Soil Moisture Retrieval Approaches Research Articles

A cross-resolution transfer learning approach for soil moisture retrieval from Sentinel-1 using limited training samples

Synthetic Aperture Radar (SAR) data is increasingly popular as a data source for global near-surface soil moisture mapping, but large-scale applications are still challenging due to the complex scattering process and the cumbersome data preprocessing. The emergence of deep learning methods has allowed advances in the remote sensing of large-scale surface parameters, but its application in SAR soil moisture retrieval has suffered from the availability of ground soil moisture measurements. Accordingly, this study proposed a cross-resolution transfer learning framework, with the assumption that sophisticated models for different spatial resolutions share a similar model architecture and trainable parameters. A robust high-resolution model can thus be trained with fewer samples by using coarse models. Accordingly, 25 deep learning models were pre-trained taking ∼387,000 Soil Moisture Active Passive (SMAP) Level-3 9 km enhanced passive soil moisture measurements as the truth, with an average validation RMSE of 0.03 m3/m3. They were then transferred to finer grids of 0.1–1 km using a small number of in-situ samples. A total of ∼190,000 daily soil moisture measurements from the international soil moisture network (ISMN) were used to evaluate the proposed framework in three scenarios. The results show that 1) 5000–6000 random samples are sufficient to achieve a target RMSE of 0.06 m3/m3; 2) training samples from a short period (2 or 4 months for Sentinel-1) of 2021 resulted in an overall RMSE of ∼0.068 m3/m3 in an independent period of 2016–2020; 3) the transfer learning also improved the retrieval accuracy (10–30% in relative) over areas without ground samples used for training but failed to yield an acceptable accuracy over mountainous areas. The promising results from this study confirmed the effectiveness of using “pre-trained models + scenario specific models” for regional to global soil moisture retrieval from Sentinel-1.

Optimizing Soil Moisture Retrieval: Utilizing Compact Polarimetric Features with Advanced Machine Learning Techniques

Soil moisture plays a crucial role in various environmental processes and is essential for agricultural management, hydrological modeling, and climate studies. Synthetic Aperture Radar (SAR) remote sensing presents significant potential for estimating soil moisture due to its ability to operate in all weather conditions and provide day-and-night imaging capabilities. Among the SAR configurations, the Compact Polarimetric (CP) mode has gained increasing interest as it relaxes system constraints, improves coverage, and enhances target information compared to conventional dual polarimetric SAR systems. This paper introduces a novel approach for soil moisture retrieval utilizing machine learning algorithms and CP SAR features. The CP SAR features are derived from a series of RADARSAT Constellation Mission (RCM) CP SAR imagery acquired over Canadian experimental sites equipped with Real-Time In Situ Soil Monitoring for Agriculture (RISMA) stations. This study employs a diverse dataset of compact polarimetric SAR features and corresponding ground truth soil moisture measurements for training and validation purposes. The results of our study achieved a Root Mean Square Error (RMSE) of 6.88% with a coefficient of determination R2 equal to 0.60, which corresponds to a correlation R between true and predicted soil moisture values of 0.75, using optimized Ensemble Learning Regression (ELR) with a decision-tree-based model. These results improved, yielding an RMSE of 5.67% and an R2 equal to 0.73 (R = 0.85), using an optimized Gaussian Process Regression (GPR) model.

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.

Integration of L-Band Derived Soil Roughness into a Bare Soil Moisture Retrieval Approach from C-Band SAR Data

