥ 10 years old reduced projected 2020–2035 removals by 79% compared to protecting all SFs (age ≥ 1 year; which could sequester 691 Tg CO₂). Conclusions Achieving net-zero targets under Pará’s “Amazon Now” State Plan (PEAA) is challenging. Offsetting current emissions would require doubling the existing regenerating (≥ 1 year) area to 112,644 km². Given the long-term degradation of successional trajectories by fire and land use, reducing annual deforestation from 2,673 to 1,341 km² by 2030 appears to be a more efficient and feasible strategy for carbon neutrality than relying solely on natural regeneration. Ultimately, understanding the role of regenerating forests in the global carbon cycle is fundamental for designing climate policies and targets aimed at achieving net-zero greenhouse gas emissions.