Research On Soil Moisture Retrieval Model Based On CYGNSS

Soil moisture,also known as soil water content,is a physical quantity that measures the dryness and wetness of the surface.As a key variable in meteorology,hydrology,agriculture and ecological applications,accurate large-scale observation of it and its changes can reflect the dryness and wetness of the land.So as to provide reliable data support for research on plant growth index,energy exchange between the ground and the atmosphere,and flood forecasting.The traditional soil moisture monitoring method mainly relies on weather stations.Although it can provide good accuracy,it consumes a lot of manpower and material resources,and the cost of monitoring equipment is high.It is difficult to meet the needs of monitoring soil moisture in a large area.Meteorological satellites can quickly obtain large-area ground information,and the L-band electromagnetic waves emitted by GNSS satellites have the advantages of all-weather and strong penetration,which provides a new means for soil moisture monitoring.However,spaceborne GNSS reflection signals(GNSS-R)are easily affected by terrestrial variable environmental factors.At present,there are few studies on the error analysis of spaceborne GNSS-R soil moisture retrieval and the analysis of the extrapolation performance of the retrieval model.In this paper,a semi-empirical inversion model for Cyclone Global Navigation Satellite System(CYGNSS)/Soil Moisture Active Passive(SMAP)data fusion is proposed to solve the problems of soil moisture model error influence and model performance.The accuracy and effectiveness of the model were verified by the soil moisture products provided by CLDAS and SMAP,and the inversion performance of the model was comprehensively analyzed.The main research contents are as follows:(1)In view of the idea of the method proposed in this paper,the application field of GNSS-R technology and the research status of GNSS-R soil moisture retrieval at home and abroad are introduced,which provides a theoretical basis for the experiment.(2)Summarize the basic theory of spaceborne GNSS-R soil moisture retrieval,including GNSSR technology,reflection signal characteristics,reflection coefficient,etc.,and analyze the functional relationship between reflection coefficient and dielectric constant through Fresnel reflection coefficient,leading to the basic principle of soil moisture detection.(3)Select the satellite data(CYGNSS,SMAP,CLDAS)used in the experiment,preprocess the reflection data,and calculate the reflectivity through the reflection surface information provided by CYGNSS.The optimal linear regression rate was obtained by fitting,and finally a semi-empirical model for CYGNSS/SMAP data fusion was established.(4)For the errors in the inversion model,the accuracy is improved by correcting the reflectivity.The errors are mainly analyzed from four aspects: incident angle,transmission power system error,surface vegetation attenuation,and surface roughness attenuation.Cloud models and other corrections reduce errors.After correcting various errors,the accuracy of the model has been improved.Bias can be increased by up to 6.80%,and RMSE can be increased by up to 3.30%.Use the soil moisture products provided by the SMAP data set and the CLDAS data set as the verification data set,bias(Bias),root mean square error(RMSE),and correlation coefficient(R)as evaluation indicators,and evaluate the performance of the model based on the inversion results.The accuracy of extrapolation and inversion is mainly related to the semi-empirical model and the feature coverage of the training data.The more data in the training model,the lower the inversion bias and the smaller the RMSE.The inversion results of the one-year training data are obviously better.According to the results of 3-month training data,Bias increased by 51.46% on average,and RMSE increased by 32.01% on average.Through timeliness analysis,it is found that the inversion model obtained by one-year data set training can extrapolate the soil moisture for one year.Compared with the SMAP soil moisture product,the inversion result has a Bias of-0.0037(8(8(8(83/(8(8(8(83 and an RMSE of 0.0264(8(8(8(83/(8(8(8(83.The correlation coefficient is 0.9636.According to the inversion of seasonal classification,the average winter Bias can be increased by 54.80%,the RMSE can be increased by 21.58%,the spring Bias can be increased by an average of 59.05%,and the RMSE can be increased by 21.05%.push precision.The SMAP data was used to train the model,and the inversion results were compared with CLDAS to verify that the inversion results were still good,which further proved the accuracy of the inversion model.