Soil Moisture Retrieval Model Research Articles

GNSS-IR Soil Moisture Retrieval Using Multi-Satellite Data Fusion Based on Random Forest

The accuracy and reliability of soil moisture retrieval based on Global Positioning System (GPS) single-star Signal-to-Noise Ratio (SNR) data is low due to the influence of spatial and temporal differences of different satellites. Therefore, this paper proposes a Random Forest (RF)-based multi-satellite data fusion Global Navigation Satellite System Interferometric Reflectometry (GNSS-IR) soil moisture retrieval method, which utilizes the RF Model’s Mean Decrease Impurity (MDI) algorithm to adaptively assign arc weights to fuse all available satellite data to obtain accurate retrieval results. Subsequently, the effectiveness of the proposed method was validated using GPS data from the Plate Boundary Observatory (PBO) network sites P041 and P037, as well as data collected in Lamasquere, France. A Support Vector Machine model (SVM), Radial Basis Function (RBF) neural network model, and Convolutional Neural Network model (CNN) are introduced for the comparison of accuracy. The results indicated that the proposed method had the best retrieval performance, with Root Mean Square Error (RMSE) values of 0.032, 0.028, and 0.003 cm3/cm3, Mean Absolute Error (MAE) values of 0.025, 0.022, and 0.002 cm3/cm3, and correlation coefficients (R) of 0.94, 0.95, and 0.98, respectively, at the three sites. Therefore, the proposed soil moisture retrieval model demonstrates strong robustness and generalization capabilities, providing a reference for achieving high-precision, real-time monitoring of soil moisture.

A novel global grid model for soil moisture retrieval considering geographical disparity in spaceborne GNSS-R

Spaceborne global navigation satellite system-reflectometry has become an effective technique for Soil Moisture (SM) retrieval. However, the accuracy of global SM retrieval using a single model is limited due to the complexity of land surface. Introducing redundant ancillary data may also result in over-reliance problems. Therefore, we propose a method for SM retrieval that considers geographical disparities using the data from Cyclone GNSS (CYGNSS) observations and Soil Moisture Active and Passive (SMAP) product. Based on the CYGNSS effective reflectivity and ancillary datasets of SMAP, we establish five models for each grid with different parameters to achieve global SM retrieval. Subsequently, an optimal model, determined by the performance indicator, is used for SM retrieval. The results show that the root mean square error SRMSE\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$S_{\\mathrm{RMSE}}$$\\end{document} with the improved method is decreased by 9.1% using SMAP SM as reference with the SRMSE\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$S_{\\mathrm{RMSE}}$$\\end{document} = 0.040 cm3/cm3 compared with using single reflectivity-temperature-vegetation method. Additionally, using the in-situ SM of International Soil Moisture Network as reference, the overall correlation coefficient R\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$R$$\\end{document} and SRMSE\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$S_{\\mathrm{RMSE}}$$\\end{document} values with the improved method are 0.80 and 0.064 cm3/cm3, respectively. The average R\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$R$$\\end{document} of the chosen sites is increased by 22.7%, and the average SRMSE\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$S_{\\mathrm{RMSE}}$$\\end{document} is decreased by 8.7%. The results indicate that the improved method can better retrieve SM in both global and local scales without redundant auxiliary data.

Differentiable modeling for soil moisture retrieval by unifying deep neural networks and water cloud model

Machine learning has been widely used in high-spatial-resolution surface soil moisture (SSM) retrieval studies, but in recent years, this purely data-driven retrieval method has been controversial due to its lack of physical interpretability and generalization ability. Physical retrieval models based on the theory of radiative transfer equations respect physical laws, but their retrieval accuracy is usually affected by many insufficient accurate inputs, the complex model structure, and parameter adjustment method. In order to explore the retrieval method of unifying these two types of models, in this study, a differentiable model (DM) was constructed to realize the soil moisture retrieval at 10 m resolution based on Sentinel data. The differentiable soil moisture retrieval model takes the water cloud model (WCM) as the skeleton, and united the WCM and neural networks by implementing differentiable programming of the WCM in a machine learning platform. The differentiability makes the retrieval model trained by the gradient descent method the same as the neural network, which allows the retrieval model to be physical while obtaining more accurate retrieval results. Luan River Basin, Shandian River Basin, Maqu, and Lake Tahoe study areas with various land cover types and climate types were selected for model evaluation, and the performances of DM were close to that of the random forest model with Pearson correlation coefficient (R) of 0.747, 0.853, 0.838 and 0.792 in four study areas, respectively. While in the assessment of extrapolation capability of retrieval models, the DM showed its strong generalization ability and retrieval performance that exceeded that of the other retrieval models, with R of 0.786, unbiased root mean square error (ubRMSE) of 5.523 vol% and bias of 0.054 vol%. The DM synthesizes the advantages of both physical and machine learning models while providing high-resolution SSM estimates with acceptable accuracy for the study areas. This study creates favorable conditions for the realization of large-scale soil moisture retrieval with high resolution and high accuracy, and provides new ideas for the combination of machine learning and physical knowledge in other retrieval studies.

