Estimates of soil carbon concentration in tropical and temperate forest and woodland from available GIS data on three continents

Aim: Concern about climate change, with the subsequent emergence of carbon markets and policy initiatives such as REDD (reducing carbon emissions by decreasing deforestation and forest degradation), have focused attention on assessing and monitoring terrestrial carbon reserves. Most effort has focused on above-ground forest biomass. Soil has received less attention despite containing more carbon than above-ground terrestrial biomass and the atmosphere combined. Our aim was to explore how well soil carbon concentration could be estimated on three continents from existing climate, topography and vegetation-cover data. Location: Peru, Brazil, Argentina, Australia, China. Methods: Soil carbon concentration and leaf area index (LAI) as well as GIS-derived climate and topography variables for 65 temperate and 43 tropical, forest and woodland ecosystems, were either directly measured or estimated from freely available global datasets. We then used multiple regressions to determine how well soil carbon concentration could be predicted from LAI, climate and topography at a given site. We compared our measurements with top soil carbon estimates from the Food and Agriculture Organization of the United Nations (FAO) harmonized world soil map. Results: Our empirical model based on estimates of temperature, water availability and plant productivity provided a good estimate of soil carbon concentrations (R² = 0.79). In contrast, the values of topsoil carbon concentrations from the FAO harmonized world soil map correlated poorly with the measured values of soil carbon concentration (R² = 0.0011). Main conclusions: The lack of correlation between the measured values of soil carbon and the values from the FAO harmonized world soil map indicate that substantial improvements in the production of soil carbon maps are needed and possible. Our results demonstrate that the inclusion of freely available GIS data offers improved estimates of soil carbon and will allow the creation of more accurate soil carbon maps.