In this session, we aim to place a particular emphasis on the dynamics of disturbance, recovery, and resilience in tropical forest ecosystems. We invite contributions that examine the consequences of these contrasting processes across multiple dimensions, including above- and belowground biomass, forest-atmosphere interactions, carbon budgets, biodiversity, and social livelihoods. We welcome studies using diverse approaches, ranging from experimental and field-based investigations to remote sensing and modeling, particularly on data poor regions like Central Africa. Our goal is to foster a holistic understanding of how disturbance and recovery shape both the current and future states of tropical vegetation and the communities that depend on it.
Orals: Mon, 4 May, 14:00–18:00 | Room N1
Tropical forests will be increasingly impacted by intensifying droughts under ongoing climate change. Droughts may have large implications for the capacity of tropical forests to store carbon in wood and act as long-term carbon sinks. To assess this possible implication, pantropical analyses of tree growth recovery from past droughts are needed.
Here we use tree-ring data from 250 tropical forest sites (>5000 trees) in the tropical tree-ring network (tropicaltreeringnetwork.org) to evaluate the capacity for woody growth to recover after droughts. We assessed impacts on stem growth of the 10% and 5% driest years since 1930 and the extent to which stem growth recovered in the subsequent years. We selected years with the lowest rainfall, highest vapour pressure deficit, and largest climatic water deficit.
We found that the pantropical impact of droughts on stem growth was a modest 2.8% reduction (CI95%: -3.0 to -2.4%) during the 10% driest years, and 2.9% (CI95%: -3.4 to -2.4%) during the 5% driest years. Yet, in a quarter of sites located predominantly in the drier tropical forest regions, growth was reduced by over 10%.
Growth recovery was generally rapid: postdrought years exhibited significantly smaller negative growth anomalies, and anomalies often shifted to positive. Thus, this rapid recovery of stem growth to predrought levels suggests that no strong or long lag effects exist at the pantropical scale. We also analysed regional, taxonomic and seasonal differences in growth recovery, and compared the results to drought responses in photosynthesis.
We conclude that stem growth of tropical forest trees has so far been able to recover from droughts, but that this capacity may diminish under aggravating climate change.
How to cite: Zuidema, P., Babst, F., Groenendijk, P., Rahman, M., and Trouet, V. and the Tropical Tree-ring Network: Disturbance-recovery cycles of tree growth after droughts in tropical forests, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-10854, https://doi.org/10.5194/egusphere-egu26-10854, 2026.
The global forest sink, particularly the tropical forest sink, plays a critical role in the global carbon cycle (Bueno et al., 2020; Pan et al., 2024). Carbon stored in forests is equivalent to almost half the carbon emitted from fossil fuels between 1990 and 2019; however, persistent patterns of deforestation and degradation threaten this carbon sink and its vital role in climate regulation (Pan et al., 2024). Abandoned or left to recover, post-disturbance landscapes can lead to naturally regenerating forests that contribute to significant carbon sequestration, biodiversity benefits, and ecosystem services (Chazdon et al., 2016). Therefore, it is crucial to understand when and where regenerating forests occur to highlight their environmental contributions and promote forest restoration (de Jong et al., 2025). Advancements in RS technologies have enabled a proliferation of RS-based land cover datasets with the potential to estimate secondary forest age and extent (Xu et al., 2024; Baker et al., 2025). However, many have not been explicitly validated for the age or extent of post-disturbance forest regrowth. First comparisons of datasets on secondary forest extent reveal substantial differences at the pixel scale, calling for the need for a reference dataset.
Creating a high-quality reference dataset is essential to ensure the reliability of remote sensing-based maps (Baker et al., 2025) and deliver meaningful results to policymakers (Xu et al., 2024; Tyukavina et al., 2025). Here, we propose a robust methodology and reference dataset to validate different secondary forest estimates and support a broader analysis of tropical forest regrowth dynamics. Using a stratified sampling design based on Potapov et al. (2022) and Zhang et al. (2023)'s land cover products, samples were extracted from areas of agreement and disagreement on forest gain, forest loss, stable forest, and stable non-forest regions. Forest trajectories, drivers of disturbance and regrowth, and years of disturbance and regrowth were interpreted using Landsat, Planet, and Google Earth Pro imagery from 2000-2020. Focusing on tropical biomes outside the Amazon Basin which are often overlooked, weighted area estimates from a preliminary 500-sample reference dataset in South America reveal 87.97Mha (5.12% ± 1.00%) of deforestation, 21.70Mha (1.24% ± 0.25%) of secondary forests, and 82.34Mha (4.72% ± 0.92%) of degraded forests. Although the preliminary sample does not indicate systematic over- or under-estimation of secondary forest age, results reveal high commission errors in the extent of naturally regenerating secondary forests outside the Amazon Basin. Furthermore, high uncertainty in the interpretation of tropical dry forest samples highlights the challenges in identifying secondary forests outside the humid tropics and emphasizes the need for more research outside of the Amazon.
This dataset has the potential to expand across the pan-tropics to harmonize essential information on regenerating forests and guide urgent action needed to protect the forest carbon sink. Learning from the preliminary study in South America, we emphasize the importance of data quality and draw attention to the uncertainty of large-scale secondary forest products.
How to cite: Graham, H., Heinrich, V., Tsendbazar, N., Carreiras, J., Beuchle, R., Bourgoin, C., Carboni, S., Freitas, S., Lastra Munoz, J., and Herold, M.: Clarifying tropical secondary forest estimates: results from a growing reference dataset on post-disturbance regrowth, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-13106, https://doi.org/10.5194/egusphere-egu26-13106, 2026.
Fire regimes and human impacts on Amazonian forests have varied over decades to centuries, resulting in disturbances and recoveries that leave lasting legacies on vegetation and soil. While intact (old-growth) forests were once thought to be largely fire-free, our work has shown these forests to have a history of infrequent but recurrent fire. Wildfires have left their signatures in the soil in the form of black carbon (pyrogenic carbon or PyC), the product of incomplete combustion of organic matter. Vegetation has also likely responded to past fire; however, the mechanistic effects of these past disturbances remain poorly understood. Here, we examine Amazon disturbance and recovery processes over space and time (ancient to modern) in relation to fire.
We utilise permanent forest plot data (soil PyC, physicochemical properties, vegetation) from two large-scale projects across the Amazon Basin, combined with remote sensing data. The analysis shows that soil texture and hydrology primarily explain the spatial variation of soil PyC at 30 cm depth, while historical climate played a relatively minor role. Furthermore, soil PyC from ancient wildfires is associated with increased soil fertility in intact forests. We also found that distinct groups of tree species in Amazonia are associated with ancient soil PyC. In contrast, modern fires increase soil PyC but result in a reduction in total SOC, degrade soil health, and reduce species richness.
These findings indicate that infrequent ancient wildfires recurring at intervals spanning several hundred years had positive impacts on soil fertility and left legacy effects on modern forest composition. Conversely, modern fires, which are extensive and have short return intervals, negatively impact Amazon soils and vegetation on decadal scales. To better assess the long-term impacts of fire on soil carbon, we are incorporating soil PyC into Land Surface Models.
How to cite: Feldpausch, T., Barbosa de Camargo, P., Carvalho, L., Bieluczyk, W., Pompeu, J., Naval, M., Alvarez, F., Phillips, O., Bird, M., Aragão, L., Leonardo, M.-S., Silva, K., Quesada, C., Schwantes Marimon, B., Brando, P., Macario, K., and Marimon Junior, B. H.: Fire-driven dynamics in Amazonia: Contrasting ancient legacies and modern degradation in soils and vegetation, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-20082, https://doi.org/10.5194/egusphere-egu26-20082, 2026.
