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.</p> <p>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 (<em>r</em> ≈ 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.