Japan Geoscience Union Meeting 2023

Presentation information

[J] Online Poster

A (Atmospheric and Hydrospheric Sciences ) » A-CG Complex & General

[A-CG39] Biogeochemical cycles in Land Ecosystem

Fri. May 26, 2023 10:45 AM - 12:15 PM Online Poster Zoom Room (7) (Online Poster)

convener:Tomomichi Kato(Research Faculty of Agriculture, Hokkaido University), Munemasa Teramoto(Arid Land Research Center, Tottori University), Takeshi Ise(FSERC, Kyoto University), Kazuhito Ichii(Chiba University)


On-site poster schedule(2023/5/25 17:15-18:45)

10:45 AM - 12:15 PM

[ACG39-P03] Environmental Factors Influencing the Emergence of Submerged Landforms by Thermokarst in Eastern Siberia

*Noriko Kurata1, Kosuke Takaya2, Takeshi Ise2, Yoshihiro Iijima3 (1.Yamaguchi Pref. Univ., 2.Kyoto Univ., 3.Mie Univ.)

Keywords:thermokarst, deep learning, satellite image, geomorphology

In the continuous permafrost areas of Eastern Siberia, there is a concern that the land surface is changing due to permafrost thawing as a result of climate change, and that topographic subsidence due to thermokarst is spreading rapidly. However, In addition to climate change, however, other factors contributing to these environmental changes include the expansion of residential areas along with population growth, and land development related to forestry and mining. Thus, the global environmental change of global warming, regional anthropogenic influences such as population growth and land development, and local topography and microclimate are thought to have a combined effect on the emergence of thermokarst. In this study, the location of thermokarst development areas detected by the artificial intelligence model is superimposed on environmental conditions such as topography to examine what conditions influence the appearance of subsidence landforms (polygons) caused by thermokarst. Detection of polygonal landforms was conducted by applying a deep learning method called the "chopped picture method" to satellite images. Data such as slope (whether sunny or shaded side) and topographic undulations are used as environmental conditions. The novelty of this study lies in the automatic detection of polygonal landforms over a relatively large area using deep learning on satellite images and comparing its occurrence with environmental conditions. The combination of satellite imagery and deep learning makes it possible to scan a wide area using a uniform method and is expected to reduce fluctuations in accuracy due to individual differences that are a concern in visual discrimination. By identifying the environmental conditions under which thermokarst is likely to develop, it may be possible to predict areas that are particularly vulnerable to the effects of climate change. Surveys of local residents in Eastern Siberia have shown that there are large differences in perceptions of environmental change among local resident groups. By quantifying complex factors such as climate change, local environmental conditions, and anthropogenic influences, it is hoped that in the future this information will be applied to comprehensive environmental measures, such as developing educational activities for local residents regarding environmental change.
This study was supported by the Arctic Challenge for Sustainability Research Projects II (JPMXD1420318865) funded by the Ministry of Education, Culture, Sports, Science, and Technology of Japan.