Guide to using telluric recipes

1. Intro

There are 3 recipes:

• obj_mk_tellu.py
• obj_fit_tellu.py
• obj_mk_obj_template.py

---

2. Adding to the telluric database

When telluric (hot) stars are observed they should be added to the telluric database.

obj_mk_tellu.py night_name filename

where night_name is the directory and filename is your extracted telluric (hot) star E2DSFF AB file

e.g.

obj_mk_tellu.py TEST2/20180802 2294972o_pp_e2dsff_AB.fits

Outputs of obj_mk_tellu.py are

• {filename}_trans.fits file in the reduced folder and if they pass the quality control TELL_MAP in the telluDB (copied to the telluDB folder)
• {wave solution master filename}_tapas_convolved.npy file (only produced once per machine) - in the telluDB folder

---

3. First guess telluric correction

When science observations are done one wants to correct the E2DS or E2DSFF for tellurics. To do this a template is needed (to measure the continuum I believe?) - before having observations corrected for tellurics this is done by estimating the template.

For every science observation one should run the following:

obj_fit_tellu.py night_name filename

where night_name is the directory and filename is your extracted science target E2DSFF AB file

e.g.

obj_fit_tellu.py TEST2/20180802 2295266o_pp_e2dsff_AB.fits

Outputs of obj_fit_tellu.py are:

• {filename}_tellu_corrected.fits file in the reduced folder and if they pass the quality control TELL_OBJ in the telluDB (copied to the telluDB folder)

As stated in the title to this section This is a first guess at the telluric correction and can be improved with more observations of the same object (See section 4 below)

---

4. Improving the telluric correction

Once one has enough science observations that are telluric corrected one can improve the telluric correction by basing our "guess" of the continuum on the already corrected spectra of the object we are interested in. *Etienne will have to tell you how many "enough" is!! This is the reason the TELL_OBJ are in the telluDB.

To improve the telluric correction "guess" we create an object template with obj_mk_obj_template.py

It is run as follows:

obj_mk_obj_template night_name filename

where night_name is the directory and filename is any E2DSFF AB file for the object (e.g. Gl699) that we have "enough" corrected tellurics for (from section 3).

e.g.

obj_mk_obj_template TEST4/20180527 2279680o_pp_e2dsff_AB.fits

Outputs of obj_mk_obj_template are:

• Template_{OBJNAME}.fits file in the reduced folder and if it passes the quality control OBJ_TEMP in the telluDB (copied to the telluDB folder)
• BigCube_{OBJNAME}.fits file in the reduced folder - this shows a stack of all observations used to make the object template "N observations" by "x pixels" in a cube (one image for each order) - shifted and corrected for wave length solution and BERV

---

5. Updating the corrections for all objects

• Once an object template is made (and probably best to check the BigCube to make sure everything is well behaved, we need to re-process all the science observations (the E2DSFF AB files) with the new telluric correction (OBJ_TEMP) to do that we simply run obj_fit_tellu.py on all science targets for the objects that have an OBJ_TEMP.

i.e. as in section 3:

obj_fit_tellu.py night_name filename

e.g.

obj_fit_tellu.py TEST2/20180802 2295266o_pp_e2dsff_AB.fits

---

Notes:

• I am basically taking what Etienne has told me above so do not take how I explained why we do things as correct just as a rough guide -- however the process is solid and works very well
• All the outputs needed by the "next steps" are copied to the telluDB and unlike the calibDB are NOT copied back into individual night folders (this is because we could have hundred of observations - also this means the data is never passed between nights and hence never between P.I. - unless they have access to the telluDB)
• These codes are much quicker than extraction so re-running is very quick but if one just copies the telluDB you cannot run obj_fit_tellu.py without extracting (or re-extracting for Gl699) the science observations first
• The telluDB we provide has all the tellurics processed (TELL_MAP) for May to Sept
• The telluDB we provide has Gl699 fully processed for May to Sept (i.e. obj_fit_tellu on all E2DSFF_AB followed by obj_mk_obj_template followed by obj_fit_tellu on all E2DSFF_AB)
• For a worked example (that I run to test these recipes) please see the test runs Tellu_Test.run and Tellu_Test2.run here: https://github.com/njcuk9999/spirou_py3/blob/master/INTROOT/SpirouDRS/spirouUnitTests/Runs/Tellu_Test. https://github.com/njcuk9999/spirou_py3/blob/master/INTROOT/SpirouDRS/spirouUnitTests/Runs/Tellu_Test2.run
• Tellu_Test.run and Tellu_Test2.run do not include the re-processing of the science observations in step 5 (As they are not needed to test the code)