The Amazon rainforest is increasingly affected by both forest fragmentation and extreme droughts. However, the interacting impacts of these two disturbances on forest carbon dynamics remain poorly understood. Here, we investigated how fragmentation-induced forest edges modulated productivity responses to the 2023–2024 “once-in-a-hundred-years” drought in the Amazon. Using high-resolution satellite observations of photosynthetic indices together with canopy height and forest cover data, we compared the differences in drought responses between edge and interior forests across the basin. We found that, at the basin scale, edge forests exhibited stronger drought-induced productivity declines than interior forests and contributed 64.0 ± 18.1% of the total productivity loss during this drought, indicating that edge effects overall amplify drought impacts on productivity. However, edge responses exhibited strong regional contrasts during this drought. In the northeastern Amazon, where water tables were deeper and droughts were more severe, edge forests exhibited 4.9 ± 2.4% greater productivity declines than interior forests. In contrast, in the southwestern Amazon, characterized by shallow water tables, edge forests showed 2.9 ± 1.8% smaller productivity reductions than interior forests. In addition, edge-related forest structural degradation, reflected by reduced canopy height, further intensified the differences in drought responses between edge and interior forests. Our findings show that edge–drought interactions substantially undermine the carbon uptake across large areas of the Amazon, highlighting the urgent need to curb further fragmentation and protect remaining interior forests, particularly in drought-prone edge regions.
How to cite: Zhao, D., Duveiller, G., Cescatti, A., Ciais, P., Cao, R., Zhang, Z., Zhu, L., Peñuelas, J., Li, W., and Zhang, Y.: Forest edge effects amplify impacts of 2023-2024 Amazon drought on forest productivity, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5962, https://doi.org/10.5194/egusphere-egu26-5962, 2026.
Tree crown damage from disturbance events can strongly influence forest demography, yet its effect on stem growth remains poorly quantified. Hurricanes provide a powerful natural experiment to examine these dynamics, as they inflict a broad range of structural damage across individuals and forest stands. Here we assess how crown damage from Hurricane Mar´ıa (2017) affected post-storm stem growth in a wet
subtropical forest in Puerto Rico by combining airborne LiDAR with field measurements for 1,082 trees. Unlike previous studies, we used a continuous, objective measure of crown damage and explicitly separated individual- from neighborhood-level effects using a causal inference framework. Across the population, stem growth rates after the hurricane were similar to pre-hurricane values. Larger and more heavily damaged trees exhibited moderately reduced growth, while neighborhood crown damage and neighborhood mortality had no detectable effect. However, these damage effects were smaller than the influence of pre-hurricane growth rates, indicating that pre-hurricane individual vigor outweighed biomass loss and competitive release
in shaping growth responses. Our findings highlight the resilience of surviving trees in sustaining carbon uptake after a severe disturbance, while challenging the assumption of a strong growth suppression following biomass loss, embedded in dynamic vegetation models.
How to cite: Boeschoten, L., Ankori-Karlinsky, R., Arellano, G., Brown, A., Fang, D., Leitold, V., Morton, D., Yuan, G., Zheng, T., Zimmerman, J., and Uriarte, M.: Tree growth after a major hurricane reflects pre-disturbance vigorrather than canopy damage, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-1615, https://doi.org/10.5194/egusphere-egu26-1615, 2026.
Tropical montane forests are globally important for biodiversity conservation, carbon storage, and regulation of hydrological and biogeochemical cycles, yet they are increasingly shaped by interacting environmental gradients and human-induced disturbances. In the Eastern Afromontane biodiversity hotspot in southwest Ethiopia, community-managed forests represent a growing governance model intended to reconcile conservation and livelihoods. To date empirical evidence on how disturbance and recovery processes operate within these systems remains limited. In this paper, we examine woody species diversity, forest structure, and regeneration dynamics in the Andaracha Participatory Forest Management (PFM) forest of, with a specific focus on how environmental heterogeneity and disturbance intensity structure forests ecology.
We conducted a field-based ecological assessment using 44 systematically distributed nested plots across 11 forest management units (Gots). Floristic inventories recorded species identity and diameter at breast height (DBH) for all woody species, alongside plot-level regeneration data based on seedling counts. Environmental variables (altitude, slope, canopy cover) and a composite disturbance index integrating logging, grazing, fuelwood extraction, and access trails were recorded. In addition, institutional context was characterised using a PFM engagement score derived from participatory monitoring of annual plan and bylaw enforcement, meeting frequency, forest patrolling, and reporting mechanisms. Multivariate analyses, including hierarchical clustering, principal component analysis (PCA), non-metric multidimensional scaling (NMDS), and PERMANOVA, were used to assess relationships between species composition, regeneration patterns, and environmental-disturbance gradients.
Across the study area, a total of 60 woody species belonging to 38 genera and 28 families were recorded, with Rubiaceae and Euphorbiaceae among the most species-rich families. Forest structure was heterogeneous, with reverse-J DBH distributions at the landscape scale indicating ongoing recruitment, but substantial plot-level variation in size-class structure. Species composition clustered into five distinct community types aligned primarily along altitudinal, slope, and canopy gradients. Regeneration dynamics were highly uneven: more than two-thirds of all seedlings were concentrated in fewer than 10% of the plots, revealing strong spatial patchiness in recovery processes.
Ordination analyses highlighted disturbance intensity and canopy cover as key axes structuring both adult community composition and regeneration assemblages. Moderate disturbance levels were associated with higher species diversity and more balanced regeneration, whereas heavily disturbed and open-canopy plots showed reduced recruitment and greater dominance by disturbance-tolerant taxa. Conversely, steep, high-altitude plots with low disturbance exhibited environmentally filtered regeneration characterised by low diversity but stable species composition. These patterns indicate that woody species composition and regeneration in the Andaracha community-managed forest are shaped not by disturbance alone, but by its interaction with topographic constraints and canopy structure.
Our findings demonstrate that community-managed Afromontane forests can sustain high woody biodiversity and active regeneration, but that recovery is highly spatially uneven and sensitive to ecological thresholds of disturbance. These findings underscore the importance of site-specific, ecologically informed management strategies to enhance regeneration resilience in tropical montane forests undergoing rapid socio-ecological change.
How to cite: Sy, T. B., Hwang, B., Snell, M., Ozigis, M. S., Wood, A., and Tolera, M.: Woody species diversity and regeneration dynamics along environmental and disturbance gradients in a community-managed Afromontane forest, southwest Ethiopia, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-22013, https://doi.org/10.5194/egusphere-egu26-22013, 2026.
Central african forests are subject to increasing anthropic disturbance, with slash and burn agriculture and wood extraction as the leading causes of disturbance. The extent of resulting fallows and secondary forests are currently poorly investigated in the region while they are crucial to map and to monitor to better understand forest state, functioning and potential recovery of the second largest tropical forest of the world.
In this study, we explore the hypothesis that recovery rate of carbon and functional diversity can be modelized by change in canopy structure. To estimate the carbon recovery rate in secondary forest, 15 permanent forest inventories of 1ha each are conducted in the Mabali research centre of the Equateur province in DRC. We use space for time substitution method to create a chronosequence covering forests of different ages and/or different types of past disturbance. UAV’s LIDAR data acquired during the field campaign over the plots and the surrounding forest are processed to extract meaningful canopy features. By exploring the relationship between forest structure and forest recovery rate, we aim to deepen the understanding of the complexity of secondary forests and provide datasets for the calibration of Land Surface Modeling and satellite-based biomass estimation.
How to cite: Collet, T., Mbavumoja, T., Vanderlinden, A., and Bastin, J.-F.: Modeling Carbon and Functional Diversity Recovery in Central African Secondary Forests Using Canopy Structure, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-17643, https://doi.org/10.5194/egusphere-egu26-17643, 2026.
Intensification of shifting cultivation practices across the Congo basin are causing a rise in secondary forest area across the African tropical belt. Nutrient exports during burning (e.g. ash pulses) and cultivation of the land (e.g. biomass export and nutrient leaching) are potentially limiting the regrowth potential of the secondary forests that reestablish after land abandonment, potentially putting a break on the C accumulation rates of these forests. Among the nutrients that are being exported from the land, cations (Ca, Mg, K) are expected to be particularly vulnerable, especially on the old, highly weathered soils that are characteristic for the majority of lowland tropical forests in the Congo basin. Past studies on forest chronosequence data have already shown that (1) the availability of cations in the soil declines over time, and (2) cations accumulate in woody biomass over time and might get lost from the system permanently when timber or fire wood is being extracted from the system.
In our study, we aim to integrate data from a large fertilization experiment and observations along a chronosequence of secondary forest stands to assess the role of cation depletion on forest regrowth in the Congo basin. Here, we aim to present our (1) preliminary data and first findings and (2) our approach to integrate these findings into a biogeochemical forest growth model PnET-BGC. Via model-based scenario analyses, we present potential impacts of cation limitation on forest regrowth in the Congo basin and discuss how management can conserve nutrients and sustain carbon uptake in regenerating tropical forests.
