USING TREE CLUSTERING METHOD FOR FORESTRY ARRANGEMENT PLANNING
Russian forests grow in various climatic and forest zones. According to this, for efficient forest planning, 42 forest areas are defined legislatively. These areas are assumed to have similar conditions for use, conservation, protection, and forest reproduction within their boundaries. Alongside natural and forestry factors, some socio-economic and infrastructural factors need to be taken into account when planning the optimal scope of fire-prevention measures in the forests. In large territories, a significant number of factors and their combinations occur. This prevents to ensure the accuracy while building numerical models for all the processes associated with the emergence, distribution, and suppression of forest fires. We present a new algorithm for optimizing the wildfire prevention arrangement standards, which is crucial for increasing the overall efficiency of forest fire protection when financial resources are limited. The new approach discovers and assesses the similar indicators of investigated forest areas based on aggregated factors and their combinations. It enables to identify abnormal deviations, which are considerably below or above the mean values of the group. For the algorithm implementation, key factors are determined, and forest areas are grouped in obedience to the maximum similarity in factors complex. Considering the research task specifics, determining key factors are the relative number of fires per area, tension of the fire season, population density, transport accessibility of territories, etc. A preprocessed data matrix is used for tree-type clustering of forest areas. The constructed dendrogram is applicable for optimizing forestry regulations, comparative efficiency assessment of forest fire protection system in particular areas, and many other perspective tasks in forestry-related decision-making systems.
1. Center of the forest pyrology, The Branch of the Federal Budgetary Organization “All-Russian Research Institute of Silviculture and Mechanization of Forestry”, Krasnoyarsk, Russian Federation. 2. Sibirskij gosudarstvennyj universitet nauki i tehnologij imeni akademika M F Resetneva, Krasnoarsk, Krasnoârskij kraj, Russian Federation.