Updates to the baseline#
Over time, the baseline survey strategy evolves due to changes in the observatory, update in our weather predictions, details of the Feature Based Scheduler code, and most importantly, due to changes in the survey strategy.
All of the updates below are (so far) pre-commissioning, pre-operations.
(Note: WFD = Wide Fast Deep; DDF = Deep Drilling Field).
v4.0#
This represents an official update to the baseline survey strategy. These simulations should be very similar to v3.6, with a few minor bugfixes related to handling the additional downtime during year that would increase the overall number of observations during year one very slightly.
v3.6#
These simulations are updated versions of the v3.5 simulations, but standardized with additional downtime in year 1. This additional downtime is the survey scheduling team’s attempt to represent expected observatory efficiency throughout year one. The actual downtime is likely to vary significantly depending on results during commissioning. In addition, ToOs are included in every simulation, which lowers the overall efficiency of observations slightly.
The DDF dither pattern was modified in v3.6 to reduce the dither to 0.2 degrees instead of 0.7 degrees. This significantly improves the cadence within the DDFs.
The near-sun microsurvey was modified to skip attempted observations when the moon is close to the desired portion of sky. This removed the instances of the near-sun twilight survey observing a single field more than ten or twenty times in during a single twilight.
v3.5#
The v3.5 simulations represent early versions of the phase 3 recommendations. The changes include:
uniform rolling (pauses in rolling cadence for year 4 and year 7 data releases)
exposure time changes in u band (38s) and other bandpasses reduced to 29 seconds
readout time changed to 2.4 seconds
bluer filter balance in LMC/SMC
updated footprint in galactic plane, including directing visits toward the Roman deep field
DDF scheduling updated, leading to more observations per DDF in year one
Notable in the changes above are the slight reduction in amount of time spent in rolling cadence – uniform rolling includes 3 cycles (6 years) of rolling cadence while v3.1 to v3.4 included 4 cycles (8 years) of rolling cadence. One of the simulations included in the v3.5 release does illustrate 4 years of rolling cadence.
Another notable difference is the shift in exposure times per bandpass. This is equivalent to shifting the survey back towards the u-band depths per visit that were present before the mirror coating and throughputs update in v3.3.
There was a minor bugfix introduced, to reduce the number of non-paired visits per night. There was also a bugfix added to ensure that morning twilight near-sun microsurvey visits include r band when appropriate.
Some simulations in v3.5 include ToOs, as indicated in their filename. Some simulations in v3.5 include additional downtime in year one - this additional downtime was adopted as standard at v3.6.
v3.4#
Updates from v3.3 are minor, primarily in rubin_scheduler code updates.
There were improvements to generalize masking of unavailable areas on the sky,
as well as preventing scripted surveys such as the DDFs from executing too close
to twilight. Cross-platform repeatability was improved.
Metrics run on baseline_v3.4_10yrs.db should present very similar
results to baseline_v3.3_10yrs.db.
v3.3#
Updates from v3.2 are significant, as the new 3xAg (‘triple silver’) throughput curves were introduced in v3.3. As of v3.3, the throughput curves in use are v1.9 of the syseng_throughputs curves, while v3.2 used v1.7. While the actual survey strategy itself does not change, metrics evaluated on v3.3 which depend on five sigma limiting magnitudes will produce noticeably different results than earlier simulations. Throughput improved in grizy bands, although decreased in u band.
There were minor changes in the weighting of various basis functions in the scheduler code. One of these also produced a noticeable change in u-band visits over time, as a bug which caused u-band visits to briefly “pause” after the end of year 1 was fixed. The result of this bug fix was slightly more visits overall in u-band (about 15%) which helped compensate for the drop in u-band sensitivity.
v3.2#
Updates from v3.0 are significant. The survey start date was updated to May 1, 2025, in accordance with updated estimates of the survey timeline.
As the camera can only hold 5 filters at a time, the scheduler plans to swap the u filter with either z or y band depending on the lunar phase. Simulations prior to v3.2 swapped u-band with z-band at ~ +/-7 nights around new moon; as of v3.2, the u-band filter is swapped with y-band instead. There was also a small update to the low-dust extinction WFD survey footprint to include a small ‘swathe’ to match the southern most edge of the Euclid footprint.
