Influence of Surface Roughness Measurement Scale on Radar Backscattering in Different Agricultural Soils

Abstract

Soil surface roughness strongly affects the scattering 1 of microwaves on the soil surface and determines the backscat-2 tering coefficient (σ 0) observed by radar sensors. Previous 3 studies have shown important scale issues that compromise the 4 measurement and parameterization of roughness especially in 5 agricultural soils. The objective of this paper was to determine 6 the roughness scales involved in the backscattering process over 7 agricultural soils. With this aim, a database of 132 5-m profiles 8 taken on agricultural soils with different tillage conditions was 9 used. These measurements were acquired coinciding with a 10 series of ENVISAT/ASAR observations. Roughness profiles were 11 processed considering three different scaling issues: 1) influence 12 of measurement range; 2) influence of low-frequency rough-13 ness components; and 3) influence of high-frequency roughness 14 components. For each of these issues, eight different roughness 15 parameters were computed and the following aspects were eval-16 uated: 1) roughness parameters values; 2) correlation with σ 0 ; 17 and 3) goodness-of-fit of the Oh model. Most parameters had a 18 significant correlation with σ 0 especially the fractal dimension, 19 the peak frequency, and the initial slope of the autocorrela-20 tion function. These parameters had higher correlations than 21 classical parameters such as the standard deviation of surface 22 heights or the correlation length. Very small differences were 23 observed when longer than 1-m profiles were used as well as when 24 small-scale roughness components (<5 cm) or large-scale rough-25 ness components (>100 cm) were disregarded. In conclusion, 26 the medium-frequency roughness components (scale of 5-100 cm) 27 seem to be the most influential scales in the radar backscattering 28 process on agricultural soils.

Investigating soil moisture dynamics for improved applications of radar satellites in agricultural water management

“…A constant value of H r ms = 1 was used over the study period. The selected H r ms was found to be well within the range of values measured by Sano et al (1998) for grass (0.67 -1.99 cm) and by Rakotoarivony et al (1996) and Martinez-Agirre et al (2017) for corn after sowing (0.8 -1.5 cm, and ∼0.8 -∼1.8 cm, respectively). In addition, Benninga et al (2019) applied values very close to H r ms = 1 for estimation of soil moisture in the Twente network.…”

Surface Roughness Estimation in the Orog Nuur Basin (Southern Mongolia) Using Sentinel-1 SAR Time Series and Ground-Based Photogrammetry

“…Instead, only the two indices TORT and ZVAL, which address for the vertical and horizontal variations, were correlated. Similarly, [29] has found best relationship between C-Band SAR data and the ZVAL for natural sites in high arctic environment, while results of [61] over different agricultural soils showed better correlations for surface roughness indices just sensitive to horizontal surface variations compared to indices combining the horizontal and vertical component. As such, the environmental setting and/or the treatment of the surface (e.g., harrowing, plowing) has major influence on the sensitivity.…”

The significance of soil properties to the estimation of soil moisture from C-band synthetic aperture radar

“…The process of estimating soil moisture from C-band SAR data is complicated by scattering from vegetation (Lang and Sidhu, 1983), scattering due to soil surface roughness (Álvarez-Mozos et al, 2009;Martinez-Agirre et al, 2017), temperature dependence (Rodionova, 2017b), scattering from man-made objects (Tadono et al, 2000) and sources of radio frequency interference (RFI) (Monti-Guarnieri et al, 2017).…”

“…H rms is equivalent in approximate magnitude to the Random Roughness (RR) (Currence and Lovely, 1970) parameter often used by agriculturalists and soil conservationists, which is the standard deviation of the soil height relative to a plane of best fit through the soil surface. The determination of these parameters by remote sensing is of interest, because surface roughness is one of the main sources of m v retrieval errors from satellite SAR sensors (Martinez-Agirre et al, 2017;Lievens et al, 2009). A number of strategies to eliminate surface roughness effects in the context of soil moisture retrieval from SAR were reviewed by McNairn and Brisco (2004), who discovered that to separate the effects of soil moisture and soil surface roughness requires diversity in measurement (Gorrab et al, 2016) comprising, multi-frequency (such as combining different SAR sensors (Zhang et al, 2018)), multi-angle (which might be achieved with different orbits of the same sensor (Wang et al, 2016)), or multi-polarisation.…”