A study of the role of wetlands in defining spatial patterns of near-surface (top 1 m) soil carbon in the Northern Latitudes

Abstract. A study of two observation-based maps (the Harmonised World Soil Database, HWSD and the Northern Circumpolar Soil Carbon Database, NCSCD) of the surface (1 m) soil carbon in the Northern Latitudes (containing the Arctic and Boreal regions) reveal that, although the amounts of carbon estimated to be present in this region are very uncertain, the patterns are robust: both maps have soil carbon maxima that coincide with the major wetlands in the region, as described in the Global Lakes and Wetlands Database, GLWD. In fact, the relationship between near-surface soil carbon and the presence of wetlands is stronger than the relationship with soil temperature and vegetation productivity.

These relationships are explored using the land surface model of the UK Hadley Centre GCM: JULES (Joint UK Land Environment Simulator). The model is run to represent conditions at the end of the 20th century. Observed vegetation and phenology are used to define the vegetation, the physical properties of organic soils are represented, the fine-scale topography of the region is included in the parameterisation of the hydrology and as a result the GPP and location of the wetlands of the region are reasonably well simulated using JULES.

Despite this, the soil carbon simulated by the model does not reveal the same patterns or the correlation with the wetland regions that are present in the data. This suggests that the model does not represent sufficiently strongly the suppression of heterotrophic respiration in saturated conditions.

A simple adjustment to the JULES model was made whereby the heterotrophic respiration was reduced by the fraction of the grid that is modelled to be saturated. In effect, for the saturated areas the respiration was zero. This adjustment represents a simple experiment to establish the role of wetlands in defining the spatial patterns of near-surface soil carbon.

The results were an improved predicted spatial pattern of soil carbon, with an increase in the correlation between soil carbon and wetlands although not as strong as suggested by the analysis of the data. This may be because the size of the wetlands was underestimated by the model.

The study suggests that land surface models in general, and JULES in particular, need to establish a stronger moderation of soil respiration in saturated conditions in order that future climate controls on wetlands in the Northern Latitudes will result in the correct changes in soil carbon and carbon emissions.