Abstract: Farmland shelterbelts are an important part of the Three-North Shelterbelt Project. Detecting changes in the network structure of farmland shelterbelts is crucial for assessing the project’s construction and shelter effects. We proposed revised landscape index that considered the characteristics of farmland shelterbelts and remote sensing data to analyze the spatiotemporal variations in their network structure. Our evaluation method was applied to a typical area of shelterbelt construction in the north of Changchun. The results showed that the revised landscape index calculation was simpler and more suitable for medium and low-resolution remote sensing data. By dividing the research area and calculating the difference in landscape index between real and ideal conditions, the spatiotemporal variations in the farmland shelterbelt network structure could be effectively evaluated from three aspects, including the number of shelterbelts, their development status, and distribution. From 1985 to 2020, the structure of the shelterbelt network in the selected areas tended to be optimized, although closed-formed shelterbelt networks were less frequent and only about 14% of the shelterbelt networks graded as excellent in circularity degree. Our method provides a novel approach for evaluating the dynamics of shelterbelt network structure and can serve as a reference for the sustainable management of shelterbelts at the regional scale.

Key words: farmland shelterbelt, shelterbelt network structure, landscape index, remote sensing, spatiotemporal variation