Maximising Data Return: Towards a quality control strategy for Managing and Processing TRDI ADCP Data Sets from Moored Instrumentation

In this paper, we evaluate, by means of case studies, methods for quality controlling data obtained from Teledyne RDI ADCP measurements (ensembles and bins or depth cells) such that the data can be checked and acceptance (pass or fail) criteria established. The QA strategy applied utilizes the quality control criteria recorded for each beam and bin by the Teledyne RDI ADCP itself and which are stored within the binary data structure of the recorded file. These stored parameters include the ´Percentage Good Pings´, ´Number of 3/4 Beam Solutions´ and ´Correlation Magnitude´ obtained from the Broadband Signal Processing. ´Error Velocity´ values reported from comparison of the two independent vertical velocity measurements from each of the ADCPs beam pairs are also used, either as a percentage of the overall magnitude of current velocity, or as absolute values specific to the instrument deployment. Using a relational database approach to the problem of managing and manipulating current profiler data a quality parameter for each measurement bin may be assigned as the data is imported, and Standard Query Language (SQL) is subsequently used as a means of manipulating the data-set, allowing time-series to be extracted for a particular vertical window or temporal extent, subject to a quality threshold. In this paper we test two such QA strategies the first which weights the error velocity of each measurement bin and a second tiered QA strategy using a series of test data sets to evaluate the performance and ease of implementation of the two approaches. Whilst it must be recognized that any QA strategy will never recover poor data, the ability to rapidly select out any bins or ensembles that return low quality data invariably allows the opportunity for the user to gain some insight into the possible causes, be they deficiencies in the instruments configuration, deployment limitations, hardware problems or simply the vagaries of the environment in which the measurement- s are being made. The user is thus able to draw conclusions as to how the measurement configuration could be improved on subsequent deployments.