Геоэкологическое картографирование и оценка хозяйственно освоенных территорий Забайкальского края, подверженных наводнениям

Floods are dangerous natural phenomena that severely impact on the economies and significantly reduce the quality of life for affected population. According to statistical data, the number of these phenomena in Russia has a noticeable upward trend. The consequences of past floods in Transbaikalia highlight a critical issue: when there’s a prolonged absence of flood treats, territories are often developed and built upon, leading to severe when floods inevitably occur. To develop preventive measures and mitigate to reduce the negative impact of floods on vulnerable areas, space-based monitoring of the Earth's surface is currently being used to monitor, assess and forecast possible flood damage, followed by mapping. However, the major challenge lies in the insufficient effectiveness of existing mapping methods and underestimation of risks in territorial planning during long periods of absence of floods. Current methods of automatic decryption of infrastructure and buildings have low accuracy and performance, and the method of visual decryption is time-consuming. The purpose of the study is to develop an operational methodology for mapping and identifying urbanized and residential areas of the Trans-Baikal Territory that are prone to flooding. The methodological basis is a comprehensive analysis of temporary changes in the ratio of land areas, built-up areas and flood zones in settlements of the Trans-Baikal Territory using remote sensing data processed using artificial intelligence algorithms (convolutional neural network). The results of the study demonstrate a steady trend of increasing the proportion of buildings in flood zones in a number of analyzed settlements in the Trans-Baikal Territory. The developed geoecological mapping technique has proven effective and can be scaled to areas with pre-existing flood zones, as well as in the absence of hydrological observations. Creating maps of the economic development within flood-prone areas enables monitoring of buildings in hazardous settlement zones. This is particularly crucial given that manual and visual decryption methods are ineffective for such tasks.