Guidelines for using the LASAIR Particle Monitor Telemetry at LBTO
The LASAIR III (310C) particle monitor is mounted in the Aux Control Room on Level 5 with an intake through the external wall to sample particles at that level (height of the observing floor). The unit is presently collecting samples of 1 minute for 1 cubic foot of air. It samples particles with sizes of 0.3 micron, 0.5 micron, 1 micron, 5 micron, 10 micron and 25 micron. All counts are cumulative for that size and larger.
These guidelines for using the particle monitor telemetry as an aid to determine when to open/close the enclosure are under development since 20200617. The good and bad thing about the second half of June 2020 was that the Bighorn Fire in the Santa Catalina Mountains north of Tucson provided ample opportunity to cross-compare the particle monitor counts against more traditional particle sampling methods (smoke smell, particles seen in a flashlight beam, observed ash accumulation on horizontal surfaces, visual opacity). The great benefit of the particle monitor is that it allows you to follow trends in the particle counts with going outside on the patio every five minutes. We emphasize that we are still learning about the various kinds of particle weather, so these guidelines are a tool to be combined with your other senses/sensors when determining whether or not to open.
The levels labelled "baseline" on the plots were determined from a windy night in June (20200617). We were NOT in the smoke plume from the Bighorn fire on in the Santa Catalina mountains that night, but being fire season in June in Arizona we had some particles from very distant fires that are generally in the air. Data from "clean" nights in spring 2019 show particle levels that can be 5x below "baseline".
In addition to having the mountain assistant OSA (during restricted C-19 observing), We have been using the 0.5 micron particle counts (navy blue, second line from the top) as a quantitative detection of smoke. We assume that the 0.3-0.5 micron particles are what we smell as smoke, and that is what we see as white smoke on daytime photographs.
The following thresholds have been determined based on when we were open this week and last week (June 17-30), and from the in-person reports from Tim and Joe (averaging over several OSAs and several noses). Smoke in the form of sub-micron-size particles is detected at very high signal-to-noise levels by the particle monitor, and can even saturate the count levels at 3.6 million in very strong smoke plumes. The main issue is how to determine the proper threshold, and how to normalize the human response of smelling smoke. Thanks to Josh, David, Steve, Guido, Robert and all the mountain assistant OSAs for their assistance in developing these guidelines.
The present threshold for opening is: IF "0.5 micron < 100k" THEN "OK to open"
The present limit for closing is: IF "0.5 micron > 200k" THEN "should close"
For 0.5 micron particles (with allowances for individual noses)
- 10k is the baseline for a windy night in June
- 20k can't really smell any smoke outside,
- 50k is a noticeable smoke smell,
- 100k is a distinct smoke smell,
- 1000k is a really strong smoke smell.
- 3600k saturates the particle counts
Smoke from very distant fires
(from JMH email on 20200824)
As you can tell by looking outside (either in Tucson or on the mountain) the air is full of smoky haze from distant forest fires. The LASAIR particle monitor on Level 5 sees this "smoke". The 0.5 micron particle count has been at or above the closing threshold (which we established in June 2020) for several days. This morning the 0.5 micron particle count is steady at 400k, while the closing threshold is 200k.
We have opened the telescope in these conditions with the following logic:
a) The sub-micron particles are not a direct threat to either the AdSec
or the Primary Mirror coatings.
b) The sub-micron particles are inside the building whether we open the shutter doors or not.
c) We had been using the sub-micron particle count as a proxy for detecting the smoke plume from the Bighorn Fire, because the small particles have much high signal-to-noise in the telemetry.
d) The 25 micron particles (what we are really worried about) have not been showing elevated counts over the last few days. Presumably they fall out of the atmosphere in the time it takes to travel here from California.
My view is that we should continue to operate the telescope in these conditions as we did last night.
Note that the fires burning just west of Globe (Salt, Gin, Griffin) have been growing in the last few days, so we have the chance of getting "local" smoke with ash particles from those fires with a West-ish wind. Thus, we should continue to pay attention to the particle monitor and especially the larger particle counts.