Ensembles of multiple models for soil moisture retrieval from remote sensing data over agricultural areas: A deep learning-based framework

In agricultural areas, most surface soil moisture (SM) retrieval models are unstable in terms of their accuracy and performance during crop growth. As a result, there is no consensus on which model performs optimally during the agricultural season. This is because of the uncertainties associated with model physics, initial conditions, model inversion processes, input data, vegetation attenuation and soil characteristics. To better deal with these practical concerns, we propose a simple, but robust SM retrieval method of using combination of multiple models based on deep learning and multi-model ensemble approach, called the DL-MME method, which makes use of the ‘collective intelligence’ and ‘wisdom of crowds’ concepts. The advantages of this method are: (1) robustness to model selection, and (2) robustness to model calibration during the growing season. In addition, this method is less dependent on one type of data across various agricultural areas compared to the single model approach. Firstly, the coupled water cloud model (WCM) and soil backscattering models (Oh model or advanced integral equation model (AIEM)) with different vegetation descriptors were calibrated and validated during the growing season in sugarcane and winter wheat fields for Sentinel-1 backscattering coefficients (VV and VH). SM was also retrieved by employing the trapezoid model (OPTRAM) with different parameters from Sentinel-2 images. To optimize SM retrieval computations, we used the outputs from optical and SAR models, auxiliary features, and reliable in situ SM measurements as inputs to a deep learning convolutional neural network (DL-CNN). For sugarcane and wheat fields in the early stages of crop growth, WCM models retrieved more accurate time-series SM than optical models. OPTRAM soil moisture retrievals showed greater accuracy in the late crop growing season. Time-series SM retrieval accuracy using DL-MME was higher than for the optical and semi-empirical SAR models. According to the results of the in situ validation for wheat (sugarcane) fields, the minimum MAE by an optimal combination of models was around 0.01 (0.02) m3m−3 (RMSE = 0.036 (0.074) m3m−3; R = 0.87 (0.71)). The findings demonstrate that our method is reliable and feasible for SM retrieval. Additionally, our method provides a way to select an optimal model for retrieving time-series SM during the crop growing season.

Improvement of CYGNSS soil moisture retrieval model considering water and surface temperature

The key to retrieving soil moisture (SM) with spaceborne Global Navigation Satellite System-Reflectometry (GNSS-R) data is to correct the effects of soil surface temperature (SST), roughness, water, and vegetation. In this study, Cyclone Global Navigation Satellite System (CYGNSS) data are used to calculate surface reflectivity and surface roughness is characterized based on the statistical moments of Delay-Doppler Maps (DDM). After removing open water, a multiple linear regression model is created to retrieve SM by combining the SST and vegetation optical depth (VOD) parameters provided by Soil Moisture Active Passive (SMAP) data. With a correlation coefficient of 0.815 and a root-mean-square error of 0.066 cm3cm−3, an experimental analysis reveals decent consistency between the CYGNSS SM and the referenced SMAP SM. In addition, a time-series analysis between the CYGNSS SM and the International Soil Moisture Network (ISMN) referenced SM data shows a good correlation. Since surface reflectivity is significantly affected by water and because there is a coupling relationship between SST and SM, the differences of CYGNSS SM when the two factors are simultaneously considered or ignored are also analyzed. The experimental results show that after removing water and incorporating SST into the linear regression model, the accuracy of CYGNSS SM has been improved significantly, with the root mean square error, mean absolute error, and Bias increasing by 6%, 7%, and 11.3%, respectively. This study demonstrates the necessity of considering water and SST in SM retrieval and provides a novel approach for SM retrieval using high accuracy and high spatial and temporal resolution data.

