Stefan Olin

Land-use history has improved the simulation of vegetation and carbon cycling in dynamic global vegetation models since the legacies of past land-use persist in soil carbon and nitrogen pools for several years. However, the influence of land-use and management history on crop yields is not well explored in large-scale modelling. This study assessed the effect of land-use change legacies on global crop yield simulated with LPJ-GUESS for wheat, maize and rice and its interaction with climatic and management drivers. A total of 56 global simulations were performed by combining three factors: (1) four global land-use setups before the crop simulation: cropland, natural vegetation, natural vegetation and pasture, and a historical land-use change reconstruction; (2) the individual contributions of atmospheric CO2, precipitation, radiation, temperature, and fertilisation; and (3) two different climate-forcing datasets, CRU-NCEP and AgMERRA. Simulations with previous land-use of natural vegetation and pastures caused higher soil nitrogen and carbon pools, increasing yields at the beginning of the simulations. Simulations assuming previous cropland cover had similar trends to those using the land-use change reconstruction. All the land-use setups tended to converge over time to give similar yields, but convergence could take several decades. Fertilisation was the main driver impacting the trend and interannual variability of yields. The main results highlight the critical role of historical land-use in simulating crop yield at subregional to local scales, particularly in locations with low fertilisation. At global to regional scales, the assumption of previous cropland cover simplifies simulations without affecting accuracy significantly.

Wood density (WD) is a crucial anatomical trait influencing forest carbon storage. However, dynamic global vegetation models (DGVMs) typically assume a fixed species-level WD, neglecting environment-driven variability. In this proof-of-concept study, we explore the potential impact of dynamic WD on tree- and forest-level carbon storage by integrating a simple temperature-response function of WD into the DGVM LPJ-GUESS from Smith et al., 2014. Simulations along a temperature gradient show that incorporating environmentally responsive WD can substantially alter simulated stand structure and carbon stocks. Overall, our model experiments illustrated that sites with higher WD had more, but smaller trees, which stored less carbon compared to the standard model. The strongest effects were predicted to appear before canopy closure, where per-tree carbon deviated by up to 32%. This exploratory study suggests the need to represent a mechanism for dynamic WD to better assess ecological feedbacks to forest carbon storage predictions, particularly in young and regenerating forests.

China plans expanding 49.5 million hectares of new forests by 2050 to strengthen carbon sequestration. However, estimates of the carbon benefits from this expansion rarely consider the effect of ‘forest edge’, where tree mortality increases under intensified stress from wind, drought, pests, and fire. Here we show that proximity to forest edges substantially reduces biomass carbon storage, and develop a spatial optimization strategy that prioritizes planting in areas that minimize edge effects. Our projections show that forestation optimized for edge effects results in a 51% increase in carbon gain (986 ± 22 Tg by 2060), with approximately half of the total gain driven by reduced edge effects. These findings demonstrate that ignoring edge effects can significantly overestimate carbon sink potential and highlight spatially optimized forestation as a pathway to maximize climate mitigation and ecological benefits.

Summary: Vegetation demographic models (VDMs) are advanced tools for simulating forest responses to climate and land‐use changes, and are essential for projecting carbon cycling and large‐scale forest management strategies. Despite their increasing incorporation into Earth System Models, VDMs differ in their demographic assumptions, with no prior quantitative comparison of their performance. We benchmarked nine VDMs against observational data from boreal, temperate and tropical sites, assessing their accuracy in predicting tree growth, carbon turnover, biomass stocks and size distributions. Models were simulated under consistent climate conditions with postdisturbance recovery monitored for at least 420 yr. Postdisturbance carbon recovery trajectories showed significant variability while remaining within observational ranges. Initial regrowth rates varied substantially (0.03–0.60, 0.18–0.70 and 0.35–1.10 kgCm ⁻² yr ⁻¹ for boreal, temperate and tropical sites, respectively), influenced by each model's initial forest state. Models captured mature forest carbon content but showed compensating effects between overestimated growth and underestimated mortality rates. This first multi‐model benchmarking identifies growth and mortality rates as critical calibration targets and highlights the need to refine postdisturbance establishment conditions for model development. We outline specific benchmarking variables needed to improve predictions of forest responses to environmental change.

