ABSTRACT Eddy covariance measurements of aerodynamic roughness length, z 0u , and thermal roughness length, z 0T , were obtained for five discrete observation periods on the Peyto Glacier tongue, where katabatic and non‐katabatic wind flow occurred. Katabatic flow during the first three periods was identified by low‐level wind speed maxima observed in wind profiles of the first 6 m above the glacier surface, most persistently near the 4 m measurement level. This feature was mostly absent from the last two periods, signifying non‐katabatic flow. Microtopographic data from 70 measurement samples distributed over the glacier tongue yielded a mean z 0u value of 2.78 mm to serve as a measurement standard. Eddy covariance sampling was restricted to stationary flow periods of near‐neutral stability, yielding 61 roughness measurement pairs in katabatic flow, 12 in non‐katabatic flow. The z 0u values obtained in non‐katabatic flow were close to the measurement standard, but z 0u values for katabatic flow were one to two orders of magnitude smaller due to vertical flux divergence error. It was also noted that Obukhov lengths, L , in katabatic flow were one order of magnitude smaller than those measured in non‐katabatic flow. Reducing the stability limit on sample selection from z/L ≤ 0.2 to z/L ≤ 0.1 eliminated most katabatic z 0u samples, while removing the stationarity restriction resulted in 120 additional non‐katabatic samples, with sample mean z 0u close to the measurement standard. The non‐katabatic z 0u and z 0T results are consistent with using either a fixed z 0u with dynamically modelled z 0T or a fixed z 0u with z 0T = z 0u /100 in bulk aerodynamic heat transfer models. The Obukhov length may be useful in identifying katabatic flow conditions from single‐level eddy covariance data.
D. Scott Munro (Wed,) studied this question.