摘要：an order of magnitude). The two remote techniques were able to differentiate between ice and snow cover and return momentum roughness lengths that were within 1–2 mm (≪ an order of magnitude) of the in situ eddy covariance values. Changes in wind direction affected the magnitude of the momentum roughness length due to the anisotropic nature of features on a melting glacier surface. Persistence in downslope wind direction on the glacier surfaces, however, reduced the influence of this variability. Scalar roughness length values showed considerable variation (up to 2.5 orders of magnitude) between locations and seasons and no evidence of a constant ratio with momentum roughness length or each other. Of the tested estimation methods, the Andreas (1987) surface renewal model returned scalar roughness lengths closest to those derived from eddy covariance observations. Combining this scalar method with the remote techniques developed here for estimating momentum roughness length may facilitate the distributed parameterisation of turbulent heat flux over glacier surfaces without in situ measurements.
