The effects of small-scale heterogeneity on the large-scale dynamics of west Siberian wetland carbon fluxes.

摘要

Wetlands are the world's largest natural source of methane and an historically large sink of atmospheric carbon. High-latitude wetlands have received increasing scrutiny recently, due to the temperature sensitivity of their carbon emissions, and the implications of ongoing and predicted warming - including implications of permafrost thawing. The centerpiece of this dissertation is a new large-scale wetland modeling scheme that accounts for the heterogeneous effects of microtopography on water table depth and carbon fluxes. I incorporated this new scheme into the Variable Infiltration Capacity (VIC) land surface model, extended it to include carbon cycle processes, and linked it to an existing wetland methane emissions model. Using this modeling framework, I simulated wetland hydrology and biogeochemistry in the West Siberian Lowland (WSL) over the period 1948-2010. Changes in temperature and precipitation influenced both water table depth and methane emissions. I found that simpler schemes used in previous studies were subject to errors of +/- 30% in their predictions of end-of-century boreal wetland methane emissions due to the nature of their simplifying assumptions. While calibrating to intensive methane flux observations in the WSL, I found a strong north-south gradient in observed methane emissions, which could only be reproduced a) accounting for sub-grid heterogeneity in water table depth and b) using spatially-varying methane emissions parameters. Because most previous studies neglected at least one of these two controls, the majority of methane emissions from the WSL have, apparently incorrectly, been attributed to permafrost wetlands in the northern half of the domain. Finally, I used the outputs of the CMIP5 global climate model projections to force simulations over the WSL for the 21st century and explored the possible responses of the soil microbial communities to climate change. End-of-century methane emissions from the WSL ranged from 6 to 120% more than historical levels, with the range primarily determined by the nature of the response of soil microbes to climate change. Crucially, under one potential scenario, the majority of methane emissions will shift from the south of the WSL to the north, where permafrost thaw is a concern. These results suggest a need to both a) account for sub-grid heterogeneity in wetland soil moisture conditions and b) constrain the response of soil microbial communities to future changes in climate and vegetation.