Deforestation remains one of the most significant sources of greenhouse gas emissions globally. Tropical regions are particularly critical to climate mitigation due to their high carbon density and ongoing conversion to agricultural uses. While earth observation (EO) has become fundamental for assessing land-use change over time and space, key gaps persist in understanding associated carbon emissions (CE) and their distribution across production systems, information required by diverse international and European policy frameworks targeting greenhouse gas reduction and deforestation. To address this, the study integrates satellite-derived aboveground biomass (AGB) with machine learning-based land-use modeling to quantify commodity-specific CE from deforestation in Colombia. Our key findings reveal that pasture expansion for livestock dominates emissions (>67% in the Amazonas and Orinoquía regions, approximately 25% nationally), while cocoa and coffee, despite smaller cultivation areas, exhibit high emission factors (>25% of subregional totals in the Andes, Pacífico, and Caribe). Smallholder croplands contribute up to 16.7% regionally, and oil palm plays a modest role nationally. These results underscore the importance of targeting both extensive pasture systems and high-emission perennial crops in deforestation-free supply chain policies, while considering the diverse regional land-use dynamics. In conclusion, our framework offers spatially explicit, crop-specific emission estimates that address essential gaps in EO-based CE quantification. It provides clear, actionable baselines for monitoring policies aimed at promoting deforestation-free supply chains, such as the EU Deforestation Regulation (EUDR), across tropical landscapes. Ultimately, this data supports the creation of a comprehensive database of country-specific emission factors, a resource critical for enhancing national carbon inventories, guiding policy decisions related to high-risk deforestation commodities, and ensuring more transparent and traceable zero-deforestation supply chains.

Oil palm (OP) expansion and associated forest clearance can significantly impact greenhouse gas (GHG) fluxes. We investigated carbon stocks and soil GHG (N2O, CO2, CH4) fluxes in a degraded forest and an adjacent 17-yearold OP plantation in Peruvian Amazonia. The plantation comprised three nitrogen (N) fertilizer treatments: 0 (OPN0), 84 (OPN1), 168 (OPN2) kg N ha-1 y-1. Carbon stocks were inventoried across all pools. GHG and environmental parameters were monitored monthly for 11 months and (bi)daily when fertilizing the OP, with measurements taken both near and far from trees/palms. Ecosystem-scale CO2equivalent losses from the conversion were computed by balancing carbon stock losses against N2O emission changes. N2O emissions (kg N ha-1 y-1) in the forest (6.7 f 1.2) where litterfall N inputs were large (213 kg N ha-1 y-1) were 11, 5, and 3 times the emissions in OPN0 (0.6 f 0.2), OPN1 (1.4 f 0.2), and OPN2 (2.3 f 0.3). In the plantation, 1 % of the N fertilizer applied was released as N2O. Across ecosystems, N inputs primarily controlled N2O emissions. Soil respiration (Mg C ha-1 y-1) was 1.4 times higher in the forest (9.1 f 0.6) than in the plantation (7.3 f 1, 5.5 f 0.5, 6.5 f 0.3 in OPN0, OPN1, OPN2). The forest was a soil CH4 (kg C ha-1 y-1) sink (-1.5 f 0.3) while all OP treatments were sources (0.2 f 0.3, 0.7 f 0.5, 0.2 f 0.4 in OPN0, OPN1, OPN2). Ecosystem carbon stock losses from forest-to-OP conversion were substantial (196.8 f 44.0 Mg CO2 ha-1 15 y-1) and partially offset (14-20 %) by decreased N2O emissions. Complementary studies for this transition are needed to improve global GHG assessments.

