Looking at Data, Episode 2

As mentioned in a previous post, one of the ways we make sure that our instruments are functioning properly is to plot the data in real time and make sure there are no bizarre glitches. This has already allowed us to catch at least one issue, which was fortunately a problem with the data transfer between computers rather than a problem with the sensors on SWIMS (and therefore more easily resolved). Another advantage of looking at the data as we collect it, even if the plotting is fairly rough and not something you would put in a paper, is that it allows us to see interesting features and keep an eye on how things are changing in time and space. One example of an interesting plot we saw during our first survey is shown below. As with many plots of oceanographic data, there’s quite a bit going on.

To orient you to the plot, we’ll start with the basic set up. Each section of the plot captures roughly two hours worth of profiles from SWIMS (the instrument we are towing behind the ship). Time is plotted horizontally (“Yday”), but since the ship is moving, this also represents change in position horizontally along the ship track. Rather than plotting the profiles themselves, as with the plot shown in the previous post about data, here we are gridding the data, which essentially means filling in the blanks between measurements to get a general idea of the three dimensional structure. These plots may not be publishable in a formal setting yet, but they are incredibly valuable for understanding approximately what we are measuring. The top plot shows the temperature from near the surface down to about 100 meters; the middle plot shows the same thing for salinity and the bottom plot shows the same thing for density (which is calculated from the temperature, salinity, and depth since there is no way to measure it directly with a sensor at the moment).

Plot of temperature, salinity, and density from SWIMS data. We look at plots in color scales like these to help make trends in the data be more visually apparent.

The colorbar to the right of each plot shows what each color represents in terms of a number. For temperature, this is degrees Celsius, with red being warmer temperatures and blue being cooler temperatures. Intuitively, looking at this plot makes sense because in some way we expect warmer things to be reddish. For salinity (in “practical salinity units”), saltier is red and fresher is blue, and with density (kilograms per cubic meter, minus 1000), heavier water is red and lighter water is blue. As we expect to see, the densest water is at the bottom, with progressively lighter water above it. We call this “stably stratified”, because there don’t appear to be any situations where heavy water is sitting on top of light water. When this does happen, we expect it to lead to an overturning circulation where the dense water sinks and the light water rises.

The most interesting feature of this plot is the bands of high and low salinity and high and low temperature that we see. In the very first profile, which would be on the left of each plot, we see a fairly typical temperature progression, where it starts out warm and gets progressively cooler. However, as we move farther to the right on the plot, we can see that the temperature goes warm-cool-warm-cool, and the salinity alternates fresh-salty-fresh in a similar way. Put together, this ends up leading to a warm and fresh bit of water at the surface, followed by a cool and fresh chunk of water, then a warm and salty chunk of water, and finally another cool and relatively fresh chunk of water. As we discussed in the last data post, there doesn’t seem to be much of a mixed layer: both temperature and salinity show substantial variations almost all the way up to the surface.

Observing the evolution of structures like this in time and space, and speaking to why they occur, is one of our goals. This could be an example of “interleaving” (or some other comparable restratification process), where two adjacent water masses with different properties and densities slide past each other horizontally, with the lighter water sliding on top and the denser water below. Cool fresh water could also be the result of a rain storm, that then got covered up with another water mass. One of the things that we will be doing once we have finished collecting data is going back to look at features like this to see if we can combine this data with that gathered from our other instruments to get a more complete picture of what’s going on. This will require some effort to separate spatial features associated with the ship moving from temporal features associated with the dynamics of the region, which is not a trivial problem.