Does erosion represent landscape-level loss or gain of carbon stocks?


The rice fields of Southeast
Asia would be a lot less
fertile without historical erosion of the associated hill slopes – does this mean that erosion can be carbon neutral?
In 1994 at the first big tropical soil carbon meeting after the Rio 1992 Convention spiked interest in the emission effect of all types of land use change, I shocked many in the audience by challenging the perceived wisdom that erosion was a major source of emissions
True, erosion does lead to lower soil-carbon stocks on the eroding slopes, but we need to know what happens with the carbon-rich material that travels through the landscape. It might, for instance, get buried in a place where decomposition is slower than in the soil profile from where it came.
The idea that erosion might be a carbon-storage scenario, paralleled by burning of forests and associated charcoal accumulation in soils, was “politically incorrect”, and in the desire to identify win-win solutions for local and global benefits, the idea got buried.
Still, the second Intergovernmental Panel on Climate Change (IPCC) report in 1995, in its chapter on soils of which I was co-author, reported that the jury was out on net gain or loss from erosion at landscape scale, and that further research was needed
Ignoring the challenge, however, the idea crept back into the literature that erosion was a bad thing not only for on-site productivity but also as cause of global C emissions.
In a new paper titled ‘Legacy of human-induced C erosion and burial on soil–atmosphere C exchange’ in the Proceedings of the National Academy of Sciences, van Oost and co-authors from Leuven University said that for a landscape (the Dijle catchment) in Belgium, a long-term C sink in colluvial sites had stored the equivalent of 43% of the eroded C, and that this sink had offset 39% (17–66%) of the C emissions due to anthropogenic land-cover change since the advent of agriculture in the area.
Of course this does not mean that stimulating erosion is a good idea, as it does have mostly negative consequences. But it does show that a landscape-scale of assessment is more than multiplying the measured change in sample locations by the total area involved, especially where the “landscape filters” where deposition occurs may have been left out of the sampling scheme that focussed on representative fields.
If we want to appraise the role of agroforestry at a landscape scale in reducing net carbon loss to the atmosphere, we may have to point to the roles of trees in stabilizing deposition sites in landscape filters, rather than in their presumed roles in reducing erosion; but in fact, critical studies in tropical landscapes on this phenomenon are yet to be made.
A study that comes close is one by Rodenburg et al., who documented high erosion rates after slash-and-burn land clearing on a steep slope but zero sediment transfer to the nearby stream, as carbon-rich material accumulated at the bottom of the slope.
Such phenomena are likely behind the fractal dimension of net erosion, estimated to be around 1.6 which implies that area-based scaling up of measured results can lead to gross errors. Net sediment loss at landscape scales is indeed a very different phenomenon than field-level erosion, as we learned in the Sumberjaya research site in Indonesia
For the time being, we had better leave erosion and its control out of the discussions of soil-based carbon emission – unless we want to get serious about landscape-level studies like van Oost and colleagues did, and be challenged by their conclusions.


Please enter your comment!
Please enter your name here