Amazonian forests function as biomass and biodiversity reservoirs, contributing to climate change mitigation. While they continuously experience disturbance, the effect that disturbances have on biomass and biodiversity over time has not yet been assessed at a large scale. Here, we evaluate the degree of recent forest disturbance in Peruvian Amazonia and the effects that disturbance, environmental conditions and human use have on biomass and biodiversity in disturbed forests. We integrate tree-level data on aboveground biomass (AGB) and species richness from 1840 forest plots from Peru's National Forest Inventory with remotely sensed monitoring of forest change dynamics, based on disturbances detected from Landsat-derived Normalized Difference Moisture Index time series. Our results show a clear negative effect of disturbance intensity tree species richness. This effect was also observed on AGB and species richness recovery values towards undisturbed levels, as well as on the recovery of species composition towards undisturbed levels. Time since disturbance had a larger effect on AGB than on species richness. While time since disturbance has a positive effect on AGB, unexpectedly we found a small negative effect of time since disturbance on species richness. We estimate that roughly 15% of Peruvian Amazonian forests have experienced disturbance at least once since 1984, and that, following disturbance, have been increasing in AGB at a rate of 4.7 Mg ha(-1) year(-1) during the first 20 years. Furthermore, the positive effect of surrounding forest cover was evident for both AGB and its recovery towards undisturbed levels, as well as for species richness. There was a negative effect of forest accessibility on the recovery of species composition towards undisturbed levels. Moving forward, we recommend that forest-based climate change mitigation endeavours consider forest disturbance through the integration of forest inventory data with remote sensing methods.

Mangrove forest plays a key role in regulating climate change, earth carbon cycling and other biogeochemical processes within blue carbon ecosystems. Therefore, mangrove forests should be incorporated into Earth system climate models with the aim of understanding future climate change. Despite multiple carbon stock and flux assessments taking place over the past couple of decades, concrete knowledge of carbon source/sink patterns is largely lacking, particularly in the biodiversity-rich Asia-Pacific (AP) region with its 68 493 km 2 of mangrove area. Thus, to understand the gaps in mangrove blue carbon research in the AP region, we summarize a recent decade-long inventory of carbon stock pools (aboveground, belowground and soil) and biogeochemical flux components (burial, export/import, soil-air and water-air CO 2 flux) across 25 AP countries to understand the current knowledge and gaps. While carbon stock assessments of individual components are available for all 25 countries, whole ecosystem carbon stocks—including live and standing dead aboveground and belowground, downed woody debris and soil carbon stocks—are often lacking, even in highly researched countries like Indonesia. There is restricted knowledge around biogeochemical carbon fluxes in 55% of the countries, suggesting poor carbon flux research across the region. Focusing on flux components, reports on sediment-to-sea carbon exports are extremely limited (coming from just nine countries in the AP region). There is notable scarcity of data on carbon export fluxes in Indonesian mangroves. Given the key role AP mangroves play in climate change mitigation worldwide, more detailed and methodologically comparable investigation of biogeochemical source/sink processes is required to better understand the role of this large carbon source in global carbon stocks and fluxes, and hence, global climate.

Flooded savannas are valuable and extensive ecosystems in South America, but not widely studied. In this study, we quantify the spatial distribution of soil organic carbon (SOC) content and stocks in the Casanare flooded savannas. We sampled 80 sites at two soil-depth intervals (0-10 and 10-30 cm), where SOC values ranged from 0.41% in the surface and 0.23% in the sub-surface of drier soils to over 14.50% and 7.51%, in soils that experienced seasonal flooding. Spatial predictions of SOC were done through two digital soil mapping (DSM) approaches: Expert-Knowledge (EK) and Random-Forest (RF). Although both approaches performed well, EK was slightly superior at predicting SOC. Covariates derived from vegetation cover, topography, and soil properties were identified as key drivers in controlling its distribution. Total SOC stocks were 55.07 Mt with a mean density of 83.1±24.3 t·ha -1 in the first 30 cm of soil, with 12.3% of this located in areas that experience long periods of flooding (semi-seasonal savannas) , which represented only 7.9% of the study area (664,752 ha). Although the study area represents only 15% of the total area of the Casanare department, the intensive pressure of human development could result in the reduction of its SOC stocks and the release of important amounts of greenhouse gases into the atmosphere. At regional level, the impact of a large-scale land use conversions of the flooded Llanos del Orinoco ecosystem area (15 Mha) could transform this area in a future source of important global emissions if correct decisions are not taken regarding the land management of the region.

