Sample	Redshift Range	UBV Shift	Instrumentation	Science Goal
A	0.35 - 0.65	VRI	ground-based optical (and NIR for gray dust), plus HST for precision optical photometry	spectral feature tests for evolution & long color baseline test for gray dust
B	0.65 - 0.90	RIZ	HST F675W,F814W & F850LP	fill-in gap in Hubble diagram
C	0.90 - 1.25	IZJ	HST F814W & F850LP; ground NIR	test for evolution/dust via deviations for a Hubble law

MOSAIC on the CTIO 4m (0.38 sq. deg.) and the 12kx8k camera on CFHT (0.33 sq. deg.) will be used for searching. The searchs have to be scheduled to mesh with spectroscopic screening, which will take place at Keck and possibly a limited amount at the VLT. Of the SNe which are confirmed as Type Ia, 3 of these in the redshift range 0.8 < z < 1.0 and 2 of these in the redshift range z ~ 1.2 will be sent to HST for photometric follow-up. For the z ~ 1.2 SNe, we want to obtain restframe V-band photometry at peak, which requires NIR observations - these are to be obtained using ISSAC at VLT, NIRI at Gemini, and possibly NIRC at Keck (and Subaru?). For several (TBD) of the 0.5 < z < 0.8 SNe which we should also find, we want to obtain photometric follow-up at VLT, NTT, and WIYN. In order to address evolution/dust questions, we would like to obtain a spectral time series and NIR photometry (this plan requires further definition).

In order to deliver the required number and redshift distribution of SNe the searches need to cover about 1 sq. deg. very deeply so that about 5 SNe Ia with 1 < z < 1.2 are expected and about 6 SNe Ia with 0.8 < z < 1 are expected. These numbers need to be refined, but note that we have to allow for uncertainties in the rates as well as statistical (Poisson) fluctuations. The searches should also cover enough area at a shallower depth so we get several SNe Ia with 0.5 < z < 0.8. The timing of the search and follow-up indicates that the searches will have to reach I=25.2 at the threshold S/N so that a SN Ia at -10 days can be detected. Based on the spurious nature of many of the S/N < 5 sources in our last several z > 1 searches, Greg advocates that threshold S/N > 7 be required (meaning S/N > 9.9 in the refs and the news). The last two effects imply that much deeper exposures than in the past be obtained in the deep fields. Indeed, at these signal levels, a subtraction errors as small as 0.4% of sky can generate an artifical source as bright as the SN!

No single field is visible all night. For safety all the HST SNe should come from one field (up to 1 degree in radius). z < 0.8 (non-HST) SNe are being followed from the ground. Ground follow-up requires visibility for up to two months after discovery, so there should be a z < 0.8 field visible in the last hours of the night during the reference runs. The HST field must be visible during the all but the last hours of the night between the time of the reference and the last NIR imaging. Note that under this arrangement, the HST field fill be searched in I-band, while the z < 0.8 field can/should be searched in R-band. Thus, up to 1st quarter moon can be tolerated for the references and searches.

There was some discussion regarding the optimal gap between the reference and the search. This needs to be simulated so not too many stale SNe are found and so that spectroscopy is taken just before peak so that HST photometry does not miss peak. Roughly speaking, we need at least 3 days from the end of the search to the start of Keck spectroscopy and HST observations can't begin until 5 days after the end of Keck spectroscopy. If we get 6 nights at Keck and all the time is optimally scheduled, this results in 14 days from the end of the search to the first good HST observations. We could obtain short screening exposures within 8 days of the end of the search. All proposals should stress that we need to be consulted before time is scheduled. We need to be in a position to reassess the optimal schedule once we know what allocations we are getting on the many telescopes for which we are proposing.

Due to some creative math, we have 10 HST orbits which are not absolutely needed for the 5 designated HST SNe Ia. These could be put towards screening candidates, longer exposures to mitigate against CTE, or to obtain final reference images for HST SNe in need. These options do not directly impact the upcoming proposals, but should be borne in mind. (Note, we can specify that the SNe be placed closer to the readout amplifier (at pixel 200,200) to mitigate CTE effects.)