How to cite: Landuyt, D., Lokana Mande, J., Bondongwe Wombe, M., Alebadwa Mombenga, S., Boeckx, P., and Bauters, M.: Combining experiments, observations and modelling to assess the role of soil cation depletion on secondary forest regrowth in the Congo basin, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-14859, https://doi.org/10.5194/egusphere-egu26-14859, 2026.
Reforestation is increasingly promoted as a nature based climate solution, yet the degree to which soil carbon recovers alongside vegetation remains uncertain, particularly in deeply weathered tropical soils. While rapid gains in above ground biomass are often interpreted as indicators of ecosystem recovery, it is unclear whether such gains translate into meaningful changes in below ground carbon pools.
We investigated soil and vegetation carbon dynamics across a thirty year gradient of forest disturbance and recovery in tropical montane forests of Kibale National Park, Uganda, spanning primary forest, passive natural regeneration, and actively replanted stands. The study integrates depth resolved soil organic carbon stocks to one metre, stable carbon isotope profiles, soil physical and geochemical properties, and long term forest inventory data from permanent monitoring plots. By combining carbon stocks, isotopic indicators, and soil mineral properties, we assess how strongly soil carbon is coupled to forest recovery and at what depths soils respond to changes in vegetation. The analysis reveals clear contrasts between above ground and below ground carbon trajectories and highlights the role of soil depth and legacy effects in shaping carbon storage in recovering tropical forests.
Our results provide new insight into the limits of using biomass recovery as a proxy for soil carbon sequestration and underline the importance of depth resolved and process oriented approaches when evaluating reforestation outcomes.
How to cite: Cooper, M., Bauters, M., Riemer, K., van Goor, W., Mugenyi, K., Summerauer, L., Cook, J., Kigenyi, R., and Doetterl, S.: Decoupled recovery of soil and vegetation carbon reveals mineral-driven stabilisation in reforested tropical montane ecosystems, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-20803, https://doi.org/10.5194/egusphere-egu26-20803, 2026.
Tropical forests play a critical role in the global carbon cycle, biodiversity conservation and climate regulation, yet their internal functioning, phenology and fine-scale dynamics remain poorly characterized. While satellite observations have revolutionized large-scale assessments of forest change, their interpretation is still limited by the scarcity of intermediate-scale observations bridging ground-based measurements and orbital sensors. This gap is particularly acute in dense African forests, where diffuse degradation, small canopy openings and climate-driven stress processes are difficult to detect and attribute, and the capacity of space-borne optical observation is limited by clouds and other atmospheric effects.
Here we present the potential of a network of UAV-based forest observatories (Canobs.net), deployed across tropical forest regions, in South America, Central Africa, South East Asia and Oceania. These observatories combine repeated drone acquisitions (RGB, multispectral photogrammetry and LiDAR), permanent forest inventories, targeted ecophysiological measurements and multi-sensor satellite time series. This integrated framework enables spatially continuous and temporally dense monitoring of canopy structure, forest functioning and biodiversity at resolutions inaccessible to satellites alone.
We show that such observatories are essential to: (i) resolve forest phenology and canopy functioning by linking UAV-based monitoring of canopy dynamics with photosynthetic capacity and satellite signals; (ii) quantify the dynamics and mortality of large trees, which dominate carbon stocks and fluxes; (iii) interpret, calibrate and validate satellite-derived biomass products, Essential Biodiversity Variables and functional forest maps. A major recent advance is the application of the Pl@ntNet AI-based species identification app to UAV imagery, allowing identification and monitoring of canopy tree diversity.
The Canobs network forms a critical link between plant- and leaf-scale ecophysiology, field inventories and continental-scale satellite studies, providing a robust framework to better understand and monitor the shifting dynamics of tropical forests under climate and land-use change.
How to cite: Barbier, N., Ploton, P., Heuret, P., Engel, J., Burban, B., Ball, J., Bastin, J.-F., Angoboy Ilondea, B., Mbock, G., Sonké, B., Endezoumou, B., Leborgne, B., Ordway, E., Lombardo, J.-C., Vermeulen, L., Fayolle, A., Midoko Iponga, D., Adet, L., Derroire, G., and Besson, N. and the Canobs team: Bridging Scales in Tropical Forest Monitoring with UAV Observatories Integrating Ecosystem Function, Biodiversity, and Satellite Data, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-13150, https://doi.org/10.5194/egusphere-egu26-13150, 2026.
Fires have long played an important role in social-ecological and agricultural systems across the tropics, but until the early 1990s, these fires rarely posed a threat to surrounding forests. Since then, the size, intensity, and frequency of fires in tropical forests in the Amazon and South East Asia have increased, particularly during periods of drought. This increase in fire activity is fueled by a combination of changing climate conditions and land-use practices and poses a significant threat to biodiversity and carbon storage. In the Congo Basin, home to Earth’s second-largest tropical forest and the largest tropical peatland complex, fire activity has increased in recent decades according to the satellite record. However, current fire regimes and drivers, as well as the long-term response of Congo Basin forests to fire, remain poorly understood. This is in part due to limitations the of current satellite-based datasets to detect fire in the region. We 1) synthesize what is known about fire in the Congo Basin, 2) examine trends in existing remotely sensing fire datasets, 3) discuss difficulties detecting fire in the Congo Basin to highlight why current methods are likely under-detecting fire, 4) explore possible social-ecological drivers of fire by highlighting changes in forest disturbances and climate, and 5) report research needs to advance understanding of changing fire dynamics in the Congo Basin. This work highlights a key knowledge gap and provides a roadmap for improving the ability to detect and monitor fire in the Congo Basin, to improve understanding of tropical carbon flux dynamics, and support local fire adaptation and management plans for communities across the Congo Basin.
How to cite: Stouter, H., Morton, D., Brando, P., Coffield, S., Egoh, B., Li, Y., Mané, L., Sonwa, D., Zouh, I., and Ordway, E.: Assessing the risk of a fire-driven tipping point in the Congo Basin , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-4631, https://doi.org/10.5194/egusphere-egu26-4631, 2026.
With their dense tropical forests and vast woody savannas, African ecosystems play a crucial role in global carbon regulation. However, recent findings show their role is shifting from carbon sink to source. Assessing changes in aboveground biomass (AGB) requires consistent time series to reflect temporal continuity and epoch comparability. Here, we leverage machine learning and satellite time series data to 1) produce temporally consistent annual maps of AGB in Africa for the period 2015-2021 and 2) provide a spatially and temporally detailed assessment AGB changes and their drivers at a resolution relevant for land management (100 m). Our retrieval algorithm estimates AGB using a combination of metrics that reflects canopy structure and cover fraction from spaceborne Synthetic Aperture Radar (SAR), Light Detection and Ranging (LiDAR) and optical data, as well as additional covariates influencing tree size and biomass (woody plant functional traits and prevailing moisture and topographic conditions). Trend analysis and breakpoint change detection from LandTrendR are used to evaluate overall AGB changes and to differentiate periods of gradual (e.g. natural growth, degradation) versus abrupt changes (e.g. disturbances, deforestation, replanting). Drivers of the changes are further inferred from temporally aligned land cover dynamics, alongside other ancillary data reflecting vegetation alterations (e.g. fire occurrence, management). AGB changes, quantified in terms of direction, magnitude, rate, and duration are finally summarized across multiple stratification levels (i.e. by driver, biome, country, etc.) to estimate carbon gains and losses. The results provide observation-based carbon stock trajectories over time, which are useful and timely to inform policy decisions on forest restoration and climate mitigation and support Measurement, Reporting and Verification (MRV) frameworks for REDD+, the new Tropical Forests Forever Facility (TFFF) and other policy instruments.
How to cite: Acil, N., Rodriguez-Veiga, P., M. Ryan, C., J. Mograbi, P., Hancock, S., Quegan, S., Disney, M., Benitez, L., Chalo, D., G. Dexter, K., Feig, G., Godlee, J., Gotore, T., Masinde, C., McNicol, I., Muledi, J., Ramaswiela, T., Van der Merwe, H., Mogonong, B., and Balzter, H. and the SEOSAW Partners / data providers: Tracking Aboveground Biomass Dynamics in Africa: Evidence for a Changing Role in Carbon Cycling, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-21337, https://doi.org/10.5194/egusphere-egu26-21337, 2026.
