Xingming Zheng,Zhuangzhuang Feng,Lei Li,Bingzhe Li,Tao Jiang,Xiaojie Li,Xiaofeng Li,Si Chen
Abstract:Both radar and optical signals are sensitive to the change of surface soil moisture (SSM) and surface roughness properties (such as root mean squared height- RMSH), and the accuracy of retrieved SSM from single radar and optical remote sensing data is influenced by the spatiotemporal change of surface roughness. Here, we attempt to explore a method to simultaneously estimate SSM and RMSH of bare soil by combining optical and radar data, so as to weaken the effect of surface roughness on SSM inversion results. To achieve this goal, two satellite synchronous ground experiments were carried out, collecting 88 sampling plots each with an area of 50 m × 50 m. Radar backscattering coefficient and spectral reflectance are uniformly corrected to a fixed observation direction and solar incident direction respectively, which can eliminate the difference of satellite signal resulted from various sun-satellite geometry. Combining radar backscattering and optical reflectance model, Sentinel-1 and Sentinel-2 data are used to simultaneously retrieve SSM and RMSH of bared soils, and some conclusions are given as below: 1) a strong correlation is observed for (radar and optical) satellite signals and soil surface parameters (SSM and RMSH); 2) a higher accuracy was obtained by the combined use of optical and radar data, indicated by the decreased root mean squared error of retrieved SSM (~0.045 cm3/cm3) and RMSH (~0.8 cm); 3) the further improvement of retrieved SSM and RMSH was achieved by introducing their initial values, revealing that the prior knowledge of soil properties is also beneficial to improve the retrieval accuracy. This study proposed an framework for simultaneous estimation of SSM and RMSH by combining optical and radar data, and its feasibility is verified by experimental data.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to simultaneously estimate soil surface moisture (SSM) and surface roughness (RMSH) of bare soil surfaces by combining optical and radar data, so as to reduce the impact of surface roughness on SSM retrieval results. Specifically, the paper aims to explore a method to improve the estimation accuracy of SSM and RMSH by using satellite optical and radar data simultaneously, and verify the feasibility of this method.
### Background and Problem Description
- **Background**: Surface soil moisture (SSM) has a direct impact on national food security, human health, and ecosystem functions. Timely acquisition of SSM is of great significance in agriculture, hydrology, and numerical weather prediction. Remote sensing technology is widely used to estimate SSM due to its large - scale, real - time, and low - cost advantages.
- **Problem**: The SSM estimation accuracy of a single radar or optical remote sensing data is affected by the spatio - temporal variation of surface roughness. Both the radar backscattering coefficient (\(\sigma_{PQ}\)) and spectral reflectance (\(r_{opt}\)) are sensitive to SSM and surface roughness parameters (such as root - mean - square height - RMSH). Therefore, how to combine optical and radar data to simultaneously estimate SSM and RMSH in order to reduce the impact of surface roughness on SSM retrieval results has become the focus of research.
### Research Objectives
- **Main Objective**: Propose a method that combines optical and radar data to simultaneously estimate SSM and RMSH and improve the estimation accuracy.
- **Specific Objectives**:
- Verify whether combining optical and radar data can improve the estimation accuracy of SSM and RMSH.
- Explore whether introducing initial values (prior knowledge of SSM and RMSH) can further improve the estimation accuracy.
### Methods
- **Experimental Design**: Two synchronous ground experiments were carried out, and data from 88 sampling plots were collected, with each plot having an area of 50m × 50m.
- **Data Processing**: Pre - process Sentinel - 1 radar data and Sentinel - 2 multispectral data, including orbit correction, radiometric correction, multi - view processing, speckle filtering, and geocoding.
- **Model Building**: Establish a bare soil reflectance model and a backscattering coefficient model, which respectively describe the relationships between spectral reflectance and radar backscattering coefficient with SSM and RMSH.
- **Joint Inversion Strategy**: Use the partial least squares method (PLS) to minimize the cost function between satellite observations and model values, and propose four inversion strategies: using only optical data, using only radar data, using optical and radar data simultaneously, and using optical and radar data simultaneously and introducing initial values.
### Results
- **Inversion Accuracy**: The inversion strategy that combines optical and radar data significantly improves the estimation accuracy of SSM and RMSH. Introducing initial values further improves the accuracy.
- **Error Analysis**: The accuracy of the estimation results was evaluated by mean bias (Bias), root - mean - square error (RMSE), and Pearson correlation coefficient (R).
### Discussion
- **Contributions**: Explore the influence of soil properties (such as organic matter content, clay content, ridge structure direction, and height ratio) on SSM and RMSH estimation errors.
- **Advantages and Limitations**:
- **Advantages**: Combining optical and radar data can reduce the possibility of falling into local optimal solutions, and introducing initial values can limit the values of SSM and RMSH.
- **Limitations**: The detection depths of optical and radar signals for SSM are inconsistent. Ideally, optical and radar data need to be obtained simultaneously, but it is difficult to achieve in practical applications.
In conclusion, this paper proposes an effective method to simultaneously estimate SSM and RMSH by combining optical and radar data, improves the estimation accuracy, and verifies the feasibility of this method.