Addressing Weather Loss
at Gemini South
Weather conditions at Gemini South
repeat themselves fairly regularly
across each semester. In the past we
didn’t make any allowance for that.
Recently, we’ve made a change to
the way we fill the queue for Gemini
South. Figure 3 shows a repetitive
five-year pattern of weather losses at
Gemini South; because we filled the
queue each month as if the weather
loss was uniform (and it wasn’t), we
forced ourselves to battle the elements at
the worst times of the year. In the 2016A
TAC process, we adjusted the way we fill
the queue: we no longer overload May-September, and allow more programs into the
southern summer months. We’ll see how it
goes now that we’re into the semester itself.
2015B: Major Challenges at
Gemini South
As mentioned, Gemini South has had some
challenging semesters of late. Semester
2015B, for instance, had plenty of adversity
to go around: the Gemini Multi-conjugate
adaptive optics System (GeMS) out of action
April 2016
due to a major earthquake that struck in September 2015; the Gemini Planet Imager (GPI)
still ramping up; and many programs either
lost completely to weather, or executed under marginal conditions. Based on the discussion above, we would therefore expect
a significant hit on programs in all Bands,
with Band 3 (able to take the worst conditions, and therefore not containing any GPI
or GeMS programs) performing reasonably
well. That’s borne out by the results shown
in Figure 4: a significant number of GPI programs were not attempted at all, hardly any
GeMS programs started, and many programs ended up in the “tail” of completions
below 100%.
GeminiFocus
Figure 3.
Percentage of time lost
to weather at Gemini
South. This pattern
is quite reproducible
from year to year.
We now take better
account of it in the time
allocation process; we
no longer overload the
mid-year months (MaySeptember), and allow
more programs into the
southern summer time.
Figure 4.
Program completion in
2015B at Gemini South.
Note the significant
numbers of
unstarted
programs
(driven by
instrument
unavailability),
and the excess
of Band 3
programs in the
100%-complete
bin (a signature
of bad weather).
17