Garry Hayman

Land use, management, and land-cover change account for approximately a quarter of anthropogenic greenhouse gas (GHG) emissions. Despite substantial progress in Earth system modelling, the CO₂, CH₄, and N₂O emissions arising specifically from land management remain incompletely accounted for, preventing closure of the feedback between management decisions, the resulting emissions, the climate response, and ecosystem impacts. Here we extend the LandSyMM modelling framework by coupling it with the intermediate-complexity climate emulator IMOGEN and by integrating IPCC Tier-1 agricultural-emission methods. The framework assembles a sector-complete, bottom-up emissions dataset by combining the SSP-RCP background emissions with socioeconomically driven land-use projections, Tier-1 agricultural emissions, and process-based land-biosphere fluxes. These emissions (not prescribed concentrations) drive the emulator, which converts them to atmospheric GHG concentrations and radiative forcing, and pattern-scales the result into climate, which in turn forces a final simulation of the resulting ecosystem trajectories. Simulating the historical period and five Shared Socioeconomic Pathway-Representative Concentration Pathway (SSP-RCP) scenarios (SSP1-2.6 through SSP5-8.5), we evaluate the framework's reproduction of historical agricultural emissions, integrated emission and concentration trajectories, as well as climate-forcing patterns against established inventories (FAOSTAT, EDGAR), observational concentration records and global budgets, and the Inter-Sectoral Impact Model Intercomparison Project phase 3b (ISIMIP-3b) climate reference. LandSyMM reproduces historical agricultural GHG emissions across major source categories within the source-level uncertainty of the Tier-1 method (1970–2020): the agricultural CH₄ aggregate runs +19 % high owing to IPCC default emission factors for the dominant livestock sources, while the N₂O aggregate is near-balanced against FAOSTAT. With the complete natural-source terms included, the integrated emissions match the observationally constrained global budgets to within a few per cent at the present day. The prognostic atmospheric CO₂, CH₄, and N₂O concentration records are reproduced to within about 7 ppm, 50 ppb, and 6 ppb (root-mean-square over the historical period), subject to small end-of-period offsets. Relative to the unmodified reference-scenario anthropogenic emissions with the emulator's default natural baseline, our bottom-up integrated emissions – substituting the IPCC Tier-1 agricultural and LPJ-GUESS land-biosphere terms – diverge increasingly over the century: by 2100 the integrated CH₄ is 14 to 71 % higher, while the explicit terrestrial carbon sink lowers the net CO₂ source by about 1 to 5 Pg C yr⁻¹. The emulated climate reproduces the ISIMIP-3b reference closely in both spatial pattern (r ≈ 0.98 for temperature) and global-mean magnitude (within about 0.3 K across scenarios and analysis windows), leaving only a small residual regional mean-state difference that is near-neutral in the global mean. Returned to LPJ-GUESS, the two configurations settle to slightly different absolute carbon pools because they respond to different observation-based baseline climatologies (an initial-condition effect rather than a coupling signal), but compared on the meaningful basis of their changes over time they track each other closely, on a geography closely shared with the reference and with small, mixed crop-yield differences. Passed through IMOGEN, this lower net CO₂ source outweighs the higher CH₄ and lowers projected global-land-mean warming by about 0.2 to 0.5 K – a cooling that is largest (about 0.5 K) under the low-forcing SSP1-2.6 – and an identical-emulator diagnostic confirms this emission-pathway feedback is independent of the baseline climatology. The framework establishes a traceable platform for investigating the climate-ecosystem consequences of land-management GHG emissions, enabling explicit attribution of ecosystem responses to specific agricultural management practices.

