Unnecessary rotation of extra-galactic sims?

[msyriac]

SO maps are made in the equatorial coordinate system since that is natural for our noise properties and scan strategy, given we are on the ground. I appreciate that pysm makes maps available that are rotated from Galactic to equatorial coordinates.

However, it seems that these are made available as rotate(extragalactic + Galactic).

This means that the extragalactic simulations have also been rotated from G->E. Since the extragalactic sky is isotropic and uncorrelated with the Galactic sky, this rotation is unnecessary. It in fact makes downstream analysis more complicated and expensive: if an analyst wants to compare the outputs of their analysis pipeline to the original extragalactic inputs (e.g. the Websky input kappa map, halo catalogs or point source catalogs), they must also have access to rotate(extragalactic).

There are two solutions to this:

1. pysm always releases maps that are simply extragalactic + rotate(Galactic). This will save a lot of compute time down the line.
2. pysm also makes available as an extra data product rotate(extragalactic, which analysts can cross-correlate with. This will also save compute time in the future, but will make each pysm data release a bit more bloated.

[zonca]

The main strength of PySM is that it builds a standard for simulations across different collaborations, i.e. if we refer to d1, we refer to one single model, whatever instrument it is being simulated for.

So even if extragalactic is isotropic and uncorrelated, if we simulate cib1, we want that to be consistent, so if we compare a cib1 map in galactic coordinates for Litebird and a cib1 map in equatorial for SO, they better show a reference frame rotation, like any other component.

Being all PySM components natively in Galactic, I decided to consider Websky to be in Galactic coordinates.

So I do not think that solution 1 is a good idea.

For solution 2, I think it would be probably of general use to have Websky producs rotated to Equatorial, I do not think this should be tied to a Simons Observatory data release. We can probably create that product and ask the Websky team to make it available on their website.

[msyriac]

Good point about compatibility across experiments! I agree then that solution 2 is the way to go forward.

[zonca]

@msyriac what products among the ones available here should we make available?

https://mocks.cita.utoronto.ca/data/websky/v0.0/

We can ask to the Websky team or the Panexp GS group if anyone volunteers to work on this.