In-Situ GNSS-R and Radiometer Fusion Soil Moisture Retrieval Model Based on LSTM

Global navigation satellite system reflectometry (GNSS-R) is a remote sensing technology of soil moisture measurement using signals of opportunity from GNSS, which has the advantages of low cost, all-weather detection, and multi-platform application. An in situ GNSS-R and radiometer fusion soil moisture retrieval model based on LSTM (long–short term memory) is proposed to improve accuracy and robustness as to the impacts of vegetation cover and soil surface roughness. The Oceanpal GNSS-R data obtained from the experimental campaign at the Valencia Anchor Station are used as the main input data, and the TB (brightness temperature) and TR (soil roughness and vegetation integrated attenuation coefficient) outputs of the ELBARA-II radiometer are used as auxiliary input data, while field measurements with a Delta-T ML2x ThetaProbe soil moisture sensor were used for reference and validation. The results show that the LSTM model can be used to retrieve soil moisture, and that it performs better in the data fusion scenario with GNSS-R and radiometer. The STD of the multi-satellite fusion model is 0.013. Among the single-satellite models, PRN13, 20, and 32 gave the best retrieval results with STD = 0.011, 0.012, and 0.007, respectively.

Toward an Improved Surface Roughness Parameterization Model for Soil Moisture Retrieval in Road Construction

In passive microwave remote sensing, the estimation of the surface roughness parameter is a significant obstacle for soil moisture (SM) retrieval. For a given SM content, the geometric soil surface roughness has been shown to have a large impact on the surface emission at L-band frequency, which affects the SM retrieval success when using the information observed from the radiometer and is represented through the so-called the surface roughness parameter ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$H_{R}$ </tex-math></inline-formula> ). Moreover, no previous study has examined the effect of this factor in the context of road construction, where the geometric soil surface roughness is affected by the compaction process, resulting in a substantial change in roughness before and after compaction. Accordingly, a series of experiments at various compaction levels and SM contents was performed for a sand subgrade material in order to identify their effects on <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$H_{R}$ </tex-math></inline-formula> . The soil brightness temperature (TB) was measured using an L-band radiometer at different incidence angles and a laser profiler was used to measure the surface roughness standard deviation ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\sigma$ </tex-math></inline-formula> ) before and after compaction. The results of this article have demonstrated that the incidence angle ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\theta$ </tex-math></inline-formula> ) and SM both affect <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$H_{R}$ </tex-math></inline-formula> and its relation to the geometric soil surface roughness. Importantly, these factors are not accounted for by existing models. Consequently, a modified surface roughness parameter ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$H_{R}$ </tex-math></inline-formula> ) model, based on the traditional Choudhury model, was developed to include the contribution of these two factors, and its impact on the accuracy of SM retrieval results tested. Specifically, it was shown that it is possible to obtain SM retrieval results with an accuracy of 0.04 cm3/cm3 at almost all incidence angles using either dual-polarization [both horizontal (H) and vertical polarization (V)] or only vertical polarization observations. The modified surface roughness parameter ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$H_{R}$ </tex-math></inline-formula> ) model has improved the performance of the SM retrieval model to achieve an accuracy of 0.04 cm3/cm3, whereas the traditional Choudhury model achieved an accuracy of only 0.05 cm3/cm3.

Soil Moisture Retrieval From Sentinel-1 and Sentinel-2 Data Using Ensemble Learning Over Vegetated Fields

Soil moisture is a valuable basic data in climate, hydrological models, and agricultural applications. The rapid development of remote sensing technology can be used to monitor changes in soil moisture at multiple spatial and temporal scales. In this article, we unfolded a soil moisture retrieval method using ensemble learning combined with the Water Cloud Model(WCM) by Sentinel-1 and Sentinel-2 with multi-source datasets. First, using the WCM, the influence of vegetation cover on the backscattering coefficient was removed, where we use three vegetation index (Enhanced Vegetation Index (EVI), Normalized Difference Vegetation Index (NDVI) and Normalized Difference Water Index (NDWI)) for analysis and comparison. Then, combined with other multi-source datasets, a soil moisture retrieval model was established based on the ensemble learning algorithm. Here we choose two familiar ensemble learning algorithms for analysis and comparison, using pearson correlation significance analysis, which are the Random Forest (RF) and the Adaptive Boosting (AdaBoost). The results revealed that the RF model performed is slightly superior to the AdaBoost model. The optimal performance mean absolute error (MAE), root mean square error (RMSE)and the unbiased RMSE (ubRMSE)of RF model are 2.289 vol%, 2.934 vol%, 2.934 vol% respectively, which are slightly better than AdaBoost model. EVI is suitable for WCM model to remove vegetation scattering effect. It shows that it is attainable to utilize the ensemble learning method to inversion of SM using radar data. The proposed framework maximizes the potential of WCM, RF model, and multi-source datasets in deriving spatiotemporally continuous SM estimates, which should be valuable for SM inversion development.