Land use, management, and land-cover change account for approximately a quarter of anthropogenic greenhouse gas (GHG) emissions. Despite substantial progress in Earth system modelling, the CO₂, CH₄, and N₂O emissions arising specifically from land management remain incompletely accounted for, preventing closure of the feedback between management decisions, the resulting emissions, the climate response, and ecosystem impacts. Here we extend the LandSyMM modelling framework by coupling it with the intermediate-complexity climate emulator IMOGEN and by integrating IPCC Tier-1 agricultural-emission methods. The framework assembles a sector-complete, bottom-up emissions dataset by combining the SSP-RCP background emissions with socioeconomically driven land-use projections, Tier-1 agricultural emissions, and process-based land-biosphere fluxes. These emissions (not prescribed concentrations) drive the emulator, which converts them to atmospheric GHG concentrations and radiative forcing, and pattern-scales the result into climate, which in turn forces a final simulation of the resulting ecosystem trajectories. Simulating the historical period and five Shared Socioeconomic Pathway-Representative Concentration Pathway (SSP-RCP) scenarios (SSP1-2.6 through SSP5-8.5), we evaluate the framework's reproduction of historical agricultural emissions, integrated emission and concentration trajectories, as well as climate-forcing patterns against established inventories (FAOSTAT, EDGAR), observational concentration records and global budgets, and the Inter-Sectoral Impact Model Intercomparison Project phase 3b (ISIMIP-3b) climate reference. LandSyMM reproduces historical agricultural GHG emissions across major source categories within the source-level uncertainty of the Tier-1 method (1970–2020): the agricultural CH₄ aggregate runs +19 % high owing to IPCC default emission factors for the dominant livestock sources, while the N₂O aggregate is near-balanced against FAOSTAT. With the complete natural-source terms included, the integrated emissions match the observationally constrained global budgets to within a few per cent at the present day. The prognostic atmospheric CO₂, CH₄, and N₂O concentration records are reproduced to within about 7 ppm, 50 ppb, and 6 ppb (root-mean-square over the historical period), subject to small end-of-period offsets. Relative to the unmodified reference-scenario anthropogenic emissions with the emulator's default natural baseline, our bottom-up integrated emissions – substituting the IPCC Tier-1 agricultural and LPJ-GUESS land-biosphere terms – diverge increasingly over the century: by 2100 the integrated CH₄ is 14 to 71 % higher, while the explicit terrestrial carbon sink lowers the net CO₂ source by about 1 to 5 Pg C yr⁻¹. The emulated climate reproduces the ISIMIP-3b reference closely in both spatial pattern (r ≈ 0.98 for temperature) and global-mean magnitude (within about 0.3 K across scenarios and analysis windows), leaving only a small residual regional mean-state difference that is near-neutral in the global mean. Returned to LPJ-GUESS, the two configurations settle to slightly different absolute carbon pools because they respond to different observation-based baseline climatologies (an initial-condition effect rather than a coupling signal), but compared on the meaningful basis of their changes over time they track each other closely, on a geography closely shared with the reference and with small, mixed crop-yield differences. Passed through IMOGEN, this lower net CO₂ source outweighs the higher CH₄ and lowers projected global-land-mean warming by about 0.2 to 0.5 K – a cooling that is largest (about 0.5 K) under the low-forcing SSP1-2.6 – and an identical-emulator diagnostic confirms this emission-pathway feedback is independent of the baseline climatology. The framework establishes a traceable platform for investigating the climate-ecosystem consequences of land-management GHG emissions, enabling explicit attribution of ecosystem responses to specific agricultural management practices.

Land use, management, and land-cover change account for approximately a quarter of anthropogenic greenhouse gas (GHG) emissions. Despite substantial progress in Earth system modelling, the CO₂, CH₄, and N₂O emissions arising specifically from land management remain incompletely accounted for, preventing closure of the feedback between management decisions, the resulting emissions, the climate response, and ecosystem impacts. Here we extend the LandSyMM modelling framework by coupling it with the intermediate-complexity climate emulator IMOGEN and by integrating IPCC Tier-1 agricultural-emission methods. The framework assembles a sector-complete, bottom-up emissions dataset by combining the SSP-RCP background emissions with socioeconomically driven land-use projections, Tier-1 agricultural emissions, and process-based land-biosphere fluxes. These emissions (not prescribed concentrations) drive the emulator, which converts them to atmospheric GHG concentrations and radiative forcing, and pattern-scales the result into climate, which in turn forces a final simulation of the resulting ecosystem trajectories. Simulating the historical period and five Shared Socioeconomic Pathway-Representative Concentration Pathway (SSP-RCP) scenarios (SSP1-2.6 through SSP5-8.5), we evaluate the framework's reproduction of historical agricultural emissions, integrated emission and concentration trajectories, as well as climate-forcing patterns against established inventories (FAOSTAT, EDGAR), observational concentration records and global budgets, and the Inter-Sectoral Impact Model Intercomparison Project phase 3b (ISIMIP-3b) climate reference. LandSyMM reproduces historical agricultural GHG emissions across major source categories within the source-level uncertainty of the Tier-1 method (1970–2020): the agricultural CH₄ aggregate runs +19 % high owing to IPCC default emission factors for the dominant livestock sources, while the N₂O aggregate is near-balanced against FAOSTAT. With the complete natural-source terms included, the