= Summary of MD04.pv2_stacktest.20130903 =
[[PageOutline]]

(back to [wiki:MediumDeepFields] summary)

This is a test sample for PV2 and deep stacks using MD04 while it is an inactive field.

PV2: update type reprocessing

Deep stacks:

 * for comparison, the last deep stack test using MD09 PV1 ([wiki:MD09.PV1_deepstack_tests])


Reference stack -- this should be useful for making the MD04 refstack for return to field early November 2013. May move section for refstack to its own page for consistency.


Other stacks:
 * in and out of season stacks -- haven't been done yet, need sample for Taiwan meeting
 * fewer N input refstack


Status:
 * prior 9/3 many issues being worked out for PV2 and MD ([wiki:MD10.refstack.20130715])
 * 9/3 PV2 updates started for MD04 after M31 reprocessing finished
 * 9/6 PV2 updates finished 
 * 10/3 test all-in deep stack skycell run
 * 10/10 started full all-in deep stack  run

----

= PV2 updates =

== MD04 ==

PV1 camera label -- data_group: mix of processing labels/data_groups, but all of PV1 type (cam_id>=434615)
|| label || data_group ||
|| MD04.pv1.20130322 || MD04.pv1.20130322 ||
|| MD04.refstack.20121125 || MD04.refstack.20121125 ||
|| MD04.nightlyscience  || MD04.<obsdate 20121129--20130525> 
|| || ||

There have been some test reprocessing that had to be avoided -- czw.MDVDtest%, czw.MDVD.MD04.%

PV2 camera (as well as chip, warp) label '''MD04.pv2.20130903''', data_group left as originally was to be able to track back.


With ipp047 data disk down, have found ~3% of exposures missing a related imfile (~half are the burn.tbl files that can be regenerated) and indicates serious need of fixing the replication of raw data. Also found some replicated but on a host (ipp064X) that no longer exists.. it is also possible some replicated products may be on the same host if push to the stsci nodes with 3 partitions..


Number totals: (not all exposures pass camera stage, FWHM>12 pixels)
{{{
| filter  | count(cam_id) | count(warp_id) |
+---------+---------------+----------------+
| g.00000 |           596 |            577 | 
| i.00000 |           719 |            713 | 
| r.00000 |           588 |            571 | 
| y.00000 |           471 |            465 | 
| z.00000 |           711 |            710 | 
}}}

PV2 warps will also be kept available (not cleaned up) for making deep stacks  but also will be a accessible via the PSS/PSPS interface more quickly (release name MD04.PV2).


== 3PI overlap ==
Unclear what want to do here still, with LAP running, and how useful it will be in the stack.


----

= Deep Stacks = 
Like with MD09, different versions of stacks will be produced for testing (not all will be full-field?)
 * all -- 
 * moderate input cuts -- 
 * best cuts >2010  and moderate input cuts -- 


== All ==
Basically all warps available with ~FWHM <13 pixels to avoid any badly measured PSF warps.

 * label, data_group = MD04.alldeeptest.20131010


== Cut ==
All warps FWHM <10 pixels and input warp cuts 

 * ZPT > ZPTmean-0.33
{{{
g: 24.11
r: 24.32
i: 24.07
z: 23.81
y: 22.79
}}}
 * elongation <1.5
 * background:
{{{
g: <400
r: <500
i: <1800
z: <2500
y: <5000
}}}

( -- still in progress --  )
 * label, data_group = MD04.cutdeeptest.20131103



== Best Cut ==
All warps FWHM <10 pixels and the same input warp cuts, only season years 2-4 (excludes 2009/2010 season)

 * label, data_group = MD04.cut2009deeptest.20131103


----

= Reference Stacks = 
In addition to the normal reference stack for a season, test samples of planned final reference stack types also being made.


{{{
| exp_name    | filter  | dateobs             | comment                         |
+-------------+---------+---------------------+---------------------------------+
| o4777g0224o | g.00000 | 2008-11-07 14:56:59 | MD04 g MJD54777 MD04 center     | 
..
| o4803g0354o | r.00000 | 2008-12-03 15:25:16 | MD04 r MJD54803 MD04 center     | 
---- pre-survey
| o5161g0222o | g.00000 | 2009-11-26 13:55:38 | MD04 g,r MJD55161 g MD04 center | 
..
| o5321g0014o | i.00000 | 2010-05-05 06:20:20 | MD04 i N5321 MD04 center        | 
---- yr1 -- 11/26/2009 -- 5/5/2010
| o5545g0436o | g.00000 | 2010-12-15 12:12:35 | MD04 g,r N5545 g MD04 center    | 
..
| o5681g0089o | i.00000 | 2011-04-30 07:34:50 | MD04 i N5681 MD04 center        | 
---- yr2 -- 12/15/2010 -- 4/30/2011
| o5892g0155o | g.00000 | 2011-11-27 14:41:57 | MD04 g,r N5892 g MD04 center    | 
..
| o6056g0114o | z.00000 | 2012-05-09 07:38:31 | MD04 z N6056 MD04 center        | 
---- yr3 -- 11/27/2011 -- 5/09/2012
| o6260g0448o | y.00000 | 2012-11-29 15:14:41 | MD04 y N6260 MD04 center        | 
..
| o6437g0061o | y.00000 | 2013-05-25 05:50:54 | MD04 y N6437 MD04 center        | 
---- yr4 -- 11/29/2012 -- 5/25 2013
---- yr5 -- actively being observed since 2013-11-08

}}}

(more details to follow)

== Basic Reference Stacks ==
Default reference stacks for nightly diffims
 * reduction MD_REF_STACK_1DG, FWHM <6 pixel and 1D convolution 
 * best ~60 input warps excluding 2009/2010 year/season
 * label, data_group = MD04.refstack.20131018, MD04.refstack.20131018_1dg (_1dg to indicate 1D convolution method, but is now default for all stacks)


== IQ Stacks ==
Basically a best reference stack possible -- used MD_DEEP_STACK_1DG so no extra cuts on FWHM made in order to look for bad warps need to be excluded in future runs
 * label, data_group = MD04.iqtest.20131118


== Reference Yearly/Seasonal Stacks ==
Basically a best reference stack possible using the specific year/season data -- used MD_DEEP_STACK_1DG so no extra cuts on FWHM made in order to look for bad warps need to be excluded in future runs

 * label, data_group = MD04.yr1.reftest.20131118
 * label, data_group = MD04.yr2.reftest.20131118 
 * label, data_group = MD04.yr3.reftest.20131118
 * label, data_group = MD04.yr4.reftest.20131118

Comparison of yr1 and yr2-4 may help show if 2009/2010 season should be excluded from other stack types due to baffle improvements.


== Reference Out-of-Season Stack == 
Two types possible, all exposures before season (default during active observing) and all exposures before/after season (or N seasons)

Not currently possible with current tools unless add warp_id manually for each skycell -- this method may even be necessary for all final stacks to exclude problematic exposure/chips in a warp skycell