♦ Experimental Design

**1.** Cluster Selection and Sample Size

**2.** Observing Strategy

**3.** Scheduling and Technical Details

# Experimental Design

## Cluster selection and sample size

To measure the populations and dynamics of cluster members as a function of *cluster-centric radius* and *stellar mass*, the sample chosen was divided into three bins of halo mass:

♦ Massive groups (M < 10^{14} M_{⊙})

♦ Virgo-progenitors (10^{14} < M/M_{⊙} < 5×10^{14})

♦ Coma-progenitors (M > 5×10^{14} M_{⊙})

Achieving the desired precision for such a large range of halo masses requires a total sample of ∼ 1000 spectroscopic members, with at least ∼ 50 members per system in the more massive (Virgo and Coma) sub-sample bins. Thus, it is not only natural but also eﬃcient to build the** GOGREEN** strategy on existing investments. The sample consequently includes ﬁve *GCLASS* clusters at z > 1, and additional clusters with deep IRAC data available from the *SpARCS* and *STP* collaborations.

## Observing Strategy

Extending the *GCLASS* and *GEEC2* surveys to lower stellar masses and higher redshifts means taking advantage of the increased red sensitivity of *GMOS*. Completing the survey with the EEV detectors already installed on Gemini would take more than ﬁve times longer – over 2300 hours. With the new upgrade, **GOGREEN** gains a factor of 5 in depth; the ﬁeld of view covers 2.6×2.6-Mpc and thus includes the full cluster out to the expected *virial radius*.

Galaxies were selected based on their 3.6-µm ﬂux (from our deep IRAC imaging), and their z−band ﬂux (from deep imaging). By optimizing the use of nod-and-shuffle technology and the amount of masks assigned to each system, it is possible to obtain spectroscopy in even the densest of cluster cores. The faintest of galaxies were assigned to every mask over the course of several semesters in order to obtain up to 15 hours of total exposure time. Massive clusters which have little or no existing data will take 5 masks of 3 hours each, to maximize the number of brighter targets. Oppositely, clusters with existing spectroscopy will take three masks of 5 hour exposures, as there are fewer bright targets available. The masks are spread over three semesters, to make adjustments between observations (e.g., replacing faint targets that have reached the desired **S/N ***– signal to noise ratio –* prematurely).

In total,** GOGREEN** will add 700 conﬁrmed cluster members to existing data, bringing the total 1 < z < 1.5 sample to over 1000 members.

## Scheduling and Technical Details

As with *GCLASS*, the* R150 grating* is to be used, with a dispersion of 1.9-Å/pixel. Both Gemini South and Gemini North have red-blocking filters that limit the spectrum to ∼ 2250 pixels. This enables for at least two tiers of spectra to be stacked per mask while still assuring detection of [OII] and Hδ absorption lines for all galaxies at 0.65 < z < 1.5. In order to reliably measure a redshift from features near the 4000-Å break, spectra must have a predetermined acceptable *S/N* at the 4200-Å rest frame.

As described above, observations will be split to multiple masks; the number of masks and integration time on each will vary, but the total amount of integration on each cluster or group is 15 hours. Given overheads due to nod-and-shuffle spectroscopy, each system will require 19.5 hours of total time.

Therefore, the total time estimated for completion of the survey is:

♦ 39×5 hours (masks) and 40×3 hours (masks plus overheads) = 409.5 hours for the spectroscopy

♦ Added another 28.8h for the imaging, for a total of 438.35 hours

♦ The survey is consequently divided into a 6-semester plan (The first observing run scheduled on Gemini’s 2014B term)

*[Note: This is a factor 5 faster than would be possible with the current EEV detector chips]*