[M-IS01] ENVIRONMENTAL, SOCIO-ECONOMIC, AND CLIMATIC CHANGES IN NORTHERN EURASIA

[MIS01-P14] Landform effects on vegetation anomalies over Mongolian grasslands

*江 嘉敏^1,2,3、Nandintsetseg Banzragch^3,4、篠田 雅人³ (1.理化学研究所、2.香川大学、3.名古屋大学、4.鳥取大学)

キーワード：Mongolia、Grasslands、Vegetation anomalies、Landforms and Landscapes

In the drylands, pastoralists need to move extensively their livestock several times a year in search of forage across different landforms. This migration behavior must thus be linked to the vegetation dynamic variations at a regional scale. To investigate the effects of landforms on the spatial patterns of vegetation anomalies over Mongolian grasslands, a satellite-based monthly normalized difference vegetation index (NDVI) at the beginning and the end of the growing season (April and August, respectively) during 2000—2021 was used. We selected 96 target sites over Mongolian grasslands. The topographic position index (TPI) with a neighborhood size of dimension 3×3 was calculated by the global elevation dataset (GTOPO30) to distinguish topographic features (e.g., valley bottom, mountain ridge, and flat plain). The interannual vegetation anomalies in the four directions of these target sites were identified using the spatial correlation analysis, and a homogeneity distance was estimated by a distance of the same sign of vegetation anomalies with a significance level of 5%. A principal coordinate analysis with Bray-Curtis dissimilarity was implemented to reveal that the distinct cluster of spatial homogeneity splits into the southern area (43°E—46°E) and the northern area (47°E—50°E) in April, however no spatial difference showed in August. This is because the homogeneity distances in April are significantly longer in the southern area than in the northern area, but the distances in both areas are comparable in August. TPI results showed that its standard deviation in the northern area is significantly higher than in the southern area when the elevation and its standard deviation have no spatial difference. This implied that the spatial changes in landforms in the northern area are more frequent than in the southern area. Thus, we assumed that, in the northern area, due to topography with greater relief, changes between snow on the top of the mountains and sparse vegetation in the valleys induce a shorter homogeneity in April. Higher precipitation in this area during the growing season caused an increase in vegetation over the entire landscape including the peaks and valleys, producing a longer homogeneity in August. In the southern area, although a flat landscape leads to longer homogeneity in April, lower precipitation limits vegetation growth, resulting in a similar homogeneity in August.