CTD data processing for S9511 Report compiled by : G. Critchley Processing for this voyage was a long and tedious undertaking, as it is one of the voyages using the Neil Brown acquisition software and the older version of SSDLS (Southern Surveyor Data Logging System). When running "ctdnav" (to extract UTC time, date, position) from the SSDLS files there was a mismatch between UTC and Local date/time giving positions that were up to 120 nautical miles from the true position. The header files (*.HDR) and bust data files (*.BTL) had the correct dates (UTC) and approximate times and positions, but they were incorrectly formatted at entry and in some cases were missing altogether. Bearing in mind the fact that local SSDLS files were in local time - 11 hours earlier than UTC time this mismatch was taken into account when manually (using Excel) the up and down cast header files were edited with correct UTC date, time, and position - in the correct and acceptable format for the files. From these files a new version of ctdnav.001 was manually edited using "vi" to create a proper and correct version of ctdnav.001. During the processing of the corrected files it was noted that some HDR and BTL files actually had the incorrect station numbers and some HDR files were found to be missing. The station numbers were edited for correct station number in the downcast and upcast HDR files and also the BTL files. Copies of uncorrected HDR and BTL files were kept in :/bulkdatadir/orig in case they were required later. The corrected files were copied to :/bulkdatadir/raw (where the path for the bulk data directory was : /altim/critchle/s9511/). The reporting and editing program "rep" would not run correctly - looking for incorrectly named files - despite recompiling of rep a number of times. Neil White corrected the problem. A full report, station list, pressure offsets and generation of the "ctdmaster .sho" file were created. The calibration coefficients for the instrument were entered : 6 pressure coefficients for the up and the downcast files (from 11/12/93), 3 temperature coefficients (from 25/5/93) and 3 conductivity cell correction coefficients (on a cast grouping basis). The station list was compared to the corrected "ctdnav.001" file and all positions were found to be true. From the files generated, it was found that no CTD data exists for : Station 92 - down cast Station 93 - down cast Station 87 - upcast station 99 - up and downcast When running "aver" to produce the 2 dBar averaged files for a preliminary examination of the data it was noted that on a number of casts there were not enough data points for the software to create the 2 dBar binned data. This is in all likelihood due to the on-going problem with the winch design on Southern Surveyor - constant tension as opposed to constant speed. This also led to the creation of large and spurious errors in the initial calibration template file e.g. Temperatures of -5.xx degrees. This was corrected by application of corrected calibration coefficients for T and P as above. This also necessitated recommencing processing from the start. It should also be noted that the record keeping on the CTD logs and entry on the acquisition software was abysmal, which did not make matters any easier in the processing: 1. No records were kept on the log sheets as to which D/A channel was assigned to which accessory - Fluorometer or radiometer - or when they were unassigned. 2. No burst sample point data were recorded on the CTD logs i.e. Pressure, Temperature or Salinity. 3. No signs (+ or -) were applied to position data in the acquisition software to discriminate between North/South and East/West. 4. No records were kept of the Rosette ramp position before and after a cast. This is invaluable as a tool in determining if there has been a misfire or not. "Crupro" and "aver" fell over looking for the wrong data files and the programs had to modified and recompiled. "Task32" was run to compare hydro to sample burst data. Some hydro was found to be missing, this was entered and a new hydro *.csv file generated. This was then merged with the calibration template file. This gave poor results until an acceptable version was created using an offset of 18.0 M and a range of 12.0 M. When processing was continued it was apparent that something had happened between or at the end of station 9 and station 11 to cause a very large offset to the conductivity sensor. This meant that the voyage had to be treated as 2 distinct entities - stations 1 - 9 and stations 11 - 105. One result of this is the loss of data for station 10, as there were no hydrology samples taken for this cast. The data for the second part of the voyage - stations 11 - 105 were very noisy which led to suspicions that the conductivity cell may have been cracked. To determine if this was the case a rough overall calibration was conducted for these stations to see if the samples agreed "roughly" to the downcast data plots. There was general agreement so calibration was proceeded with. Bad data points were removed from the calibration template file as in the usual way. The noisiness in the data is probably due to two things, one is that the area chosen for the work is a region of high dynamics and secondly the sampling strategy was one that involved a lot of shallow stations with few deep stations with deep samples in. Sea conditions experienced in this region could also account for some of the noise. When all undesirable data were removed from both sections of the voyage, appropriate station groupings of consistent incremental and decremental offsets in conductivity were chosen and the voyage CTD data were calibrated. It should be made aware to people using this data that there are minor density inversions in the upper 10 -16 meters of water on many stations. There were too many minor inversions to successfully track down and delete from the raw data so these have been left in the data set. They are more than likely due to either the sea conditions creating a "wake effect" or sensor response rates upon immersion. Upon calibration, the standard deviation of salinity was 0.0058. Given the noisy nature of the data bursts/samples this is not too bad. It should also be noted that this voyage has 1 dBar averaged files not the usual 2 dBar averaged files. In summary no CTD data exists for : Station 92 - down cast Station 93 - down cast Station 87 - upcast station 99 - up and downcast station 10 - downcast Gary Critchley 21/11/1996. Oxygen Sensor Calibration for cruise S9511 The oxygen calibration for this voyage was carried out according to protocols used by the Divisional Data Products Group. Processing was commenced using the calibrated 1 dBar averaged files that were generated for the C, T, and D data. The Oxygen calibration software could not handle this so the 1 dBar files were saved with an extension of *.1dB and the oxygen data for these files were stripped out. New files of 2 dBar averaged data (corrected for CT & D) were generated after encountering and rectifying problems with some of the software - "crupro and aver". From the voyage information, the first 9 stations showed no D.O. sensor response - flat line only. A new sensor was installed but did not appear to respond until station 13. Thus there are no Oxygen profile data for stations 1-12 inclusive. All stations were calibrated as singlets rather than as groups due to the noisy nature of the data. Fit parameters were tried for each station (where there were enough samples taken) in order to obtain the minimum value of "dox" . When it was not possible to obtain a good fit with high "dox" or "tau" values the number of parameters was reduced from 6 to 5 and a default value of 8.000 was used for tau. This was the first voyage where I have had to calibrate data for Dissolved Oxygen. There were a number of problems encountered on the way and a sharp learning curve to go through. I found the DPG processing manual lacking in detail and will endeavour to assist in modification of the manual. For the final calibration of the Dissolved Oxygen profile data a total of 335 out of 450 samples were used. The standard deviation of the residual is equivalent to 0.127 ml/l. The following stations contain NO calibrated Oxygen data: Station 01 - 12 Station 15 Station 19 - 22 Station 25 Station 29 Station 31 - 32 Station 34 Station 36 Station 40 - 41 Station 44 Station 46 - 49 Station 52 Station 55 - 57 Station 59 - 60 Station 62 - 64 Station 66 Station 68 Station 73 Station 75 - 77 Station 79 - 80 Station 93 - 90 Station 92 - 93 Station 99 - 105 Reasons for there being no calibrated Dissolved Oxygen profiles for these CTD casts is that there were too few, or no samples taken on these stations to determine fit co-efficients. When the data files were placed in the archive "EXPORT" the oxygen calibrated files were saved with an extension of *.all (2dBar averaged). The 1dBar averaged files (no dissolved oxygen) were saved with an extension of *.1db.all. *.cdf files were created for the 2dBar files. Gary Critchley 24/03/97