Orion Sii Tere

Anything related to Teresa's thesis work on the [S II] 6716,6731 lines.

Older stuff is contained in OrionSiiJune

Contents:

25 Oct 2004

30 Oct 2004

01 Nov 2004

Improvements to =wisomom_smooth=

02 Nov 2004

What Tere needs to do next
More problems

29 Dec 2004

25 Oct 2004

I have been making maps of sii emission and derived electron density in order to compare with the hydro models of photoevaporation flows.

There are still some remaining problems with the reduced data:

The JW3 region of KPNO slits still has too high an electron density compared with the surrounding slits. The problem seems to be that you can't simultaneously match up to the other slits along the whole slit length. This suggests that the intensity calibration of the slits hasn't been done properly. What Teresa should have done is to calibrate the velocity-integrated emission of each pixel of each slit using the HST SII images. However, I have feeling that all she has done is to apply a global correction to each slit instead of adjusting the relative brightness along the slit (presumably because I never told her to do that). This needs sorting.
When I take a narrowish slice in velocity (in particular, I am looking at the +010 to +020 range), there are big slit-to-slit variations in the 6731/6716 ratio, which give horizontal stripes in the n_e map. This should have been fixed ages ago!

I used Cloudy to work out how the [S II] line ratio changes with n_e, and then fit a simple formula to it. What I get is

  elemental function nefunc(r) result(ne)
    real, intent(in) :: r
    real :: ne
    real, parameter :: n0 = 2489.37802355335, r1 = 0.697116291016009, r2 = 2.33801218015265
    ne = merge( n0*((r-r1)/(r2-r)), NODATA, r>=r1.and.r<r2 )
  end function nefunc

where r is the ratio of 6731/6716.

More about comparison with the numerical models can be found at OrionFlowComparison.

30 Oct 2004

Unfortunately, it turns out that the data that Tere has given me have not been reduced properly. They are missing most of the callibration steps that are outlined in OrionSiiJune. So, I am going to have to go back to my versions of the dat and sort out the few remaining problems.

I have been looking at the data sii[ls]_-100+100.wisomom-{sum,mean,fwhm}-raw that are in the ~/BobKPNO directory. These have date Jun 18 2004, which was actually before most of the work in OrionSiiJune, so that was a bad idea!

So, I will redo it with the -010+040 data, which seems to be the latest (Jun 21).

Commands to calculate the difference map of the two mean velocities and to display it on the range [-2,+2] km/s:

printf '%s\n' siil_-010+040.wisomom-mean-raw \
              siis_-010+040.wisomom-mean-raw \
              1.0 diff_-010+040.wisomom-mean-raw | src/isodiff

printf '%s\n' diff_-010+040.wisomom-mean-raw '-2 2' | src/isopgm

xv -raw -expand 8:2 diff_-010+040.wisomom-mean-raw.pgm &

Note that this is the other way round from what I was doing in OrionSiiJune.

These actually look pretty smooth with no major problems. The worst bit is the top and bottom of the JW6 slits, where you are seeing the effects of velocity shear along the slit. I have done the same but with s/mean/peak/g and the results look a lot noisier. There are also a few problems that only show up on the peak versions. For instance, the JW6 slits appear generally darker. This is maybe because the mean velocities are affected by the spurious red wing in the SPM data, in which case it would have been better to use the peak velocities to constrain the ufiddle values.

Now I see why I got rubbish results when I originally tried to use my data in the eden program. I was using the smoothed fits files, which dated from Jun 19 and so hadn't incorporated the later improvements. I have now regenerated the fits files and get much better results. There is still room for improvement with the ratio for the KPNO JW3 slits, though. This shows up on the -010+040 ratio images, so can be dealt with by adjusting fluxfiddle.

Bad slits are:

#57,58: too bright in siis by 20%
#64: too bright in siis by 30%
#66,67: too faint in siil by 30%

Note: I have to delete the smooth.fits files of the slits where I have changed fluxfiddle in order that they get regenerated.

Am now regenerating all the velocity slice files as per the last script in the 17 Jun 2004 of OrionSiiJune.

01 Nov 2004

Actually, need to modify the scripts slightly now that I am using the 6731/6716 ratio instead of the other way up:

id=wisomom-sum-raw
for i in $(seq -24 4 56); do 
  vv=$(printf '%+3.3i%+3.3i' $i $((i + 4)))
  printf '%s\n' siil_$vv.$id siis_$vv.$id ratio_$vv.$id | src/isoratio
  printf '%s\n' ratio_$vv.$id 0.697 2.338 | src/isopgm
  convert ratio_$vv.$id.pgm ratio_$vv.$id.png
  rm ratio_$vv.$id.pgm
done

And here is an example of how to display a set of these:

for i in $(seq -24 4 56); do
    printf "ratio_%+3.3i%+3.3i.$id.png " $i $((i + 4))
done | xargs xv -raw -expand 8:2 -gamma 2 &

These are the some of the raw ratio images that result


`-004+000`	`+004+008`	`+016+020`	`+028+032`	`+040+044`

Note that there are still a couple of problems:

The red wing of the SPM data suffers from the contaminating line, resulting in too low a ratio for the +040+044 panel.
There are a couple of noisy slits on the E side of the KPNO jw3 section. Unfortunately, this coincides with the region where the density is highest and if I were to remove those slits it would leave a gaping hole in the data.

In general though things are looking an awful lot better now. So, it's time to get back to generating the figures for the paper. This is discussed in OrionFlowComparison.

Improvements to `wisomom_smooth`

We now have user control over the variable interpolation_interval that sets how big a circle we draw when looking for neighbouring slits (in units of the smoothing kernel sigma).
There is now an option to choose uniform weighting of all slits, instead of weighting by the slit flux. Although the latter would seem more "correct", it produces sharp discontinuities in the maps when switching from the KPNO to the SPM slits because the S/N is so much better in the SPM slits so the prog uses them as far as it can.
With the 2D smoothing kernel we now properly mask out any slit pixels that had previously been set to NODATA. This allows the smoothing to interpolate over the bits in some siil slits where we had bad reflections in the spectrograph.

02 Nov 2004

What Tere needs to do next

Make sure she is up to date with all the codes
Rerun all the wisomom_raw jobs with a clean slate to make sure there are no surprises
Remove the red leak line
Check the densities against the horizontal slits

29 Dec 2004

I just spotted something in the SII spectra: SPM-SII/RawData-2003-Jan/spm099.fits shows nicely a contaminant line at about 6740Å that looks very similar to the contaminant line under the red wing of the 6716 line.

Updated: 2004-12-29

Index