Soil Moisture Retrieval During Crop Growth Cycle Using Satellite SAR Time Series

Satellite SAR-based soil moisture retrieval over agricultural fields, under crop overlain conditions, is a challenging exercise. This is so since the overlying crop volume interacts with both the incoming and the backscattered radar signal. Therefore, the soil moisture linked solely to the top layer ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$0-5$</tex-math></inline-formula> <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$cm$</tex-math></inline-formula> ) of the soil cannot be reliably retrieved under such conditions without avoiding the obscuring effect of growing crop volume. In this investigation, we demonstrated a proof of concept for a time-series approach to retrieve soil moisture during crop growth cycle. Contrary to the use of the single-scene approach, the novelty of the proposed approach lies in exploiting the satellite SAR time-series acquired during a cropping cycle. The proposed time-series approach is effective for capturing the nuances in the crop phenological stages while calibrating the Dubois-WCM soil moisture retrieval model. By employing this approach, we achieved the 0.04 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$m^{3} m^{-3}$</tex-math></inline-formula> soil moisture retrieval RMSE benchmark at a high spatial resolution and addressed the issue of solving for the Dubois-WCM model constants under data constrained conditions. Furthermore, we observed that combination of temporally non-overlapping vegetation descriptors (optical and SAR) resulted in degradation in the performance of the retrievals and under such circumstances single polarimetric descriptors performed better.

GNSS-IR multisatellite combination for soil moisture retrieval based on wavelet analysis considering detection and repair of abnormal phases

Signal-to-noise ratio (SNR) data received with standard geodetic instrumentation can be used to retrieve near-surface soil moisture (SM). However, low-quality SNR data usually cause abnormal phases in the fitting of nonlinear least squares (LLS) algorithms. This is not conducive to the effective use of multisatellite phases. In this paper, an SM retrieval method based on multisatellite combinations considering the detection and repair of abnormal phases is proposed. This method is aimed at using wavelet transform to separate the trend and modulation terms in SNR data, followed by detecting and repairing the abnormal phase for all satellites, finally constructing the multisatellite linear regression (MSLR) model for SM retrieval, and analysing the variation of accuracy with the increase of model testing days. The results indicate that with the coif5 wavelet, the trend and modulation terms of the SNR (compared to the traditional low-order polynomial) can be better separated. The abnormal phases can effectively be detected and repaired by combining the interquartile range and moving average filter, and further, the quality of the phases for each satellite can be improved. Furthermore, MSLR can fully combine the multisatellite phases to improve the accuracy of SM retrieval, and it is suitable for SM retrieval over different time periods.

Effect of surface temperature on soil moisture retrieval using CYGNSS

In this paper, a soil moisture (SM) retrieval model from spaceborne Global Navigation Satellite System Reflectometry (GNSS-R) observations that incorporates soil surface temperature (SST) for the first time is evaluated. Here, based on the grid scale, Cyclone GNSS (CYGNSS) reflectivity, SST and vegetation optical depth (VOD) are employed to estimate SM by a trilinear regression, while the other influence factors such as soil roughness and texture are regard as static. The results are compared with globally Soil Moisture Active Passive (SMAP) SM and in-situ measurements from International Soil Moisture Network (ISMN) over the year of 2018 respectively, showing a good consistency (R = 0.929 and RMSE = 0.043 cm3cm−3 against SMAP SM; R = 0.927 and RMSE = 0.042 cm3cm−3 against in-situ SM). Although the sensitivity of reflectivity to SST is found to be much smaller than that to SM from the simulation, the incorporation of SST is demonstrated to be effective in SM estimation for its coupling relationship with SM. In the comparison with SMAP SM, the improvements of RMSE by incorporating SST are varying degrees globally, and significant in many arid areas with an improvement of over 40%. In the in-situ validation, the overall RMSE decreases from 0.047 to 0.042 cm3cm−3 with an improvement of 10.6%. This work demonstrates the necessity and improvement for incorporating SST into SM retrieval for GNSS-R. Moreover, the findings provide a potential method to obtain global SST dataset from CYGNSS observations.