Dense humid forests in Central Africa hold vast biomass carbon stocks but face increasing pressure from logging, clearings, and fire disturbances. Many of these disturbance events have a small spatial scale and are missed by current satellite observations. To address this knowledge gap we generated annual 10 meter resolution maps of canopy height from 2019 to 2022 using a deep learning model fusing spaceborne lidar height measurements with Sentinel radar and optical imagery. The maps reveal widespread previously undetected small disturbance patches, with 87% of all the disturbances being of less than 1 ha in size. These disturbance patches smaller than 1 ha account for 48% of the biomass carbon gains in regrowing forests and for 37% of carbon losses from deforestation. We found a net carbon loss of −58 ± 9 Tg C yr−1 composed of a gross loss of −126 ± 7 Tg C yr−1 partially offset by a gain of 68 ± 6 Tg C yr−1, implying a turnover of biomass carbon from disturbances of 0.52% per year. The Democratic Republic of the Congo is a small net source of carbon (−46 ± 6 Tg C yr−1) due to degradation, despite having the largest carbon gains in young secondary forests. Across the Congo Basin, protected areas show a net biomass loss of −4 ± 1 Tg C yr−1, with a gross loss of −14 ± 1 Tg C yr−1, highlighting uneven conservation outcomes. Our remote-sensing data aggregated into national carbon budgets align well with country-level inventories and bookkeeping model estimates, paving the way for spatially explicit and transparent carbon monitoring.
How to cite: Wan, L., Ciais, P., de Truchis, A., Xu, Y., Brandt, M., Chave, J., Bourgoin, C., Wigneron, J.-P., Bastin, J.-F., Li, W., Ryu, Y., Liu, S., Purnell, D., Fayad, I., Sagang, L. B., Vander Linden, A., Besisa, T., and Ploton, P.: Net carbon losses in Central African forests revealed by high-resolution biomass change maps, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-7498, https://doi.org/10.5194/egusphere-egu26-7498, 2026.
Understanding how African tropical forests respond to climate and land use change requires dense, reliable time series of vegetation dynamics, yet persistent cloud cover and atmospheric variability strongly limit existing satellite products over the Congo Basin. As a result, key aspects of forest phenology, seasonality, and short term variability remain poorly resolved, constraining our ability to detect early signs of functional change or destabilisation.
Here, we develop and evaluate a high quality vegetation time series for Central African forests by combining hyper temporal resolution observations from the MSG SEVIRI geostationary sensor with high spatial resolution Sentinel 2. At a later stage, these results will also be compared with drone data and ground surveys from the CoForFunc international project. To ensure that observed variability reflects changes in vegetation rather than atmospheric fluctuations, we implement a dedicated atmospheric correction for the SEVIRI data adapted from recent developments for Himawari geostationary satellites and further optimised for humid tropical forest conditions. The near continuous sampling of SEVIRI is exploited to reduce cloud related artefacts and improve temporal consistency, while Sentinel 2 and drone observations provide spatial detail and validation at finer scales. The study establishes a robust observational baseline of forest canopy dynamics against which future climate and land use impacts can be more reliably assessed.
Initial results indicate that the combined dataset captures vegetation dynamics and seasonal transitions more consistently than commonly used products such as MODIS, revealing phenological patterns that are otherwise obscured by cloud contamination and atmospheric noise. By improving the accuracy of functional signals in one of the world’s most data limited tropical regions, this work provides a critical foundation for assessing carbon dynamics, ecosystem resilience, and potential tipping behaviour in African tropical forests under ongoing environmental change.
How to cite: Vermeulen, L. M., Li, W., Ichii, K., Ploton, P., Barbier, N., and Duveiller, G.: Clearing the air: atmospherically corrected hyper temporal observations reveal Central African forest dynamics, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-9146, https://doi.org/10.5194/egusphere-egu26-9146, 2026.
Wood density stands as one of the most integrative functional traits of trees, reflecting fundamental trade-offs in adaptive strategies and closely connected to ecosystem history and dynamics. Wood density is linked to multiple key processes including carbon accumulation and mechanical support, determining how plants allocate resources between growth and survival. Yet, despite its recognized importance for understanding forest ecology and improving carbon stock estimates, spatially explicit knowledge of local wood density variation remains severely limited, particularly in tropical rainforests.
Community wood density integrates ecological processes operating at the stand scale, primarily through species composition. Regeneration guild strategies drive much of this variation: fast-growing pioneer species exhibit low wood density while shade-tolerant species show substantially higher values. These compositional differences are also expressed through distinct forest structure patterns —early succession stages with pioneer-dominated composition generally develop lower, more uniform canopies, whereas mature forests with non-pioneer light demanding and shade-tolerant species build taller, more complex vertical architectures. We hypothesized that vertical canopy opening profiles, capturing the proportion of gaps at successive height aboveground, contain the ecological signatures of floristic composition and successional stages that ultimately determine community wood density.
These ecological relationships create opportunities to leverage high-resolution airborne LiDAR for detecting wood density variation at large scale through canopy structure. To test these hypotheses, we modeled community wood density at the stand level across 76 one-hectare plots in 18 Central African forests. We derived canopy stratification metrics from cumulative gap proportion curves extracted from Canopy Height Models, characterizing vertical opening patterns.
We show that canopy opening profiles effectively capture structural signatures associated with community wood density variation. Canopy openness at approximately 20 m and overall canopy stratification emerged as the strongest predictors. Variance partitioning and structural equation modelling reveal that this structure-wood density relationship is entirely mediated by floristic composition and successional stage, which jointly determine both forest structure and wood density. Canopy structure thus acts as a proxy for species composition.
These findings have direct implications for remote sensing applications in tropical forests such as Central African ones. The strong covariation between vertical stratification and species assemblages opens a pathway to account for local wood density variation when mapping AGB through LiDAR-derived indicators. This approach could substantially reduce uncertainties in carbon stock estimates and improve our technique for monitoring forest degradation and successional dynamics across this critical biome.
How to cite: Vander Linden, A., Bastin, J.-F., and Saatchi, S.: Airborne LiDAR Reveals the Wood Density and the Ecological Succession of Central African Forests through Canopy Gaps, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-11112, https://doi.org/10.5194/egusphere-egu26-11112, 2026.
Small-scale shifting cultivation is the main cause of disturbance in African tropical forests. As a consequence, being able to monitor and precisely quantify deforestation and secondary forest regrowth remains a challenge compared to large scale deforestation processes observed in South American and South-East Asian forests. Remote sensing data has been widely used to identify spatio-temporal variability in forest regrowth. However current approaches primarily rely on optical imagery, which is known to be subject to multiple limitations (e.g.cloud cover) in tropical area . The Synthetic Aperture Radar (SAR) is a promising way for overcoming these limitations. In this study, we developed an approach based on SAR signal (Sentinel-1 and PALSAR-2) to monitor forest regrowth. Our approach is based on a recent change detection technique relying on the cumulated sum of the signal anomalies (CuSum algorithm) that has been developed for detecting deforestation. Here, we show that this method is also able to monitor, not only forest regrowth, but also various land use dynamics and land use changes. Our approach was tested on a small area, east of Kisangani in the Democratic Republic of the Congo. We quantified the number of changes that could be attributed to increased vegetation, for which we compared plots occupied by different vegetations and transition types: 'Agroforestry', 'Cropland' and ' Forest Regrowth'. We showed that each vegetation type can be defined by very specific signal change. These preliminary results suggest that the CuSum method applied on SAR data is promising for monitoring land-use dynamics at a small spatial scale, and specifically for identifying secondary forest regrowth.
How to cite: Renaudineau, S., Ygorra, B., Sollier, V., Bauters, M., Lokana Mande, J., Motte, S., Alebadwa, S., Hubau, W., Van de Velde, V., Wigneron, J.-P., Peaucelle, M., and Frappart, F.: Monitoring forest regrowth using SAR images: The Cusum approach, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-11997, https://doi.org/10.5194/egusphere-egu26-11997, 2026.