Surface soil moisture (SSM) is a key variable for many environmental studies, including hydrology and agriculture. Synthetic aperture radar (SAR) data in the C-band are widely used nowadays to estimate SSM since the Sentinel-1 provides free-of-charge C-band SAR images at high spatial resolution with high revisit time, whereas the use of L-band is limited due to the low data availability. In this context, the main objective of this paper is to develop an operational approach for SSM estimation that mainly uses data in the C-band (Sentinel-1) with L-bands (ALOS/PALSAR) as additional data to improve SSM estimation accuracy. The approach is based on the use of the artificial neural networks (NNs) technique to firstly derive the soil roughness (Hrms) from the L-band (HH polarization) to then consider the L-band-derived Hrms and C-band SAR data (VV and VH polarizations) in the input vectors of NNs for SSM estimation. Thus, the Hrms estimated from the L-band at a given date is assumed to be constant for a given times series of C-band images. The NNs were trained and validated using synthetic and real databases. The results showed that the use of the L-band-derived Hrms in the input vector of NN in addition to C-band SAR data improved SSM estimation by decreasing the error (bias and RMSE), mainly for SSM values lower than 15 vol.% and regardless of Hrms values. Based on the synthetic database, the NNs that neglect the Hrms underestimate and overestimate the SSM (bias ranges between −8.0 and 4.0 vol.%) for Hrms values lower and higher than 1.5 cm, respectively. For Hrms <1.5 cm and most SSM values higher than 10 vol.%, the use of Hrms as an input in the NNs decreases the underestimation of the SSM (bias ranges from −4.5 to 0 vol.%) and provides a more accurate estimation of the SSM with a decrease in the RMSE by approximately 2 vol.%. Moreover, for Hrms values between 1.5 and 2.0 cm, the overestimation of SSM slightly decreases (bias decreased by around 1.0 vol.%) without a significant improvement of the RMSE. In addition, for Hrms >2.0 cm and SSM between 8 to 22 vol.%, the accuracy on the SSM estimation improved and the overestimation decreased by 2.2 vol.% (from 4.5 to 2.3 vol.%). From the real database, the use of Hrms estimated from the L-band brought a significant improvement of the SSM estimation accuracy. For in situ SSM less than 15 vol.%, the RMSE decreased by 1.5 and 2.2 vol.% and the bias by 1.2 and 2.6 vol.%, for Hrms values lower and higher than 1.5 cm, respectively.

Retrieval of Soil Moisture Content Based on a Modified Hapke Photometric Model: A Novel Method Applied to Laboratory Hyperspectral and Sentinel-2 MSI Data

Soil moisture is the crucial carrier of the global hydrologic cycle and the dynamic energy balance regulation process. Therefore, it is of great significance to monitor surface soil moisture content (SMC) accurately for the study of the natural ecological environment. The Hapke model is the most widely used photometric model in soil remote sensing research, but the development of this model is limited by the lack of valid multi–angular data. The main innovations of this paper have two aspects: (1) A novel soil moisture retrieval approach based on the Hapke (SMR–Hapke) model is derived by exploring the relationship between single scattering albedo (SSA) and SMC on the optical bands from 400 to 2400 nm. The performance of the proposed model was verified on a dataset consisting of four different soil samples, and the experimental results indicated that the inverted soil moisture from SMR–Hapke model coincided with the measurement values, with the R2 being generally more than 0.9 in the solar domain. (2) The SMR–Hapke model has been reduced to a linear form on the SWIR field and a physically-based normalized difference soil moisture index N D S M I H a p k e has been proposed. Based on the laboratory-based hyperspectral data, we compared the performance of N D S M I H a p k e with other traditional soil moisture indices using linear regression analysis, and the results demonstrate that the proposed N D S M I H a p k e had a great potential for estimating SMC with R2 values of 0.88. Finally, high–resolution SMC map was produced by combining the Sentinel–2 MSI data with N D S M I H a p k e . This study provides a novel extended Hapke model for the estimation of surface soil moisture content.

Least Square Based Iteration Approach for Agricultural Soil Moisture Retrieval Using Multi-Sensor Data

The main objective of this study is to develop a robust soil moisture retrieval approach using multi-sensor remote sensing data. Firstly, the water cloud model was employed to eliminate the vegetation effects on SAR observations over vegetated areas, thus to obtain the bare soil backscatter associated with soil moisture. Then, against the underdetermined system for soil moisture retrieval, the advanced integral equation model and calibrated integral equation model were integrated to construct soil moisture retrieval scheme in combination with multi-sensor SAR observations. Through the above processing the influences of vegetation and surface roughness can be minimized in the developed soil moisture retrieval scheme. Finally, the least square based iteration approach was applied to derive soil moisture with multi-sensor SAR observations as inputs. Based on the field measurements and multi-sensor SAR data (C-band and X-band). acquired on June 2015 and October 2015 over Hebei agricultural areas, quantitative evaluation of the developed approach was implemented. The results indicate that accurate soil moisture was obtained by the developed approach over corn covered area and bare agricultural area. In comparison with the results obtained by LUT method, the least square based iteration approach achieved better estimations for soil moisture.