integrated emissions match the observationally constrained global budgets to within a few per cent at the present day. The prognostic atmospheric CO₂, CH₄, and N₂O concentration records are reproduced to within about 7 ppm, 50 ppb, and 6 ppb (root-mean-square over the historical period), subject to small end-of-period offsets. Relative to the unmodified reference-scenario anthropogenic emissions with the emulator's default natural baseline, our bottom-up integrated emissions – substituting the IPCC Tier-1 agricultural and LPJ-GUESS land-biosphere terms – diverge increasingly over the century: by 2100 the integrated CH₄ is 14 to 71 % higher, while the explicit terrestrial carbon sink lowers the net CO₂ source by about 1 to 5 Pg C yr⁻¹. The emulated climate reproduces the ISIMIP-3b reference closely in both spatial pattern (r ≈ 0.98 for temperature) and global-mean magnitude (within about 0.3 K across scenarios and analysis windows), leaving only a small residual regional mean-state difference that is near-neutral in the global mean. Returned to LPJ-GUESS, the two configurations settle to slightly different absolute carbon pools because they respond to different observation-based baseline climatologies (an initial-condition effect rather than a coupling signal), but compared on the meaningful basis of their changes over time they track each other closely, on a geography closely shared with the reference and with small, mixed crop-yield differences. Passed through IMOGEN, this lower net CO₂ source outweighs the higher CH₄ and lowers projected global-land-mean warming by about 0.2 to 0.5 K – a cooling that is largest (about 0.5 K) under the low-forcing SSP1-2.6 – and an identical-emulator diagnostic confirms this emission-pathway feedback is independent of the baseline climatology. The framework establishes a traceable platform for investigating the climate-ecosystem consequences of land-management GHG emissions, enabling explicit attribution of ecosystem responses to specific agricultural management practices.

Land use, management, and land-cover change account for approximately a quarter of anthropogenic greenhouse gas (GHG) emissions. Despite substantial progress in Earth system modelling, the CO₂, CH₄, and N₂O emissions arising specifically from land management remain incompletely accounted for, preventing closure of the feedback between management decisions, the resulting emissions, the climate response, and ecosystem impacts. Here we extend the LandSyMM modelling framework by coupling it with the intermediate-complexity climate emulator IMOGEN and by integrating IPCC Tier-1 agricultural-emission methods. The framework assembles a sector-complete, bottom-up emissions dataset by combining the SSP-RCP background emissions with socioeconomically driven land-use projections, Tier-1 agricultural emissions, and process-based land-biosphere fluxes. These emissions (not prescribed concentrations) drive the emulator, which converts them to atmospheric GHG concentrations and radiative forcing, and pattern-scales the result into climate, which in turn forces a final simulation of the resulting ecosystem trajectories. Simulating the historical period and five Shared Socioeconomic Pathway-Representative Concentration Pathway (SSP-RCP) scenarios (SSP1-2.6 through SSP5-8.5), we evaluate the framework's reproduction of historical agricultural emissions, integrated emission and concentration trajectories, as well as climate-forcing patterns against established inventories (FAOSTAT, EDGAR), observational concentration records and global budgets, and the Inter-Sectoral Impact Model Intercomparison Project phase 3b (ISIMIP-3b) climate reference. LandSyMM reproduces historical agricultural GHG emissions across major source categories within the source-level uncertainty of the Tier-1 method (1970–2020): the agricultural CH₄ aggregate runs +19 % high owing to IPCC default emission factors for the dominant livestock sources, while the N₂O aggregate is near-balanced against FAOSTAT. With the complete natural-source terms included, the integrated emissions match the observationally constrained global budgets to within a few per cent at the present day. The prognostic atmospheric CO₂, CH₄, and N₂O concentration records are reproduced to within about 7 ppm, 50 ppb, and 6 ppb (root-mean-square over the historical period), subject to small end-of-period offsets. Relative to the unmodified reference-scenario anthropogenic emissions with the emulator's default natural baseline, our bottom-up integrated emissions – substituting the IPCC Tier-1 agricultural and LPJ-GUESS land-biosphere terms – diverge increasingly over the century: by 2100 the integrated CH₄ is 14 to 71 % higher, while the explicit terrestrial carbon sink lowers the net CO₂ source by about 1 to 5 Pg C yr⁻¹. The emulated climate reproduces the ISIMIP-3b reference closely in both spatial pattern (r ≈ 0.98 for temperature) and global-mean magnitude (within about 0.3 K across scenarios and analysis windows), leaving only a small residual regional mean-state difference that is near-neutral in the global mean. Returned to LPJ-GUESS, the two configurations settle to slightly different absolute carbon pools because they respond to different observation-based baseline climatologies (an initial-condition effect rather than a coupling signal), but compared on the meaningful basis of their changes over time they track each other closely, on a geography closely shared with the reference and with small, mixed crop-yield differences. Passed through IMOGEN, this lower net CO₂ source outweighs the higher CH₄ and lowers projected global-land-mean warming by about 0.2 to 0.5 K – a cooling that is largest (about 0.5 K) under the low-forcing SSP1-2.6 – and an identical-emulator diagnostic confirms this emission-pathway feedback is independent of the baseline climatology. The framework establishes a traceable platform for investigating the climate-ecosystem consequences of land-management GHG emissions, enabling explicit attribution of ecosystem responses to specific agricultural management practices.

