| Snow cover is an important component of the cryosphere,and one of the most active natural elements of the surface.Seasonal snow cover is one of the main sources of mountainous rivers in Northwest China and the precious fresh water resources in arid and semi-arid regions.The accumulation and ablation of snow,along with the storage and release of energy and water,play the very important role in the climate system.Snow water equivalence(SWE),which refers to the vertical depth of water when the snow completely melts.SWE is an important indicator for charactering the snow water resource quantity,and its temporal and spatial variation directly reflects the process of snow accumulation and ablation.As the principal means for earth observation,remote sensing makes estimation of SWE in large-scale,high-frequency and high-precision possible.It is also the only effective means of obtaining the temporal and spatial variation of SWE for harsh mountainous area lack of meteorological data.Synthetic aperture radar(SAR),which can get information inside the snow layers and has high spatial resolution,plays a key role in the quantitative retrieval of SWE in basin scale.Therefore,it has important significance for the water resource management in river basin and regional climate research to retrieve SWE by using SAR technology.This paper is funded by the project named "Retrieval of snow water equivalent based on SAR and high spatial resolution optical remote sensing" supported by the National Natural Science Foundation of China(Grant No.:41271353).According to the project requirement of SWE retrieval in mountainous area,based on the typical study area in Manas River Basin of Tianshan Mountain and the snow simultaneous field observation data,the feasibility of SWE retrieval in mountainous area by using C-band SAR and EQeau(EQuivalence eau)model was discussed.During this process,parameter optimization of EQeau model for complex terrain condition was the focus of this discussion.The main research contents and conclusions are as follows:(1)Acquisition of snow parameters.During the period of Dec.13 to Dec.16 2013,87 test points of snow depth,snow density,SWE and snow thermal resistance were achieved in simultaneous field observation.The observation results show that:the study area was covered by shallow snowpack,and the average snow depth was less than 10 cm;SWE were primarily distributed around 1?2 cm;the mean density of snow was 0.202 g cm-3 and the snow density distribution was concentrated,for more than 86%of observation point values were between 0.154?0.243 g cm-3,and the standard deviation was 0.032 g cm-3;the average of snow thermal resistance was 0.75 m2 K W-1.The acquisition of snow parameters provides the data support for parameter optimization of EQeau model and precision evaluation of the final retrieval results.(2)Parameter optimization of retrieval model.EQeau model converts the snow density and snow depth which determine SWE into the snow density and snow thermal resistance through formulas,and SWE retrieval is achieved by establishing a half relationship between the backscattering coefficient ratio of the winter and autumn and the snow thermal resistance.In this paper,AIEM(Advanced Integral Equation Method)model was used to analyze the influence factors of EQeau model,which are identified as polarization mode,soil surface roughness,local incidence angle,autumn soil moisture,and snow density.The backscattering coefficient ratio increases with the increasing of the local incidence angle,the increasing extent of the ratio under cross polarization mode is up to 18 dB,while the increasing range of the radio of the copolarization mode is only 3 dB.It indicates that the optimal polarization mode of EQeau is copolarization.The backscattering coefficient ratio decreases as surface root-mean-square height increasing,and it increases with the increasing of correlation length.The total influence of the surface roughness is about 5 dB,which cannot be ignored.The backscattering coefficient ratio increases with the autumn soil moisture decreases,and the influence is about 5 dB,which cannot be ignored.Accordingly,this study discussed the method of parameter optimization of EQeau model under the mountainous conditions.To optimize the model parameters in the absence of the surface roughness data and soil moisture data of autumn,the study area can be divided into cultivated land,grass in sunny slope,grass in shady slope with<40° local incidence angle and grass in shady slope with>40° local incidence angle on the basis of land cover,aspect and local incidence angle.Using snow parameters obtained from synchronous observation to fit the model for coefficients revising in each of the four types,and finally compose them to form the optimized EQeau model.The overall fitting goodness of the optimized EQeau model is 0.78,which indicates the fitting degree is sound.(3)Retrieval of SWE.According to the optimized EQeau model expression,the parameters including the backscattering coefficient ratio of winter and fall,local incidence angle were calculated and snow density was interpolated.Then,considering difference aspects,land cover types and local incidence angles,the corresponding EQeau model expression was chosen to retrieve the SWE,and obtained the distribution map of SWE in the study area.Finally,compared with ground data,the precision evaluation and error analysis of the retrieval results were completed.The result of SWE retrieving is of high precision with the mean absolute error of 0.70 cm,which illustrates that the method of SWE retrieving is feasible and effective based on multi-temporal C-band SAR data and parameter-optimized EQeau model.After analyzing the distribution of SWE in the study area,it is discovered that the SWE were generally small,mainly of 1-2 cm and the average was only 2.73 cm in December 13,2013.Aspect is the main topographical influence factor of the SWE distribution.The average of SWE in north slope was maximum,which was minimum in south slope.Aiming at the problem of SWE retrieval in mountainous area,based on the snow simultaneous field observation data,a set of research ideas,which are influence factors analyzation,parameter optimization,coefficients revision for EQeau model,retrieval process improvement,retrieval,and accuracy assessment,were proposed by using C-band SAR data and EQeau model.It has the theoretical innovation.In addition,for the complex terrain and underlying surface condition in mountainous area,a parameter optimization method of adding topographical factors,land cover type and local incidence angle as the EQeau model parameters was proposed by analyzing the influence factors of EQeau model.It improved the fitting goodness of model,and obtained a high precision retrieval result of SWE.The method has significant technical innovation. |
PDF Full Text Request