As the international hydrographic community continues to address the question of the irreducible uncertainty in modern surveys, we must ask how we do the same with archived Vertical Beam Echosounder (VBES) and leadline datasets. The ONR funded Strataform project surveyed an area of the New Jersey shelf around 39◦12’N 72◦50’W using an EM1000 Multibeam Echosounder (MBES). This area is also covered by NOAA surveys from 1936- 38 (assumed to be leadline) and 1975-76 (VBES). By comparison of the archival soundings to the MBES data, estimates of measurement error for the archival surveys are constructed as a function of depth. The analysis shows that archival leadline smoothsheets are heavily biased in deeper water because of ‘hydrographic rounding’ and may be unrecoverable, but that the VBES data appear approximately unbiased and may be used to construct products compatible with modern surveys. Estimates of uncertainty for a surface model generated from the archive data are then constructed, taking into account measurement, interpolation, and hydrographic uncertainty (addressing the problems of unobserved areas and surface reconstruction stability). Finally, the paper addresses the generality of the method, and its implications for the community’s duty to convey our uncertainty to the end user.
Journal or Conference Title
Canadian Hydrographic Conference
May 28 - May 31, 2004
Toronto, Ontario, Canada
Canadian Hydrographic Association
Calder, Brian, "On the Uncertainty of Archive Hydrographic Datasets" (2004). Canadian Hydrographic Conference. 301.