National-scale assessments of post-deforestation land-use are crucial for decreasing deforestation and forest degradation-related emissions. In this research, we assess the potential of different satellite data modalities (single-date, multi-date, multi-resolution, and an ensemble of multi-sensor images) for classifying land-use following deforestation in Ethiopia using the U-Net deep neural network architecture enhanced with attention. We performed the analysis on satellite image data retrieved across Ethiopia from freely available Landsat-8, Sentinel-2 and Planet-NICFI satellite data. The experiments aimed at an analysis of (a) single-date images from individual sensors to account for the differences in spatial resolution between image sensors in detecting land-uses, (b) ensembles of multiple images from different sensors (Planet-NICFI/Sentinel-2/Landsat-8) with different spatial resolutions, (c) the use of multi-date data to account for the contribution of temporal information in detecting land-uses, and, finally, (d) the identification of regional differences in terms of land-use following deforestation in Ethiopia. We hypothesize that choosing the right satellite imagery (sensor) type is crucial for the task. Based on a comprehensive visually interpreted reference dataset of 11 types of post-deforestation land-uses, we find that either detailed spatial patterns (single-date Planet-NICFI) or detailed temporal patterns (multi-date Sentinel-2, Landsat-8) are required for identifying land-use following deforestation, while medium-resolution single-date imagery is not sufficient to achieve high classification accuracy. We also find that adding soft-attention to the standard U-Net improved the classification accuracy, especially for small-scale land-uses. The models and products presented in this work can be used as a powerful data resource for governmental and forest monitoring agencies to design and monitor deforestation mitigation measures and data-driven land-use policy.

Collective Payments for Ecosystem Services (PES), where forest users receive compensation conditional on group rather than individual performance, are an increasingly used policy instrument to reduce tropical deforestation. However, implementing effective, (cost) efficient and equitable (3E) collective PES is challenging because in-dividuals have an incentive to free ride on others' conservation actions. Few comparative studies exist on how different enforcement strategies can improve collective PES performance. We conducted a framed field experi-ment in Brazil, Indonesia and Peru to evaluate how three different strategies to contain the local free-rider problem perform in terms of the 3Es: (i) Public monitoring of individual deforestation, (ii) internal, peer-to-peer sanctions (Community enforcement) and (iii) external sanctions (Government enforcement). We also examined how inequality in wealth, framed as differences in deforestation capacity, affects policy performance. We find that introducing individual level sanctions can improve the effectiveness, efficiency and equity of col-lective PES, but there is no silver bullet that consistently improves all 3Es across country sites. Public monitoring reduced deforestation and improved the equity of the program in sites with stronger history of collective action. External sanctions provided the strongest and most robust improvement in the 3Es. While internal, peer enforcement can significantly reduce free riding, it does not improve the program's efficiency, and thus par-ticipants' earnings. The sanctioning mechanisms failed to systematically improve the equitable distribution of benefits due to the ineffectiveness of punishments to target the largest free-riders. Inequality in wealth increased group deforestation and reduced the efficiency of Community enforcement in Indonesia but had no effect in the other two country sites. Factors explaining differences across country sites include the history of collective action and land tenure systems.

Miombo woodlands are extensive dry forest ecosystems in central and southern Africa covering ≈2.7 million km2. Despite their vast expanse and global importance for carbon storage, the long-term carbon stocks and dynamics have been poorly researched. The objective of this paper was to present and summarize the evidence gathered on aboveground carbon (AGC) and soil organic carbon (SOC) stocks of miombo woodlands from the 1960s to mid-2018 through a literature review. We reviewed the data to find out to what extent aboveground carbon and soil organic carbon stocks are found in miombo woodlands and further investigated if are there differences in carbon stocks based on woodland categories (old-growth, disturbed and re-growth). A review protocol was used to identify 56 publications from which quantitative data on AGC and SOC stocks were extracted. We found that the mean AGC in old-growth miombo (45.8 ± 17.8 Mg C ha−1), disturbed miombo (26.7 ± 15 Mg C ha−1), and regrowth miombo (18.8 ± 16.8 Mg C ha−1) differed significantly. Data on rainfall, stand age, and land-use suggested that the variability in aboveground carbon is site-specific, relating to climatic and geographic conditions as well as land-use history. SOC stocks in both old-growth and re-growth miombo were found to vary widely. It must be noted these soil data are provided only for information; they inconsistently refer to varying soil depths and are thus difficult to interpret. The wide range reported suggests a need for further studies which are much more systematic in method and reporting. Other limitations of the dataset include the lack of systematic sampling and lack of data in some countries, viz. Angola and Democratic Republic of the Congo.

