Retrieval Of Seasonal Snow Depth In Tianshan Mountains Based On Active Microwave Data

Tianshan mountain known as "the water tower in Central Asia",is the birthplace of Silk Road economic belt of water resources.Snow melt water is a valuable renewable water resource in Tianshan Mountain area,but at the same time snowmelt flood is also the main factors which restrict the development of the Tianshan Mountain area economy.Therefore,it is of great practical significance and value to accurately quantify the amount of snow and to find the critical value of positive and negative feedback.The snow depth is the main paremeter used to estimate total snow cover?study on snow situation spatialization?simulate of snowmelt runoff?study on water balance and flood prediction.Therefore,it is necessary to improve the accuracy of snow depth information in order to obtain the information of snow cover in the region.The research content of this paper is to select Sentinel1 A satellite C band data and retrieve the snow depth of the study area by using two rail differential interferometry.Finally,the snow depth data are retrieved for further data assimilation.The basin Juntang lake is located in Hutubi County as the study area because in Juntanghu basin is a typical independent development,and complete a semi closed basin,the basin often outbreak of spring snowmelt flood,therefore suitable for snow depth research.The research methods of this paper include the field experiment ? two pass differential interferometry?the Kriging interpolation method and Ensemble Kalman filter.Field experiment: 18 ultrasonic snow depth measurement sites and 3 artificial stations were set up in the study area.The snow depth data from January 16,2016 to February 26,2016 were measured;At the same time,other meteorological and snow performance data were measured to study the dynamic variation of snow depth;Kriging interpolation method:with the weather and meteorological data,the ultrasonic snow depth data of 47 days were processed by Kriging interpolation method,R2 is 0.7720.The above research lays the foundation for the accuracy verification of snow depth data.Two track differential interferometry: in order toobtain snow depth data with higher temporal and spatial resolution,Sentinel-1A satellite C band data are selected to retrieve snow depth by using two track differential interferometry.Ensemble Kalman filter: Kriging interpolation obtained snow depth data and two rail differential interferometric snow depth data obtained by inversion measurement collection Calman filter,finally obtained the controlled by measured data and the closest to the true value of the snow depth data.The results show that the inversion results are consistent with the measured data,the error is 12 cm.The snow distribution in the study area is largely affected by terrain,because of the barrier of warm and wet air,it is forced to rise and cool down,and the snow cover depth is higher than that in other areas.Secondly,strong coherence in high altitude,no human life interference area of snow depth inversion results is close to measured value,the low coherence low altitude areas and close to the results of city population living area inversion and the measured value of the deviation of the inversion results is obviously low,the total volume of snow depth inversion of R20.7720.Snow depth correction based on Ensemble Kalman filter,the result shows that the snow depth of long time controlled by measured data R2 is 0.8148.From the data source level,the original image quality of the original data source(Sentinel-1A)is higher,which improves the spatial resolution of the final retrieved data;from the aspect of the method,the semi physical model is updated from the geometric point of view.Finally,the calculation time is reduced,and the inversion of snow depth becomes simple.In a word,the depth of snow cover in the study area based on the inversion of D-In SAR has obtained a large area and long time series of snow depth data,and ultimately improve the accuracy of snow depth data.