Research On Vegetation Phenology Changes And Scale Sensitivity In The Yangtze River Basin Based On Multi-spatial Resolution

In the context of global change,the impact of global climate change on the environment is increasing.Vegetation is the connector between climate and environment.Changes in vegetation phenology can profoundly reflect changes in climatic factors and are the most intuitive response of terrestrial ecosystems to environmental changes.In addition,in the process of remote sensing vegetation phenology research,vegetation shows different phenological characteristics and changing rules under different spatial resolution data,so the scale sensitivity of vegetation phenology is studied.The paper uses MODIS NDVI data and dynamic threshold method with spatial resolutions of 250m,500m,1km,and 8km to simulate and extract three key phenological parameters of vegetation growth season in the Yangtze River Basin from 2001 to 2016,including the beginning,end and length of the growing season.At the same time,Landsat 4-5TM data was used to interpret the three representative vegetation types and their distribution information of Pinus massoniana,dryland and deciduous broad-leaved and evergreen broad-leaved mixed forest.Based on the distribution information of vegetation types in the Yangtze River Basin,the phenological grid maps of the beginning,the end and the the length of the growing season of three vegetation types in the past 16 years under four spatial resolutions were extracted from the simulated grid maps of the phenological parameters.The phenological characteristics and spatial pattern of the three vegetation types under different spatial resolutions were studied.The trends of the three vegetation types in the past 16 years were analyzed under different spatial resolutions,and the differences of the trends in different spatial resolutions were compared.Finally,based on the CA-Markov model,the scale sensitivity of vegetation phenological trend change is explored and the accuracy of the simulation results is evaluated.This paper has a certain reference value for the study of ecological environment in the Yangtze River Basin.The main conclusions are as follows:(1)Study on phenological characteristics of vegetation.The results showed that the growth season of Pinus massoniana,dryland,deciduous broad-leaved and evergreen broad-leaved mixed forests in the plains of the Yangtze River Basin and the surrounding low-altitude mountain areas started earlier,and later in the Western Alpine mountain areas and the high-altitude mountain areas.At the end of the growing season,it is earlier in low mountains and plains,but later in rugged hills and mountains.The length of the growing season is shorter in mountainous hills and plains with lower altitudes and longer in high altitude mountains.(2)Study on Phenological Trend Change of Vegetation.The results show that the main trends of the three vegetation types under different spatial resolutions are the early beginning of the growing season,the delayed end of the growing season and the extension of the length of the growing season,which are concentrated in the mountains with high altitude or rugged terrain.(3)Study on scale sensitivity of vegetation phenological trend change.The results show that the trend characteristics of Pinus massoniana,dryland and deciduous broad-leaved and evergreen broad-leaved mixed forest decreased with spatial resolution in the scale of 250m,500m and 1km,the simulation accuracy of CA-Markov model also shows a decreasing trend,that is,with the increase of cell size,the simulation accuracy decreases gradually.Moreover,there is a cell-scale threshold in the model for different vegetation types.When the threshold is exceeded,the simulation precision drops quickly.When the cell size increases to 8 km,the Kappa coefficient of simulation accuracy increases,but there is a big difference with the actual effect.Therefore,we should pay attention to the existence of cell size limit in the study.