Prediction of organic carbon content in upland forest soils of Quebec, Canada

Soil organic carbon (C) is an important component of forest carbon pools and should be taken into account in sustainable forest management. However, there is a need to derive indicators for this attribute, as organic C content (Mg(.)ha-1) in forest soils is generally not available in Quebec survey data. Thus, we developed models to predict organic C accumulation in the forest floor and in mineral soils of upland forest soils (i.e., with forest floor thickness less than or equal to 30 cm) using soil survey data as input variables. The best-fit model for predicting forest floor organic C content consisted of the explanatory variables forest floor thickness, latitude, and longitude. The model R(2) was 0.76, and its CV was 28%. The second best-fit model, excluding geographical coordinates, included forest floor thickness and mean growing season precipitation as explanatory variables (R(2) = 0.71, CV = 29.5%). The model for predicting mineral soil organic C content was composed of two submodels: (i) organic C concentration of a mineral horizon as a function of its colour and (ii) bulk density of that horizon as a function of its estimated C concentration (obtained from the former model). The R(2) of the model for predicting organic C content in mineral soils was 0.57, and its CV was 29%. The models were used to predict organic C contents in 5547 pedons, dispersed throughout the commercial forest of Quebec and for which basic soil profile description data were available. It was then possible to evaluate and compare mean soil C accumulation in different forest stand types and to construct two maps of soil organic C accumulation in upland forest soils of southern Quebec. The results pointed out that forest floor thickness combined with mineral soil horizon colour, texture class, and pH would be useful sustainable forest management indicators of organic C accumulation in upland forest soils.