Simulation of extreme storm effects on regional forest soil carbon stock

Abstract

This study aims to project changes in soil carbon stocks under different frequencies of storm and climate change scenarios. We calibrated and validated the dynamic process-based soil carbon model “Yasso07” for a wind-throw prone forest area of 1.4 M ha in the State of Baden Württemberg in central Europe, Germany. We fitted climate-biomass models using retrospective climate data and biomass measurements from three consecutive national forest inventories for six forest growth regions. Three IPCC scenarios (RCP2.6, RCP6.0 and RCP8.5), three storm frequencies (10, 20, and 50 years interval), and three post-storm harvest strategies (business-as-usual, full retention, and 50% retention), in combination with a total of 27 scenarios, were applied for projections into the 21st century. We could reduce the uncertainty of YassoBW parameter values significantly by up to 30%, by applying Bayesian calibration, although the absolute value of most parameters did not deviate very much from the original Yasso07 parameters. The projection results showed that forest soil organic carbon (SOC) may lose approximately 30 and 10 t C/ha during the first and the second half of this century, respectively. Three storm frequencies led to a larger range of annual SOC reduction (-0.34, -0.49) t C/ha than climate and harvest strategies (-0.41, -0.42) t C/ha. If no storms occur, the total carbon stock would increase to over 200 t/ha with 258 t/ha under RCP8.5. Considering storm impacts, total forest carbon was reduced from -20 to -90 t C/ha, regarding 10 and 50 year storm frequencies respectively. The largest reduction of forest soil carbon stock originated from the loss in non-solubles (N), followed by acid-solubles (A), humus (H), water-solubles (W) and ethanol-solubles (E).