关键词: 农林交错区, 土地利用, 生境质量, InVEST模型, 地理探测器

Abstract: The ecosystem in agroforestry ecotone of Northeast China is complex and highly disturbed by human activities. Exploring the changes and driving forces of habitat quality will provide scientific guidance for revealing temporal and spatial variations and driving mechanisms of regional ecological environment and guiding natural resource protection and management. Arun Banner, a typical county in the agroforestry ecotone of Northeast China, was taken as the research area. Based on multi-period data of land use of Arun Banner, we quantitatively analyzed the spatiotemporal variations of landscape pattern and habitat quality using the transfer matrix, landscape metrics and InVEST model, and identified the dominant factors for habitat quality by geographic detector method. From 1980 to 2020, the area of farmland increased by 626.36 km², which was mainly transferred from forests, grassland, and marsh land, with the transfer area of 330, 190, and 170 km² respectively. The land use transfer varied greatly in different periods, among which the transfer was the most intense from 1990 to 2000. From 1980 to 2000, the landscape patch density showed a trend from decreasing to increasing, the number of patches decreased, the patch connectivity and concentration became better, and the fragmentation decreased. There was no significant change in landscape pattern index after 2000. Habitat quality decreased from northwest to southeast, with the annual average of 0.794. The area with superior and medium level habitat quality accounted for 52.07% and 32.14%, respectively. The habitat quality showed a declining trend of less than 0.2 from 1980 to 2020. Land use change had a significant impact on habitat quality. The transfer to farmland accounted for 54.42% of habitat quality deterioration, while the transfer to forests accounted for 81.08% of habitat quality improvement. Habitat quality was driven by potential multiple factors. Different potential factors mainly exhibited a dual factor enhancement effect, with the main potential factors being population density and elevation (individual interpretation rate of 51%). Our results could provide a scientific reference for habitat restoration and ecological security pattern optimization of agroforestry ecotone in Northeast China.

Key words: agroforestry ecotone, land use, habitat quality, InVEST model, geographic detector