Modeling Soil Organic Carbon Pool Weight Associated with Soil Physico-chemical Properties within Glandrood Forest in Northern Iran

Soil organic carbon (SOC) storage estimation with high accuracy is very important due to having such prominent role corresponding to climate change and global warming reduction. Therefore، the current study was carried out in the beech mixed-stands forest of reserved compartment in Glandrood forests، in the northern parts of Iran. Stepwise regression method was used in order to present the SOC pool weight modeling considering the prominent soil physic-chemical properties at 3 different soil depths (0-10، 10-30، and 30-50 cm). The results of ANOVAs showed that exchangeable calcium content، lime and clay percentage (P < 0. 05) and also carbon content، soil organic matter، nitrogen and C/N ratio (P < 0. 01) were significantly different among the three soil depths. Moreover، on average، the total soil carbon stored considered as SOC pool weight in the ecosystem equaled 543. 87 ± 22. 07 t. C ha-1. According to the stepwise regression analysis C/N ratio was the effective content as explanatory variable to predict SOC pool (R2adj = 0. 44; SEE = 4. 8). In the following models، P، N and clay percentage، respectively، were the effective contents which improved significantly SOC pool prediction (R2adj = 0. 64-0. 83; SEE = 3. 8-2. 6). C/N ratio، P، N and clay were included 46. 3، 22. 41، 14. 71 and 9. 44 percent respectively with respect to the variations of SOC pool. Regression analysis indicated that introducing electrical conductivity، pH، and Mg as input variables in the following steps did not ameliorate the precision of SOC pool prediction clearly (R2adj = 0. 85-0. 89; SEE = 2. 5-2. 1). Furthermore، during the modeling and according to the collinearity diagnostic test، the result indicated that the maximum variance inflation factor of the models was less than 10 (VIF < 10)، validating them for application.