Large amount of organic carbon is stored in high latitude soils. A substantial proportion of this carbon stock is vulnerable and may decompose rapidly due to temperature increases that are already greater than the global average. It is therefore crucial to quantify and understand carbon exchange between the atmosphere and subarctic/arctic ecosystems. In this paper, we combine an arctic-enabled version of the process-based dynamic ecosystem model, LPJ-GUESS (version LPJG-WHyMe-TFM) with comprehensive observations of terrestrial and aquatic carbon fluxes to simulate long-term carbon exchange in a subarctic catchment comprising both mineral and peatland soils. The model is applied at 50 m resolution and is shown to be able to capture the seasonality and magnitudes of observed fluxes at this fine scale. The modelled magnitudes of CO 2 uptake generally follow the descending sequence: birch forest, non-permafrost Eriophorum , Sphagnum and then tundra heath during the observation periods. The catchment-level carbon fluxes from aquatic systems are dominated by CO 2 emissions from streams. Integrated across the whole catchment, we estimate that the area is a carbon sink at present, and will become an even stronger carbon sink by 2080, which is mainly a result of a projected densification of birch forest and its encroachment into tundra heath. However, the magnitudes of the modelled sinks are very dependent on future atmospheric CO 2 concentrations. Furthermore, comparisons of global warming potentials between two simulations with and without CO 2 increase since 1960 reveal that the increased methane emission from the peatland could double the warming effects of the whole catchment by 2080 in the absence of CO 2 fertilization of the vegetation. This is the first process-based model study of the temporal evolution of a catchment-level carbon budget at high spatial resolution, integrating comprehensive and diverse fluxes including both terrestrial and aquatic carbon. Though this study also highlights some limitations in modelling subarctic ecosystem responses to climate change including aquatic system flux dynamics, nutrient limitation, herbivory and other disturbances and peatland expansion, our application provides a mechanism to resolve the complexity of carbon cycling in subarctic ecosystems while simultaneously pointing out the key model developments for capturing complex subarctic processes.