The Arctic is especially vulnerable to climate change and is a source of large feedback mechanisms. One such mechanism is the release of stored carbon from peat. Arctic soils contain about 1600 Pgs of carbon, which is about twice as much as the atmospheric burden. The Arctic had previously been a net sink for carbon, however recent measurements over the last several decades indicate that the Arctic has transitioned to a net source of carbon, presumably due to increased microbial activity linked to temperature increases. The leading hypothesis is that previously frozen carbon stored in peat has thawed and is being respired by microbes in the soil. Measurements of seasonal carbon fluxes show substantial release to the atmosphere in the autumn and winter when surface air and soil temperatures are below zero. This suggests that the microbial activity in the active layer plays an important role in carbon release. This positive feedback cycle could potentially be balanced by increased vegetation growth due to CO2 fertilization and warmer temperatures. However, current models do not even agree on the direction of the Arctic carbon feedback (McGuire et al. 2018). This large uncertainty can be attributed to several mechanisms such as changes in the water table, CO2 fertilization, and shrub and tree expansion that are poorly resolved in current climate models.
We are developing an arctic version of a terrestrial biosphere model (ED-Peat) to simulate the current and future carbon budget of the Arctic. Recent eddy covariance tower measurements have highlighted the importance of winter heterotrophic respiration (HR) in the annual carbon budget. Soil carbon and HR is now a vertically resolved property in ED2-Peat. Litter is added as fast and structural soil carbon at the surface. As carbon accumulates, its depth is tracked and it affects the soil properties. It also is converted to slow carbon and/or lost to the atmosphere via HR. The HR rate in the model is a function of soil carbon pool and abundance as well as temperature and moisture. These model developments allow ED2-Peat to track HR through time and soil depth and attribute changes to temperature and moisture fluctuations in the soil column. Increases in HR due to spring snowmelt are evident in the figure below, as is the zero-curtain period where the surface is frozen but deeper soil layers remain at or above freezing and active with regards to microbial activity.
To close the carbon budget, accurate estimates of net primary productivity (NPP) need to be calculated. We have developed tundra specific PFT parameterizations for deciduous shrubs, evergreen shrubs, and graminoids. Trait data for plant species characteristic to these functional types comes from the TTT and TRI databases. These parameterizations have been evaluated with data at Toolik, Atqasuk, and Barrow tower sites. These new PFTs reasonably simulate NPP, which combined with the HR estimates provide a simulated net ecosystem production (NEP) to evaluate against tower measurements.
This project is funded as part of the NASA ABoVE campaign. We are collaborating with the Wofsy-Munger group and other ABoVE members to validate the model carbon flux estimates with tower measurements at specific sites. The goal is to understand and quantify the competing processes of carbon uptake at the surface and carbon release from the subsurface, and how these processes change in a warming environment.