We report preliminary results of the analysis of crowd- sourcing data on fan-shaped deposits that are produced by CO2 gas jets depositing sand and dust on top of the seasonal CO2 ice layer at the Martian south pole A database of over 98,000 HiRISE subframes has been produced that is embedded into a web-based graphical interface for marking the position and outline of the fan deposits. Clustering techniques (DBSCAN) are used to combine these data into coordinates and orientations to provide data on the intra- and inter-seasonal developments of the fan-shaped deposits in several active south polar regions.

{"references": ["Scikit-learn: Machine Learning in Python, Pedregosa et al., JMLR 12, pp. 2825-2830, 2011"]}