Forests and savannas frequently coexist as patches within tropical landscapes, yet the mechanisms controlling their spatial configuration remain unclear. The presence of both vegetation states under similar climatic conditions is often attributed to fire–vegetation feedbacks, but could also reflect variation in overlooked external drivers. In Central Africa, forest–savanna coexistence becomes more common with increasing topographic roughness, but how well topographic heterogeneity explains the forest–savanna configuration within coexistence landscapes is unknown.
Here we address this question and examine the role of individual topographic variables that may influence tree cover by, for instance, changing water availability and fire spread. We identify coexistence landscapes from remotely sensed tree cover data and derive topographic variables from a digital elevation model. We use these variables to develop machine learning algorithms predicting vegetation state within coexistence landscapes.
Models achieved an average prediction accuracy of 0.75, with local elevation (relative to the surrounding 500 m or 5000 m) emerging as the strongest predictor of vegetation state. Both model accuracy and the role of topographic predictors varied strongly among landscapes, reflecting the diverse pathways by which topography can influence tree cover. This highlights the need to consider local context when analysing the distribution and stability of tropical forest and savanna ecosystems. Overall, our findings indicate that topographic heterogeneity is a major driver of forest–savanna coexistence in Central Africa. Coexistence landscapes are more deterministic than previously assumed, suggesting that their response to disturbances and climate change will be spatially heterogeneous, thereby reducing the likelihood of abrupt large-scale shifts between forest and savanna states.
How to cite: Zwaan, A., Staal, A., te Beest, M., and Rietkerk, M.: Topography is a major determinant of forest–savanna distributions in coexistence landscapes in Central Africa, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-9821, https://doi.org/10.5194/egusphere-egu26-9821, 2026.
Posters on site: Mon, 4 May, 10:45–12:30 | Hall X1
The Maasai Mau Forest is part of Kenya’s Mau Forest Complex, which is one of East Africa’s most vital tropical water towers. Over the recent decades, it has experienced significant disturbances from encroachment, unauthorised livestock grazing, and illegal abstraction of forest resources. These pressures have significantly altered the structural composition of forests, reduced aboveground biomass, and weakened critical ecosystem services. Currently, countries have engaged in national initiatives aimed at forest ecosystem restoration, including AFR100, REDD+, and SDG15, and Kenya has an ambitious objective to plant 15 billion trees by 2032. These offer a significant opportunity to assess the large-scale recovery of tropical forests. However, monitoring efforts remain limited by weak indicators, a lack of baseline data, and inconsistent reporting systems.
This study examines disturbances and recovery in restored areas of the Maasai Mau Forest using an integrated remote sensing and machine-learning approach. The analysis focuses on assessing vegetation growth, area restored in hectares and carbon sequestered during the process using Sentinel-1 radar, Sentinel-2 optical imagery, GEDI lidar, and ground measurements. They are used to quantify spatial, temporal, and structural (3D) changes in vegetation following disturbance and during recovery.
Data from 2019 to 2025 were processed to develop fused satellite products for the region of interest. Sentinel-1 (VV, VH) was corrected and speckle-filtered using the refined Lee method, while Sentinel-2 imagery was cloud-masked and reduced to relevant spectral bands. From these datasets, radar-based indices, VV/VH ratios, and optical vegetation indices (NDVI, EVI, SAVI, PSSRa) were derived. The indices and selected bands were fused, and principal component analysis (PCA) was performed to generate harmonized inputs for classification.
K-means clustering was applied to the PCA outputs and subsequently labelled as forest and non-forest classes. NDVI was also used to derive annual indices and assess time-series trends. Comparing the classified outputs over time enabled a change detection of forest loss and gain. NDVI-based thresholds and temporal metrics were combined with classified outputs to identify restored areas and map vegetation recovery trajectories.
The results show a clear pattern of forest regeneration. NDVI analysis and satellite-based classification indicate a stable increase in forest cover and a decline in non-forest areas between 2019 and 2025. Dense vegetation increased after 2023, while moderate vegetation declined and sparse vegetation remained relatively stable, with a trend of y = 2336.2x + 30453 (R² = 0.510). PCA-based classification shows forest cover increasing from 32,424 ha to 36,791 ha, while non-forest areas decreased from 13,751 ha to 9,385 ha. Linear trend analysis supports this positive trajectory (forest: y = 1202.7x + 31066, R² = 0.643; non-forest: y = –1202.7x + 15110, R² = 0.643), suggesting a progressive transition from non-forest to forested conditions.
This research shows how tropical forests regenerate after disturbance and enhances understanding of vegetation response to structured efforts. The findings offer valuable evidence for policymakers, conservation planners, and climate practitioners aiming to strengthen restoration outcomes across tropical landscapes.
How to cite: Mutwiri, F. and Vitti, A.: Leveraging Geospatial Techniques to Monitor Restoration Efforts and Assess Associated Forest Ecosystem Services: Case Study of Maasai Mau Forest in Kenya, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-525, https://doi.org/10.5194/egusphere-egu26-525, 2026.
Fires are very important proponents of disturbance within an ecosystem. The direct impacts of fires on the forests cause mortality in biodiversity such as birds, reptiles, and other organisms which may use the forests as a means of survival and livelihood. However, critical information on the level of forest fire risk in most forest ecosystem in Nigeria is scares. However, this study was designed to determine the susceptibility of Ise Forest Reserve to fire hazard.
Climatic variables such as temperature (°C), relative humidity (%), and precipitation (mm/day) for
1994, 2004, 2014, and 2024. were assessed using data obtained from NASA Power. Vegetation and moisture status was spatially obtained from Vegetation indices: Normalized Difference Vegetation Index (NDVI), Normalized Difference Moisture Index (NDMI). Landsat satellite imagery (1994–2024) to assess trends. Fire risk maps were developed, after observing the fuel load throughout the area. Areas of risk were identified and classified into high risk, moderate risk, and low risk areas.
Temperature increased steadily from 24.59 °C in 1994 to 25.34 °C in 2024. The precipitation of the study area had an overall decrease from 5.08mm/day in 1994 to 3.61mm/day in 2024. There was a slight increase in relative humidity from 67.6 in 1994 to 68.4% in 2024 which was an overall 1.26% change within the period. Further variations were observed between the years. NDVI, and NDWI revealed moderate to high tree vigour with the reserve occasionally experiencing regions with robust vegetative growth (SAVI > 0.5). However, the existence of low-SAVI patches highlights persistent issues with changing land use, degraded soil, or climate stresses.
The fuel load index showed that the reserve is largely low risk of fire ranging from 8 to 0. However, the reserve may generally be of moderate risk by 2034 since the area of high risk has increased.
This study highlights the increasing degradation on the fringe areas of the forest which may have adverse effects on the conserved Pan African Chimpanzee, which is conserved within the reseerve.
Keywords: Fire risk, fuel load, fire prediction, climate variations Ise forest reserve
How to cite: Akintunde-Alo, D. A., Ade-Onojobi, T. T., and Alegbeleye, O. M.: Fire Risk Assessment of Ise Forest Reserve, Ekiti State, Nigeria, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-642, https://doi.org/10.5194/egusphere-egu26-642, 2026.
The recently adopted EU Deforestation Regulation (EUDR) requires fine-scale verification of deforestation and forest degradation associated with timber supply chains. In Central Africa, however, region-specific data remain scarce, limiting the ability to distinguish the impacts of different logging practices and to differentiate regulatory definitions of degradation from actual ecological outcomes. This study addresses this gap by proposing a drone-based methodology to assess post-harvest impacts at the tree level. Focusing on FSC-certified selective logging, which typically removes only one to two trees per hectare, we aim to establish direct relationships between harvested tree characteristics and resulting canopy openings. Using high-resolution UAV imagery acquired before and after logging, combined with forest inventory data, canopy gaps are delineated and linked to individual trees. Gap size is analyzed in relation to species identity, tree diameter, felling conditions, forest type, topography, and crown-related traits derived from allometric equations. By identifying the key drivers of canopy opening at the individual level, this approach seeks to provide operational tools to better characterize forest degradation and support more robust monitoring frameworks under the EUDR.
How to cite: Rousseau, Z., Bastin, J.-F., and Doucet, J.-L.: Quantifying selective logging impacts in Central African forests using UAV imagery: from individual trees to canopy gaps, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-2587, https://doi.org/10.5194/egusphere-egu26-2587, 2026.
