Remote Sensing Monitoring And Drivers Analysis Of Biocrusts And Vegetation In Mu Us Sandy Land

Mu Us Sandy Land is a typical arid and semi-arid desert area located in the core of the agricultural and pastoral ecotone in northern China.Biocrusts and vegetation Mosaic distribution in Mu Us Sandy Land coexist and interact.Both habitats are relatively fragile,sensitive to changes in environmental factors,and strongly affected by human activities.In the context of global warming,combined with the implementation of a series of ecological and environmental protection projects led by the Chinese government,the surface landscape of the region has undergone significant changes and vegetation recovery is obvious.However,this area is still a fragile area,and the biocrusts and vegetation on the surface show obvious dynamic changes with the change of environmental factors.Biocrusts is a surface soil community formed by mosses,lichens,bacteria,fungi,algae and cyanobacteria,which plays an extremely important role in the carbon and nitrogen cycle of dryland ecosystems and the retention of surface soil.In recent years,the recognition and monitoring of biocrusts by remote sensing have received widespread attention,and some exploratory studies have been carried out,but the spatial distribution and change dynamics of biocrusts at the regional scale are rarely studied.In order to explore the dynamic changes of biocrusts and vegetation in Mu Us Sandy Land,this paper collected years of Landsat 8 OLI multispectral data,Sentinel-2A multispectral data and MODIS data,and extracted biocrusts based on spectral index method and mixed pixel decomposition method.The type,spatial distribution and dynamic changes of biocrusts in recent 10 years were monitored by remote sensing.Based on vegetation coverage,leaf area index and net primary productivity,the vegetation changes in Mu Us Sandy Land were monitored and the vegetation restoration in recent 20 years was studied.A number of factors affecting the distribution of biocrusts and vegetation were selected,such as meteorological,topographic,and cultural factors.The driving factors were analyzed by using geographic detector model and random forest model to determine the main driving factors of biocrusts and vegetation change.The main research results are as follows:1)It is more stable to identify biocrusts using remote sensing images from non-vegetation growth period(March-May).In 2020,the area proportions of three biocrusts in Mu Us Sandy Land were lichen crusts(21.75%)> moss crusts(18.36%)> algal crusts(15.46%).The dominant biocrusts were lichen crusts.The main land cover types in Mu Us Sandy land are grassland and sparse vegetation.The coverage of algal crust,lichen crust and moss crust was12.85%,12.01% and 12.85%,respectively.The coverage area of biocrusts changed greatly in recent 10 years,with the largest distribution area in 2015 and 2021,and a small distribution area in 2013,2018 and 2022.The algal crust was mainly distributed in the grassland in the northwestern region of Mu Us Sandy Land,and a few were distributed in sparse vegetation and shrubs.Lichen crust and moss crust are mainly distributed in the southwest and central parts of Mu Us Sandy Land.2)The biocrusts showed obvious spatial fluctuation characteristics during the 10-year period.The mean coefficients of variation CV of algal crust and lichen crust were 0.0181 and0.0967,respectively,which were relatively stable,but the stability had significant spatial fluctuations: The low fluctuation(CV< 0.15)was distributed throughout the Mu Us Sandy Land,while the high fluctuation(0.15< CV< 0.2)and high fluctuation(CV > 0.2)were concentrated in the waters around the Gushui River,Wuding River and Duwei River,and the areas where rain and snow mingled.The mean CV of moss crust was 0.3848,the high fluctuation area accounted for 78.7% of the Mu Us Sandy Land,and the high fluctuation area accounted for 10.3%.The low fluctuation area accounted for only 3.9% of the whole study area.The spatial distribution of lichen crust obtained by BSCI may contain more algal crust.Based on the spatial superposition of algal crust and lichen crust obtained by CI and BSCI,it can be seen that the variation of overlapping area over time is very similar to the spatiotemporal variation of algal crust area proportion.3)In the Mu Us Sandy Land,the vegetation restoration effect was obvious in the past 20 years.The spatio-temporal dynamics of vegetation indicators(vegetation coverage,leaf area index,net primary productivity)tended to be consistent and gradually increased,and the increase rate became larger after 2011.The annual average growth rate of vegetation coverage was 0.59% per year.The vegetation coverage in different seasons was successively from high to low: summer,autumn,spring and winter,and the change rate was the fastest in summer.The spatial distribution characteristics of vegetation index were higher in the eastern and southern margins,and lower in the northwestern and central regions.The areas with the highest vegetation coverage and the most obvious restoration were located in the eastern and southern margins of Mu Us Sandy Land(in Shaanxi Province).4)The comprehensive interpretation ability of natural and socioeconomic factors on vegetation coverage in Mu Us Sandy Land was 58.38%.Land cover type,vegetation coverage,distance from river system and GDP of grid were the main driving factors affecting biocrusts.Land cover type and aridity index had the most significant effect on vegetation distribution,followed by precipitation and GDP of kilometer-grid.The two methods(factor detector method and random forest method)obtained highly consistent ranking results of factor importance.In terms of natural factors,higher vegetation coverage was positively correlated with precipitation,temperature,humidity and drought index,but negatively correlated with altitude.In terms of human factors,GDP is positively correlated with vegetation coverage.Among the land cover types,forest land had the highest vegetation coverage and bare land the least.The growth and restoration of vegetation are affected not only by natural conditions but also by regional economic development and land use.In this paper,remote sensing monitoring of biocrusts and vegetation in Mu Us Sandy Land was carried out.The temporal and spatial distribution characteristics of biocrusts and vegetation in Mu Us Sandy Land were revealed,and the main driving factors affecting biocrusts and vegetation coverage were identified.The research results can provide a powerful theoretical basis and technical support for the remote sensing monitoring of biocrusts and vegetation and ecological restoration on a large scale.