Towards a quantitative understanding of the effects of wind motion on airborne and satellite SAR imagery of vegetation

The monitoring of environmental land resources with airborne and satellite Synthetic Aperture Radar (SAR) is time-critical and relies on the availability of good quality images often acquired within specific time windows when there are marked differences in backscatter between and among vegetation. Wind disturbance can differentially alter the backscatter and image statistics of different vegetation types in a scene, and cause rapid temporal decorrelation between images. Recently, this temporal behavior has been utilized as a discriminant between vegetation types. However, to date, almost all this work has been based on empirical inspection and supposition, with little or no validation by field observation or modeling. In this paper we present results from a model developed to provide realistic wind-induced motions based on extensive field observations. The results allow us to make a quantitative assessment of the behavior and quality of SAR imagery according to wind conditions, vegetation response, and imaging scheme. We also look at ways of alleviating wind effects by use of appropriate data collection and processing schemes. It is found that the effect of the motion on the imagery is a complex interaction between the motion that describes the characteristic period of oscillation of the target and the way the imaging process samples that motion. There are also critical dependencies on radar frequency, range and resolution. With consideration of the latter dependency we look to likely future SAR platforms which promise markedly increased spatial resolution, but for which motion effects will become more acute.