Predicting the spatial distribution of soil organic carbon stockin Swedish forests using remotely sensed and site-specificvariables

Abstract. The status of the SOC stock at any position in the landscape is subject to a complex interplay of soil-state factors operating at different scales and regulating multiple processes resulting either in soils acting as a net sink or net source of carbon. Forest landscapes are characterized by high spatial variability and key drivers of SOC stock might be specific for subareas compared to those influencing the whole landscape. Consequently, separately calibrating models for subareas (local models) that collectively cover a target area can result in different prediction accuracy and SOC stock drivers compared to a single model (global model) that covers the whole area. The goal of this study was therefore to (1) assess how global and local models differ in predicting the humus layer, mineral soil and total SOC stock in Swedish forests, (2) identify the key factors for SOC stock prediction and their scale of influence. We use the Swedish National Forest Soil Inventory (NFSI) database and a digital soil mapping approach to evaluate the prediction performance using Random Forest modelling calibrated locally for the northern, central and southern Sweden (local models) and for the whole Sweden (global model). Models were built by considering (1) only site characteristics which are recorded on the plot during NFSI, (2) remotely sensed variables and (3) both site characteristics and remotely sensed variables. Local models are generally more effective for predicting SOC stock after testing on independent validation data. Using remotely sensed variables together with NFSI data indicates that such covariates have limited predictive strength but that site specific variables from the NFSI covariates show better explanatory strength for SOC stocks. The most important covariates that influence the humus layer, mineral soil and total SOC stock were related to the site characteristic covariates and include the soil moisture class, vegetation type, soil type and soil texture. Future studies could focus in mapping these influential site covariates which have potential for future SOC stock prediction models.