Hydrography has no well-established, mathematically rigorous, objective model for data quality. Survey effort can therefore sometimes be applied inefficiently, in that there can be as much effort expended on 1m high rocks in 30m depth as in 5m (which should be of lesser concern), or on areas where there is light to non-existent traffic. A computational model of return-on-investment associated with survey effort might allow assessment of where to survey first in a given area, and to determine when extant survey effort was \“good enough\” to meet survey specification, leading to survey efficiencies.

Mathematical models of risk associated with ship passage have previously been demonstrated as models for generalized end-user chart uncertainty, and re-survey priority estimation. Here, they are proposed as a model for survey completion which can be applied incrementally in order to rationalize the effort being applied in each area of the working grounds, and to determine when the area is sufficiently well surveyed to be considered complete.

This paper demonstrates the requirements for, and implications of, adopting such an approach for survey completeness prediction, focusing particularly on the data dependencies, and model calibration.\ The methods are illustrated with historical survey data from Hampton Roads, VA.

}, author = {Brian R Calder} }