Clear Sky Detection experiment

Date: 2010-02-25, updated 2015-03-27

We are currently experimenting with a technique to determine a "clear sky" using the SQM-LE/LU.

A description by Dr. Doug Welch is here.

Basically, we have noticed that the light fluctuates more when clouds are overhead.

Here is an annotated example where the Moon was visible during the night and set at 4:08am and hence the reason for the non-zero slope:

Here is a live experimental plot from the Unihedron office:

Using the SQM-LE/LU like this will most easily detect low clouds when either the moon is up or there is some light pollution present.

The formula we are using for the above plot is fairly simple. We use an array of the last 10 readings which have been taken one minute apart. A sum of the absolute differences between each reading is a pretty good indicator of a clear sky (0=clear, above 1 = cloudy).

Send comments

Thomas Posch and Johannes Puschnig from the Institut für Astronomie Universität Wien have created a website which includes active CSD calculations where plots from previous nights can be examined:
Monitoring Light pollution
Plotted results (in German)