Surface and aerodynamic roughness in arid and semiarid areas and their relation to radar backscatter coefficient

Béatrice Marticorena, Mouldi Kardous, G. Bergametti, Yann Callot, Patrick Chazette, Houcine Khatteli, Sylvie Le Hégarat‐Mascle, M. Maillé, Jean‐Louis Rajot, D. Vidal‐Madjar, Mehrez ZribiJournal of Geophysical Research Atmospheres 2006 · DOI

TL;DR

This paper investigates the relationship between surface roughness and radar backscatter in arid and semi-arid regions. By combining field measurements of geometric roughness (lateral cover) and aerodynamic roughness (roughness length) with C-band radar data, the authors propose an empirical method to estimate aerodynamic roughness from satellite observations, crucial for modeling dust emissions.

What Problem Does This Paper Address?

Surface roughness is a critical factor influencing the exchange of mass and energy between the atmosphere and the Earth's surface, particularly in arid and semi-arid environments where it plays a significant role in aeolian (wind-driven) erosion and dust emission. However, detailed field measurements of surface roughness in these regions are scarce, hindering accurate modeling of dust transport. Existing empirical relationships between geometric and aerodynamic roughness, often derived from wind tunnel experiments with artificial surfaces, are not always directly applicable to natural, complex terrains. Furthermore, the potential of satellite-based radar systems to map this crucial surface parameter over large areas has not been fully explored for operational applications. This paper aims to bridge this gap by:

  1. Quantifying both geometric and aerodynamic roughness in diverse natural arid/semi-arid landscapes.
  2. Validating existing relationships between these two types of roughness for natural surfaces.
  3. Investigating the capability of C-band radar backscatter coefficients to retrieve aerodynamic roughness length.

Methods

The study employed a multi-faceted approach:

  • Field Campaign: An intensive field campaign was conducted in southern Tunisia over 10 sites with varying surface roughness characteristics.
  • Geometric Roughness Measurement (Lateral Cover, $L_c$): $L_c$ was determined by measuring the frontal silhouette area of roughness elements (vegetation, pebbles, gravel) per unit of horizontal surface. This involved a line-intercept method to cover sufficient area and measure the dimensions of individual elements. Assumptions were made about the shapes of roughness elements (half-ellipsoid for vegetation, rectangular for pebbles/boulders) to calculate their frontal areas.
  • Aerodynamic Roughness Length ($Z_0$) Measurement: $Z_0$ was experimentally determined using high-precision wind velocity and air temperature profiles. This involved deploying a 10-m telescopic mast with multiple anemometers and temperature probes. Data were processed using a gradient method based on Monin-Obukhov similarity theory, with rigorous data selection criteria to ensure reliable inversions.
  • Radar Data Acquisition and Processing: C-band SAR/ERS images were acquired over the study sites. Preprocessing included incidence angle corrections and a correction for the variation of backscatter coefficient ($\sigma_0$) with incidence angle using the Integral Equation Model (IEM).

Datasets

  • Field Data: Measurements of lateral cover ($L_c$) and aerodynamic roughness length ($Z_0$) were collected over 10 experimental sites in southern Tunisia. This dataset represents a novel collection of in-situ roughness measurements for diverse arid and semi-arid natural surfaces.
  • Remote Sensing Data: SAR/ERS images were used. This dataset is a standard operational dataset for Earth observation.
  • Literature Data: The study also referenced and compared its findings with existing datasets from previous studies, including airborne radar data (AIRSAR, SRL) and other spaceborne radar data (SIR-C SLR) from different arid and semi-arid regions (North Africa, South Africa, North America).

Key Insights and Ideas

  • Natural Surfaces Follow Wind Tunnel Trends: The study found that empirical relationships established from wind tunnel experiments with artificial roughness elements generally hold true for natural arid and semi-arid surfaces. This suggests that the aerodynamic behavior of natural elements like vegetation and pebbles can be approximated by simpler geometric shapes.
  • Geometric Roughness Predicts Aerodynamic Roughness: The research confirms that geometric characteristics of the surface, particularly lateral cover ($L_c$) and the ratio of aerodynamic roughness to mean height ($Z_0/h_w$), are strong predictors of aerodynamic roughness length ($Z_0$). This is a crucial finding for remote sensing applications where direct wind profile measurements are not feasible.
  • Radar Sees Roughness: C-band radar backscatter coefficients ($\sigma_0$) are shown to be significantly correlated with both geometric roughness ($L_c$) and aerodynamic roughness ($Z_0$). This indicates that radar can effectively sense and quantify surface roughness in arid and semi-arid environments.
  • Operational Mapping Potential: The strong correlation between radar backscatter and aerodynamic roughness, consistent with previous studies using more specialized sensors, highlights the potential for using operational spaceborne radar systems (like ERS/SAR) to map aerodynamic roughness over large areas.

