A Comparison of Several “Single-Pass” Estimators of the Standard Deviation of Wind Direction

Abstract

Computation of the standard deviation of wind direction σ? generally requires repeated consideration of the individual measurements of wind direction. This need for multiple passes through the data sample can create a storage problem for small or remote data acquisition systems. Statistical quantities that can be computed during a single pass through incoming wind speed and direction data are first reviewed, and relationships between these quantities are discussed. A single-pass estimator for σ? is then developed and found to correspond to within ±2% of conventional, two-pass σ? computations for unimodal angular distributions spanning the full σ? range of from 0 to 103.9°. Similar comparisons are presented for other, recently proposed estimators of σ?.