Abstract:

Bio-crust is a widely distributed ground cover plant in arid and semi-arid regions， which plays an important role in controlling soil erosion and regulating the stability of the ecological environment in the region. Currently， the study of bio-crust mainly focuses on the observation of sample area scale， and the large-scale extraction method based on remote sensing technology has significant deficiencies， which， to some extent， hinders the exploraition of the spatial and temporal changes of bio-crust and their ecohydrological effects. In this study， we take Huangfuchuan， a typical drainage basin in the arid area， as the study area， and consider the complexity and spatial and temporal heterogeneity of the ground cover， and use the synergistic combination of multiple remote sensing indices to construct a remote sensing model based on step-by-step extraction of bio-crust. The method uses a stepwise identification model to extract bare soil， vegetation， and bio-crust pixels， develops 30 m-resolution spatial distribution datasets of bio-crust in the Huangfuchuan drainage basin in a typical arid area， and explores the effects of topography， climate， and land use practices on the spatial distribution of biocrust. The results show that there were significant differences in the thresholds between different vegetation index in the stepwise extraction model of bio-crust in arid regions constructed based on the soil adjusted vegetation index （SAVI）， Modified soil adjusted vegetation Index （MSAVI）， Normalized difference vegetation index （NDVI）， Near-infrared reflectance of vegetation （NIR_V）， and biological soil crusts index （BSCI）. （SAVI<0.13 and ≥0.19； MSAVI<0.11 and ≥0.17； 0≤NDVI<0.44； 0≤NIR_V<0.11； 5.89≤BSCI<7.28）， but the overall accuracy of the model to extract the bio-crust in the arid regions was reliable. The spatial distribution data of bio-crust in Huangfuchuan drainage basin in 2021 were developed， and it was found that the bio-crust in Huangfuchuan drainage basin was widely spread， and the area of bio-crust like element accounted for 21.51% of the total area， and the bio-crust was mainly distributed in the area of grassland， cropland， and other land-use types； There is a high correlation between the spatial distribution of bio-crust and the magnitude of precipitation in each region of the study area， and the larger the precipitation， the more bio-crust are distributed. The focus of this study is to develop a large-scale remote sensing extraction method of bio-crust in arid and semi-arid areas， and to lay down a research paradigm for the study of bio-crust in the Huangfuchuan drainage basin and arid and semi-arid regions， with a view to providing scientific references for the study of ecological and hydrological cycles in the basin.

Key words: bio-crust, spatial distribution, Huangfuchuan drainage basin, multispectral remote sensing, progressive extraction