AAG Cloudwatcher Settings

The weather has been good in Massachusetts the last few days, but when I checked this morning PAN006 had not recorded any data for 4 days. It appears that the cloud sensor is reporting too low a difference between the sky and ambient temperatures (e.g. about -4.1C aloft and 16.4C at the sensor last night at 2am).

We don’t make use of the AAG software (it runs on Windows), and instead have software for estimating from those temperatures whether it is cloudy, using a very simple model related to the difference between those two temperatures: > 25C is clear, 25C to 15C is cloudy, and less than 15C is very cloudy. In the example above, at 2am the difference was 20.5C, but the skies were clear, as you can see here:

Wilfred tells me that he adjusted the thresholds by modifying POCS/conf_files/peas_local.yaml so that >20C is clear. That would have worked last night, but would be very close to failing.

It seems that we need a better model than a simple linear model. The AAG software uses a linear model by default, but they include the option of providing settings for a non-linear model.

If anyone has pointers to background on this problem, I’d appreciate learning more about how to solve this.

I just wanted to bump this by saying that here in Sydney the AAG we have does a terrible job with the linear model, especially once it’s over 30°C.

I had tried (about 8 months ago) to use the AAG GUI to change their settings or understand what was going on and it was generally none helpful.

I’m hoping to get back to hardware work on PAN008 soon and once that is done I’ll need to worry about the AAG there again. Probably not for a month or two unfortunately.

I proposed the project of developing a better model to Prof. Maitra at Wheaton, who was supportive of this being a formal project for a student this fall. There are 6 months or more of data collected by PAN006 (i.e. sky and ambient temps plus humidity), and Wheaton has several years of all sky camera images which could be used to compute a ground truth value.

Great! That’s probably about the time I’ll be getting back to this.

If given time this could also expand somewhat to look at image quality during different weather conditions, which might be interesting. I keep wanting to do that during the high wind (> 90 kph) observing that PAN001 does. But also if you have nightly aerosol maps, which are probably available online.

FYI, @james.synge Huntsman has an AAG at Siding Springs Observatory that we will probably need to get working before your Fall semester so make sure to check back before starting any student project. I will also try to keep updated here.

I’m also adding @joshw to the conversation as he had done the original work on the AAG for PAN001 and might have some familiarity.

Those settings are for Mauna Loa and won’t necessarily work at other sites. That’s why they are in a config file. Also, there may be a temporal drift as the sensor gets less sensitive (or gets contaminated somehow). I’ve not seen that for AAG, but I did see that over ~5+ years with a Boltwood and I think the same IR sensor is in both.

The sky transparency (at IR wavelengths) will be a factor. At high altitude, dry sites like Mauna Loa, the sky temperature should get very cold, that may not be the case at all sites, so the thresholds should be set based on experience with that specific site and perhaps even that specific AAG unit.

Sorry, I should have written all this up as part of the instructions. We should review the AAG instructions to make sure I’m remembering correctly and then summarize how to do this.

I don’t believe we have any instructions related to the AAG, so I just hooked it up and it worked… until it didn’t. I suppose if I’d deployed it in June I would have immediately been unhappy because it would never have run.

And I do think that the Wheaton College project would be worthwhile even if Huntsman does their own version of the same. As Josh says, sites with very different conditions (topography, climate, etc.) will likely also need different settings for the AAG.

Thanks for the answer @joshw. My understanding is that what is in the config file is acting on the temperatures as it receive them from the AAG. Do you have any experience with the settings that go into the AAG? For us it is not so much an issue of the difference for sky and ambient, which we can learn, as it is an issue that the readings that come out of the AAG are wildly incorrect. The AAG gui in windows has coefficients you enter to control this but it’s very unclear how those coefficients are getting stored or used on the AAG itself. I’ve found a few websites in the past but nothing that makes it clear.

Yes, there are serial commands to change those settings. I’m not sure I changed them for the MLO unit though. I don’t think I added any methods in the python code for them.