Key Results

  • Empirical Relationship for $Z_0$ and $L_c$: The study established a relationship between aerodynamic roughness length ($Z_0$) and lateral cover ($L_c$) for natural surfaces, which aligns well with previous findings.
  • Empirical Relationship for $\sigma_0$ and $L_c$: A logarithmic relationship was found between ERS/SAR backscattered signal ($\sigma_0$) and lateral cover ($L_c$), with $\sigma_0 = 3.31 \ln(L_c) – 3.74$ ($r = 0.87, n = 14$).
  • Empirical Relationship for $\sigma_0$ and $Z_0$: A significant logarithmic relationship was established between ERS/SAR backscattered signal ($\sigma_0$) and aerodynamic roughness height ($Z_0$), with $\sigma_0 = 2.24 \ln(Z_0) - 0.11$ ($r = 0.84, n = 23$). This relationship was further validated by combining their data with previous studies, resulting in $\sigma_0 = 2.73 \ln(Z_0) + 2.05$ ($r = 0.88, n = 28$).
  • Validation of C-band Radar: The study demonstrated that C-band radar measurements can reliably retrieve aerodynamic roughness length, even for semi-arid areas with low permanent vegetation cover (<25%).

Limitations and Future Work

  • Vegetation Porosity and Flexibility: While the study found good agreement with existing models, it notes that the assumption of identical aerodynamic contributions from vegetation and boulders weighted by their lateral cover is only valid if vegetation porosity is negligible. The authors suggest that for sparse vegetation, the contribution of vegetation to the radar signal might be attenuated by the soil backscattering.
  • Site 3 Data: For site 3, aerodynamic roughness length ($Z_0$) could only be estimated under stable or unstable conditions, leading to its exclusion from some analyses to ensure consistency with near-neutral conditions.
  • Site 1 and 9 Exclusion: Site 1 (olive tree plantation) was excluded from radar analysis due to geometric complexity, and Site 9 (Chott Jerid) due to variable soil moisture.
  • Future Work: The authors propose extending their findings to other operational spaceborne radar sensors like ENVISAT/ASAR and RadarSAT, which have similar configurations. They also suggest that their detailed field measurements of surface characterization and wind profiles can serve as a valuable resource for calibrating future radar-based roughness retrieval models. The proposed empirical relationship is expected to improve the assessment of wind threshold friction velocities for aeolian erosion, thereby enhancing dust cycle modeling.

Summary updated

Abstract

Surface roughness is a key parameter for surface‐atmosphere exchanges of mass and energy. Only a few field measurements have been performed in arid or semiarid areas where it is an important control of the aeolian erosion threshold. An intensive field campaign was performed in southern Tunisia to measure the lateral cover, L c , and the aerodynamic roughness length, Z 0 , over 10 sites with different surface roughnesses. L c was determined by combining field measurements of the geometry of the roughness elements and simple assumptions on their shapes. Z 0 was experimentally determined from high‐precision wind velocity and air temperature profiles. The resulting data were found to be in good agreement with the existing relationships linking the geometric and the aerodynamic roughness. This suggests that for natural surfaces, Z 0 can be estimated on the basis of the geometric characteristics of the roughness elements. This data set was then used to investigate the capabilities of radar backscatter coefficients, σ 0 , to retrieve L c and/or Z 0 . Significant relationships were found between σ 0 and both L c and Z 0 . The SAR/ERS data set is in agreement with the SIR‐C SLR data set from Greeley et al. (1997). On the basis of these two data sets including data from different arid and semiarid areas (North Africa, South Africa, North America), we propose an empirical relationship to retrieve Z 0 using radar observations in the C band from operational sensors.
BibTeX
@article{marticorena2006surface,
  author = {Béatrice Marticorena and Mouldi Kardous and G. Bergametti and Yann Callot and Patrick Chazette and Houcine Khatteli and Sylvie Le Hégarat‐Mascle and M. Maillé and Jean‐Louis Rajot and D. Vidal‐Madjar and Mehrez Zribi},
  title = {{Surface and aerodynamic roughness in arid and semiarid areas and their relation to radar backscatter coefficient}},
  journal = {Journal of Geophysical Research Atmospheres},
  year = {2006},
  doi = {10.1029/2006jf000462},
}
Figure