| Snow is an important parameter in the global climate and hydrological system,so it is very important to obtain the snow depth information timely and accurately.With the rapid development of global satellite navigation systems,Global Navigation Satellite System Reflectometry(GNSS-R)technology has gradually become a research hotspot.GNSS-R technology is widely used in the inversion of soil moisture,sea surface wave height,sea surface wind speed,snow depth and other parameters due to its advantages of multiple data sources,low cost and high spatial and temporal resolution.In recent years,the research on snow depth inversion based on GNSS-R has not paid enough attention to the low data utilization rate and the loss of inversion accuracy caused by complex terrain.Based on the signal analysis method,this thesis studies the data quality and inversion accuracy of the reflected signal,and further considers the influence of terrain errors,and integrates multi-system GNSS data to study the high-precision inversion of GNSS-R snow depth.The main research contents are as follows:(1)The method of data quality analysis of reflection signal is studied.In this thesis,wavelet decomposition method,complementary ensemble empirical mode decomposition(CEEMD)method and variational modal decomposition(VMD)method are introduced to extract the signal-to-noise ratio(SNR)reflection signal,and the snow depth inversion experiment is carried out respectively.The experimental results show that the three methods can solve the problem of multi-frequency peaks in the spectrum analysis of signals processed by traditional methods to a certain extent,improve the utilization rate of data and improve the accuracy of snow depth inversion,and the CEEMD method is the best.(2)Based on the trajectory clustering algorithm,the SNR data of GPS,GLONASS,GALILEO and BDS at P351 station of PBO(Plate Boundary Observatory)are used to invert the snow depth of azimuth clustering.The accuracy of GNSS-R snow depth inversion in different azimuth intervals of multi-system and multi-band is analyzed,and the accuracy of multi-frequency and multi-system fusion inversion is compared.(3)Aiming at the influence of terrain error on snow depth inversion,a terrain error correction inversion algorithm based on track clustering is proposed.The multi-system multi-frequency GNSS data of P351 station in PBO network and the L1 band data of GPS system of P350 station and P682 station are selected for experiments.The experimental results show that the terrain error correction method in this thesis can be effectively applied to data of different systems and different frequencies,and has certain universality.(4)A high-resolution multi-system GNSS snow depth fusion inversion method with distance weighting is proposed.On the basis of the previous research,the grid modeling of the reflection area around the site is carried out based on the effective reflection area.According to the distance between the focus of Fresnel reflection area and the center point of the grid,the multi-system GNSS snow depth inversion results are fused to generate a high-resolution snow depth map.The experimental results show that compared with the traditional equal weight fusion algorithm,the proposed algorithm can effectively reflect the difference of snow depth around the site,improve the spatial resolution of snow depth inversion,and has higher inversion accuracy. |