Refining a Polarimetric Decomposition of Multi-Angular UAVSAR Time Series for Soil Moisture Retrieval Over Low and High Vegetated Agricultural Fields

The model-based polarimetric decomposition under multi-angular condition is refined to estimate soil moisture over agricultural fields covered by different crops from Uninhabited Aerial Vehicle Synthetic Aperture Radar (UAVSAR) time series. The approach allows to disentangle the vegetation and ground scattering components in order to invert the latter for the retrieval of soil moisture. For the vegetation volume separation, the crop orientations were estimated from SAR observations acquired at different incidence angles, and the associated volume scattering component was subtracted from each acquisition individually. Afterward, the soil moisture was retrieved from both ground scattering components (surface, dihedral), using the developed multi-angular cost functions comprised of dominant Bragg surface ( β ) or Fresnel dihedral ( α ) scattering parameters. Compared to former soil moisture retrievals from model-based decomposition of multi-angular polarimetric SAR data, the present refined approach that integrated both ground components, surface and dihedral, is independent of the power attenuation and loss during the microwave propagation through the vegetation. In this way, the ambiguity in the dihedral scattering component (most prominent around 45° incidence angle) was overcome, enabling a more robust retrieval methodology by clearly decoupling the soil and vegetation dielectric constants. The proposed multi-angular approach for soil moisture retrieval was validated with respect to the ground measurements conducted during the Soil Moisture Active Passive Validation Experiment in 2012. Due to the increased number of valid dominant surface/dihedral components which are used to retrieve the soil moisture in the multi-angular approach, an overall retrieval rate of 90%, significantly higher than that of the single-angular condition (50%), is obtained. The results indicate an overall retrieval rmse of 0.07–0.09 m3/m3 for the early crop growth stage, and a rmse of 0.09–0.12 m3/m3 for the later crop development until mature stage. However, the retrieval performance is highly dependent on the crop structure and phenological development stages, but the multi-angular rmse range is mostly lower than all single-angular rmses, indicating better quality of the multi-angular inversion than the single-angular one.

Investigating vegetation water dynamics and drought using Metop ASCAT over the North American Grasslands

In this study, we examined the ASCAT backscatter data from Metop-A from 2007 to 2016 to characterize spatial and temporal variability in the vegetation parameters of the TU Wien Soil Moisture Retrieval approach (TUW SMR) across the North American Grasslands. The vegetation parameters are the slope and curvature of a second order Taylor polynomial used to describe the incidence angle dependence of backscatter σ°. A recent development allows the vegetation parameters to be determined dynamically using the local slope values within a prescribed temporal window. Seasonal, interannual and diurnal variations in the vegetation parameters were found to vary across grassland cover types, reflecting variations in soil moisture availability and growing season length. While the slope has always been considered a measure of vegetation density, our results show that curvature also contains information about vegetation. Drought events in 2011 and 2012 resulted in extensive negative σ40∘ and soil moisture anomalies during the maximum biomass period. Contiguous anomalies in slope and curvature were observed where the severity and persistence of the drought were enough to impact vegetation. Observed diurnal differences in slope and curvature suggest that daily moisture transport within the vegetation influences the relative dominance of scattering from the vegetation and soil surface.

An Information fusion approach for PALSAR data to retrieve soil moisture

Estimation of vegetation covered soil moisture with satellite images is still a challenging task. Several models are available for soil moisture retrieval in which water cloud model (WCM) is most common. But, it requires an estimation of accurate vegetation parameterization. Thus, there is a need to develop such an approach for soil moisture retrieval which minimize these limitations. Therefore, this paper deals with the soil moisture retrieval using fully polarimetric SAR data by fusing the information from different bands. Various polarimetric indices and observables were critically analysed, and found that the index; SPAN (total scattered power) gives better information of vegetation cover as compared to other indices/observables. Based on this, WCM model has been modified using SPAN as parameter and soil moisture content were retrieved.

Comparison of SMOS and SMAP soil moisture retrieval approaches using tower-based radiometer data over a vineyard field

