Modeling of Green Canopy Cover of Marjan Rangelands, Boroujen During Growing Season Using Spectral Indices of OLI Sensor

Vegetation is considered as an important measuring indicator in rangelands, which play a significant role in ecological processes. In this regard and due to the lack of models in estimating green vegetation canopy, this research aimed to evaluate temporal changes of canopy cover of vegetation in Marjan rangelands of Broujen. As the completion of the past investigated studies in other areas, it is tried to present the growth stages and time effect on destroyed green canopy cover through a model which has an appropriate estimation power throughout the growing season and for a time such that the education sample do not exist. To this end, the canopy of green vegetation was measured in 19 sampling points over a length of 10 km transects. In each of the points, there are about five quadrats as one quadrat in center and four quadrats at the four directions around the central quadrat (a total of 95 quadrats in a period). Sampling points were placed with distance of 400-1000 meter apart from each other. Measurements were repeated during 4 periods of field operations (in total 380 quadrats). Thus regarding to the index type (Atmospheric Resistant Vegetation Index and the moderating effect of soil line) 12 vegetation indices due to the use of multi-temporal images and the effect of soil reflectance were calculated in this arid region. In the next stage, the model of estimating green plant canopy was prepared using the regression equations between obtained canopy cover from five sampling periods and derived vegetation indices of Landsat 8 image data. Thus this model was also generated to prepare the vegetation cover map to other 4 periods which have not education data. Finally canopy cover map of vegetation was prepared at the 8 periods during the growth season (April to mid-September). Using image differencing technique, the changes of the vegetation coverage of area was estimated. Validation of cover estimating model represents no significant difference between estimated cover and land cover data. It is concluded that achieving to the growth model which can be used for all seasons are possible by using Landsat 8 images. The results showed that the indicators of ARVI, SARVI and EVI (with coefficients of determination about 8.0) have the best correlation relationship with canopy cover. So not only there is a significant relationship between vegetation indices and vegetation in arid areas, also it is possible to prepare only a unique data model for estimating all different time of growing seasons. According to the results of this research, the vegetation is in its peak on May among the eight growth times in the Mrjan rangelands and the canopy cover has heterogeneous distribution. In terms of coverage changes during the growing season, most of area is categorized in small variations class. In the period from April to mid-May we are witnessed for increased class in the region, but after this period, reduction in area were the dominant phenomena, as well as amplification of increased class area from 3 to 19 September is related to the emergence of new autumn species in the highlands region.