Estimates of point-cloud positional accuracies in terrestrial laser scanning (TLS) datasets are currently limited to rudimentary combinations of GPS position error and manufacturer precision specifications. However, rigorous error propagation techniques can be applied to the three-dimensional TLS points and potentially integrated into software visualization and analysis products. Beyond the immediate value of qualitatively observing the distribution of expected TLS errors within a point cloud, rigorously estimated point errors can be further propagated to quantify expected errors in derived products such as point-to-point distance measurements, best-fit planes or volume computations. We review TLS error sources, detail their propagation through a rigid registration and illustrate the application of estimated TLS point errors to propagated snow volume uncertainties for a large and small TLS dataset. The resulting volume errors are of negligible size compared to the volume magnitudes, in no case exceeding 0.007% of the computed snow volume. For a dataset generating a large snow volume, the method of surface representation (e.g. grid or triangulated mesh) was more influential than the estimated TLS point errors on volume uncertainty. This suggests the random errors inherent in TLS measurement techniques are not a limiting factor in achievable snow volume accuracies.