this will require running the DM pipeline and will take a lot of effort and time hence the short cut suggested in #16 is being investigated as an alternative
Due to the likely time and effort costs, the recipient group, i.e. the PZ WG, are not as keen on pursuing this work (i.e. applying simulation tools to generate Y1-like observational properties)
feedback received from Markus Rau (WG convener) who says Y1 systematics might be quite significant and so might want to consider including in the forward model
@hangqianjun has therefore sent Markus (and @joachimi) a more detailed update but as yet has had no reply. Their replies will determine the next steps for this. These next steps could be.
The Y1 data used in the test is a sub-sample of full Y1 data but this appears to be too small for the precision of the above test. A small dataset was chosen due to the lengthy run-times for larger data sets. So a possible next step is to consider whether it is feasible to run a test containing more Y1 data.
Re. forward modelling: the mitigation at photoZ level might not be feasible so perhaps instead could include systematics as a package in the real degrader .
@hangqianjun has sent a message to Markus and @joachimi to see if they agree if these are a way forward.
If no response soon from Markus and @joachimi will likely proceed with these plans anyway
@joachimi replied to say the 2nd option will be followed that is will provide a package in the RAIL degrader to incorporate the effects of the observation conditions into the photoZ.
26/MAY/22