Journal Highlights

Modeling Permafrost’s Role in the Global Carbon Cycle

From Eos.org: Research SpotlightsA team of international scientists surveyed an array of Earth ecosystem models, recommending several ways to reduce uncertainties.

About half of the carbon that exists beneath Earth’s surface is stored within regions of permanently frozen ground, called permafrost. However, as global temperatures rise, the planet’s permafrost is beginning to thaw, releasing carbon into the atmosphere. It is critical for scientists to study this process, as it disrupts what is known as the global carbon cycle—the natural flow of carbon from plants into the air and back again—and could have future climate impacts.

Given permafrost’s importance for understanding global climate, Xia et al. reviewed the Earth ecosystem models currently being used to represent and evaluate carbon processes in permafrost regions. Unfortunately, the international team of researchers found that current models do not represent these processes well, contributing to large uncertainties in the projected levels of carbon being released from (and, ultimately, returned to) permafrost regions.

Past studies have mostly focused on the net primary productivity, which is the total amount of carbon entering an ecosystem through photosynthesis minus any carbon lost through plant respiration. However, scientists have shown that the net primary productivity, as simulated by these models, can differ depending on whether the region contains primarily boreal forest or tundra. Quantities such as gross primary productivity (all of the carbon that is assimilated through photosynthesis in an ecosystem) and carbon use efficiency (an ecosystem’s ability to build plant tissue with the assimilated carbon), which are based on scientists’ understanding of how carbon is allocated to and respired from plant tissues, can also have sizable uncertainties.

Pinpointing and fixing these systematic biases would improve Earth ecosystem models’ ability to consistently and accurately simulate processes in the global carbon cycle. For this study, the team looked at 10 such models. Running a series of retrospective simulations on data from 2000-2009, the researchers examined the models’ abilities to estimate an ecosystem’s net primary productivity, as well as monitor modeled responses to climate change, in permafrost regions of the Arctic and sub-Arctic.

The team found that, across the board, the models simulated a net primary productivity that was about 20% higher than data collected by NASA’s Moderate Resolution Imaging Spectroradiometer sensors for the same years. This uncertainty likely stems from a limited understanding of the mechanisms underlying both carbon use efficiency and gross primary productivity.

Reducing these uncertainties may be difficult, the researchers noted, because each model has a different structure and set of parameters. However, the authors recommended that modeling groups start by focusing on improving simulations of relevant processes, such as plant respiration, and by assembling a global database of estimated productivity. They also advised that scientists seek to better understand environmental regulation of these dynamics, and link carbon processes to environmental factors.

In combination, all of these steps could further improve global carbon cycle models, and allow scientists to better predict carbon feedback in permafrost regions as it relates to climate change. 

-- Sarah Witman, Freelance Writer,

Article Category
Research Articles

Terrestrial ecosystem model performance in simulating productivity and its vulnerability to climate change in the northern permafrost region

Jianyang Xia, A. David McGuire, David Lawrence, Eleanor Burke, Guangsheng Chen, Xiaodong Chen, Christine Delire, Charles Koven, Andrew MacDougall, Shushi Peng, Annette Rinke, Kazuyuki Saito, Wenxin Zhang, Ramdane Alkama, Theodore J. Bohn, Philippe Ciais, Bertrand Decharme, Isabelle Gouttevin, Tomohiro Hajima, Daniel J. Hayes, Kun Huang, Duoying Ji, Gerhard Krinner, Dennis P. Lettenmaier, Paul A. Miller, John C. Moore, Benjamin Smith, Tetsuo Sueyoshi, Zheng Shi, Liming Yan, Junyi Liang, Lifen Jiang, Qian Zhang, Yiqi Luo
First Published:
| Vol:
122,
Pages
430–446
| DOI:
10.1002/2016JG003384
Free