Estimating The Desertification Boundary Of Remote Sensing Using Vegetation Abundance Derived From MODIS Data Based On Pixel Unmixing

Desertification has been defined as a land degradation in arid,semi-arid and dry sub-humid areas resulting from various factors,including climatic variations and human activities,it has caused serious harm to the ecological environment,social economy and human life.Consequently,identifying the desertification boundary(i.e.the desert/non-desert boundary)is essential to understanding the dynamic process of desertification.The objective of this study is to explore the possibility of using vegetation abundance derived from MODIS data to estimate the desertification boundary at a large regional scale.Its significance in application is that it can quickly monitor the process of desertification by the movement of the boundary and improve accuracy of desertification assessment.In this study,the Shanxi-Gansu-Ningxia region was selected as the study area,and MODIS data at resolution of 1000 meters from 2001 to 2015 was chosen as data source.Firstly,the pixel-based vegetation abundance was selected as an important indicator for desertification study,and the cumulative curve based on the vegetation abundance was created,representing vegetation growth status in different land surface conditions.Secondly,the morphological characteristics and gradient changes of the cumulative curves were analyzed,and a certain index that represented the desertification boundary was determined by using the linear fitting method.Finally,in order to further prove the rationality of the method proposed in this paper,field survey data and satellite images were used to show land cover types between different vegetation abundance levels.The results showed that:(a)Compared with the methods based on precipitation data and NDVI,the method developed in this paper allowed the easy and scientific determination of the desertification remote sensing boundary.Based on the extracted pixel vegetation abundance information and using the annual maximum vegetation abundance as the vegetation coverage in the study area,we found that the overall vegetation coverage in the Shanxi-Gansu-Ningxia region decreased from southeast to northwest.(b)The annual cumulative curve based on the vegetation abundance can reflect the growth state and variation characteristics of vegetation in desert and non-desert environments:the vegetation coverage in desert area is low and the vegetation is difficult to grow,the vegetation coverage in non-desert area is high and the vegetation is easy to survive.By extracting the gradient characteristic points of the cumulative curve,we set a method to extract desertification boundary,and using satellite images,the field investigation for further verification.The results show that the cumulative curve gradient method is reliable for the evaluation of desertification.(c)When using the estimated 15%vegetation abundance isoline as the desertification remote sensing boundary,we found that the desertification area shows a general trend of constant improvement with some volatility during the period from 2001 to 2015.Desertification is divided into 3 stages,2001 to 2004 decreased significantly,2004 to 2011 remained stable,and since 2011,it has been significantly affected by human activities.The boundary of the desert-non-desert transitional zone in the Shanxi-Gansu-Ningxia region is expanding and contracting,and the whole transitional zone is moving toward the northwest desert.Generally speaking,the areas with frequent desertification changes are concentrated in the areas where agriculture and animal husbandry are interlaced.Due to the interaction of different surface landscapes and human activities,the area is highly sensitive and fragile.