The objective of this study was to compare several approaches to soil moisture (SM) retrieval using l-band microwave radiometry. The comparison was based on a brightness temperature (TB) data set acquired since 2010 by the L-band radiometer ELBARA-II over a vineyard field at the Valencia Anchor Station (VAS) site. ELBARA-II, provided by the European Space Agency (ESA) within the scientific program of the SMOS (Soil Moisture and Ocean Salinity) mission, measures multiangular TB data at horizontal and vertical polarization for a range of incidence angles (30°–60°). Based on a three year data set (2010–2012), several SM retrieval approaches developed for spaceborne missions including AMSR-E (Advanced Microwave Scanning Radiometer for EOS), SMAP (Soil Moisture Active Passive) and SMOS were compared. The approaches include: the Single Channel Algorithm (SCA) for horizontal (SCA-H) and vertical (SCA-V) polarizations, the Dual Channel Algorithm (DCA), the Land Parameter Retrieval Model (LPRM) and two simplified approaches based on statistical regressions (referred to as ‘Mattar’ and ‘Saleh’). Time series of vegetation indices required for three of the algorithms (SCA-H, SCA-V and ‘Mattar’) were obtained from MODIS observations. The SM retrievals were evaluated against reference SM values estimated from a multiangular 2-Parameter inversion approach. As no in situ SM data was used, the evaluation made here is relative to the use of this specific reference data set. The results obtained with the current base line algorithms developed for SMAP (SCA-H and -V) are in very good agreement with the ‘reference’ SM data set derived from the multi-angular observations (R2≈0.90, RMSE varying between 0.035 and 0.056m3/m3 for several retrieval configurations). This result showed that, provided the relationship between vegetation optical depth and a remotely-sensed vegetation index can be calibrated, the SCA algorithms can provide results very close to those obtained from multi-angular observations in this study area. The approaches based on statistical regressions provided similar results and the best accuracy was obtained with the ‘Saleh’ methods based on either bi-angular or bipolarization observations (R2≈0.93, RMSE≈0.035m3/m3). The LPRM and DCA algorithms were found to be slightly less successful in retrieving the ‘reference’ SM time series (R2≈0.75, RMSE≈0.055m3/m3). However, the two above approaches have the great advantage of not requiring any model calibrations previous to the SM retrievals.

Spatial and temporal soil moisture estimation from RADARSAT-2 imagery over Flevoland, The Netherlands

► Assess the ability of RADARSAT-2 for soil moisture retrieval over Flevoland. ► Evaluate both spatial and temporal soil moisture retrieval techniques. ► Circumvent surface roughness field measurements. ► Spatial and temporal retrievals have accuracy around 4 vol% for medium rough soils. ► Temporal retrieval performs better for smoother fields covered by stubbles. Both spatial and temporal soil moisture dynamics have a large impact on hydrologic processes in agricultural catchments. As acquiring ground measurements of soil moisture is labor intensive, it is often limited in space (to a few locations) and time (to a small number of campaigns). Remote sensing offers a useful alternative, as it allows for observing both across time and space. This study analyses the potential to retrieve spatial and temporal soil moisture information from a series of RADARSAT-2 HH- and VV-polarized C-band (5.405 GHz) backscatter observations over a large number of (nearly) bare soil agricultural fields in Flevoland, The Netherlands, acquired in the framework of AgriSAR 2009. The RADARSAT-2 backscatter observations are found to be strongly correlated with continuous soil moisture measurements recorded at two agricultural sites during the 2009 growing season, as well as with spatial soil moisture measurements conducted during three intensive field campaigns in August and September 2009. Furthermore, two different soil moisture retrieval approaches have been evaluated: a physically-based modeling approach and change detection technique. The physically-based approach makes use of the Integral Equation Model with input of effective surface roughness parameters that are updated every acquisition for each field. The soil moisture retrieval accuracy of this technique is calculated based on leave-one-out cross-validation, yielding an accuracy (RMSE) about 4 vol% for fields with medium surface roughness at both polarization schemes. For smoother fields covered by stubbles, the retrieval accuracy slightly deteriorates. The change detection technique relies on a normalization of the backscatter between dry and wet reference soil moisture measurements, yielding accuracies about 4 vol% both for fields characterized by medium roughness and stubble conditions. This study confirms the large potential of RADARSAT-2 data for the retrieval of spatial and temporal soil moisture dynamics.

Overview of Soil Moisture Retrieval from Passive Microwave Remote Sensing

The surface soil moisture is a quantity resulting from the water and energy exchanges at the land-surface/atmosphere interface. Measuring this variable has potential application in hydrology and meteorology. Passive microwave remote sensing techniques have great potential for providing estimates of soil moisture with good temporal repetition on a daily basis and on a regional scale (~10 km).Compared with optical and thermal-infrared sensors, although passive microwave measurements are of lower spatial resolution than optical and infrared measurements, they are less influenced by solar illumination, aerosols, and clouds, and they are responsive to different dynamic ranges of vegetation structure and biomass. Compared with active microwave sensors, the passive microwave systems include frequent coverage, low data rates, simpler data processing. However, the effects of vegetation cover, soil temperature, snow cover, topography, and soil surface roughness also play a significant role in the microwave emission from the surface. Different soil moisture retrieval approaches have been developed to account for the various parameters contributing to the surface microwave emission. In this paper, approaches for monitoring soil moisture using passive microwave remote sensing were discussed and compared, especially the retrieval approaches which will have to be used in the future operational application.

