I will try to describe the situation I had.
I was working in the weather service and was returning 30 years of
observations to a station. It was 30 years with 24 observations each day
(I think 23 christmas eve). They got the data on a pc with a program for
extracting statistics. To test the program I chose to count the number
"total cloud cover" when it was <= 3/8.
For every hour of the day each month I got a number of satisfied
conditions. The maximum number was around 900 (30 years*30 days/month).
The reference was another program that was operational on another
platform. It gave the same result.
This was (partly) before modern ceiolometertechnology.
Meteorologists say that when dark falls you have difficulty observing
incoming cloud covers. The dawn often is a surprise regarding cloud cover.
At this site sun sets between 15 hours and 21 hours local time.
In every month there was a maximum in the number of "clear" events at
midnight. It had comparable slopes on either side of midnight.
As I recall it, it was maybe 30-50% higher than at sunset. This was 10
years ago, and I have not saved the numbers. I am not working there now
so I can not repeat it.
If you say that the observer is "tired", why is he not so x hours after
sunset and not at midnight?
There was a program on TV some months ago about the suns impact on earth
One meteorologist could prove that when the sun had many spots, the
earth became warmer. (Few sunspots in 17:th century, which here is
called "little ice age")
I do not remember exactly how it worked, but the result was that
radiation (cosmic or solar) created condensation nuclei that produce clouds.
My interpretation is that at midnight the particles that hit my site has
a minimum which explains why there are fewer clouds at that time.
It would be interesting if someone reading this has access to a similar
database and can check if it can be repeated elsewhere.
With a wish for Clear Skies
Ling,Alister [Edm] wrote:
>sounds to me like a problem either with instrumentation or with the
>observations. It is highly unlikely that a minimum at midnight exists year
>round. Are you looking at cloud base? For example if there is overcast
>cirrus, and the machine/person reporting uses a ceilometer that is not
>capable of detecting cloud above 10,000 feet, then the observation may say
>"Clear below 10,000" which is obviously neither "clear", nor reports
>A less likely possibility is selective observation, where the human observer
>is just plain tired and didn't do a proper observation on conditions.