design-observational-campaign

Design Observational Campaign

Plan an observing campaign by scheduling targets within sky and telescope constraints, calculating required exposure times for target signal-to-noise, designing calibration frames, and building a contingency plan for variable conditions.

Why This Is Best Practice

Adopted by: ESO (European Southern Observatory), NOAO, STScI (Hubble Space Telescope), and all major observatories use formal Phase 1 (proposal) and Phase 2 (scheduling) observation planning frameworks. ALMA requires detailed observing parameter justification including expected sensitivity, angular resolution, and atmospheric constraint specification. Gemini, Keck, and VLT all enforce systematic observation design as a prerequisite for time allocation. Impact: Howell (2006) demonstrates that exposure time calculation is the single most important pre-observation step — an under-exposed observation is useless (insufficient S/N to detect the target), while over-exposure wastes precious telescope time and saturates the detector. Proper scheduling to minimize airmass maximizes data quality; Kitchin (2013) shows that observing at airmass 2.0 (30° elevation) rather than 1.0 (zenith) increases sky background by a factor of 2 and attenuates signal by a further factor of ~2-3 due to increased atmospheric extinction.

Steps

1. Define the science case and observational requirements

Before any scheduling:

• What are you measuring? Photometric brightness, spectrum, polarization, astrometry, timing
• What signal-to-noise ratio (S/N) is required? Photometry: S/N ≥ 50-100 for 1-2% precision; spectroscopy: S/N ≥ 20 per resolution element for line detection
• What resolution is required? Angular (arcsec), spectral (R = λ/Δλ), temporal (cadence in s/min/days)
• What are the critical constraints? Moon phase (dark/grey/bright time), seeing requirement (FWHM ≤ 0.7"?), time sensitivity (transient event, orbital phase coverage)

2. Calculate required exposure time (signal-to-noise budget)

For CCD photometry, the signal-to-noise ratio is:

S/N = N_star / √(N_star + n_pix × (N_sky + N_dark + N_read²/G²))

Where:

• N_star = star photons detected per second × exposure time
• N_sky = sky background photons per pixel per second × exposure time
• N_dark = dark current (e⁻/pixel/sec) × exposure time
• N_read = readout noise (e⁻ RMS per pixel)
• n_pix = number of pixels in the aperture
• G = gain (e⁻/ADU)

Simplified for bright-sky-background-limited case:

t_exp ≈ (S/N)² × N_sky / N_star²

Estimate N_star using the instrument's quantum efficiency, filter transmission, and magnitude of the target (convert to flux using Vega or AB zero-points).

Use ESO Exposure Time Calculator (ETC), SIGNAL (CFHT), or ITC tools for your telescope/instrument — always use the official ETC before proposing.

3. Compile the target list and prioritize

For each target, record:

• Coordinates (RA, Dec, J2000)
• Magnitude in relevant band(s)
• Size (for extended objects — impacts sky subtraction annulus)
• Time-critical constraints (transit timing, orbital phase, outburst activity)
• Minimum required sky conditions (seeing, transparency)

Prioritize targets:

1. Time-critical (transient, eclipse, conjunction) — schedule first
2. Sky-condition-sensitive (faint, small angular size) — schedule in best seeing windows
3. Bright calibrators — can observe in any conditions

4. Schedule to minimize airmass and sky brightness

Airmass X = sec(z) where z is the zenith angle:

• X = 1.0: at zenith (best; minimum extinction and seeing)
• X = 1.41: 45° altitude
• X = 2.0: 30° altitude (acceptable; atmospheric extinction ≈ 0.6 mag worse than zenith)
• X > 2.5: generally not scientifically useful (high extinction, poor seeing)

Rule: observe each target within 30° of the meridian (minimum airmass point), but not closer than 1h before or after transit for safety.

Moon avoidance:

• Dark time: moon below horizon or illumination < 30%
• Grey time: moon illumination 30-70%
• Bright time: moon illumination > 70%
• Angular separation from moon: ≥ 20° absolute minimum; ≥ 45° for faint blue objects (Rayleigh scattering from moon)

Use Stellarium, pyephem, astropy, or online visibility tools (e.g., ESO Visibility Tool) to schedule.

5. Plan calibration frames

For every science observation, acquire:

• Bias frames: minimum 10-20; captures readout pattern and electronic offset; take at beginning and end of night
• Dark frames: match exposure time to science exposures; required if dark current is non-negligible (>10 e⁻/pixel/hr)
• Flat fields: minimum 10 per filter used; sky flats (at dusk/dawn) or dome flats; flat-field correction removes pixel-to-pixel QE variation and vignetting
• Standard star observations: photometric standards from Landolt or Smith (SDSS) catalog; observe at 2-3 different airmasses to solve for extinction coefficient; acquire within 30 min of science target observations

6. Build contingency and backup plans

Observing conditions are variable:

• Backup targets: prepare a list of brighter/larger targets observable in poor seeing or thin clouds; sort by target brightness and sky area
• Abort criteria: define minimum acceptable conditions before starting (e.g., seeing > 1.5" → abort spectroscopy, switch to imaging; thin clouds → abort photometric standard observations)
• Recovery plan: if interrupted mid-sequence, document at what orbital phase / cadence point you stopped; partial sequences are still scientifically useful for most programs

Common Mistakes

• Not accounting for overheads (readout, slew, telescope setup): Only ~60-75% of allocated time is "on-sky shutter-open" time for typical programs. Plan realistic exposure counts accounting for 3-5 min overheads per target.
• Scheduling targets at high airmass because they fit the time slot: A target at airmass 3.0 may produce data too poor to analyze. Better to skip and revisit the next night than take useless data.
• Observing photometric standards only at one airmass: Solving for extinction requires standards at two or more airmass values. A single airmass observation cannot determine the extinction coefficient.

When NOT to Use

• Space-based telescopes with fixed scheduling windows: HST, Chandra, JWST proposals follow different (proposal-driven) scheduling frameworks managed by STScI/CXC with dedicated scheduling software.