Terrestrial ecosystems currently absorb one quarter of the carbon dioxide released by fossil fuel burning into the atmosphere, and thus reduce the rate of climate change. As conditions become more favourable for plant growth, most models predict that high latitudes will take up more carbon during the 21st century. However, vast stores of carbon are frozen in boreal and arctic permafrost, and warming may result in some of this carbon being released to the atmosphere.
The recent inclusion of permafrost thaw in large-scale model simulations has suggested that the permafrost feedback is potentially so significant that it could reduce substantially the predicted global net uptake of carbon by terrestrial ecosystems during the 21st century, with major implications for the rate of climate change. Large uncertainties remain in predicting rates of permafrost thaw and in determining the impacts of thaw in contrasting ecosystems, with many of the key processes missing from carbon-climate models. Firstly, the role that different plant communities play in insulating soils and protecting permafrost is poorly quantified, with key groups such as mosses absent in most models. In addition, fire disturbance can substantially accelerate permafrost thaw, and hence the ability of permafrost-protecting plant communities to recover from fire may play a key role in determining permafrost resilience. Secondly, different ecosystems may respond differently to thaw with contrasting effects on release of greenhouse gasses. In free-draining ecosystems, thaw may result in the net release of carbon due to increased decomposition of previously frozen organic matter. On the other hand, when thawing takes place in peatlands, soil subsidence can effectively raise the water table, which could result in carbon accumulation. However, this potential negative feedback may be offset by enhanced release of the more powerful greenhouse gas, methane. Importantly, the full range of feedbacks to permafrost thaw in these contrasting ecosystems is not currently reflected in process-based models. To address these issues, we will undertake directed fieldwork campaigns to determine (1) the role that different plant communities play in protecting permafrost within different soil types, and in unburned and fire-disturbed ecosystems, and (2) the impacts of permafrost thaw on fluxes of carbon dioxide and methane in free-draining versus peatland systems. Through links to Canadian partners, data will be collected from a range of field sites where permafrost monitoring is ongoing, including: (i) two contrasting boreal peatlands differing in permafrost extent, and where there is permafrost degradation; (ii) burnt and unburned sites within three important forest types in boreal Canada. Data will be provided from burnt and unburned moist acidic tundra within the continuous permafrost zone in Alaska by our US partners. The spatially variable vegetation recovery at the fire sites allows relationships between vegetation and permafrost to be tested in detail, while comparisons between the tundra, forest and peatland sites provide insights into the impacts of permafrost thaw in contrasting ecosystems. Critically, these data will be used to develop, parameterise and evaluate a detailed process-based model of vegetation-soil-permafrost interactions. The in-depth representation of vegetation-permafrost linkages will improve predictions of rates of permafrost thaw. The model will be the first to simulate the full range of biogeochemical feedbacks (methane and carbon dioxide) in free-draining versus wetland ecosystems. Furthermore, through links with Met Office scientists, our model will be coupled to the Joint UK Land Environment Simulator (JULES), allowing regional simulations to be run, coupled to a climate model. Ultimately, our project will improve predictions of both the rates and consequences of permafrost thaw, and help determine the potential impacts on 21st century climate change.