Preparation of Distribution and Health map of Afforestation in Langroud County using Sentinel-2 Sensor Images and Field Data

Background and aim: Monitoring the current status of existing afforestation’s in management decisions is very important for the development of afforestation’s in the future. This study was conducted in order to monitor the area, distribution and health of afforestation’s in Langroud County, Guilan province. Sustainable management of current and future annual afforestation’s requires studies on the status of the afforestation’s with focusing on their health, and healthy afforestation’s can have greater environmental performance compared to unhealthy ones. The aim of this study is to prepare a map and survey of the areas of afforestation’s in Langroud County using ground surveys, Android GPS software (GFAMP), and Google Earth for the protection, enrichment, and development of afforestation’s in the future. The health of the stands is also assessed using Sentinel 2 sensor images and vegetation cover indices of SAVI, TNDVI, NDVI, and RVI.
Materials and methods: For this purpose, first, field surveys were done in the form of land points from the existing afforestation, and the distribution map of afforestation was prepared using the land surveys, GPS Fields Area Measure PRO application (GFAMP), and Google Earth system. Then the Sentinel 2 sensor image related to the growing season in Langroud County was prepared from the Copernicus site. From the Sentinel 2 sensor images, various vegetation indices such as NDVI, TNDVI, SAVI and RVI related to the growing season were extracted and their maps were prepared in the area of afforestation. In the following, the amount of each vegetation index was extracted at the points of land harvesting and the correlation of the values of each vegetation index (resulting from images of the growing season of Sentinel 2 sensor) with the health of afforestation (resulting from field harvesting) was investigated. For this purpose, Pearson's correlation coefficient was used. Then, the index that showed the highest correlation with the health of afforestation in Langroud County (NDVI) was selected as the most important index to estimate the health of afforestation and its regression relationship with the health of afforestation was obtained. In the following, using the map of the most favorable vegetation index, information of field survey and the relationship between these two cases, a health map of afforestation in Langroud County was prepared.
Results and discussion: Based on the findings of this study, a total of 66 afforestation plots were identified in this county that the total area of these plots was obtained using the 100% inventory method and ground control as 746.2 ha, which are mainly distributed in the southwest of the county. In addition, the results showed that the NDVI Index is the most favorable vegetation index for estimating the health of afforestation in Langroud County, which indicates the ability of this index to assess the health of trees during the growing season. After that, the SAVI index showed the highest correlation with tree health, which also shows the good ability of this index in monitoring tree health. While the two indices TNDVI and RVI showed a much lower correlation with tree health within the Langroud afforestation area, their use for assessing tree health in future studies is not recommended. Of the current total area of afforestation, 400.56 ha are in full health status, 305.60 ha are in medium health status, and 40.04 ha are in unhealthy status. The overall accuracy (OA) of the afforestation’s health map in this study was 80 %, the producer accuracy was 79 %, the user accuracy was 78 %, and the kappa coefficient was 0.73. These results indicate the optimal classification of the afforestation’s health map into different tree health and vigor classes. In general, the afforestation’s in this county are mostly in perfect health, but some areas of Kachlebon, Larzian, Gandom Bijaran, Ghazi Dasht, Khorsh Sara, and Chaksar (Chakdasht) are less successful due to the lack of conservation operations and the lack of guards in the early years.
Conclusion: The results of this research will help the forest managers for quantitative and qualitative monitoring of afforestation and its continuity in certain time periods as well as future afforestation development plans. The presented regression model, based on the high correlation of the NDVI index with the health of afforestation, allows for rapid estimation, quantitative, low-cost, affordable, and economic assessment of the health status of afforestation on a large scale and in inaccessible areas. The results of this research, by understanding the current status of afforestation, provide a good view of the capacity and potential of Langroud County for afforestation, considering the environmental and ecological conditions of this county.