Penitentes are a common feature of snow and ice surfaces in the semi-arid Andes where very low humidity, in conjunction with persistently cold temperatures and sustained high solar radiation favour their development during the ablation season. As penitentes occur in arid, low-latitude basins where cryospheric water resources are relatively important to local water supply, and atmospheric water vapor is very low, there is potential value in understanding how penitentes might influence the runoff and atmospheric humidity. The complex surface morphology of penitentes makes it difficult to measure the mass loss occurring within them because the (i) spatial distribution of surface lowering within a penitente field is very heterogeneous, and (ii) steep walls and sharp edges of the penitentes limit the line of sight view for surveying from fixed positions and (iii) penitentes themselves limit access for manual measurements. In this study, we solved these measurement problems by using a Microsoft Xbox Kinect sensor to generate small-scale digital surface models (DSMs) of small sample areas of snow and ice penitentes on Tapado Glacier in Chile (30°08'S; 69°55'W) between November 2013 and January 2014. The surfaces produced by the complete processing chain were within the error of standard terrestrial laser scanning techniques. However, in our study insufficient overlap between scanned sections that were mosaicked to cover the studied sites can result in three-dimensional positional errors of up to 0.3 m. Mean surface lowering of the scanned areas was comparable to that derived from point sampling of penitentes at a minimum density of 5 m−1 over a 5 m transverse profile. Over time the penitentes become fewer, wider, deeper and the distribution of slope angles becomes more skewed to steep faces. These morphological changes cannot be captured by the interval sampling by manual point measurements. Roughness was computed on the 3D surfaces by applying previously published geometrical formulae; one for a 3D surface and one for single profiles sampled from the surface. For each method a range of ways of defining the representative height required by these formulae was used, and the calculations were done both with and without using a zero displacement height offset to account for the likelihood of skimming air flow over the closely spaced penitentes. The computed roughness values are in the order of 0.01–0.10 m during the early part of the ablation season increasing to 0.10–0.50 m after the end of December, in line with the roughest values previously published for glacier ice. Both the 3D surface and profile methods of computing roughness are strongly dependent on wind direction. However, the two methods contradict each other in that the maximum roughness computed for the 3D surface coincides with airflow across the penitente lineation while maximum roughness computed for sampled profiles coincides with airflow along the penitente lineation. These findings highlight the importance of determining directional roughness and wind direction for strongly aligned surface features and also suggest more work is required to determine appropriate geometrical roughness formulae for linearized features.