Land use, management, and land-cover change account for approximately a quarter of anthropogenic greenhouse gas (GHG) emissions. Despite substantial progress in Earth system modelling, the CO₂, CH₄, and N₂O emissions arising specifically from land management remain incompletely accounted for, preventing closure of the feedback between management decisions, the resulting emissions, the climate response, and ecosystem impacts. Here we extend the LandSyMM modelling framework by coupling it with the intermediate-complexity climate emulator IMOGEN and by integrating IPCC Tier-1 agricultural-emission methods. The framework assembles a sector-complete, bottom-up emissions dataset by combining the SSP-RCP background emissions with socioeconomically driven land-use projections, Tier-1 agricultural emissions, and process-based land-biosphere fluxes. These emissions (not prescribed concentrations) drive the emulator, which converts them to atmospheric GHG concentrations and radiative forcing, and pattern-scales the result into climate, which in turn forces a final simulation of the resulting ecosystem trajectories. Simulating the historical period and five Shared Socioeconomic Pathway-Representative Concentration Pathway (SSP-RCP) scenarios (SSP1-2.6 through SSP5-8.5), we evaluate the framework's reproduction of historical agricultural emissions, integrated emission and concentration trajectories, as well as climate-forcing patterns against established inventories (FAOSTAT, EDGAR), observational concentration records and global budgets, and the Inter-Sectoral Impact Model Intercomparison Project phase 3b (ISIMIP-3b) climate reference. LandSyMM reproduces historical agricultural GHG emissions across major source categories within the source-level uncertainty of the Tier-1 method (1970–2020): the agricultural CH₄ aggregate runs +19 % high owing to IPCC default emission factors for the dominant livestock sources, while the N₂O aggregate is near-balanced against FAOSTAT. With the complete natural-source terms included, the integrated emissions match the observationally constrained global budgets to within a few per cent at the present day. The prognostic atmospheric CO₂, CH₄, and N₂O concentration records are reproduced to within about 7 ppm, 50 ppb, and 6 ppb (root-mean-square over the historical period), subject to small end-of-period offsets. Relative to the unmodified reference-scenario anthropogenic emissions with the emulator's default natural baseline, our bottom-up integrated emissions – substituting the IPCC Tier-1 agricultural and LPJ-GUESS land-biosphere terms – diverge increasingly over the century: by 2100 the integrated CH₄ is 14 to 71 % higher, while the explicit terrestrial carbon sink lowers the net CO₂ source by about 1 to 5 Pg C yr⁻¹. The emulated climate reproduces the ISIMIP-3b reference closely in both spatial pattern (r ≈ 0.98 for temperature) and global-mean magnitude (within about 0.3 K across scenarios and analysis windows), leaving only a small residual regional mean-state difference that is near-neutral in the global mean. Returned to LPJ-GUESS, the two configurations settle to slightly different absolute carbon pools because they respond to different observation-based baseline climatologies (an initial-condition effect rather than a coupling signal), but compared on the meaningful basis of their changes over time they track each other closely, on a geography closely shared with the reference and with small, mixed crop-yield differences. Passed through IMOGEN, this lower net CO₂ source outweighs the higher CH₄ and lowers projected global-land-mean warming by about 0.2 to 0.5 K – a cooling that is largest (about 0.5 K) under the low-forcing SSP1-2.6 – and an identical-emulator diagnostic confirms this emission-pathway feedback is independent of the baseline climatology. The framework establishes a traceable platform for investigating the climate-ecosystem consequences of land-management GHG emissions, enabling explicit attribution of ecosystem responses to specific agricultural management practices.

