Abstract
The effects of spatial, combined spatial and temporal sampling errors, and wind measurement errors on profiler-derived divergence estimates computed using the linear vector point function method are examined. Analysis indicates that divergence errors are minimized when the ratio between the spacing of the profilers and the sampled wavelength (Δx/Lx) is between 0.15 and 0.24 and the ratio between the profiler sampling time to the timescale of the weather system (Δt/T) is less than 0.055.
When Δx/Lx ≤ 0.24, synoptic-scale divergence smaller than ±1.0 × 10−5 s−1 cannot be measured, because the error in the profiler wind estimates is larger than the horizontal velocity gradients. The expected errors in divergence calculations given typical profiler spatial and temporal sampling strategies are examined.