Deforestation-driven forest loss substantially alters the global carbon budget and degrades ecosystem services, while subsequent forest regrowth is critical for ecosystem recovery and carbon sequestration. However, comprehensive datasets explicitly characterizing post-deforestation forest regrowth remain lacking. Here, we integrate multiple remote sensing products to develop the first spatially explicit dataset quantifying forest structural regrowth following deforestation across globally important deforestation regions at 30 m resolution. The dataset characterizes regrowth dynamics of forest height, aboveground biomass (AGB), leaf area index (LAI), and the fraction of photosynthetically active radiation (FPAR). For each structural attribute, regrowth ratios and rates are provided at 5-year intervals, primarily spanning 1985–2020. This dataset enables a detailed assessment of post-deforestation forest regrowth across spatial, temporal, and structural dimensions, supporting improved quantification of forest carbon budgets and enhanced evaluation of forest ecosystem services.
How to cite: Zang, J.: Forest regrowth dataset for globally key deforestation regions, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-7220, https://doi.org/10.5194/egusphere-egu26-7220, 2026.
Multi-AXis Differential Optical Absorption Spectroscopy (MAX-DOAS) uses trace gas absorptions in spectra of scattered sun light recorded at different elevation angles to retrieve vertical profiles of trace gas concentrations and aerosol extinctions in the lower troposphere as well as their total tropospheric vertical column densities. A major advantage of MAX-DOAS is the possibility to observe multiple trace gases, such as formaldehyde and glyoxal, simultaneously for the same air mass. We operate two MAX-DOAS instruments at the Amazon Tall Tower Observatory (ATTO) at altitudes of 80 m (since 2017) and 298 m (since 2019) above ground. Besides the full profile retrievals for both instruments, this setup allows the determination of (small-scale) vertical gradients of trace gas abundances in the altitude range between both instruments (ca. 200 m) by directly comparing their measurements.
Located in a pristine rain forest region in the central Amazon Basin about 150 km north-east of Manaus, the ATTO site offers a unique opportunity to study the chemical processing of tropospheric trace gases far from major anthropogenic emission sources. Further, the site hosts long-term and campaign-based measurements of a large variety of different atmospheric constituents and parameters. Combining these measurements allows investigating chemical processes at the canopy-atmosphere-interface and directly above it. Comparisons with model data yield further insights, e.g. the identification of processes that are not (fully) represented by the simulations or the confirmation of surprising observational results.
In this study, the MAX-DOAS results of formaldehyde and glyoxal are compared to measurements of their major precursor substances, i.e. isoprene and monoterpenes. This includes assessments of their respective seasonal and diel variations as well as their (small-scale) vertical gradients. For selected time periods, the results of these atmospheric measurements are also compared to model simulations performed with WRF-Chem, using the MOZART-4 chemical mechanism, in order to investigate whether the characteristic variations and vertical gradients found for the measurement data are also reflected in model simulations.
How to cite: Donner, S., Lauster, B., Ziegler, S., Artaxo, P., Beirle, S., Edtbauer, A., Kuhn, L., Machado, L. A. T., Pozzer, A., Ringsdorf, A., Williams, J., and Wagner, T.: Linking MAX-DOAS measurements of formaldehyde and glyoxal to precursor substances at the canopy-atmosphere-interface at ATTO, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-11562, https://doi.org/10.5194/egusphere-egu26-11562, 2026.
The tropical forests of Central Africa, storing approximately 10% of the world’s terrestrial carbon play a vital role in the global carbon cycle. These forests also constitute a hotspot of biodiversity and a source of livelihood and ecosystem services for local communities. The need to monitor forest loss and gain in the region has led to the development of a variety of forest maps through the use of orbital sensors in recent years.
In this study, we map the extent of the tropical moist forest of Central Africa at 10m resolution using an advanced Sentienel-2 processing technique. Initial calibration is performed with existing dataset (GFW, TMF) before refining with VHR data. Sentinel-2 data are processed into spatially coherent cloud-free annual composites. Classification into a forest/non forest map is then performed with XGBoost and the addition of ancillary variables, yielding yearly maps of the forest’s extent. This allows forest extent to be monitored with unprecedented resolution, which is crucial for the Congo Basin complex landscape, dominated by small-scale agriculture.
How to cite: Delecluse, S., De Maet, T., and Defourny, P.: Mapping of the forest extent in the Congo Basin at 10m-resolution with Sentinel-2, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-19399, https://doi.org/10.5194/egusphere-egu26-19399, 2026.
Village relocations from protected areas provide natural experiments for understanding forest recovery dynamics. We investigated vegetation recovery across 14 villages relocated from Panna Tiger Reserve, India, spanning 8-30 years post-relocation, to test whether recovery follows classical successional trajectories or diverges into alternative stable states. Using 198 plots, we surveyed trees, seedlings, saplings, shrubs, herbs, grasses, and invasive species, and compared communities with adjacent buffer forest. Contrary to succession theory predictions, time since relocation had no significant effect on woody vegetation recovery (p = 0.607). Instead, landscape position determined recovery trajectory: all riverine villages (n = 7) achieved forest state while all interior villages (n = 7) remained grassland, regardless of time elapsed. Riverine sites supported 4.7 times higher seedling abundance and 2.5 times higher sapling abundance than interior sites. Beta diversity partitioning revealed turnover-dominated differentiation (>92%) rather than nestedness, indicating species replacement between states rather than progressive accumulation. NMDS ordination showed discrete forest-grassland clusters, and indicator species analysis identified state-specific assemblages: fire-adapted Themeda-Heteropogon grasses dominated grassland while shade-tolerant Dichanthium-Oplismenus characterized forest state. Critically, seedling-to-tree ratios were identical between states (10.1 vs 10.3), demonstrating that recruitment limitation occurs post-germination rather than at seed dispersal. Invasive species declined autonomously (10.4%/year, p = 0.008), suggesting competitive exclusion by native grasses. These findings demonstrate that grass-fire feedbacks maintain alternative stable states, with landscape position determining initial trajectory. Passive restoration is insufficient for interior sites; active intervention breaking grass-fire feedbacks is required. Village relocation alone does not guarantee forest recovery as outcome depends fundamentally on landscape context.
How to cite: Singh, A., Krishnamurthy, R., and Page, N.: Vegetation recovery following disturbance removal revealed forest-grassland alternative stable states in a tropical dry forest of central India, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-20788, https://doi.org/10.5194/egusphere-egu26-20788, 2026.
Tropical forests play a central role in regulating the Earth's climate and the global carbon cycle, yet their accurate representation in terrestrial biosphere models (TBMs) remains a challenge: High species diversity, strong climatic variability, and lack of long term observations lead to high parameter uncertainty and hinder model calibration. Consequently, many questions regarding the long-term carbon storage capacity of tropical forests and especially the Amazon, as well as their vulnerability to extreme events, particularly droughts, remain open.
In this study, we apply the TBM QUINCY (Thum et al., 2019), with a new implementation of plant hydraulics, to simulate seasonal and interannual vegetation dynamics at the Amazon Tall Tower Observatory (ATTO, http://attoproject.org), central Amazon, Brazil. The model is evaluated using eddy covariance data (NEE, GPP, ET) from 2014 to 2023 and complementary observational data including time series of soil water content, sap flow, and dendrometer measurements. This evaluation allows us to assess the representation of soil and plant water dynamics and to identify model limitations.
Specifically, our objective is to identify systematic mismatches between modeled processes and observations, in order to support targeted model development and parameterization. By linking uncertainties in carbon and water fluxes to specific model components and processes, we aim to establish a structured pathway toward improving TBM performance at ATTO, and to better understand the ecosystems sensitivity to drought under future climate change.
How to cite: Eisele, G., Papastefanou, P., van Asperen, H., Botía, S., Dias-Júnior, C., Durgante, F., Horna, V., Rammig, A., Sabot, M., Alves de Souza, C., and Zaehle, S.: Process-Based Evaluation of Tropical Forest Responses to Drought at the Amazon Tall Tower Observatory, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-19253, https://doi.org/10.5194/egusphere-egu26-19253, 2026.
