Evaluation and Prediction of Decline of Oak Forests in Middle Zagros (Lorestan Section) with a Climate Change Approach

The oak trees decline in Zagros forests is one of the main problems of the ecosystem of this region. Therefore, in this research, estimation and prediction of oak forests decline in the middle Zagros (Lorestan section) were performed to evaluate the RS data and decline indices including DVI, GEMI, IPVI, MSAVI2, NDVI, EVI, OSAVI, Sarvi2, SR and VHI. To this end, these indices were extracted using ArcGIS for 110 sites of oak forests. Finally, a matrix of 345 × 110 dimensions during 2001 to 2016 was created. Then, fuzzy and neural network methods were used in the prediction strategy, and the method with less error was considered as the prediction basis for the period 2017-2030. The cluster analysis and the discriminant analysis were also used to identify the decline areas, and spectral analysis was used to investigate the oscillation. The results showed that the decline indices had an increasing trend in the most cases. The results of cluster analysis also revealed that oak forest decline can be divided into four areas (no decline, normal decline, high decline and very high decline areas). Investigating the fluctuations of the dominant areas also indicated short-term cycles of 2 to 4 years fluctuations on drying indices. Neural network evaluation and fuzzy method also demonstrated that neural network method is a more suitable method for predicting decline indices. The prediction results indicated that in the future, most of the decline indices in the region will be increased.