The near-sun twilight microsurvey underwent several changes, pushing observations closer to the sun. This improved metrics related to discovery of interior-to-earth asteroids. The use of a distribution of filters in the microsurvey was also updated to be more uniform.
The ‘triplets’ (long blobs) survey mode was also modified; v3.0 triggered this survey mode every 6 nights, multiple times within those nights. As of v3.2, the ‘triplet’ mode is only triggered once at the start of a night, but more frequently (every 3 nights). This results in the triplet observations being acquired on more independent nights, although at more limited times within each night.
There were additional changes in the underlying codebase, which should not significantly change evaluations of the simulated survey.
v3.1#
This was an un-released set of simulations, for internal use only.
v3.0#
Updates from the v2 series are significant. The v3 series of simulations responds to recommendations from the SCOC in PSTN-055.
Major survey strategy changes can be summarized as follows:
Visits in u-band moved from 2x15s snaps to a single 1x30s exposure. Visits in other bandpasses remain at 2x15s. This improves u-band depth per visit by shifting these visits toward the sky-noise dominated rather than readnoise-dominated regime.
The survey footprint was updated by a small addition to the WFD at the Virgo Cluster. The Galactic Plane coverage (in particular, the area covered at WFD levels) was significantly updated, to add coverage at a wider range of galactic longitudes and for some stellar clusters. The filter balance of the footprint in the Galactic Plane was modified to spend more time in bluer filters, but unchanged in other areas.
Time spent in Deep Drilling Fields increased from slightly less than 5% of the overall survey time, to over 6.5% of the survey time. The COSMOS DDF now receives additional coverage, in order to reach the expected 10 year DDF depth within the first 3 years of the survey, serving as a pathfinder for later processing requirements.
A ‘triplet’ survey mode was introduced, such that every 6 nights, pointings observed in pairs early in the night will acquire a later (2-7 hours later) third visit (in one of the same filters as the pair). While a small fraction of total observations, this provides opportunities for short-timescale time-domain science.
A near-sun twilight microsurvey was introduced, taking observations within a band approximately +/-20 degrees of the ecliptic, during -12 to -15 degree twilight. These visits are only 15 seconds long but repeat 4 times within the short (about 20 minute) period of twilight, in order to enable discovery of interior-to-earth asteroids. The visits are at low solar elongation, high airmass towards the direction of the sun.
The overall effect of these changes is to fulfill additional science goals that were not met previously, however an additional effect is to reduce the number of visits per pointing in the WFD portion of the survey footprint.
v2.0#
Updates from the v1 series are significant. The v2 series of simulations responds to recommendations from the SCOC in PSTN-053.
Major survey strategy changes can be summarized as follows:
The survey footprint is significantly updated, placing more area into the WFD portion of the survey footprint. The low-dust-extinction area is increased by approximately 15%, allowing more useful area for extragalactic science. To maintain coverage of important galactic plane areas, the survey footprint now includes ‘dusty plane’ areas, observed with a few hundred visits per pointing. Additional coverage of the Galactic Bulge, at WFD-level (~800 visits per pointing) has been introduced to enable time-domain science in this area. The overall survey footprint is increased, which results in fewer visits per pointing.
To maintain and improve on the cadence of visits within the WFD, a
rolling cadenceis introduced. Therolling cadencesplits the WFD survey footprint into 4 different bands, based on declination, then alternately “activates” 2 of these bands in successive seasons. When a band is “active” it receives more visits, then in the next season when it is “inactivate”, it receives fewer visits. A pointing in the WFD would receive approximately 82 visits (825 visits over the survey, divided by 10 years) in a standard season – in an “active” season it might receive 145, while in an “inactive” season, it would only receive about 20, depending on how ‘strong’ the rolling in the rolling cadence is.
The overall effect of these changes is to improve science metrics, but reduce the number of visits per pointing in the WFD.