Paleo-Shorelines Detection And Environmental Evolution Research Of Juyanze Paleo-Lake Basin Based On Remote Sensing

Paleo-lake shoreline is a special landform in arid western China.It is a good indicator of lake water level variations during lake regression.Effective extraction of paleo-lake shoreline is of great significance to global climate change and environmental evolution researches.Juyanze is located in the southeastern part of the Ejina Banner of the Alxa League-Inner Mongolia,China.With a typical dry continental climate,this region is sensitive to temperature and precipitation changes.As a closed lake in arid western China,Juyanze gradually shrank and dried up under the influence of global climate change,preserving distinct remains of multiple paleo-shorelines with an obvious "cardiopulmonary" image feature.Taking Juyanze paleo-lake basin as the study area,fully polarimetric Radarsat 2 image was used to detect and extract the paleo-shorelines by means of various filtering algorithms and polarimetric target decomposition technique firstly.Then,the south part of western Juyanze,the north part of eastern Juyanze and the north-south part of western were used to analyze the relationship between the soil parameters,such as soil water content,salinity and soil surface roughness,at the sampling points and the spectral reflectance.The soil dielectric model with high simulation accuracy was selected to discuss the dielectric properties of the medium in Juyanze lake basin,so as to deeply understand the formation mechanism of the“cardiopulmonary" image features.Finally,based on present SRTM DEM data,GIS spatial analysis tool is used to calculate the area and volume(water storage)of the lake area in historical periods,and the environmental evolution processes of this region were quantitatively reconstructed by the combination of precise elevation data and OSL(Optical Stimulated Luminescence dating)data.The main conclusions are as follows(1)Evaluation results of Radarsat 2 filter window sizes and filtering algorithms in paleo-lake Basin show that the 5x5 filter window is the best in this study.Lee filter edge retention is higher,and image distortion is lower,but the filtering algorithm has poor suppression of speckle noise.Gamma filtering and Frost filtering is capable in smoothing,and the image quality is good,but the image has certain distortion and the texture is fuzzy In general,the non-local mean filtering effect is the best,the median filtering is the worst,and the global threshold method and the wavelet decomposition method are moderate(2)Polarization target decomposition of fully polarizes Radarsat 2 images using Freeman-Durden and Cloude target information decomposition technique.The results show that the spatial information of the false color synthetic image combined by single scattering,bulk scattering and single scattering in Cloude decomposition is more abundant and can clearly show the multi-level continuous or intermittent paleo-lake shoreline formed in the process of retreating the Juyanze Lake.(3)A comprehensive analysis of the relationship of the soil parameters such as moisture,salinity and surface roughness and spectral reflectance show that from the lakeshore to the lake,the moisture and salinity show an overall upward trend;the spectral reflectance of the internal paleo-lake is higher than that of the peripheral region;the spectral reflectance are afrected by surface roughness,salinity,and water,etc.It is believed that the difference of salinity and the surface structure are the main reason for the "cardiopulmonary" image characteristics,and that the local groundwater content has a certain effect on its.(4)The Dobson-S model with high simulation accuracy is selected to analyze the sensitivity of soil moisture and salt content to the real and imaginary parts of the dielectric constant.The results show that the real and imaginary parts of soil dielectric constant are afrected by soil moisture,and the moisture directly determines the real part of the soil dielectric constant.The soil salinity is closely related to the imaginary part of the dielectric constant.Under the condition of low moisture,salt content determines the imaginary part of the dielectric constant.(5)GIS spatial analysis method is used to analyze the evolutionary history of paleo-lake by combining the typical paleo-lake shoreline data and OSL data.It is believed that Juyanze Lake began to appear in the middle and late Pleistocene period.The water level has fluctuated many times and it has reached a peak in the late Holocene,that is,the lake water level is higher than the bottom of the modern lake basin(about 890 meters),about 37 meters.During the 2.0?1.0 ka period,the lake water level fluctuates,about 0.3 ka.After that,the lake in the Juyanze Basin gradually shrinks and disappears.