In this post Peter Manning discusses his recent paper ‘Simple measures of climate, soil properties and plant traits predict national-scale grassland soil carbon stocks‘

Soils contain more carbon than the atmosphere and vegetation combined, but the future of this reserve is uncertain – will it remain in the ground or be released into the atmosphere, potentially amplifying climate warming by several degrees in a worst case scenario? The answer to this question is not simple to obtain; the dynamics of the soil organic matter that contains this carbon are complex and a full understanding of them is still out of reach.

One key step in acquiring this knowledge is to identify the most important drivers of soil carbon storage, as the identity of these is still hotly debated by ecologists and soil scientists. Related to this is an ability to accurately predict the amount of carbon that is stored in soils, as measuring this directly can be an expensive and laborious task that is outside the capabilities of most landowners and managers.

In our study ‘Simple measures of climate, soil properties and plant traits predict national-scale grassland soil carbon stocks’ we attempted to address both of these issues by using data from a large survey of 180 English fields, covering a wide range of grassland types and farm management practices. In these we measured soil organic carbon in a range of soil particle size fractions that differ in their likelihood of releasing carbon to the atmosphere, and several environmental factors that are known to influence carbon dynamics. These included climate, farm management, soil pH and the abundance of different plant species. These plant measures were combined with database values of their ‘functional traits’, properties that influence the growth and metabolism of plants and the organisms that consume them, to give a measure of the overall functioning of the vegetation in each field. All this data was then used to build statistical models that explain how these factors relate to the amount of carbon found in each field.

What we found is that carbon stocks, in the different particle size classes, showed consistent relationships with mean annual temperature and precipitation, soil moisture and pH and the plant trait measures across the whole of England, and this suggests that these factors are the primary drivers of soil carbon stocks in English grasslands. We confirmed these relationships by measuring the same factors in a new, previously unmeasured, region and finding that the models we built were generally good at predicting the amounts of carbon found there.

The approach we used here needs to be thoroughly tested before coming into general use, but it has the potential to be adopted far more widely, and in a range of ecosystems throughout the world. If it is found to be successful then it would open up the possibility of low cost surveys in which simple soil measures and a few vegetation records could be taken at a site, combined with freely available climate data and fed into models to generate detailed maps of soil carbon stocks over large areas. Such maps would be immensely useful for accounting in future carbon trading schemes, and also as the starting point in climate models that project future changes to soil carbon and its feedback to climate.