The industry's recognized authority for design, engineering and application of equipment and services in the global ocean community
Issue link: http://sea-technology.epubxp.com/i/477967
www.sea-technology.com March 2015 / st 33 clear indication of system health. If the total number of radials decreases be- low a threshold value, it may indicate some problem with the system, which may affect the data quality. The Averaged Bearing of all Radial Vectors. This value should typically be stable over time; hence, a sharp change in this value may indicate an antenna problem. Comparison Between Averaged Radial Bearing for the Measured and Ideal Radials. If the distortion of the environment is low or the antenna is in good condition, the average bearing of radials obtained from the ideal pat- tern and the average bearing obtained from the measured pattern should be very similar. If the difference between measured and ideal radials is above a certain value, then the radial data is fagged as wrong data. If this high- difference value is persistent, it may indicate a possible malfunctioning of the antenna. Threshold values for each one of these parameters have been selected according to a radial diagnostic sta- tistic for each antenna. The temporal evolution of the signal-to-noise ratio, number of radial vectors and average radial bearing diagnostics every 10 minutes for the Ibiza antenna during 2013 gives an indication of the sys- tem performance. During that year, signal-to-noise ratio was consistently larger than 20 decibels. However, the number of radial vectors had an abrupt change during February and March 2013, when radial maps had many gaps, which indicated system malfunctioning. That abrupt change is in accordance with abrupt changes in both average radial bearing for mea- sured and ideal radial vectors, indica- tive of an antenna hardware problem (the antenna was replaced afterward). Consequently, the radial data (and the total vectors) for that period is fagged as "bad data." The diagnostic analysis concludes that various radial param- eters have to be analyzed to ensure the high quality of the radial and total sur- face SOCIB HF radar data. Additionally, an automated bat- tery of tests for post-processing quality control is applied to every single vec- tor of the hourly combined total vec- tors feld, in which temporal spikes, anomalous gradients and out-of-range values in vector module or direction are found and fagged. This automated quality check is based on acceptable ranges for temporal variations in sur- face ocean current velocities, dealing each HF radar gridpoint current node as a single-point-wise current meter. It should be noted that this automated quality check is also implemented and automated for current-meter devices at SOCIB. After the battery of quality tests, each HF radar data is associated with a quality fag in order to indicate the data quality based on the procedures above. These fags indicate "good data," "probably good data," "prob- ably bad data," and "spyke" or "bad data." We should note that when data are fagged as "good data" or "prob- ably good data," that is based solely on radar system performance and individ- ual time series data analysis consider- ations, but further data validation stud- ies are needed to increase data quality and reliability. Validation and Comparison Against Surface Drifters In order to ensure the highest quali- ty of HF radar data, a comparison exer- cise was performed with the use of La- grangian drifters in the Ibiza Channel. The HF radar system (operating at a fre- quency of 13.5 megahertz) measures an average of the frst meter depth of the water column. Lagrangian drifters are highly infuenced by wind effect. In order to have a correct measurement of the surface currents, consistent with the HF radar measurements, the drift- ers used in this study carried different kinds of drogues, which prevent the impact of wind on the data. The com- parison between HF radar and drifter trajectories and instantaneous surface velocities is not obvious, due to the im- pact of the smoothing effect in the HF radar data, and the wind stress effect in the drifters, depending on the drogue depth and shape. The differences between both data sources were quantifed using root- mean square differences and corre- lation coeffcients. Drifters from two oceanographic campaigns were used: three drifters from the TOSCA (Track- ing Oil Spill and Coastal Awareness Network) campaign and one from G- ALTIKA. The TOSCA mission was carried out in the framework of the European Union-funded TOSCA project. Six dif- ferent drifters were launched at differ- ent sites. Three of them were drifting in