Climate change calls for adaptive strategies to manage land system across governance levels, as differing multi‐level policies distinctly shape land system and long‐term ecosystem resilience. This study proposes an iterative approach for optimizing land‐use pathways that balance competing policy objectives across national, provincial, and local levels without compromising ecosystem integrity in a changing climate. This approach was applied to the Huangshui River Basin on China's Qinghai‐Tibet Plateau, a region facing significant challenges from climate change and human activities. We integrated the land‐use change model CLUMondo with the dynamic vegetation model LPJ‐GUESS to compare our sustainable development pathway against scenarios based on plans prioritizing national, provincial, and local governance objectives. The analysis revealed considerable mismatches in management goals across governance levels within the Huangshui River Basin, emphasizing the necessity of multi‐scale coordination to align planning objectives for achieving desired goals. This study presents an optimization framework to quantitatively evaluate trade‐offs and balance between sustainability objectives and ecosystem integrity in response to system feedbacks, offering critical insights into reconciling potentially conflicting sustainability goals across multiple scales within socio‐ecological systems.

Multiple cropping, the simultaneous cultivation of several crops in space or time, is a global practice essential for intensifying and diversifying agriculture. Despite its substantial impact on environmental and socioeconomic outcomes of farming, multiple cropping is hardly accounted for in assessments of global food production, sustainability, and climate impacts. Such studies, often relying on modelling of cropping systems, land use change, and eventually the Earth system, are of growing importance in decision-making and policymaking. However, they primarily assume monocropping, neglecting carryover effects between crops and their implications for land use. This limitation compromises the representativeness of these studies and the conclusions they draw, essentially overlooking a substantial option space for sustainable intensification, nature-based solutions, and resulting land-atmosphere feedback. Herein, we outline the relevance of multiple cropping, reflect on its consideration in land-use models, and identify development requirements to enhance their inclusion in informing policymaking for sustainable food systems.

Nitrogen (N) transformation processes by soil microbes account for significant nitrous oxide (N₂O) emissions from natural ecosystems and cropland. However, understanding and quantifying global soil N₂O emissions and their responses to changing environmental conditions remain challenging. Here, we implemented a soil nitrification–denitrification module into the dynamic vegetation model LPJ-GUESS to estimate N₂O emissions from global lands. The performance of this new development is examined using observed N₂O fluxes from natural-soil and cropland field trials and independent global-scale estimates. LPJ-GUESS broadly reproduces the cumulative N₂O emissions under different climate conditions and N fertilizer applications that are observed in the field experiments, with some deviations in emission seasonality. Globally, simulated soil N₂O emissions from terrestrial ecosystems increase from 5.6±0.2 Tg N yr⁻¹ in the 1960s to 9.9±0.3 Tg N yr⁻¹ in the 2010s, with croplands contributing about two-thirds of the total increase. East Asia and South Asia show the fastest growth rates in N₂O emissions over the study period due to the expansion of fertilized croplands. On a global scale, N fertilization (including synthetic fertilizer and manure use), atmospheric N deposition, and climate change contribute 58 %, 46 %, and 24 %, respectively, to the simulated soil N₂O emissions in the 2010s. Rising CO₂ levels in the atmosphere reduce the simulated emissions by 32 % through increased plant N uptake, whereas land use changes have varied spatial effects on emissions depending on N management intensity after land cover conversion. Our estimates only account for the direct soil N₂O emissions, excluding those from fertilized pastures. This study highlights the importance of environmental factors in influencing global soil N₂O emissions, particularly for assessing greenhouse gas mitigation potential in agricultural ecosystems.

