Analysis of TerraSAR-X data and their sensitivity to soil surface parameters over bare agricultural fields Nicolas Baghdadi a, ⁎, Mehrez Zribi b , Cécile Loumagne c , Patrick Ansart c , Thais Paris Anguela b a CEMAGREF, UMR TETIS, 500 rue François Breton, F-34093 Montpellier, France b CETP/CNRS,10/12 avenue de l'Europe, 78140 Vélizy, France c CEMAGREF, UR HBAN, Parc de Tourvoie, BP 44, 92163 Antony Cedex, France abstract article info Article history: Received 6 May 2008 Received in revised form 12 August 2008 Accepted 12 August 2008 Keywords: TerraSAR-X ASAR/ENVISAT ALOS/PALSAR Surface roughness Soil moisture The sensitivity of TerraSAR-X radar signals to surface soil parameters has been examined over agricultural fields, using HH polarization and various incidence angles (26°, 28°, 50°, 52°). The results show that the radar signal is slightly more sensitive to surface roughness at high incidence (50°–52°) than at low incidence (26°– 28°). The difference observed in the X-band, between radar signals reflected by the roughest and smoothest areas, reaches a maximum of the order of 5.5 dB at 50°–52°, and 4 dB at 26°–28°. This sensitivity increases in the L-band with PALSAR/ALOS data, for which the dynamics of the return radar signal as a function of soil roughness reach 8 dB at HH38°. In the C-band, ASAR/ENVISAT data (HH and VV polarizations at an incidence angle of 23°) are characterised by a difference of about 4 dB between the signals backscattered by smooth and rough areas. Our results also show that the sensitivity of TerraSAR-X signal to surface roughness decreases in very wet and frozen soil conditions. Moreover, the difference in backscattered signal between smooth and rough fields is greater at high incidence angles. The low-to-high incidence signal ratio (Δσ°= σ 26°–28° /σ 50°–52° ) decreases with surface roughness, and has a dynamic range, as a function of surface roughness, smaller than that of the backscattering coefficients at low and high incidences alone. Under very wet soil conditions (for soil moistures between 32% and 41%), the radar signal decreases by about 4 dB. This decrease appears to be independent of incidence angle, and the ratio Δσ° is found to be independent of soil moisture. © 2008 Elsevier Inc. All rights reserved. 1. Introduction Soil surface characteristics, namely the soil moisture content and roughness, play an important role in hydrological studies. Floods, excess runoff, and soil erosion are, among others, key factors controlled and influenced by soil surface conditions. Spatial remote sensing is a vital tool for measuring and mapping soil parameters, thanks to its capacity to provide permanent coverage over large areas. In active microwave remote sensing, Synthetic Aperture Radars (SAR) provide measurements day and night, whatever the meteorological conditions, with high to medium spatial resolutions (1 m to 1 km), whereas scatterometers provide data with a spatial resolution of a few tens of kilometers (ERS and ASCAT). Moreover, the SAR technique is very sensitive to soil surface conditions, especially roughness and moisture content in the first few centimeters of soil (Bruckler et al., 1988). The radar signal, which depends on various radar parameters (polarization, incidence angle and frequency), is also correlated with (1) the slope, i.e., the topography; (2) the surface roughness; and (3) the dielectric properties related mainly to soil moisture (e.g. Dobson & Ulaby, 1986; Ulaby et al., 1986; Fung, 1994). In recent years, many researchers have demonstrated the potential of microwave remote sensing for the estimation of surface roughness and soil moisture (e.g. Baghdadi et al., 2002, 2006, 2007; Le Hégarat et al., 2002; Satalino et al., 2002; Zribi & Dechambre 2002; Oh, 2004; Alvarez et al., 2005; Mattia et al., 2006; Rahman et al., 2008; Zribi et al., 2008). For natural surfaces, the statistical properties of surface roughness can be summarized using two parameters: the standard deviation of surface height (root mean square, rms) which specifies the vertical scale of roughness, and the correlation length (L) representing the horizontal scale over which similar (correlated) roughness conditions are detected. The relation between the radar wavelength λ and the statistical roughness parameter rms is given by k · rms (k =2π/λ). Thus, with increasing wavelength, the roughness parameter is decreasing. Also, the local incidence angle plays an important role for defining the roughness condition of a soil. At low incidences, the surface appears rougher than at high incidences. The Frauenhofer criterion proposed in Ulaby et al. (1982) considers a soil surface as rough when the phase difference between two rays scattered from separate points on the surface (Δϕ =2· k · rms · cosθ) exceeds π/8 (rms N λ/(32 · cosθ)). Remote Sensing of Environment 112 (2008) 4370–4379 ⁎ Corresponding author. Tel.: +33 4 67 54 87 24; fax: +33 4 67 54 87 00. E-mail address: baghdadi@teledetection.fr (N. Baghdadi). 0034-4257/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.rse.2008.08.004 Contents lists available at ScienceDirect Remote Sensing of Environment journal homepage: www.elsevier.com/locate/rse