Improving ground cover monitoring for wind erosion assessment using MODIS BRDF parameters

Measuring and monitoring controls on wind erosion can facilitate detection and prediction of soil degradation important for food security. Ground cover is widely recognised as an important factor for controlling soil erosion by wind and water. Consequently, maintaining ground cover (e.g., vegetation, crop canopy, crop residue) is a recommended management practice which is widely adopted by farmers and land owners. Wind erosion is a lateral or horizontal process and the amount of ground cover needed to maintain lateral cover (L-c = nbh/S where n roughness elements occupy ground area S and have b and h mean breadth and height, respectively) is not well established. Soil may be removed from beneath or between crop and natural vegetation canopies depending on the width, height and distribution of cover types relative to wind direction and strength. Monitoring by repeated measurement or estimation of ground cover provides information to develop an understanding of its spatial and temporal variation. Fractional cover (f(c)) retrieved from optical satellite remote sensing (e.g., Moderate Resolution Imaging Spectroradiometer; MODIS) provides a consistent and repeatable measure of ground cover when viewed from above. Therefore, f(c) provides an areal assessment of components of ground cover. Fractional cover is consequently not the most appropriate approximation of the protection of the soil from wind erosion. Extant wind erosion model parameterisations of L-c already benefit from the use of satellite-derived cover data (L-fc). However, the parameterisations are not well developed. Here, we address the need for a dynamic (multi temporal), moderate resolution and global metric for wind erosion assessment and modelling. We demonstrate the benefits of using L-c within the context of monitoring ground cover for the assessment of wind erosion and review the basis for estimating ground cover using L-c. We describe a new method for an albedo-based approximation of aerodynamic sheltering (L-omega). We use ray-casting of rough surfaces from an existing wind tunnel study to establish a relation between measured L, and directional hemispherical reflectance omega(dir)(0(center dot), lambda), the so-called 'black-sky albedo' and its inverse to estimate shadow. The relation is confirmed to be dependent on the solar zenith angle (theta) and spectral (lambda) confounding factors (e.g., soil moisture, soil mineralogy). We reduced the A. dependency of omega(dir)(0(center dot), lambda) by normalising with the MODIS (MCD43A1) BRDF parameter A to estimate albedo-based lateral cover (L-omega) globally over space (500 m pixels) and time (every 8 days). We compared L-omega with f(c) and L-fc over time for selected locations in Australia and examined L-omega across Australia and the USA using national biogeographic regions. Consistent with current approaches to estimating L-c our results were not field validated due to the dearth of ground-based measurements. However, our results demonstrate that L-omega will improve wind erosion models particularly over large areas and L-omega, is likely to be a valuable source of decision-support information to guide policy makers and land managers on where, when and how to reduce wind erosion.