On November 22nd, attention was brought to the fate of a mysterious island located on various maps and virtual globes between Australia and New Caledonia. Recent explorations by the Australian Hydrographic Office showed that in fact this Pacific Ocean Island did not exist.
As often, our team was asked by various media to provide satellite imagery "proofs". We did so, but emphasizing that it was fairly often that openly available low scale data base like WVS, DCW or VMAP1 were full of shit, because of their production process and the genealogy of the original sources.
We summarized our analysis in the following figure :
Except that if you do not know the complete impact of all the information included in the metadata, you are in deep trouble. In this particular Sandy Island exemple, if you do not know what is DLMB (a geospatial file for AWACS radar surveillance planes) and how it was produced (especially in remote pacific areas), you cannot imagine that the information included in the World Vector Shoreline that Michelin, Google Earth or Bing use, comes directly from Pepe Boyington and Captain Hook. Fortunately, we did produce DLMB for the french AWACS in the 90's and we remember how we specified the production on Mediterranean or West Indies Islands.