IDENTIFYING TIME PATTERNS AT THE FIELD SCALE FOR RETRIEVING SUPERFICIAL SOIL MOISTURE ON AN AGRICULTURAL AREA WITH A CHANGE DETECTION METHOD: A PRELIMINARY ANALYSIS

Abstract. A preliminary analysis based on the application of a change detection method for remote sensed soil moisture retrieval at high resolution is presented. Sentinel-1 SAR images are used for studying agricultural areas in Spain, where in situ soil moisture data are available through the International Soil Moisture Network. The total backscattered SAR signal is modelled as the sum of vegetation and soil contributions. At first, the relationship between soil moisture and the co-polarized band of Sentinel-1 was analyzed for all the measurement stations of the area, and the ones with stronger relation were selected. Time series analyses were then conducted at the field scale for studying the interactions between some SAR parameters and the in situ data. The two polarizations and the polarization ratio were analyzed with respect to in situ soil moisture observations and precipitation data in order to identify homogeneous time domains in which the method can be applied in a consistent manner. Analyses show that the main driver of wide range SAR signal variations is the presence of precipitation events. Moreover, SAR coherence and polarization rate manifest specific behaviors that can be exploited either for deepening the knowledge on the role of model parameters and identifying suitable time and space extends in which operate separate estimations of vegetation, soil moisture and soil roughness parameters. Identification and isolation of precipitation driven patterns, as long as the selection of homogeneous time spans and space regions is the basis for improving the capability of satellite based soil moisture retrieval models.

SOIL MOISTURE RETRIEVAL MODEL BASED ON DIELECTRIC MEASUREMENTS AND ARTIFICIAL NEURAL NETWORK

SOIL MOISTURE RETRIEVAL MODEL BASED ON DIELECTRIC MEASUREMENTS AND ARTIFICIAL NEURAL NETWORK

Retrieving the Soil Moisture in Bare Farmland Areas Using a Modified Dubois Model

Soil moisture is an important parameter for global soil moisture transport, environmental evaluation, and precision agricultural research. The accurate retrieval of soil moisture in farmland areas using Synthetic Aperture Radar (SAR) depends on the accurate description of surface and SAR parameters. In these parameters, surface roughness and incidence angle are the key factors that affect the accuracy of the soil moisture retrieval model. This article proposes a modified Dubois model to retrieve soil moisture suitable for the bare surface of farmland area. The model eliminates the incidence angle parameters and uses polarization parameters to depict the surface roughness parameters in the Dubois model. To eliminate the incidence angle, the backscattering coefficients gamma0, which eliminates the effect of the incidence angles, are used to replace the sigma0. Under rain and no rain condition, the trend of backscattering coefficients (VH and VV) and cross-polarization ratio (VH-VV) of different soil texture with the soil moisture are compared. Then, the polarization parameter based on VH backscattering coefficients is used to describe surface roughness. The model is evaluated with time-series soil moisture observation data in situ of the study area. The results indicate that the modified model can retrieve soil moisture with high accuracy, and the total RMSE can reach 0.064 cm3cm−3 while the Dubois model is 0.124 cm3cm−3. Under rain and no rain condition, the retrieval accuracy of the modified model is 0.066 cm3cm−3 and 0.063 cm3cm−3. The retrieval accuracy is 0.060 cm3cm−3 and 0.067 cm3cm−3 under high and low incidence angles conditions, respectively. These results indicate that the modified Dubois model can retrieve soil moisture with high accuracy under different conditions.

Soil moisture retrieval under wheat crop using RISAT-1 hybrid polarimetric SAR data

The study aims to retrieve soil moisture from RISAT-1 hybrid polarimetric SAR data. Although the use of linear polarimetric SAR data has been well understood and documented, but hybrid polarimetric SAR data is grossly under explored and under reported for this purpose. Regression analysis has been carried to develop soil moisture retrieval models and validated the same. The retrieval models have been developed from back scattering coefficients (RH &amp; RV) and m- space decomposition parameters (even bounce, odd bounce, and volume component) generated from RISAT-1 hybrid polarimetric SAR data. A total of three models are analyzed in this work, (i) using both RH &amp;RV, (ii) volume component, and (iii) using even bounce, odd bounce and volume component. The study results showed that the model using m- decomposition derived parameters can provide better accuracy with R2 and RMSE of 0.92 and 2.45 per cent respectively in comparison to other two models.&#x0D;

A Calibration/Disaggregation Coupling Scheme for Retrieving Soil Moisture at High Spatio-Temporal Resolution: Synergy between SMAP Passive Microwave, MODIS/Landsat Optical/Thermal and Sentinel-1 Radar Data.