Assessing land-use following deforestation is vital for reducing emissions from deforestation and forest degradation. In this paper, for the first time, we assess the potential of spatial, temporal and spatio-temporal deep learning methods for large-scale classification of land-use following tropical deforestation using dense satellite time series over six years on the pan-tropical scale (incl. Latin America, Africa, and Asia). Based on an extensive reference database of six forest to land-use conversion types, we find that the spatio-temporal models achieved a substantially higher F1-score accuracies than models that account only for spatial or temporal patterns. Although all models performed better when the scope of the problem was limited to a single continent, the spatial models were more competitive than the temporal ones in this setting. These results suggest that the spatial patterns of land-use within a continent share more commonalities than the temporal patterns and the spatial patterns across continents. This work explores the feasibility of extending and complementing previous efforts for characterizing follow-up land-use after deforestation at a small-scale via human visual interpretation of high resolution RGB imagery. It supports the usage of fast and automated large-scale land-use classification and showcases the value of deep learning methods combined with spatio-temporal satellite data to effectively address the complex tasks of identifying land-use following deforestation in a scalable and cost effective manner.

Background Amazon palm swamp peatlands are major carbon (C) sinks and reservoirs. In Peru, this ecosystem is widely threatened owing to the recurrent practice of cutting Mauritia flexuosa palms for fruit harvesting. Such degradation could significantly damage peat deposits by altering C fluxes through fine root productivity, mortality, and decomposition rates which contribute to and regulate peat accumulation. Along a same peat formation, we studied an undegraded site (Intact), a moderately degraded site (mDeg) and a heavily degraded site (hDeg) over 11 months. Fine root C stocks and fluxes were monthly sampled by sequential coring. Concomitantly, fine root decomposition was investigated using litter bags. In the experimental design, fine root stocks and dynamics were assessed separately according to vegetation type ( M. flexuosa palm and other tree species) and M. flexuosa age class. Furthermore, results obtained from individual palms and trees were site-scaled by using forest composition and structure. Results At the scale of individuals, fine root C biomass in M. flexuosa adults was higher at the mDeg site than at the Intact and hDeg sites, while in trees it was lowest at the hDeg site. Site-scale fine root biomass (Mg C ha −1 ) was higher at the mDeg site (0.58 ± 0.05) than at the Intact (0.48 ± 0.05) and hDeg sites (0.32 ± 0.03). Site-scale annual fine root mortality rate was not significantly different between sites (3.4 ± 1.3, 2.0 ± 0.8, 1.5 ± 0.7 Mg C ha −1 yr −1 at the Intact, mDeg, and hDeg sites) while productivity (same unit) was lower at the hDeg site (1.5 ± 0.8) than at the Intact site (3.7 ± 1.2), the mDeg site being intermediate (2.3 ± 0.9). Decomposition was slow with 63.5−74.4% of mass remaining after 300 days and it was similar among sites and vegetation types. Conclusions The significant lower fine root C stock and annual productivity rate at the hDeg site than at the Intact site suggests a potential for strong degradation to disrupt peat accretion. These results stress the need for a sustainable management of these forests to maintain their C sink function.

This infobrief summarizes results from analyzing scientific articles across all fields of studies to answer three main questions: (i) What is the state of science in Key messages• "Transformational change" is needed for meeting ambitious 2030 climate mitigation, sustainability and development goals.Definitions of transformational change in scientific and grey literature have the following in common: (i) movement away from the current status, opening new pathways; (ii) sustained change, through institutionalization or deep changes; (iii) focus on root causes; and (iv) harnessing knowledge and learning.• Depth, speed and scale are three dimensions of transformational change.Four types of drivers of transformational change: Processes, Resources, Norms, and Legitimacy.Empirical examination on these drivers' efficacy is urgently needed.• The scientific literature is dominated by authors in high-income countries and rarely results from large collaborative efforts.This is in contrast with the drivers of transformational change that the literature itself has identified: inclusiveness, collaboration and cross-learning.• Specific research is needed on transformational change in land use and climate change, drawing on rich insights from health, education and business sectors.• Scientific knowledge and practical needs must be reconciled, e.g., by providing guidelines and tools for monitoring and evaluation, programme and project management, and financing mechanisms adapted to complex, multilevel and long-term, 'transformational' endeavours.