A Model Analysis Using a Combined Active/Passive Microwave Remote Sensing Approach for Soil Moisture Retrieval

This study develops a new algorithm for estimating bare surface soil moisture using combined active/passive microwave remote sensing on the basis of TRMM(Tropical Rainfall Measuring Mission).To accurately estimate bare surface soil moisture,the algorithm simultaneously uses a precipitation radar(PR) and microwave imager(TMI) for observation.In this paper,through an analysis of the numerically simulated surface backscattering and emissivity database by Advanced Integral Equation Model(AIEM) with a wide range of soil moisture and surface roughness conditions,the authors have found that the surface roughness parameter has an important impact on both backscattering coefficient and emissivity.Based on the simulated database by AIEM,the authors use theoretical model(Geometrical Optical Model) and semi-empirical model(Qp Model) through the analysis of surface roughness as well as TRMM/PR and TMI data to develop the algorithm for estimating bare surface soil moisture.

Retrieving surface soil moisture over a wheat field: Comparison of different methods

The main objective of the Soil Moisture and Ocean Salinity (SMOS) mission over the land surface is to monitor surface soil moisture (SM) from measurements of the microwave emission at L-band. These measurements will be done at many incidence angles and two polarizations. The multiconfiguration system of SMOS provides the possibility to retrieve one or several surface parameters in addition to soil moisture. Such an approach, referred to as N-parameter retrievals (where N is the number of retrieved parameters), could make the retrieval process more efficient. In this study, this possibility was investigated using an experimental dataset acquired over a wheat field. First, a detailed analysis of the attenuation properties of the wheat canopy, a vertical-stem dominated crop, was made. The results showed that the polarization and incidence angle dependence of attenuation was very significant and could be accurately modeled using only two parameters: the optical depth τ at nadir ( τ( θ=0)) and a polarization parameter ( C pol). The need for this latter parameter results from the difference in the dielectric properties of wheat for the vertical and horizontal polarizations. Based on these results, a detailed comparison of different soil moisture retrieval approaches was made. One-, two-, and three-parameter retrieval approaches were tested. In this latter method, soil moisture and the two vegetation parameters ( τ( θ=0) and C pol) were retrieved simultaneously. This method was found to be the most efficient: it provided the best accuracy in the SM retrievals and it did not require ancillary information on the vegetation characteristics, which cannot be easily estimated for operational SM mapping applications.

Retrieving near-surface soil moisture from microwave radiometric observations: current status and future plans

Surface soil moisture is a key variable used to describe water and energy exchanges at the land surface/atmosphere interface. Passive microwave remotely sensed data have great potential for providing estimates of soil moisture with good temporal repetition on a daily basis and on a regional scale (∼10 km). However, the effects of vegetation cover, soil temperature, snow cover, topography, and soil surface roughness also play a significant role in the microwave emission from the surface. Different soil moisture retrieval approaches have been developed to account for the various parameters contributing to the surface microwave emission. Four main types of algorithms can be roughly distinguished depending on the way vegetation and temperature effects are accounted for. These algorithms are based on (i) land cover classification maps, (ii) ancillary remote sensing indexes, and (iii) two-parameter or (iv) three-parameter retrievals (in this case, soil moisture, vegetation optical depth, and effective surface temperature are retrieved simultaneously from the microwave observations). Methods (iii) and (iv) are based on multiconfiguration observations, in terms of frequency, polarization, or view angle. They appear to be very promising as very few ancillary information are required in the retrieval process. This paper reviews these various methods for retrieving surface soil moisture from microwave radiometric systems. The discussion highlights key issues that will have to be addressed in the near future to secure operational use of the proposed retrieval approaches.

Soil moisture retrieval from AMSR-E

The Advanced Microwave Scanning Radiometer (AMSR-E) on the Earth Observing System (EOS) Aqua satellite was launched on May 4, 2002. The AMSR-E instrument provides a potentially improved soil moisture sensing capability over previous spaceborne radiometers such as the Scanning Multichannel Microwave Radiometer and Special Sensor Microwave/Imager due to its combination of low frequency and higher spatial resolution (approximately 60 km at 6.9 GHz). The AMSR-E soil moisture retrieval approach and its implementation are described in this paper. A postlaunch validation program is in progress that will provide evaluations of the retrieved soil moisture and enable improved hydrologic applications of the data. Key aspects of the validation program include assessments of the effects on retrieved soil moisture of variability in vegetation water content, surface temperature, and spatial heterogeneity. Examples of AMSR-E brightness temperature observations over land are shown from the first few months of instrument operation, indicating general features of global vegetation and soil moisture variability. The AMSR-E sensor calibration and extent of radio frequency interference are currently being assessed, to be followed by quantitative assessments of the soil moisture retrievals.