It is a challenge for landscape ecology research to apply the principles of landscape ecology to the ac-curate and rapid assessment of desertification status(DS).Based on the land use type data in the sandy area of the central Naiman Banner in 2000 and 2018,the maximum patch index(LPI),boundary density(ED),aggre-gation index(AI)and Shannon diversity index(SHDI)were selected,and the influence of two-factor interaction and grading of landscape pattern on DS was analyzed with the supports of Geo-detector Model and Moving Win-dow function in Fragstats software.DS analysis showed that the desertification process was reversed from 2000 to 2018 in the study area because the DS is reduced from 0.393 to 0.288.There are two forms of two-factor inter-action in landscape pattern:nonlinear enhancement and two-factor enhancement,and the influence on DS was greater than that of each factor.There were differences in the influence of different pattern factors on DS,and the differences reasons were related to the area size and distribution pattern of different dune types.In the years when DS was strong(2000),mobile dunes predominated(30.6％-42.9％)in the areas with significant DS caused by LPIand ED classifications,while fixed sand dunes predominated(32.0％-44.0％)in insignificant areas.In the years when DS was weak(2018),both significant and insignificant DS areas caused by SHDI classification were dominated by mobile dunes(both over 79.2％).This study shows that the influence of two-factor interaction of landscape pattern on DS is related to both temporal and spatial scale,and the influence mechanism of interaction on DS has nonlinear characteristics.The study area size directly affects the interaction form and spatial pattern of landscape factors(scale dependent).The results of any typical case study should't be deduced across scales,but the methods can be unified.
geo-detector modelmoving window methodlandscape patterndesertification statusKerqin Sandy Land
NETL
NSTL
万方数据