Soil moisture (SM) data are required at high spatio-temporal resolution—typically the crop field scale every 3–6 days—for agricultural and hydrological purposes. To provide such high-resolution SM data, many remote sensing methods have been developed from passive microwave, active microwave and thermal data. Despite the pros and cons of each technique in terms of spatio-temporal resolution and their sensitivity to perturbing factors such as vegetation cover, soil roughness and meteorological conditions, there is currently no synergistic approach that takes advantage of all relevant (passive, active microwave and thermal) remote sensing data. In this context, the objective of the paper is to develop a new algorithm that combines SMAP L-band passive microwave, MODIS/Landsat optical/thermal and Sentinel-1 C-band radar data to provide SM data at the field scale at the observation frequency of Sentinel-1. In practice, it is a three-step procedure in which: (1) the 36 km resolution SMAP SM data are disaggregated at 100 m resolution using MODIS/Landsat optical/thermal data on clear sky days, (2) the 100 m resolution disaggregated SM data set is used to calibrate a radar-based SM retrieval model and (3) the so-calibrated radar model is run at field scale on each Sentinel-1 overpass. The calibration approach also uses a vegetation descriptor as ancillary data that is derived either from optical (Sentinel-2) or radar (Sentinel-1) data. Two radar models (an empirical linear regression model and a non-linear semi-empirical formulation derived from the water cloud model) are tested using three vegetation descriptors (NDVI, polarization ratio (PR) and radar coherence (CO)) separately. Both models are applied over three experimental irrigated and rainfed wheat crop sites in central Morocco. The field-scale temporal correlation between predicted and in situ SM is in the range of 0.66–0.81 depending on the retrieval configuration. Based on this data set, the linear radar model using PR as a vegetation descriptor offers a relatively good compromise between precision and robustness all throughout the agricultural season with only three parameters to set. The proposed synergistical approach combining multi-resolution/multi-sensor SM-relevant data offers the advantage of not requiring in situ measurements for calibration.

Evaluation of Radar/Optical Based Vegetation Descriptors in Water Cloud Model for Soil Moisture Retrieval

The accurate consideration of vegetation descriptors in water cloud model (WCM) is necessary for precise SM retrieval. Most of the vegetation descriptors are sourced from optical remote sensors. The acquisitions from optical sensors are largely hampered by bad weather conditions. For all-weather monitoring, Synthetic Aperture Radar (SAR) based vegetation descriptors are needed to identify and evaluate their performance for SM retrieval. The present study evaluates the various sources/combinations of SAR based vegetation descriptors in WCM to identify the better alternatives of optical-based vegetation descriptors. The performance of three radar-based vegetation descriptors, namely VH polarized backscattering coefficients, depolarization ratio and radar vegetation index (RVI) along with the one optical-based vegetation descriptor, namely leaf area index (LAI) from MODIS were utilized in WCM. The WCM for each vegetation descriptor has been performed using Sentinel-1 VV polarized backscattering coefficients and in-situ SM. The in-situ SM measurements were carried out in the fields around Varanasi District in India during the winter season sown with the wheat crop. The correlations coefficient (r), root mean square error (RMSE) and bias were used to evaluate the performances of vegetation descriptors in WCM for SM retrieval. The study showed that the depolarization ratio is the best for SM retrieval with accuracy of 0.096 m <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> m <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-3</sup> followed by RVI, cross-polarized and LAI with 0.100 m <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> m <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-3</sup> , 0.124 m <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> m <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-3</sup> and 0.124 m <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> m <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-3</sup> , respectively. Thus, the depolarization ratio can be used for the retrieval of SM using Sentinel-1 VV polarized backscattering coefficients over the wheat crop.