Ash and Dust
Of course, the larger 25 microns particles of ash are what we really care about for the AdSec
(particles close to the size of the gap behind the thin shell), but small number statistics (only a few particles per sample) makes them harder to evaluate. The plots include a running average line as an attempt to make the trend more clear. We assume that the 10 and 25 micron particles are what we see as dark smoke on daytime photographs. I believe a
"good" number for 25 micron particles is less than 1 per cubic foot per minute sample, but we've been open with up to 5 particles per sample. I (JMH) still have doubts about the particular value for the large particle threshold, but we need a separate threshold for large particles in the absence of smoke (e.g. the morning of 20200628).
The present threshold for opening is: IF "25 micron < 5" THEN "OK to open"
The present limit for closing is: IF "25 micron > 10" THEN "should close"
A limitation to the particle monitor is that it doesn't sample enough volume of air to detect really large particles of ash (0.5 mm) that we sometimes see falling. You still need to remain aware of those larger particles in person (or via radar).
We are presently unable to distinguish between mineral dust and ash using the particle monitor. We will be studying that more as more data is accumulated (i.e. is the particle size distribution function different? Is dust from White Sands or the Sahara different than local Pinaleno forest dust?)
We don't yet have much information on detecting pine/spruce/fir pollen with the particle monitor. Pollen tends to be large particles.
Based on limited experience, fog seems to suppress the small (sub-micron) particle counts, and to make those counts jittery. The particle monitor likely isn't the best tool for detecting fog, but the small particle counts do seem to show the effects of fog drifting through (20190927). It appears that the ratio of 1 micron particles and 25 micron particles approaches 1 when fog is present.
Snow behaves similar to fog in the sense that the smallest particle counts go down and the largest particle counts go up (20210218).
Particle plots from the 'pms' telemetry stream are included in LBTplot (python2) on obs2-obsN (CentOS
), and are included in LBTplot (python3) on obs1/robs (Fedora).
Particle plots in real-time can also be created on the mountain DMS web page. http://dms.mountain.lbto.org
Particle plots (including historical data from Spring 2019) are now included in the Nightly Summary plots at http://abell.as.arizona.edu/~hill/xplot/
The latest nighttime particle plot is at: http://abell.as.arizona.edu/~hill/xplot/latest_pms.cgi
20190616 - baseline levels
20200627 - smoke plume
20200620 - smoke saturation
20190927 - fog
20210218 - snow
20190421 - clean
Additional details about the machinery of collecting particle monitor telemetry at LBTO can be found at ParticleMonitoringSoftware
Telemetry (converted to CSV format) looks like this:
time_stamp, tai_offset,locations->selected_loc, laser_ok,
+sample_interval,volunits->selected_units, sample_volume, pt_3um_counts,
+pt_5um_count, 1um_count, 5um_count, 10um_count, 25um_count
The overhead weather display now includes the particle monitor plot along with warning states. The particle monitor plot rotates with the NWS weather plot. The warning states are as follows:
- Yellow if 3+ samples in last 10 samples reading 5+ counts (warning state)
- Red if 3+ samples in last 10 samples reading 10+ counts (error state)
- "Dust pollution monitoring at the TNG telescope", Porceddu, Zitelli, Buffa, Ghedina, SPIE 4844, 2002, https://www.spiedigitallibrary.org/conference-proceedings-of-spie/4844/0000/Dust-pollution-monitoring-at-the-TNG-telescope/10.1117/12.456725.full
- "Looking for Correlations between Dust Events and Weather at Observatories in New Mexico", C. Villanova and M. Creech-Eakman, PASP 119, 2007, https://www.jstor.org/stable/10.1086/522209
- "Astronomical site survey report on dust measurement, wind profile, optical turbulence, and their correlation with seeing over IAO-Hanle", https://link.springer.com/article/10.1007/s10686-017-9525-6
- "Effects of Airborne Particles on Optical Equipment", I. Gorgievska and C. M. Steinebach, SPIE 5491, 2004,
- 13 Jul 2020
- 30 Jan 2021