The data set shows videos from experiments described in the paper, where water jets (2 mm or 3 mm diameter) impinge vertically downwards on to a Perspex or glass plate coated with a layer of a soft solid soil. The jet impinges normally, so the soil is removed in a roughly cricular pattern.

The plates are transparent. Removal of the soil was monitored from above using a Go Pro Hero digital camera operating in video mode: removal of the soil was monitored from below using a Canon digital SLR, again operating in video mode, focused on a mirror located beneath the target plate and angled at 45 degrees to the horizontal. These videos were used to characterise removal behaviour, and to calculate the size of the cleared area. For the latter, individual frames were digitised in Matlab and the average radius of the cleared region calculated. The average radius viewed from above was labelled 'b' and that from below was labelled 'a': the labels were chosen for consistency with previous studies which only considered removal monitored from below.

The three sets show examples of different removal behaviour: in 'cratering', material is cleared away from the point of impingement with well defined top and bottom boundaries. In 'blistering', observed with thick layers, the liquid initially penetrates the soil layer and generates a cavern, with a confining shell of removed soil - this 'blister' grows with time and eventually the confining shell ruptures. In 'walling', following blister rupture, the confining shell retains its coherence and moves outwards with the removal front.