Sensitivity of C-Band SAR Polarimetric Variables to the Directionality of Surface Roughness Parameters

Surface roughness is an important factor in many soil moisture retrieval models. Therefore, any mischaracterization of surface roughness parameters (root mean square height, RMSH, and correlation length, ʅ) may result in unreliable predictions and soil moisture estimations. In many environments, but particularly in agricultural settings, surface roughness parameters may show different behaviours with respect to the orientation or azimuth. Consequently, the relationship between SAR polarimetric variables and surface roughness parameters may vary depending on measurement orientation. Generally, roughness obtained for many SAR-based studies is estimated using pin profilers that may, or may not, be collected with careful attention to orientation to the satellite look angle. In this study, we characterized surface roughness parameters in multi-azimuth mode using a terrestrial laser scanner (TLS). We characterized the surface roughness parameters in different orientations and then examined the sensitivity between polarimetric variables and surface roughness parameters; further, we compared these results to roughness profiles obtained using traditional pin profilers. The results showed that the polarimetric variables were more sensitive to the surface roughness parameters at higher incidence angles (θ). Moreover, when surface roughness measurements were conducted at the look angle of RADARSAT-2, more significant correlations were observed between polarimetric variables and surface roughness parameters. Our results also indicated that TLS can represent more reliable results than pin profiler in the measurement of the surface roughness parameters.

Soil Moisture Retrieval Model Design with Multispectral and Infrared Images from Unmanned Aerial Vehicles Using Convolutional Neural Network

This paper deals with a soil moisture retrieval model design with airborne measurements for remote monitoring of soil moisture level in large crop fields. A small quadrotor unmanned aerial vehicle (UAV) is considered as a remote sensing platform for high spatial resolutions of airborne images and easy operations. A combination of multispectral and infrared (IR) sensors is applied to overcome the effects of canopies convering the field on the sensor measurements. Convolutional neural network (CNN) is utilized to take the measurement images directly as inputs for the soil moisture retrieval model without loss of information. The procedures to obtain an input image corresponding to a certain soil moisture level measurement point are addressed, and the overall structure of the proposed CNN-based model is suggested with descriptions. Training and testing of the proposed soil moisture retrieval model are conducted to verify and validate its performance and address the effects of input image sizes and errors on input images. The soil moisture level estimation performance decreases when the input image size increases as the ratio of the pixel corresponding to the point to estimate soil moisture level to the total number of pixels in the input image, whereas the input image size should be large enough to include this pixel under the errors in input images. The comparative study shows that the proposed CNN-based algorithm is advantageous on estimation performance by maintaining spatial information of pixels on the input images.

Soil Moisture Retrieval in Farmland Areas with Sentinel Multi-Source Data Based on Regression Convolutional Neural Networks.

As an important component of the earth ecosystem, soil moisture monitoring is of great significance in the fields of crop growth monitoring, crop yield estimation, variable irrigation, and other related applications. In order to mitigate or eliminate the impacts of sparse vegetation covers in farmland areas, this study combines multi-source remote sensing data from Sentinel-1 radar and Sentinel-2 optical satellites to quantitatively retrieve soil moisture content. Firstly, a traditional Oh model was applied to estimate soil moisture content after removing vegetation influence by a water cloud model. Secondly, support vector regression (SVR) and generalized regression neural network (GRNN) models were used to establish the relationships between various remote sensing features and real soil moisture. Finally, a regression convolutional neural network (CNNR) model is constructed to extract deep-level features of remote sensing data to increase soil moisture retrieval accuracy. In addition, polarimetric decomposition features for real Sentinel-1 PolSAR data are also included in the construction of inversion models. Based on the established soil moisture retrieval models, this study analyzes the influence of each input feature on the inversion accuracy in detail. The experimental results show that the optimal combination of R2 and root mean square error (RMSE) for SVR is 0.7619 and 0.0257 cm3/cm3, respectively. The optimal combination of R2 and RMSE for GRNN is 0.7098 and 0.0264 cm3/cm3, respectively. Especially, the CNNR model with optimal feature combination can generate inversion results with the highest accuracy, whose R2 and RMSE reach up to 0.8947 and 0.0208 cm3/cm3, respectively. Compared to other methods, the proposed algorithm improves the accuracy of soil moisture retrieval from synthetic aperture radar (SAR) and optical data. Furthermore, after adding polarization decomposition features, the R2 of CNNR is raised by 0.1524 and the RMSE of CNNR decreased by 0.0019 cm3/cm3 on average, which means that the addition of polarimetric decomposition features effectively improves the accuracy of soil moisture retrieval results.