Produce Fermi all-sky HEALPIX high-energy images

[cdeil]

For the http://gamma-sky.net/ webpage we want to produce all-sky HEALPIX high-energy images.

We'll then convert those to http://aladin.u-strasbg.fr/hips/ and serve them as default image instead of the one from http://alasky.u-strasbg.fr/Fermi/Color/ which we use now.

My suggestion would be to make RGB images for the energy bands

• R = 10 -- 30 GeV
• G = 30 -- 100 GeV
• B = 100 GeV -- 1 TeV

These are the energy bands I used many years ago (2011 I think) and the maps were interesting. See https://github.com/cdeil/fermi-ts-maps for Galactic plane images ... I also have old all-sky images in CAR and AIT, but I think we should just re-do them.

As for which quantity to show, my experience is that sqrt(TS) is good.
Other options would be smoothed counts or excess or flux.

Having catalog model images and residual images would also be very cool, to see where the catalog represents the data well and where it doesn't.

I don't have a working set of Python scripts any more to produce those TS images. So if someone else has them or is willing to write them and share scripts and / or maps, that would be great!

cc @woodmd @tburnett @dankocevski @adonath @vorugantia

[dankocevski]

Hi Chris,
Are you sure the images that you have on the guthub repository are TS maps? I looked at them only briefly, but they look more like counts maps. Also, generating a TS map (i.e. a model significance map) for the entire sky over the entire lifetime of the mission would be an incredibly computationally intensive operation. I think for your purposes, a simple all sky counts map, cut into three energy bins, might be a better (and far easier) way to go. There are several people in the LAT collaboration that have the tools to easily generate these kinds of images, so I can ping them and see if they can provide the images that you’re looking for. Do you have a preferred projection that you’d like to use?

-Dan

On Jun 29, 2016, at 3:06 AM, Christoph Deil [email protected] wrote:

For the http://gamma-sky.net/ http://gamma-sky.net/ webpage we want to produce all-sky HEALPIX high-energy images.

We'll then convert those to http://aladin.u-strasbg.fr/hips/ http://aladin.u-strasbg.fr/hips/ and serve them as default image instead of the one from http://alasky.u-strasbg.fr/Fermi/Color/ http://alasky.u-strasbg.fr/Fermi/Color/ which we use now.

My suggestion would be to make RGB images for the energy bands

R = 10 -- 30 GeV
G = 30 -- 100 GeV
B = 100 GeV -- 1 TeV
These are the energy bands I used many years ago (2011 I think) and the maps were interesting. See https://github.com/cdeil/fermi-ts-maps https://github.com/cdeil/fermi-ts-maps for Galactic plane images ... I also have old all-sky images in CAR and AIT, but I think we should just re-do them.

As for which quantity to show, my experience is that sqrt(TS) is good.
Other options would be smoothed counts or excess or flux.

Having catalog model images and residual images would also be very cool, to see where the catalog represents the data well and where it doesn't.

I don't have a working set of Python scripts any more to produce those TS images. So if someone else has them or is willing to write them and share scripts and / or maps, that would be great!

cc @woodmd https://github.com/woodmd @tburnett https://github.com/tburnett @dankocevski https://github.com/dankocevski @adonath https://github.com/adonath @vorugantia https://github.com/vorugantia
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub #11, or mute the thread https://github.com/notifications/unsubscribe/AEcS19hNPe1pOuAf38mNZZ0PEHkTino8ks5qQhlngaJpZM4JA1RV.

[cdeil]

Are you sure the images that you have on the guthub repository are TS maps?

It's either TS or SQRT_TS.
I did split into small patches and use a cluster to compute it.

Do you have a preferred projection that you’d like to use?

I'd prefer HEALPIX to avoid the issues at the pole.
Of course if someone has images with any projection, that's mostly OK and I can reproject on HEALPIX and then there will be small errors near the poles.
The tool to generate HIPS (http://aladin.u-strasbg.fr/hips/#tools) does reproject to HEALPIX as first step if normal images are given.

[cdeil]

I just talked to @OlgaVorokh (GSoC student for Gammapy this summer). She will write a script to make Gauss-smoothed HEALPIX counts images from event lists using healpy, as a start on this task.

[cdeil]

We started a script here: https://github.com/gammapy/gammapy-extra/blob/master/experiments/healpix/fermi_counts_hpx.py

[cdeil]

Note to self: I moved the example dataset and script to here:
https://github.com/gammapy/gammapy-extra/tree/master/datasets/fermi_2fhl

[cdeil]

Just a note: if HiPS generation in Python doesn't become available soon (hipspy/hips#119), then doing it via the Java hipsgen for Fermi-LAT is also possible. That's what we did for HGPS, see example here:
https://github.com/cdeil/hgps_webpage_public/blob/master/make_hips_survey_maps.py

[cdeil]

@adonath - as discussed, please move https://github.com/gammapy/gamma-sky/blob/master/gammasky/maps.py to your repo, and also remove the FErmi map fetch here:

gamma-sky/gammasky/fetch.py

Line 31 in 22074c2

def fetch_all_maps():

and then we simply access Fermi-LAT HiPS from gamma-sky.net, but producing them is part of the other project / repo.