Automatic data checking for stream data

Eric Hatfield - Eric Hatfield , 22 August, 2012

Abstract

Hydrometric data collection is a time-consuming, resource-hungry, activity. Because of the dynamic environments in which data are obtained, errors are inevitable. Many are obvious when the dataset is examined in an appropriate way, but there is often insuffcient time to do the checking required to make the dataset as reliable as possible. Hydstra routines can do much of the work, but they still need to be initiated and assessed personally. It would therefore be helpful to have an automated suite of checks that would try to perform a similar analysis as a competent hydrographer would do and provide an objective assessment of possible. Such a suite has been developed, tested and found generally effective in identifying likely errors, and so enabling the hydrographer to optimise the time available for date verifcation and to zero in on the records where checking is most likely to lead to worthwhile improvements.