The enormous increase of remote sensing data from airborne and space-borne platforms, as well as ground measurements has directed the attention of scientists towards new and efficient retrieval methodologies. Of particular importance is the consideration of the large extent and the high dimensionality (spectral, temporal and spatial) of remote sensing data. Moreover, the launch of the Sentinel satellite family will increase the availability of data, especially in the temporal domain, at no cost to the users. To analyze these data and to extract relevant features, such as essential climate variables (ECV), specific methodologies need to be exploited. Among these, greater attention is devoted to machine learning methods due to their flexibility and the capability to process large number of inputs and to handle non-linear problems. The main objective of this paper is to provide a review of research that is being carried out to retrieve two critically important terrestrial biophysical quantities (vegetation biomass and soil moisture) from remote sensing data using machine learning methods.

This paper presents an approach for retrieval of soil moisture content (SMC) by coupling single polarization C-band synthetic aperture radar (SAR) and optical data at the plot scale in vegetated areas. The study was carried out at five different sites with dominant vegetation cover located in Kenya. In the initial stage of the process, different features are extracted from single polarization mode (VV polarization) SAR and optical data. Subsequently, proper selection of the relevant features is conducted on the extracted features. An advanced state-of-the-art machine learning regression approach, the support vector regression (SVR) technique, is used to retrieve soil moisture. This paper takes a new look at soil moisture retrieval in vegetated areas considering the needs of practical applications. In this context, we tried to work at the object level instead of the pixel level. Accordingly, a group of pixels (an image object) represents the reality of the land cover at the plot scale. Three approaches, a pixel-based approach, an object-based approach, and a combination of pixel- and object-based approaches, were used to estimate soil moisture. The results show that the combined approach outperforms the other approaches in terms of estimation accuracy (4.94% and 0.89 compared to 6.41% and 0.62 in terms of root mean square error (RMSE) and R²), flexibility on retrieving the level of soil moisture, and better quality of visual representation of the SMC map.

Due to its relation to the Earth’s climate and weather and phenomena like drought, flooding, or landslides, knowledge of the soil moisture content is valuable to many scientific and professional users. Remote-sensing offers the unique possibility for continuous measurements of this variable. Especially for agriculture, there is a strong demand for high spatial resolution mapping. However, operationally available soil moisture products exist with medium to coarse spatial resolution only (≥1 km). This study introduces a machine learning (ML)—based approach for the high spatial resolution (50 m) mapping of soil moisture based on the integration of Landsat-8 optical and thermal images, Copernicus Sentinel-1 C-Band SAR images, and modelled data, executable in the Google Earth Engine. The novelty of this approach lies in applying an entirely data-driven ML concept for global estimation of the surface soil moisture content. Globally distributed in situ data from the International Soil Moisture Network acted as an input for model training. Based on the independent validation dataset, the resulting overall estimation accuracy, in terms of Root-Mean-Squared-Error and R², was 0.04 m³·m⁻³ and 0.81, respectively. Beyond the retrieval model itself, this article introduces a framework for collecting training data and a stand-alone Python package for soil moisture mapping. The Google Earth Engine Python API facilitates the execution of data collection and retrieval which is entirely cloud-based. For soil moisture retrieval, it eliminates the requirement to download or preprocess any input datasets.

The successor to the current series of Metop advanced scatterometers (ASCATs), the Metop-SG SCA, will be able to record data in dual-polarization, at C -band. The aim of this study is to investigate if the information contained in the added cross-polarization measurements can improve the vegetation parameterization for the estimation of the soil moisture content. In case of the operational Hydrology Satellite Application Facility Metop ASCAT soil moisture product, vegetation dynamics are characterized by the relationship between radar backscattering intensity and the incidence angle, the so-called SLOPE parameter. Building on findings from previous studies, the assumption is that the polarization ratio, i.e., VH/VV, could improve this characterization. To verify this assumption, flexible approaches, able to integrate a combination of ASCAT VV data and AQUARIUS (NASA) VH/VV data were required. Two machine learning methods were chosen: Support-vector-regression and artificial-neural-networks, and one statistical approach, the Bayesian-Regression. Each of these methods were used to derive models with different input configurations, with and without characterization of vegetation. The results show that the information contained in the SLOPE parameter and in PR are similar. Based on a global average, almost identical SMC retrieval accuracies were achieved. Despite that, analysis of the temporal dynamics of SLOPE and PR revealed certain location specific differences, which affect the spatial distribution of SMC retrieval accuracies. As a result, improvements based on the combination of the two parameters are minor overall, but they can be significant locally.