Limiting global warming to 1.5°C above preindustrial levels is central to the Paris Agreement, yet temporary overshoots of this target are increasingly likely. While the global mean surface temperature could return to 1.5°C through mitigation and carbon dioxide removal, the regional consequences of overshoot remain uncertain. We use the UK Earth System Model (UKESM1.0-LL) to compare a policy-relevant moderate overshoot scenario—peaking at 1.8°C before cooling to 1.5°C by 2100—with a no-overshoot pathway. We show that, despite similar global temperatures by 2100, the Arctic exhibits significant hysteresis, with ~ 18% less sea ice and ~ 1°C higher near-surface temperatures under overshoot. Using a zonal energy balance model and radiative kernel analysis, we identify sea ice–albedo and lapse rate feedbacks as dominant drivers of this, while heat transport plays a minor role. These findings provide the first mechanistic attribution of Arctic hysteresis under a moderate overshoot scenario, highlighting the risks of delayed mitigation and the importance of avoiding even temporary exceedance of temperature targets.

Non-technical summary: Greenhouse gas emissions and land use change – from deforestation, forest degradation, and agricultural intensification – are contributing to climate change and biodiversity loss. Important land-based strategies such as planting trees or growing bioenergy crops (with carbon capture and storage) are needed to achieve the goals of the Paris Climate Agreement and to enhance biodiversity. The integrated Land Ecosystems Atmospheric Processes Study (iLEAPS) is an international knowledge-exchange and capacity-building network, specializing in ecosystems and their role in controlling the exchange of water, energy and chemical compounds between the land surface and the atmosphere. We outline priority directions for land–atmosphere interaction research and its contribution to the sustainable development agenda. Technical summary: Greenhouse-gas emissions from human activities and land use change (from deforestation, forest degradation, and agricultural intensification) are contributing to climate change and biodiversity loss. Afforestation, reforestation, or growing bioenergy crops (with carbon capture and storage) are important land-based strategies to achieve the goals of the Paris Climate Agreement and to enhance biodiversity. The effectiveness of these actions depends on terrestrial ecosystems and their role in controlling or moderating the exchange of water, heat, and chemical compounds between the land surface and the atmosphere. The integrated Land Ecosystems Atmospheric Processes Study (iLEAPS), a global research network of Future Earth, enables the international community to communicate and remain up to date with developments and concepts about terrestrial ecosystems and their role in global water, energy, and biogeochemical cycles. Covering critically important topics such as fire, forestry, wetlands, methane emissions, urban areas, pollution, and climate change, the iLEAPS Global Research Programme sits center stage for some of the most important environmental questions facing humanity. In this paper, we outline the new challenges and opportunities for land–atmosphere interaction research and its role in supporting the broader sustainable development agenda. Social Media Summary: Future directions for research into land–atmosphere interactions that supports the sustainable development agenda

In the summer of 2018, Europe experienced an intense heatwave which coincided with several persistent large-scale ozone (O₃) pollution episodes. Novel satellite data of lower-tropospheric column O₃ from the Global Ozone Monitoring Experiment-2 (GOME-2) and Infrared Atmospheric Sounding Interferometer (IASI) on the MetOp satellite showed substantial enhancements in 2018 relative to other years since 2012. Surface observations also showed ozone enhancements across large regions of continental Europe in summer 2018 compared to 2017. Enhancements to surface temperature and the O₃ precursor gases carbon monoxide and methanol in 2018 were co-retrieved from MetOp observations by the same scheme. This analysis was supported by the TOMCAT chemistry transport model (CTM) to investigate processes driving the observed O₃ enhancements. Through several targeted sensitivity experiments we show that meteorological processes, and emissions to a secondary order, were important for controlling the elevated O₃ concentrations at the surface. However, mid-tropospheric (∼ 500 hPa) O₃ enhancements were dominated by meteorological processes. We find that contributions from stratospheric O₃ intrusions ranged between 15 %–40 %. Analysis of back trajectories indicates that the import of O₃-enriched air masses into Europe originated over the North Atlantic, substantially increasing O₃ in the 500 hPa layer during summer 2018.