Study On The Relationship Between Phenology And Soil Salinization Based On Spatial And Temporal Sequence Reconstruction

Soil salinization is one of the major reasons for the low yield of farmland crops in the inland irrigation area of Northwest China.Removal of natural factors such as climate,water and heat,irrational irrigation methods and land development and utilization patterns of human factors are the main reasons for the increase of soil salinization.Xinjiang as an important grain reserve area in China,has a long way to go in the sustainable development of cultivated land resources.At present,the research on soil salinization in the inland irrigation area of Northwest China is very rich,and the observation of soil salt is the focus of research.Salinization classification and mapping technology has been developing rapidly in the background of rapid enrichment of remote sensing satellite data both at home and abroad.However,the commonly used indicators of salinization are widely used in the salinity index,and the applicable conditions are not unified,so the reliability is still to be improved.Phenological parameters extracted from the remote sensing image throughout the year of the vegetation index sequence,reflecting the growing season of vegetation law.Salinization,as the main hazard factor affecting the crop and vegetation growth in the irrigation area,has salt stress on the growth of vegetation,so the phenological parameters of vegetation can indirectly indicate the soil salinization.The application of vegetation phenology further excavates the application space of vegetation index,the peak of vegetation index extracted from phenological curve,the time point of occurrence of growth peak and the time point of the end point and the growth characteristics of growth peak are now evaluated in the evaluation of crop productivity and vegetation type classification Some have been widely recognized for research results,but there is little research combined with soil salinization.In recent years,some scholars have tried to study the relationship between salinization and vegetation phenology in the Yellow River Delta,and found that the phenological parameters in the farmland can explain more than 80% of the soil salinity,which explores the relationship between salinization and phenology in the inland irrigation The study provides some research basis.But the current limitation is that the spatial resolution of the phenological results calculated using MODIS data is too low.This is because the satellite load can not take into account both the spatial resolution and temporal resolution,MODIS data space resolution is only 250 m,for the mesoscale study area which will cause the sampling point data can not accurately match the remote sensing reflectivity data There is a lack of spatial detail resolution in the salinization assessment results.Based on this,this thesis combines the Landsat with MODIS images with StarFM fusion algorithm to reconstruct the vegetation index time series data with spatial resolution up to 30 m.The phenological parameters of the 30 m spatial resolution were obtained from the TIMESAT package.In order to obtain more accurate phenological parameters,use the image quality control file provided by MODIS to set the time points of each time point data to participate in the smoothing calculation of the weight setting to eliminate or reduce those affected by cloud,snow,shadow pollution caused by the impact of the data.At the same time based on the classification information to set different types of filtering smoothing parameters.Finally,the relationship between phenological parameters and salinization information was studied in SPSS and JMP software.A linear fitting model of a variety of phenological parameters and measured salinity data was established and verified by statistical analysis.The conclusions of this study are the following three points:1)By observing the statistical analysis of the phenological parameters extracted from the fusion image and establishing the measured soil salinity data and phenological parameters,the results show that there are negative correlations with the soil salinity And the seasonal integral index is the best,and the coefficient of linear regression model is 0.72.2)Based on the existing study,the difference between the vertical distribution of soil salinity and the two representative phenological stages(peak growth period and growth start period)was considered.By analyzing the correlation coefficient and establishing the extracted phenological parameters and soil conductivity Linear regression model.In the case of more than 95% confidence level,the statistical results show that the fitting effect of the beginning of the growing season is better than that of the growing season.During the growing season,0-10 cm,10-20 cm and 20-40 cm soil depth,Phenological parameters can show the role of indicators of soil salinity.During the beginning of the growing season,the soil salinity prediction capacity was observed at 0-60 cm soil depth.3)Using the StarFM algorithm to reconstruct the fusion data based on Landsat8 data and MODIS data,The spatial resolution of the generated phenological parameters of the vegetation is increased from 250 m to 30 m,which makes the sampling data more accurate with the final vegetation phenogram.So that the accuracy of statistical analysis can be improved,and the future use of phenological properties involved in the mapping of soil salinity is of great significance.