Tropical forests provide essential ecosystem services, playing a critical role in climate regulation, biodiversity conservation, and regional hydrological cycles while also supporting livelihoods. However, they are increasingly threatened by deforestation and land-use change. Accurate land cover (LC) mapping is vital to monitor these changes, but mapping tropical forests is challenging due to complex spatial patterns, spectral similarities, and frequent cloud cover. This study aims to improve LC classification accuracy in such a heterogeneous tropical forest region in Southeast Asia, namely Kulen, Cambodia, which is characterized by natural forests, regrowth forests, and agricultural lands including cashew plantations and croplands, using Sentinel-2 imagery, recursive feature elimination (RFE), and Random Forest. We generated 65 variables of spectral bands, indices, bi-seasonal differences, and topographic data from Sentinel-2 Level-2A and Shuttle Radar Topography Mission datasets. These variables were extracted from 1000 random points per 12 LC classes from reference polygons based on observed GPS points, Uncrewed Aerial Vehicle imagery, and high-resolution satellite data. The random forest models were optimized through correlation-based filtering and recursive feature elimination with hyperparameter tuning to improve classification accuracy, validated via confusion matrices and comparisons with global and national-scale products. Our results highlight the significant role of topographic variables such as elevation and slope, along with red-edge spectral bands and spectral indices related to tillage, leaf water content, greenness, chlorophyll, and tasseled cap transformation for tropical land cover mapping. The integration of bi-seasonal datasets improved classification accuracy, particularly for challenging classes like semi-evergreen and deciduous forests. Furthermore, correlation-based filtering and recursive feature elimination reduced the variable set from 65 to 19, improving model efficiency without sacrificing accuracy. Combining these variable selection methods with hyperparameter tuning optimized the classification, providing a more reliable LC product that outperforms existing LC products and proves valuable for deforestation monitoring, forest management, biodiversity conservation, and land use studies.

Droughts present a significant global challenge, particularly to forest ecosystems in regions such as eastern New South Wales, Australia, which is known for its dry climate and frequent, intense droughts. Recent studies have indicated a notable increase in tree mortality and canopy browning across this area, especially during the recent extreme drought period culminating in the Black Summer of 2019–2020. Our study investigates the impacts of drought on eucalypt forests by leveraging remote sensing and field observation data to detect and analyse vegetation health and stress indicators. Utilising data from Sentinel-2, alongside historical Landsat observations, we applied multiple spectral vegetation indices, namely the Normalized Difference Vegetation Index (NDVI), Normalized Difference Moisture Index (NDMI), Normalized Burn Ratio (NBR), and Tasseled Cap Transformation, to assess the extent of drought impacts. We found NBR to show the most consistent agreement with ground-based observations of drought-related tree mortality. Additionally, by integrating ground-based data from the “Dead Tree Detective” citizen science project, we were able to validate the remote sensing outcomes with a 90.22% consistency, providing confirmation of the extensive spatial distribution and severity of the inferred impacts. Our findings reveal that 13.16% of eucalypt forests and woodlands across eastern New South Wales experienced severe stress associated with drought during the 2019–2020 Black Summer drought. This study demonstrates the utility of satellite-derived drought indicators in monitoring forest health and highlights the necessity for continuous monitoring and research to understand the factors that trigger tree vitality loss.

Climate change calls for adaptive strategies to manage landscapes across governance levels, as differing policy frameworks distinctly shape landscape patterns and long-term ecosystem resilience. This study presents a novel adaptive framework that integrates planning objectives, landscape patterns, and ecosystem resilience through iterative feedback loops. The framework aims to achieve the desired scenario, SUS-RES, which balances sustainable planning objectives across national, provincial, and local levels while maintaining ecosystem resilience in a changing climate. The framework was applied to the Huangshui River Basin on China’s Qinghai-Tibet Plateau, a region facing significant challenges from climate change and human activities. By integrating the CLUMondo and LPJ-GUESS models, we compared SUS-RES with scenarios focusing on national (NAT), provincial (PRO), and local (LOC) governance priorities. The analysis revealed considerable mismatches in management goals across governance levels within the same socio-ecological system, emphasizing the necessity of cross-scale coordination to align planning objectives with ecosystem resilience. This study offers a practical framework for adaptive governance under global climate change, providing valuable insights into the intricate dynamics of resilience, sustainability, and governance authority in socio-ecological systems during the Anthropocene.