Soil Moisture Retrievals Using GPS Signal-to-noise Ratio Observations

Soil moisture plays an important role in natural environment.Soil moisture provides an important observation window for monitoring regional hydrological fluctuations.Soil moisture changes are one of the factors that affect the response of ecosystems to climate change.The stable and correct soil moisture monitoring provides an important monitoring tool for hydrological,climatic,agricultural and disaster prevention and control in a region.Common soil moisture monitoring methods,such as on-site sampling,are accurate but time consuming and unfriendly to environmental protection.Electromagnetic probes are difficult to assimilate due to different equipment.These two methods can only obtain small-scale soil moisture.A wide range of humidity monitoring remote sensing satellite data may have limited temporal resolution and lack of ground data validation.This thesis aims at reducing manpower and data costs and providing long-term stable soil moisture monitoring,and researches methods and influencing factors for inversion of soil moisture using ground-based GPS receiver SNR observations.The algorithm can be used to monitor the soil moisture within a radius of 25 meters from the monitoring station.The maximum difference between the inversion value and the known value is 0.12cm~3/cm~3,and the maximum standard deviation of the inversion value is 0.131cm~3/cm~3 within the need for soil moisture monitoring.According to the analysis of the amplitude,phase,and vibration frequency of the SNR data under different observation environments,we can get conclusions:1)Only considering the soil moisture,the change of soil moisture will cause the phase shift of SNR,and the amplitude will increase with the increase of soil moisture.2)Only considering the vegetation,the vegetation growth will also cause the phase shift of the signal-to-noise ratio,and the signal reflection surface will be closer to the receiver.Because the vegetation absorbs energy,the amplitude will decrease.3)Data with large topographic changes in the reflection trajectory,there is a strong noise background through spectrum analysis.4)Snow will change the frequency of the reflected signal and we can use this feature to perform snow depth inversion.Through analysis of time series changes of environmental parameters,We can get conclusions:1)the reflector absorbs signal energy,there is a case where the amplitude is abnormally low.In addition,the vegetation,temperature,and humidity have a common effect on the amplitude.Among them,vegetation has the greatest impact on amplitude.2)Effective reflection height is affected by the terrain,signal transmission depth,and ground roughness.When the soil moisture is less than 0.1cm~3/cm~3,the signal transmission depth is about 6~8cm.3)Phase parameter is strongly correlated with the soil moisture.Therefore,a linear model can be established.There are three reasons for limiting the accuracy of the model:First,the phase parameter at DOY132 rises in a leaping manner during the rainy period,there is a big gap between the known value and phase parameter.Second,the model does not consider the influence of vegetation on the phase.Third,there is a certain margin of error as known soil moisture data itself.In this thesis,the terrain,the length of the original signal selection,the selection of the elevation angle range and the influence of the signal frequency on the inversion quality are analyzed.Through terrain analysis experiments,a 1%change in slope results in significant differences in the inversion quality of different segments of each satellite.The correlation coefficient between G18 satellite's two-stage data and known humidity was reduced from0.830 to 0.712,and that of G12 satellite from 0.826 to 0.680.At the same time,the correlation coefficient of the satellites whose reflection points passed through large terrain changes was lower and the standard deviation was greater than the satellites whose reflection points passed through flat terrain.In the original data selection on inversion quality impact experiments:S2 is the better original data.If the data is too long or too short,the retrieval quality will be degraded.In the experiment,The signal-to-noise ratio of the angle of 5°~25°in the experiment results in a standard deviation of the phase parameter of0.059 and a correlation coefficient of 0.755.