Modeling the Forest Degradation Degrees of Masal Watershed NO: 12 in Guilan Province, Using Logistic Regression

Planning for management of the forests in the future is not possible without having sufficient information about the forest degradation conditions in the past. The present study was conducted to model the forest degradation levels using logistic regression in the Masal watershed NO: 12. In this study, the factors of slope, altitude, geographic direction, distance from rivers, distance from the road networks, distance from residential areas, population centers and distance from barns as independent variables and different degrees of forest degradation entered into the regression model as dependent variables. Extraction of independent variables is done by using the digital maps of the area and recording the forest degradation degrees is done through terrestrial field work. To determinate the impact of each factor a hierarchical analysis method and to model the factors the logistic regression with five connects functions including: Cauchit, Negative log-log, Complementary log-log, Logit and Probit were used. The statistical tests of similarity and Wald were used to examine the significance of the model and its coefficients. The results showed that the model constructed with the Probit connection function has the appropriate capability in modeling the forest degradation. The results also showed that 20760.32 hectares of studied forests (equivalent to 83 percent of the total forest area) were degraded from very low to moderate levels. As a result, the conservation plans for 83% of the watershed forest areas and the implementation of recovery operations for 17% of the forest area (according to the forest degradation map presented in this study) have a favorable effect on improvement of the forest condition in the studied watershed.