| Abstrakt: |
Since dead trees contain significant amounts of organic matter, they play an essential role in forest dynamics. In the present study, the changes in soil carbon stock were studied in relation to the dead tree features and its soil beneath in mixed broadleaf stands of Loveh forest in Golestan province, Iran using the Principal Component Analysis (PCA) and random forest model. For this purpose, the characteristics of 15 dead trees consisting of hornbeam, oak, Cappadocian maple, and ironwood were recorded in three decay classes 3 to 5, and the soil beneath them was collected from a depth of 0 to 15. The random forest model was implemented based on the relationship between the dependent variable (carbon stock) and the independent variables (effective factors) using the randomForest package in R software. The results showed that using the studied parameters, the random forest model justifies about 54% of the changes in soil carbon stock. According to the final model, the carbon stock of the soil beneath of dead tree is most affected by the three parameters of C/N ratio, nitrogen (%), and soil pH, so in a general trend with the increase of C/N ratio and nitrogen (%) and the decrease of pH, the amount of soil carbon stock around the dead tree has been increased. The present study's findings indicate that dead trees provide great potential for soil carbon storage, especially when faced with an oldgrowth forest ecosystem. [ABSTRACT FROM AUTHOR] |
