Subject: weather statistics at Flagstaff, 1979-1995

From: Stupendous Man

Submitted: Wed, 14 Feb 1996 10:59:57 -0500

Message number: 49 (previous: 48, next: 50 up: Index)

  Brian Skiff of Lowell Observatory has been keeping track
of the weather at Lowell for many years.  He recently posted
a summary of the past 17 years of weather there, which might
be useful as a check/comparison for APO.

  Here's Brian Skiff's message:

------------------ Brian Skiff's analysis ------------------------------
     Just to finish off the discussion about quality of observing sites, here's
some yak-yak about cloudiness.  The following is adapted from a note I prepared
for the Saguaro Astronomy Club (Phoenix) newsletter a few years ago.  Members
of the "Friends of Lowell Observatory" will also hear some echoes from a past
newsletter.  Unlike the previous general discussions, this one deals more
specifically with Arizona weather.
     I have been noting the nighttime cloudiness in Flagstaff since June 1978,
and from 1979 I have daily data on hand.  (I regret now having not started
this business the day I arrived in 1973.)  I categorize each night as follows:

photometric - cloud-free dusk to dawn
partial - more than 3 consecutive photometric hours during the night
spectroscopic - less than about 1 magnitude of obscuration by thin clouds
cloudy - cloudy

     Since clouds absorb light at all visible wavelengths equally, it is usuallynot necessary when taking spectra of astronomical objects for the sky to
actually be cloudfree.  It merely means the observer must expose longer to get
enough signal.  Thus spectroscopists are going full blast on nights that those
of us who do photometry are out of business merely from a few cirrus clouds.
Thus it is worth noting those nights for statistical purposes, since our
telescopes are working.
     The common definition of a "clear" night at most observatories is one
with at least six consecutive cloudfree hours.  This means that on a 12-hour
winter night it can be cloudy for half the night, yet it still gets counted as
a fully "clear" night.  My "photometric" definition requires that the entire
night be clear, which makes our numbers look less impressive.  But the
"partial" nights are accounted for in the tally, so the total clear hours for
any period can be estimated accurately.
     For what it's worth, I reckon the duration of the night to be the
interval between _nautical_ twilights minus 30 minutes, which corresponds to
a solar elevation of -15 degrees at our latitude.  This is about the time when
the rate of change in the twilight sky brightness slows from its exponential
decline (or starts up in the morning).  It is also about the time when the
twilight sky toward the Sun loses its blue color (or gains it at dawn).  When
doing CCD observing, one is often taking twilight flat-field frames within a
few minutes after sunset, so if you consider this to be "data" (you gotta have
it get decent CCD results, so why not?), we're observing well before/after the
traditional time of astronomical twilight (-18 deg solar elevation).
     Some highlights from the 17-year dataset are given in the table below.
This shows the total number of "photometric" nights (completely cloudfree
dusk-to-dawn) for each year along with the total number clear hours, which
includes time from partly clear nights.  In addition, the best month(s) for
each year is noted along with the number of clear nights in that month(s).
The averages for photometric nights/hours are given at the bottom.

Year    Photometric     Best
       nights  hours    Month/# clear nights
1979     121    1060    Jun/16
1980     141    1320    Oct/20
1981     142    1356    Jun/21
1982     108    1076    May/18
1983      99     969    May/19
1984      95    1025    Feb,Jun/13
1985      83     953    Jun/14
1986      72     997    Jun/11
1987      61     910    Jun/10
1988      88    1062    May/15
1989     106    1288    Nov,Dec/14
1990      91    1223    Oct/19
1991      94    1079    May/17
1992      82    1036    Jun/15
1993     109    1156    Sep/21
1994      83    1030    Jun/12
1995      80    1036    Jun,Oct/15
         <97>  <1093>

----------------------- end of Brian Skiff's message ---------------------

  It is an interesting coincidence that, in 1994-1995, Skiff reports
that 22 percent of the nights were photometric at Lowell; and
during the period Nov 7, 1994 - Feb 13, 1996, the weather
logs from APO also show 22 percent of the nights as photometric.
We might take solace in Brian's tally: it appears that
over the long term, we can expect slightly more photometric

                                    Michael Richmond
APO APO APO APO APO  Apache Point Observatory 3.5m  APO APO APO
APO  This is message 49 in the apo35-general archive. You can find
APO  the archive on
APO  To join/leave the list, send mail to
APO  To post a message, mail it to