<![CDATA[Since the establishment of the Bogor Botanical Garden (BBG) in 1817, the protection of the tree collections, even the loss of aging trees (> 100 years old), has been one of its most important tasks. Abiotic factors such as intense extreme events, i.e., heavy rainfall and strong winds, as well as biotic factors from human activities, pests and diseases, and the deterioration of the health of the plant collection with age, has threatened the survival of the old tree collections. As the BBG has many functions for conservation and human ecological activities, tree fall accidents have become a primary concern in preventing the loss of biodiversity and human life. Therefore, disaster map zonation is required to prevent and minimize such accident together with a prediction of which individual specimen is likely to fall. We examined the health of 154 to determine the falling probability of 1106 aged trees based on several factors that might cause the fall in the past and to make model predictions generated by nine supervised machine learning algorithms to get a binary value of falling probability and then classified into four categories (neglectable, low, moderate, and high probability of falling). Inverse Distance Weighted interpolation method was used to depict a zone map of trees prone to fall in BBG. We found 885 susceptible trees, of which 358 individual trees were highly susceptible to fall (red zone color), dominated by families from Fabaceae, Lauraceae, Moraceae, Meliaceae, Dipterocarpaceae, Sapindaceae, Rubiaceae, Myrtaceae, Araucariaceae, Malvaceae, and Anacardiaceae. This result was based on Random Forest model due to its highest accuracy among algorithms and its lowest false negative (FN) value. The FN value was important to minimize error calculation on aged trees that were not prone to fall but turned out to be prone to fall. The dominant factor contributing to high falling intensity was hollow and brittle on the tree trunks where many were found to have pests inside damaged parts such as termites, wood-borers, and bark-eaters. Several trees were found to have combined damages with more than a single causative factor that exacerbated tree’s health and increased falling probability.]]>
