The aim of this study is to retrieve the NDVI values during the wheat cycle using the radar data over reference fields located in the Kairouan plain in the center of Tunisia. The developed approach is based on the use of C-band Sentinel-1 acquisition specifically the cross-polarization ratio and the estimated coherence as features of curve fitting equations and machine learning algorithms such as the random forest and the support vector machine regressors. The NDVI retrieve according to the wheat growth stage, at a field scale, was marked by RMSE values lower than 0.13 and bias values under -0.03

In this paper, we investigate the potential of the modified water cloud model to estimate soil moisture in pepper crop fields with drip irrigation in a semiarid area in Tunisia using cross-polarized L-band data (ALOS-2) and C-band data (Sentinel-1) data in Horizontal-Horizontal (L-HH) and Vertical-Vertical (C- VV) polarization, respectively. Within the context of spatially heterogeneous soil moisture, the total backscattering is the sum of pepper row scattering weighted by the vegetation fraction cover (Fe) and the inter-row soil scattering weighted by (1-Fc). The vegetation row contribution is calculated as the sum of volume scattering contribution of pepper and underlying soil components attenuated by the vegetation cover. Due to the presence of drip irrigation, the underlying soil zone is divided into two parts: irrigated and non-irrigated parts. To assess the calibrated model performance, various simulations are performed under different conditions of soil moisture and vegetation biophysical properties. Under various conditions of soil moisture, the results revealed the potential of the suggested model to simulate SAR signal where cover fraction and pepper height values are under 0.4 and 0.5 m, respectively, using L-HH and cover fraction value under 0.3 and vegetation height value 0.3 m, using C-VV data.

Radar remote sensing has shown a high potential for soil surface parameters estimation in different pedo-climatic context. In the present study, we investigated Sentinel-l radar signal in order to analyze its behavior as function of soil moisture and soil roughness. In addition, we evaluated the approach combining the modified Integral Equation Model (IEM-B) and the Water Cloud Model (WCM) for estimating soil moisture in western France. Soil surface parameters were acquired over 4 campaigns during which composite soil samples were collected simultaneously to Sentinel-l acquisition dates. The dates of those campaigns were defined according to the evolution of the soil surface condition, during the agricultural season. The sensitivity of radar signal$\sigma 0$to soil moisture was studied over the 22 reference fields and over the Thiessen polygons created around the measurement points. Linear relationships are observed between the radar signal and volumetric soil moisture less than 35 vol. % with higher sensitivity for VH polarization (0.41 dB/vol.% in VH against 0.26 dB/vol.% in VV). The best correlation coefficients (R) were observed for the VH polarization with the Zs roughness parameter$(\mathrm{R}={}$0.53 and 0.29 for reference fields and Thiessen polygons, respectively). Following that, a comparison of in situ soil moisture with that predicted based on approach proposed by [1], using Neural network algorithm with a training using the two models IEM-B and Water Cloud Model (WCM) allowed an accuracy with an RMSE ranging between 6.1 and 6.5 vol. % for reference fields and Thiessen polygons respectively. These results confirm that the proposed algorithm is accurate to estimate soil moisture.

In this paper, we discuss the potential of L-band Advanced Land Observing Satellite-2 (ALOS-2) images for retrieving soil moisture over cereal fields in a semi -arid area (Merguellil- Tunisia). SAR signal sensitivity was studied as function of in-situ measurements: roughness and soil moisture. Sensitivity to soil moisture was illustrated for three classes of Normalized Difference Vegetation Index (NDVI). Results reveal the impact of soil moisture on L-band data even in dense vegetation class (NDVI > 0.6). High correlations characterize linear relationships between radar signal and vegetation biophysical properties (Leaf Area Index, vegetation height and Vegetation Water Content). Signal modeling over bare soils was evaluated through empirical equation, modified Dubois model (Dubois-B) and modified Integral Equation Model (IEM-B). For covered fields, Water Cloud Model (WCM) was parametrized for HH and HV polarizations (with and without soil-vegetation interactions component) coupled with the best accuracy bare soil backscattering models: IEM-B for co-polarization and empirical models for the entire dataset. WCM coupled to IEM - B illustrates the best performance to estimate soil water content in HH polarization. The integration of soil-vegetation interaction component provides a stable accuracy of soil moisture estimation in HH polarization and improve soil moisture accuracy in HV polarization mode.

Soil texture estimation is important in several applications. However, field sampling or laboratory analyzes are very expensive and not very representative. Besides, existing soil maps are neither exhaustive nor sufficiently precise for modeling and meeting needs at field scale. This study aimed to explore the potential of Sentinel satellites to predict topsoil texture, and more precisely to produce clay content map at fine spatial resolution. With its components and its porosity, soil texture is directly linked to soil moisture. In this context and with the arrival of Sentinel constellation, data are acquired with high spatial and temporal resolution. And soil moisture is retrieved from a synergetic use of Sentinel-l (S-1) and Sentinel-2 (S-2) data between July and early December 2017, over a semi-arid area in central Tunisia. Relationship between soil moisture and clay content is studied and used to produce texture map. Classification algorithm based on random forest (RF) is used for the mapping of clay content classes. The results showed the potential of S-1 and S-2 products to predict soil texture.

Identifying the irrigated areas is essential for waters managers who are in charge of distributing this resource over a large scale. The monitoring of water soil content and irrigation is a powerful tool for water resource management. The potential of Sentinel-1 (S1) and Sentinel-2 (S2) data for estimating the soil moisture and irrigation is studied over covered surfaces. An inversion algorithm of the Water Cloud Model (WCM) was developed after calibrating and validating the model over the Kairouan plain, a semi-arid region in Tunisia. The aim is to restitute soil moisture over the whole region. The developed algorithm used a synergy between S1, radar data in VV polarization, and NDVI derived from S2 optical data at high spatial resolution. The results showed good accuracy between retrieved and measured soil moisture with a Root Mean Square Error (RMSE) lower than 6 vol.%. Then, the resulting soil moisture maps were used for irrigation mapping. The process used a combination of Support Vector Machine (SVM) and Decision Tree classifications to distinguish between irrigated and non-irrigated agricultural fields. Results from the annual irrigation map show that the overall accuracy on the classification is about 77%.