Tropical secondary forests, i.e. regrowing forests after clearcutting through human activities, cover a larger area than primary tropical forests. This increasingly dynamic tropical landscape warrants a comprehensive understanding of carbon (C) sequestration and nutrient cycling in secondary tropical forest succession, particularly on the soil microbiome governing these processes. Yet, our current knowledge of soil microbial dynamics in secondary successions of tropical forests and their role in C turnover and sequestration is limited, particularly in deeper soil layers which are seldomly explored.
Here, we investigated microbial activity and growth along a soil depth gradient in an established space-for-time-substitution experiment in the Yoko forest reserve, Democratic Republic of the Congo. We sampled soils from four differently aged secondary forests (3 years, 8 years, 15 years and 63 years), and a primary forest acting as a control at six depths (10 cm, 30 cm, 50 cm, 100 cm, 175 cm, 250 cm). We assessed microbial growth through the incorporation of deuterium-labelled water into phospholipid fatty acids (2H-SIP), allowing us to distinguish growth and biomass of distinct microbial groups.
Microbial growth was highest in early successions, 3 to 15 years after the last biomass removal, whereas microbial respiration steadily increased with succession resulting in a decrease in soil microbial carbon use efficiency with forest successional age. Primary forests showed significantly lower microbial growth rates than early and middle-aged successions. This trend, while also present in shallow soil depths, was most evident at the depth of 100 cm.
Our results suggest an increase of labile C availability for soil microorganisms in early and middle-aged successions, most likely through higher quality and/or quantity of C inputs of regrowing plant biomass compared to the climax plant community. We further show that the response of microbial activity during secondary succession was seen beyond soil depths that are commonly considered, highlighting the importance of sampling deeper soil layers when assessing responses to land use changes. Contrary to our expectations, microbial growth decreased with successional age, and primary forests fostered lower microbial activity compared to secondary forests. Our results therefore demonstrate a complex response of soil microorganisms to secondary succession in Afrotropical forests that is decoupled from aboveground plant biomass.
How to cite: Ranits, C., Fuchslueger, L., Van de Velde, V., Schmidt, H., Prosser, H., Makelele, I., Ewango, C., Bauters, M., Boeckx, P., and Richter, A.: Microbial activity of a secondary forest succession after shifting agriculture in the Afrotropics, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-18838, https://doi.org/10.5194/egusphere-egu26-18838, 2026.
The Amazon rainforest stores ~150–200 Pg of carbon and plays a central role in the global carbon cycle, yet it is highly vulnerable to land-use change, fire, and climate change, leaving its future carbon balance uncertain. Net Ecosystem Exchange (NEE) quantifies the balance between ecosystem carbon uptake and release, but long-term NEE assessments in Central Amazonia remain scarce due to limited observational coverage from eddy covariance towers and CO₂ profile measurements. The Amazon Tall Tower Observatory (ATTO; https://www.attoproject.org) helps overcome this limitation by providing more than a decade of continuous measurements and by capturing two extreme drought events (2015/2016 and 2023/2024). In this work, we analyzed 11 years of NEE estimates at ATTO (2014-2024) and we propose a methodology for selecting the friction velocity threshold (u*), based on the identification of a plateau in the NEE-u* relationship, where NEE becomes independent of increasing turbulence, and all periods with u below the threshold are filtered due to insufficient turbulence*. We estimate monthly u* thresholds ranging from 0.18 to 0.21 m s⁻¹. We found that, on average, the forest in the ATTO flux footprint generally acted as a net carbon sink. However, in years following severe droughts, such as 2016 and 2024, we detect a temporary reversal, with the ecosystem becoming a CO₂ source during the wet season. We also quantified the effect of environmental drivers modulating NEE across seasons. We find that higher air temperature reduces carbon uptake during the wet season. In contrast, soil moisture shows opposite relationships depending on season: during the dry season, increasing soil moisture (10 cm depth) reduces net carbon uptake, whereas during the wet season, increasing it enhances net carbon uptake. Our findings deliver critical observational evidence to refine model parameterizations of tropical carbon-water interactions and to reduce uncertainty in predictions of the Amazon carbon balance under future climate scenarios.
How to cite: Souza, C., Dias-Júnior, C., Komiya, S., van Asperen, H., Zaehle, S., Santana, R., Cordeiro, L., Rizzo, L., Cely, I. M., de Araújo, A., Quesada, C., Hildebrandt, A., and Botía, S.: Long-term NEE at ATTO (2014–2024): drought legacy effects and seasonal controls on Amazon carbon uptake, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-17350, https://doi.org/10.5194/egusphere-egu26-17350, 2026.
How to cite: Martínez de Velasco, A., Delecluse, S., Meunier, F., Defourny, P., Verbeeck, H., and Bauters, M.: Analysis of Congo Basin rainforest regrowth trajectories by land use history, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-2826, https://doi.org/10.5194/egusphere-egu26-2826, 2026.
The continued degradation of the Amazon rainforest, exacerbated by rising air temperatures and increased hydric stress in recent decades, is altering its capacity to absorb carbon. However, the scarcity of local observations and the difficulty of representing the vast and diverse Amazon biome in numerical models make it challenging to accurately quantify carbon exchanges across the region. Consequently, determining whether the Amazon as a whole currently functions as a carbon source or sink remains a key priority for future research.
In this study, we conducted high-resolution simulations of CO2 concentrations over the Amazon rainforest using the Weather Research and Forecasting Greenhouse Gas (WRF-GHG) model for April and September of selected years through the end of the century. Initial and boundary conditions for
meteorological variables and background CO2 concentrations were derived from projections of the Intergovernmental Panel on Climate Change (IPCC) worst-case climate scenario based on the Coupled Model Intercomparison Project 6 (CMIP6) under the Shared Socioeconomic Pathway 5-8.5 (SSP5-8.5). To ensure consistency, SSP5-8.5 land-use projections from the Land-Use Harmonization version 2 (LUH2) dataset were reclassified to match WRF-GHG land-use categories. All simulations included a 15-day spin-up period, followed by a two-day rolling simulation framework for the target month.
The results indicate that by 2050, CO2 concentrations over the Amazon are projected to reach approximately 550-650 ppm, exceeding the global, Northern Hemisphere, and Southern Hemisphere annual mean concentrations for that year, which are estimated at about 563 ppm, 567 ppm, and 558
ppm, respectively. Notably, the simulations also suggest a slight reduction in Net Ecosystem Exchange (NEE) fluxes between 2030 and 2050.
How to cite: Vara-Vela, A. L., Rojas, N., Botía, S., Rizzo, L., and Toledo Machado, L. A.: Projected High-Resolution CO2 Concentrations over the Amazon Rainforest under the SSP5-8.5 Pathway through the End of the Century, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-2827, https://doi.org/10.5194/egusphere-egu26-2827, 2026.
The net carbon exchange between ecosystems and the atmosphere (Net Ecosystem Exchange, NEE) is determined by the balance between Gross Primary Production (GPP) and Ecosystem Respiration (Reco). However, in tropical South America (TSA), the spatial variability of these factors and the factors that influence them are not well understood, especially in areas affected by forest degradation and secondary forest recovery. Here, we use a new implementation of the Vegetation Photosynthesis and Respiration Model (pyVPRM) to generate hourly, spatially explicit estimates of NEE, GPP, and Reco across TSA. The model is constrained by eddy-covariance measurements from 22 flux towers spanning Amazonian upland and lowland forests, Andean-influenced forests, tropical wetlands, and Orinoco savannas, combined with remotely sensed vegetation indices (EVI), meteorological forcing, and annually varying land-cover maps from MapBiomas. We extend the standard pyVPRM land-cover classification to explicitly represent forest disturbance and recovery states, distinguishing undisturbed, degraded, deforested, and regenerating forests. This allows us to quantify how forest degradation and regrowth alter the magnitude and spatial distribution of gross carbon fluxes compared with simulations that do not distinguish between degradation classes. We expect that resolving these disturbance states will reduce systematic biases in both GPP and Reco over human-modified landscapes and improve the attribution of carbon sources and sinks across the TSA region beyond what is captured by climate forcing alone. By separating disturbance-driven from climate-driven flux variability, this framework provides a more realistic prior for regional atmospheric inverse modelling and a stronger basis for assessing the carbon consequences of tropical forest degradation and recovery.
