Anisotropic Error Distributions in a Bistatic Doppler Radar System

Abstract

Anisotropic errors that contaminate two-dimensional wind vectors determined using a bistatic Doppler radar system composed of a transceiver (traditional Doppler radar) and a receiver are characterized. Two kinds of errors are identified and their influences on the retrieved wind vectors described. Type-1 errors are caused by the finite resolution of the Doppler velocity observed at the transceiver and receiver. Type-2 errors are associated with the finite resolution in detecting azimuth angles at the transceiver and ranges at the receiver. The two errors show remarkably anisotropic behavior. Error vectors for type-1 or type-2 errors have lengths that are a unique function of direction. These error vectors that can appear with equal probability form an ellipse at observation points. The larger the ratio of the length of the longer axis to that of the shorter axis, the higher the anisotropy. For type-1 errors, the longer axis is approximately perpendicular to the radar beam; the shorter axis is parallel to it. The directions of the corresponding axes for type-2 errors depend on the location of the observation point and the wind there. Retrieved wind vectors contaminated with type-1 and type-2 errors are simulated. Random noise is superimposed on a uniform wind with standard deviations of half the resolutions of four parameters. The resultant type-1 and type-2 error vectors both show anisotropic distributions consistent with the error ellipses mentioned above. This suggests that spurious divergence/convergence and spurious rotation dominate cases wherein the wind is parallel or perpendicular to the direction of the longer axis.