The aim of this paper is to estimate geometric, water and physical surface soil parameters from typical semi-arid regions made over bare study area (North Africa) using multi-temporal X-band SAR images (TerraSAR-X). For spatial and temporal surface roughness estimation, empirical relationships between radar and soil roughness parameters (rms height “Hrms”, and Zg parameter) were proposed. Two roughness classes are identified through radar signal inversion (smooth and ploughed soils). For the retrieval of surface soil moisture at a high spatial resolution, an algorithm combing TerraSAR-X images with continuous thetaprobe measurements was proposed. Two assumptions were studied: (1) roughness variations during the radar acquisition campaigns were not accounted for; (2) a simple correction for temporal variations in roughness was included. Finally, an empirical relationship was established between the mean moisture values retrieved from the SAR images and the percentage of clay over several test fields. Results showed that highly accurate clay estimations can be achieved.

The goal of this paper is to analyze the potential of COSMO-SkyMed and TerraSAR-X SAR measurements over bare soils in order to estimate correctly soil parameters. We analyzed statistically the relationships between X-SAR backscattering signals function of soil moisture and different roughness parameters (the root mean square height Hrms, the Zs parameter and the Zg parameter) at HH polarization and for an incidence angle about 35.5°. Results have shown a high sensitivity of real radar data to the two soil parameters: roughness and moisture. A linear relationship is obtained between volumetric soil moisture and radar signal with the strongest correlation observed with gravimetric moisture measurements. A logarithmic correlation is observed between backscattering coefficient and all roughness parameters. The highest dynamic sensitivity is obtained with Zg parameter.

The aim of this letter is to discuss the influence of radar frequency on the relationship between surface soil moisture and the nature of radar backscatter over bare soils. In an attempt to address this issue, the advanced integral equation model was used to simulate backscatter from soil surfaces with various moisture vertical profiles, for three frequency bands, namely, L, C, and X. In these computations, we investigated the influence of the vertical heterogeneity of soil moisture on the characteristics of the backscattered signals. The influence of radar frequency is clearly demonstrated. A database produced from Envisat ASAR and TerraSAR-X data, which was acquired over bare soils with in situ measurements of moisture content and ground surface roughness, was used to validate the utility of taking the soil moisture heterogeneity into account in the backscatter model.

In this paper, it is proposed to use TERRASAR-X data for analysis and estimation of soil surface texture. Our study is based on experimental campaigns carried out over a semi-arid area in North Africa. Simultaneously to TERRASAR-X radar acquisitions, ground measurements (texture, soil moisture and roughness) were made on different test fields. A strong correlation is observed between soil texture and a processed signal from two radar images, the first acquired just after a rain event and the second corresponding to dry soil conditions, acquired three weeks later. An empirical relationship is proposed for the retrieval from radar signals of clay content percent. Soil texture mapping is proposed over the study site, which includes bare soils and olive groves.

In this letter, it is proposed to use TerraSAR-X data for analysis and estimation of soil surface texture. Our study is based on experimental campaigns carried out over a semiarid area in North Africa. Simultaneously with TerraSAR-X radar acquisitions, ground measurements (texture, soil moisture, and roughness) were made on different test fields. A strong correlation is observed between soil texture and a processed signal from two radar images, with the first acquired just after a rain event and the second corresponding to dry soil conditions, acquired three weeks later. An empirical relationship is proposed for the retrieval from radar signals of clay content percent. Soil texture mapping is proposed over the study site, which includes bare soils and olive groves.

Spatial and temporal variations of soil moisture strongly affect flooding, erosion, solute transport, and vegetation productivity. Their characterization offers numerous possibilities for the improvement of our understanding of complex land-surface–atmosphere interactions. In this paper, soil moisture dynamics at the soil's surface (the first centimeters) and in its root zone (at depths down to 1 m) are investigated using$25 \times 25\ \hbox{km}^{2}$scale data (Advanced Scatterometer (ASCAT)/METorological OPerational (METOP) scatterometer), for a semiarid region in North Africa. Our study highlights the quality of the surface and root-zone soil moisture products, derived from ASCAT data recorded over a two-year period. Surface soil moisture tends to be highly variable because it is strongly influenced by atmospheric conditions (rain and evaporation). On the other hand, root-zone moisture is considerably less variable. A statistical drought-monitoring index, referred to as the “moisture anomaly index,” is derived from ASCAT and European Remote Sensing (ERS) time series. This index was tested with ERS and ASCAT products during the 1991–2010 study period. A strong correlation is found between the proposed index and the standardized precipitation index.

Soil salinization is one of the most hazardous phenomenon accelerating the land degradation processes. Map ping and tracking soil salinity changes is fundamental for anticipating natural disaster, such as desertification, in arid and semi-arid regions. In this work, we establish an empirical model for soil salinity mapping based on a gaussian mixture and using field electrical conductivity (EC) measures. The developed model is tested on saline soil samples collected from the semi-arid region of Kairouan located in central Tunisia. It is based on statistical moments derived from multiband (HH and VV) intensity synthetic aperture radar (SAR) data of the Envisat satellite. The resulting salinity map is composed of three classes of salinity (Low, Medium and High) with respect to the EC measurements. The developed model is validated for low salinity distribution, whereas, it needs more samples to be generalized for medium and high soil salinity content.