How to cite: Cárdenas-Vélez, M., Glauch, T., Q. Dias-Junior, C., Souza, C., Komiya, S., van Asperen, H., de Paula Cordeiro, L., Rojas, N., Rizzo, L., Machado, L., Stern, R., Cosio, E., Jimenez, R., Morales, L., Gerbig, C., Trachte, K., and Botía, S.: Improving net carbon flux estimates in Tropical South America by accounting for forest degradation and recovery in a Vegetation Photosynthesis and Respiration Model, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-3087, https://doi.org/10.5194/egusphere-egu26-3087, 2026.
The Amazon rainforest has a large area occupied by rivers and wetlands, responsible for the majority of biogenic methane (CH4) emission in the region. Biogenic emissions of CH4, which is a potent greenhouse gas, are a key source of uncertainty for the global budget. Atmospheric transport models can be useful to assess the spatial distribution of CH4 concentrations and the influence of land use and climate change. It is important to use computational models with updated emission data to obtain more accurate results in atmospheric gas transport simulations. Here we present an analysis performed with the Weather Research and Forecasting model coupled with the Greenhouse Gases module (WRF-GHG), using updated maps of wetland and fast carbon pool to calculate biogenic emissions of CH4 in the Amazon region. Three wetland maps (wetmaps) and two fast carbon pool (CPOOL) maps were used in the simulations. The resulting methane concentrations were compared to in situ observations at the ATTO tower (Amazon Tall Tower Observatory). The simulations were performed from 1st to 13th January 2023 (considering the first seven days as spin up), in a single domain with 6 km resolution and grid of 212 x 121 centered at ATTO. Boundary conditions were provided by ERA5 and CAMS. The simulation using the default emission model along with the minimum inundation wetmap and the updated CPOOL map showed results closer to observational data (bias of 15 ppb) than the other simulations (bias in the range 20-320 ppb). Using the default maps resulted in an overestimation of 4.1% in CH4 concentrations at ATTO. The modeled CH4 concentrations time series showed a pronounced diurnal variability, likely driven by boundary layer dynamics and advection. On the other hand, observations showed rather constant concentrations, suggesting that background regional emissions dominate the CH4 signal at ATTO. Considering a model grid cell over a section of the Amazon River near the ATTO site (150 km southeast), simulated emissions ranged between 42 (minimum inundation and updated CPOOL map) and 623 (maximum inundation and default CPOOL map) mg CH4 m^-2 day^-1, while WetCHARTs emission inventories are in the range 153-276 mg CH4 m^-2 day^-1 and the literature reports averages of 18-21 mg CH4 m^-2 day^-1 for the Amazon River, based on field measurements. Overall, the results show a high sensibility of the WRF-GHG model towards the choice of wetland and CPOOL maps in Amazonia. Also, the correct representation of CH4 background concentrations is key to improve the simulations of near surface concentrations in areas less impacted by local wetland emissions, like ATTO.
How to cite: Martins Portella, B., Varanda Rizzo, L., Rojas Benavente, N., Botía, S., van Asperen, H., Augusto Toledo Machado, L., and Vara-Vela, A.: Modeling methane biogenic emissions using different wetland and soil carbon pool maps in Amazonia, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-13904, https://doi.org/10.5194/egusphere-egu26-13904, 2026.
Tropical forests of Central Africa play a fundamental role in the global carbon cycle, the regional moisture recycling, and also act as biodiversity hotspots. Yet, afrotropical forests experience important anthropogenic pressure, including slash-and-burn agriculture that is particularly widespread in the region. Despite their significance, the post-disturbance recovery dynamics of these forests remain poorly understood, particularly compared to the other tropical regions. With increasing anthropogenic pressures and projected shifts in rainfall regimes in the future, an improved understanding of forest regrowth processes is critical to anticipate the future of the Congo Basin carbon sink and simulate the land carbon sink of secondary forests/
This study presents a comprehensive model intercomparison of forest regrowth trajectories in the Democratic Republic of the Congo, combining ground inventory data with outputs from multiple Dynamic Global Vegetation Models (DGVMs). We compiled a harmonized dataset of field sites, representing dozens of site/age combination, along wide climatic gradients in the country. Multiple DGVMs were benchmarked against empirical regrowth curves derived from plot networks, with additional models currently under evaluation to extend the model ensemble. Each model was forced by consistent climate and land-use datasets but exhibited heterogeneous process representations and carbon allocation schemes.
Results reveal a systematic overestimation of above-ground biomass accumulation across models, particularly during the first decades of succession. Modelled forests typically regained 80–100% of their pre-disturbance biomass within 50 years, whereas inventory data indicate substantially slower recovery rates, often below 60%. Sensitivity analyses showed that the divergence between simulated and observed regrowth trajectories could be linked to differences in parameterization of turnover rates and demography. Furthermore, the influence of climatic drivers varied markedly across models: while some exhibited strong sensitivity to precipitation seasonality, others were dominated by temperature and radiation effects. Such discrepancies highlight structural uncertainties in how models capture key processes controlling regrowth, including recruitment limitation, and resource constraints.
Our findings underscore the need for process-based improvements based on existing field data. By confronting models with empirical data from the Congo Basin, this intercomparison provides an essential step toward reducing uncertainties in projections of African forest resilience under climate and land-use change.
How to cite: Meunier, F., Van de Velde, V., De Hertog, S., Boeckx, P., Bauters, M., Verbruggen, W., Peaucelle, M., and Verbeeck, H.: Forest regrowth in the Democratic Republic of the Congo: A field data/vegetation model comparison, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-14967, https://doi.org/10.5194/egusphere-egu26-14967, 2026.
Posters virtual: Thu, 7 May, 14:00–18:00 | vPoster spot 2
EGU26-12372 | ECS | Posters virtual | VPS6
From native forests to planted tree crops: disentangling tree cover transitions driven by deforestation, fire, and plantation development using satellite observationsThu, 07 May, 15:06–15:09 (CEST) vPoster spot 2
Land-use change in tropical regions has profoundly altered forest structure, disturbance regimes, and recovery pathways over recent decades, with important implications for fire activity and land–atmosphere interactions. Medium-resolution global land cover products are widely used to analyze these dynamics in climate and Earth system studies, yet their capacity to distinguish native forests from planted tree crops across disturbance and recovery phases remains limited. This raises critical questions about how forest loss, recovery, and resilience are inferred from satellite-based observations in human-modified tropical landscapes.
Here, we examine how forest-to-tree-crop transitions are represented in the ESA CCI Land Cover medium-resolution land cover (MRLC) product at 300 m spatial resolution over the period 1992–2022, and how these transitions relate to fire occurrence across Southeast Asia. We combine MRLC land cover maps and land cover transition layers with a high-resolution global dataset of planted tree crops providing spatial extent and year of establishment (Descals et al., 2024), together with fire information from FireCCI v5.1 at 250 m resolution (2001–2022) and fire polygon data from FRY v2.0. This integrated framework allows us to place land-cover changes associated with plantation establishment and maturation in the context of disturbance–recovery processes.
Our analysis focuses on land cover trajectories from native evergreen broadleaf forest to mosaic classes during plantation establishment, followed by reclassification to broadleaf tree cover as plantations mature. We examine the timing and duration of these transitions and their association with fire occurrence during land-use change. Preliminary results show systematic patterns in which oil palm expansion is linked to transient forest loss and elevated fire activity during early plantation stages, followed by reduced fire occurrence as plantations develop before being mapped again as tree cover.
These results demonstrate that confusion between native forests and planted tree crops in medium-resolution land cover products can lead to misleading interpretations of post-disturbance recovery and forest resilience. In particular, apparent forest recovery detected by satellite products may in some cases reflect a land-use replacement rather than true ecosystem recovery with important implications for the interpretation of disturbance–recovery dynamics, as well as for climate modeling and projections in human-modified tropical landscapes. This highlights the need for complementary high-resolution land cover information, such as that developed within the ESA CCI High Resolution Land Cover (HRLC) project, to better disentangle recovery from land-use change.
How to cite: San Martín, R., Ottlé, C., Peylin, P., Lamarche, C., and Mouillot, F.: From native forests to planted tree crops: disentangling tree cover transitions driven by deforestation, fire, and plantation development using satellite observations, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-12372, https://doi.org/10.5194/egusphere-egu26-12372, 2026.
