Processing LADCP data with LDEO Matlab code

Note

This document is out of date. LDEO Matlab processing can be downloaded and used for processing, but this is not a definitive guide. It assumes Matlab was installed on a UH LADCP computer (which is probably incorrect)

The processing of data with the LDEO software is straightforward once you have written the setting MATLAB script set_param_cast.m. The LDEO program comes with a well written documentation LDEO How-To [PDF] and you are invited to read and consult it for details. The routines themselves are in /home/currents/programs/ladcp_thurnherr_clone/.

Write set_param_cast.m

The MATLAB script set_param_cast.m sets some of the default parameters used by the LDEO software –the rest of the default parameters can be seen in /home/currents/programs/ladcp_thurnherr_clone/default.m. Especially, it gives the path to the CTD and GPS data. Your final script set_param_cast.m should be close to the one used in I5S 2009 (see set_cast_params.m) and you might need to change a couple of parameters only. In Section How to write set_cast_params.m, we explain line by line the script used in I5S 2009 to enable you to write your own. The script default.m in /home/currents/programs/ladcp_thurnherr_clone/ contains also details on more default parameters that can all be overwritten in set_param_cast.m. An example is given at the end of set_param_cast.m used during I5S 2009 (see How to write set_cast_params.m).

Process the data

Once set_param_cast.m is written, the data can be processed. To process a cast, say of station #56, open MATLAB in /home/science/current_cruise/proc/ladcp_ldeo:

cd /home/science/current_cruise/proc/ladcp_ldeo
matlab -nojvm
addpath /home/currents/programs/ladcp_thurnherr_clone/

and simply execute:

process_cast(56)

Notice that, in CLIVAR cruises, a station number always comes with a cast number –for instance station #56 and cast #2. The LDEO software only needs the station number as input.

The processing itself is entirely automatic and you can follow its progression by watching the different output messages. It can last up to 10-15 minutes for a 5000-m deep cast. Once the processing is done, you should have a set of figures appearing on the screen.

Check the processed data

The documentation LDEO How-To [PDF] describes the content of each figure. Here is what you should look at:

Figure 1

• It contains the profile of U (red) and V (green). Check if they sound physically reasonable, given the location (Are we near a western boundary with a strong surface current? Do we expect deep bottom flow? etc). Look for interesting features, such as eddies or waves. This is the fun and important part of the whole procedure!

Figure 2

• The second panel shows the time series of the depth of the rosette: Does the depth go down and up as expected? Is the curve of the upcast stair-like –the rosette is stopped every time a bottle needs to be closed?
• The third and fourth panels contain information on the tilting (rotation around a horizontal axis) and heading (rotation around the vertical axis) of the rosette. The resident technician and the master might be interested in these data.

Figure 4

• In the middle panel, the red (from LADCP) and black (from CTD) curves should match.
• In the lower right panel, the blue (from LADCP) and red (from CTD) curves should match.

Figure 11

• This tells if the processing went without problem and gives you some warnings otherwise. See a description of the set of warnings can you might encounter.

Figure 14

• This gives the raw signal-strength time series (upper) and the signal strength after editing (lower)

The most important checks are Figures 1 and 11. All figures are saved as .ps files in

/home/science/current_cruise/proc/ladcp_ldeo/processed.

List of warnings

Here is a list of warnings (from Fig. 11) that have been encountered so far:

• “increased error because of shear-inverse difference”: This is a serious problem and it should be treated. Look at some notes on this issue taken during I5S 2009.
• “removed pressure pikes”: This message can be ignored.
• “198 horizontal velocities > 3m/s during the middle of cast”: It may happen in an area with a strong current, such as along western boundary. It might probably be ignored.
• “large U bottom-track bias”: Try without bottom tracking (set p.btrk_mode==0 for the problematic station in set_cast_params.m; see an example at the end of set_cast_params.m used in I5S 2009).

See /home/currents/programs/ladcp_thurnherr_clone/default.m that contains a description of the available defaults, if some of them need to be changed.

Report any problem to hummon@hawaii.edu, and efiring@hawaii.edu