Use (and Potential Abuse) of Uncertainty in Hydrography
|Title||Use (and Potential Abuse) of Uncertainty in Hydrography|
|Publication Type||Conference Abstract|
|Authors||Brian R Calder|
|Conference Name||6th International Conference on High-Resolution Survey in Shallow Water|
|Conference Dates||Feb 21 - Feb 26|
|Keywords||Data Processing, Hydrographic Uncertainty, Uncertainty Estimation, Uncertainty Management|
The evaluation and use of uncertainty as a component of hydrographic data processing systems has grown considerably in the last decade. Uncertainty models for sounding data are now common, and progress has been made in developing models, methods and implementations for preserving this uncertainty in intermediate hydrographic data products. Less progress has been made in dealing with expressing the uncertainty in hydrographic data products to the user, however, which we contend should be our ultimate aim.
We draw here a distinction between the uncertainty assessed for observed sounding (and auxiliary) data and uncertainty as expressed to the user, and observe that the former is at best only a component part of the latter. That is, the uncertainty of soundings describes what we observe, where the uncertainty that we might typically express to the user may be more about what we did not observe: completeness instead of accuracy.
We consider the progress in development of uncertainty models for data, and highlight a number of outstanding questions in their implementation. We then discuss the problem of estimating uncertainty of older data, and motivate from it an alternative model of uncertainty for user-specific hydrographic products, outlining the questions that remain to be answered for this model to be effective. Finally, we consider the limits of uncertainty estimation, and their implications for our ability to communicate adequate descriptions of the limits of our knowledge for users.