Depth dependence of heading errors

[richardsc]

More detail to come, but briefly, we are seeing a very persistent depth-dependent heading error for SEA021 M33. Basically, the glider consistently varies over a yo, with a consistently negative heading error at a depth of about 70m. Sometimes the magnitude of the variation (between the surface and 70m) is as much as 90-100 degrees!

This can be seen a couple ways. First, just in the time series plots of heading:

But then also in a "composite" view, which compiles the heading error (difference between measured and desired heading) into bins:

There is a clear "peak" in heading error (of up to -50 degrees!) around 70m depth.

[richardsc]

Just a note that this doesn't appear to affect only sea021. Here's some plots from the beginning of sea024's first mission: