I am going to try removing two slits from the OI dataset. This is because they give peak velocities that are far too red.
The offending slits are #32 and #35
The second one certainly looks rather ropey but there doesn't seem much wrong with the first one. I don't really see why they are bad. They have substantial values of ufiddle. The full list will be saved in jw4-full.gnumeric.
That improved things quite a lot! The results can be seen in OrionFlowComparison:vewplots
Trying to do the moment graphs of the horizontal slits in order to
compare with the moments that come out of wisomom for the vertical
slits. None of the programs seem to work properly wit the horizontal
slit data at the moment (that is, wmomplot and wspecplot2d). When
I ask for smooth versions of the lines, it makes a fits file full of
Nans. When I don't, it gives me zero or Nan for all the moments.
Another problem is that the background continuum images look really ropey. Looks like it would be better fitting a linear continuum and averaging several rows to improve the s/n.
The backgrounds for the vertical slits look OK - they are linear.
The material from README.oi-fit is now in OrionOiFit
Aagh! I have discovered I was using the wrong parameter file for the
OI slits - I should have been using jw4-sky.csv instead of
jw4.csv. As a result, the sky line wasn't being subtracted, which
explains why we were getting blue velocities in the faint parts. Also,
the ufiddles weren't right. Now that I have corrected that, the mean
and peak velocities look pretty similar for OI, although the mean
velocities tend to be blue-shifted by 2 or 3 km/s. There is also a bit
around 40 arcsec E where the mean dips a lot to the blue. This could
well be due to undersubtracting the sky line in that spectrum (#26:
spec192-OI.fits), which I sould correct sometime. Tried it - didn't
make the slightest bit of difference. Probably did something
wrong. This is something Teresa should sort out.
Did the same check on the EW variation with horizontal slit for SIII - also seemed to work out pretty similar to the wisomom map, but I haven't done a quatitative comparison.