Glacier surface roughness from turbulence and microtopographic observations

Surface roughness is a key control on turbulent exchange over glaciers, yet it is still commonly treated as a constant or tuning parameter in surface energy balance models. This simplification stands in contrast to the strong spatial heterogeneity and temporal evolution of glacier surfaces, particularly during the ablation season. A better understanding of how roughness varies, and how this variability translates into aerodynamic properties, is therefore essential for improving representations of surface–atmosphere exchange in mountainous terrain.Here, we investigate glacier surface roughness using a combination of turbulence measurements from eddy covariance systems and high-resolution digital elevation models (DEMs) derived from close-range photogrammetry, acquired during recent field campaigns on Silvrettagletscher and Hintereisferner. Turbulence data are filtered to retain periods that approximately satisfy the assumptions required for applying the bulk aerodynamic method, allowing for consistent estimates of aerodynamic roughness length. In parallel, roughness length is derived from the DEMs using a microtopographic approach based on the Lettau formulation, explicitly accounting for the orientation of surface elements relative to the flow.Both turbulence-based estimates and microtopographic analysis consistently show that roughness length depends on wind direction. In particular, lower roughness lengths are found for along-glacier (katabatic) flow compared to cross-glacier flow. This behaviour is present on both glaciers, reflecting the anisotropic structure of the glacier surface with roughness elements that tend to be aligned with the flow direction at larger scales. As a result, the effective aerodynamic roughness experienced by the flow varies with wind direction relative to the glacier flowline, and cannot be represented by a single, direction-independent parameter. In addition, a clear temporal evolution is observed over the ablation season: Roughness sharply decreases during the transition from snow-covered to ice surfaces, followed by a progressive increase as melt features such as channels and cavities develop and organize the surface.These results show that glacier surface roughness is highly variable in space, time, and direction. This variability has direct implications for the parametrization of turbulent fluxes and suggests that commonly used approaches based on a single roughness length may be insufficient to capture the complexity of glacier–atmosphere interactions.