A Brief Dust Study for 1997

Eric Deutsch 97-10-14

Since the situation regarding the optimum closure conditions based on the dust counter currently seems confused, I have conducted a quick study of the dust situation for 1997 (Jan 1 - Oct 8). I have assembled a database of the logged weather information at APO for 1997 and cleaned it fairly well, i.e. missing and bogus information (and there's a lot of it!) has been properly masked out. Here are two plots and a few comments.

Fig. 1. --- Selected 1.0 micron dust counts for 1997. Data points are selected based on the criteria below. The Y-axis units are raw data numbers as received from the machine (particles per 0.1 cubic foot). The X-axis merely the index of each data point, and therefore indicates the total number of data points available is about 9000. The current closure condition, 1600 DN, is indicated with red horizontal line.

From all 1.0 micron dust measurements during the year, data points which meet the following criteria are discarded:

This, therefore, approximates dust counts for times when the dome would normally be open. Note that if the dew point or wind speed are not available at the time, the dust measurements are kept, rather than rejected.

Fig. 2. --- Percent telescope time lost and dust accumulation avoided as a function of closure condition. The two dashed curves represent percent telescope time lost, while the two solid curves represent percent dust accumulation avoided. the abscissa is the closure condition in raw data numbers, where the current closure condition of 1600 is marked with red vertical line. The upper set of (purple) solid and dashed lines indicate closure condition in absolute counter DN. The lower set of (blue) solid and dashed lines indicate a closure condition of DN relative to a 7 night median of dust counts. This means that during dusty times of the year (or weeks, actually), the closure condition is higher.

This plot seems to imply that with the current closure condition, we lost 7% of telescope time that might otherwise have been used, and in doing so avoided 23% of dust accumulation that we might otherwise have suffered. Note that I assume that dust accumulation rate is proportional to aerial dust density, which may or may not be true. Also, note that the closure condition of 1600 DN may not have been in effect for the whole year, and that there are probably a few hundred points which might have been rejected, but weren't because, e.g., the wind speed was not being logged (the anemometer was broken for many months this year!). Finally, it is important to realize that this is probably a lower bound as the dome is probably not instantly opened when dust, wind, humidity fall below the closure threshold, but rather, a significant portion of an hour of good conditions is probably required before the dome is opened.

Fig. 3. --- Ratio of Percent dust accumulation avoided to Percent telescope time lost as a function of closure condition. If there were a peak here, that would be wonderful and would probably yield a closure condition, but there isn't.

So what's the bottom line? I have not thought about what this all really means extensively, but I think the conclusion is that there is no easily-definable closure condition which just makes obvious sense. Rather, the percent telescope time lost and percent dust accumulation avoided increases exponentially as the closure threshold is lowered. Somehow, we must simply decide what percentage of telescope time lost is acceptable or what fraction of avoided dust is desirable and choose a corresponding level. Finally, it looks to me that there seems to be little difference in the effectiveness of absolute or relative closure conditions. Since it appears we have little idea what kind of contamination we're suffering, setting the closure level to 3500 DN absolute might be a good temporary solution: we would have lost only 0.6% of telescope time, yet avoided over 5% of contamination. It rejects only the worst dust spikes.

For purposes of comparison with some previous studies of dust, I provide the identically-derived plots as above for June and July of 1997, which were quite dusty months. While the 1600 level achieved 23% dust avoidance in 7% lost telescope time for the entire year (had it been applied the entire time), for these two months, the corresponding values are 35% and 16%.
Eric Deutsch: deutsch@astro.washington.edu
Back to My Main APO Weather Info Page
Back to My Main IR Sky Camera Page
Back to My APO Info Page
Back to My Home Page
Back to UW Astronomy Home Page