Identification of Damage to Coniferous Stands Based on Comprehensive Analysis of the Results of Remote Sensing of the Earth and Ground Surveys

Damage to forest stands and their death as a result of outbreaks of destructive insects and diseases occur over significant areas and represent an important forestry problem. To solve this problem, it is advisable to use materials from remote sensing of the Earth, since damaged and dead forests can occupy significant areas, from remote еach other. The article presents a methodology for assessing the state of forest stands and the impact of forest pests and diseases on them based on the joint processing of data from a regular grid of sample plots and materials from remote sensing of the Earth via geoinformational technologies, variance and regression analyses. The sample plots have been laid according to the ICPForests methodology and have characterized the general background of the state of the forests in the study area of the North-Western and central parts of the Leningrad Region, where the foci of forest pests and diseases reproduction have been identified at the same time. The applicability of 2 vegetation indices – normalized difference (NDVI) and short–wave (SWVI), as well as the normalized difference moisture index (NDMI), the normalized burn ratio (NBR) and the spectral characteristics of the corresponding channels of the Sentinel-2B image to identify the foci of forest damage by pests and diseases has been assessed. Statistically significant relationships between the state of the stands and the values of the listed indices have been established. It has been shown that the foci of pests and diseases reproduction are reliably detected using the materials from remote sensing of the Earth against the general background of the state of the forests represented by the regular grid of sample plots. A thematic map of the stands has been developed dividing them into 2 groups of state classes: healthy and weakened stands; suppressed stands and deadwood. Its accuracy was assessed via the support vector machine (SVM) on the basis of the combination of ground and remote sensing data. To assess the adequacy of the automatic classification, the error matrix and the calculation of the Cohen’s Kappa coefficient have been used. The coefficient has turned out equal to 0.878, which indicates a high quality of the classification. It has been shown that the developed thematic map is applicable for detecting potential foci of forest pests and diseases reproduction in the study area.