S.D. Maley writes of an interesting and somewhat surprising NASA program trying to fill in observation gaps where satellites are getting cloud cover wrong:
NASA is in the process [1] of checking two satellites [2] against ground observer reports. Student groups seem to be the primary focus, but adults can participate via:
http://science-edu.larc.nasa.gov/SCOOL/Rover/
The Database can be searched [1], and some of the ground observer reports have comparisons with what the satellite passing over was seeing. For example, one where the satellite saw 3 levels, but ground observer only noticed high level clouds:
http://scool.larc.nasa.gov/cgi-bin/en_rover_flash_matches.cgi?obs_no=2714
I have taken a few digital photos while doing ground observations for SCOOL Rover. When the sky overhead appeared to be clear, the satellite sometimes picked up 2 or 3 levels of “clouds”. As the satellite listed Opacity as Transparent, how would a ground observer be able to detect them?
Nonetheless, the Database might be a resource for helping train ground spotters about cloud classification, … unless the CERES sensors are so good that ground spotters will only be useful to fill in the gaps between satellite overpasses, or for fast-moving weather events.
—————-
[1] Database search goes back into 2008, but only a scattering of reports have both ground and satellite, even though the satellites were launched in 1999 and 2002. The reports by user Cloudie for Mar2008 suggest how hard it is for a ground observer to discern cloud altitude (though one confounding factor is that ground observer reports AGL(AboveGroundLevel), and satellites are likely showing MSL (MeanSeaLevel)).
[2] Aqua and Terra, polar-orbiting satellites carrying CERES (Cloud and the Earth’s Radiant Energy System) sensors.
www.ioccg.org/sensors/terra.html
www.ioccg.org/sensors/aqua.html
Transparent clouds. Somebody, somewhere, needs a dictionary.
I often wonder why the live Doppler radar shows rain blobs going north
when the clouds I see overhead are going west.
These NASA observers would have to know the line of sight to the satellite
when they count the number of cloud layers.
Large zenith angles reduce the probability of a cloud-free line of sight (PCLOS)
and increase the number of layers likely to be seen.
Who you going to trust?
Really sophisticated computer stuff that needs adjustments, fudge factors and guessing or slack jawed yokels that really read the temps??
Seriously, it will be interesting to see the outcome and what NASA actually does with it.
@interstellar bill:
I can explain the apparent anomalous motion of those `rain blobs’. Remember that the rain doesn’t actually move with the rain blob. It falls to the ground. What you see moving is a region of precipitation. Wind from the west carries a band of clouds which impacts on colder air causing precipitation in much the same way that waves break on a beach. The area of precipitation moves along the band just as the break moves sideways across the wave.
There are transparent clouds. And, there are a lot of them. These are clouds made of relatively large but few ice particles. They don’t scatter or absorb much sunlight and therefore appear to be transparent. However, if you use an IR or microwave detector they are readily visible in these wavelengths.
Heard of sub-visible cirrus clouds for the first time at AGU last week. S’cool seems like a good program.
I have noticed what may be a similar thing when flying. When starting descent, it is not unusual for the ground to be obscured by what looks like low cloud, which we seem to pass though on final approach, but looking up on landing seeing a clear sky.
Before they actually went to almost exclusively automatic weather stations, every observation station in the US would report all high, mid and low clouds. First in Airways, then in Metar. Now they only report significant layers below 12,000 feet…
Except for stations that are manned by military observers or contract observers. They still report all sky conditions. Just pull up KNGU and look at their ob. That’s Navy Norfolk.
And there was one set rule that we always abided by when I was an observer:
In the morning when the sky becomes light enough to see, suddenly there are reports of what we called “Cirrus Appearrus”.
At night, that suddenly became “Cirrus Disappearrus” and the reports read “SKC”
I think this exercise might prove useful in improving the ability of climate models to predict future climate. I have suggested else where that climate models should be tested by inputting climate conditions from say 1970 and running the model for the year 2000 and comparing results with real world observations. It should then be possible to apply a “confidence factor” to each model.
Just my thoughts FWIW. 🙂
Thanks! Dr. Joel Norris of the Scripps Institution gave an excellent presentation to Fermilab that discussed the many challenges of determining cloud cover from space & determining what effects clouds may or may not have on climate:
http://vmsstreamer1.fnal.gov/VMS_Site_03/Lectures/Colloquium/100512Norris/index.htm
Ground observer: “I see some invisible clouds!”
Satellite: “No, you don’t.”
Ground observer: “I don’t see any invisible clouds.”
Satellite: “Yes, you do!”
Richard111 says:
December 18, 2011 at 12:52 am
======================================
I think that’s called “hindcasting” and the models don’t do very well at it
Matthew W;
About the only thing the models do well is fundraising.
😉
NSIDC might like to check some of their purported ice extents in the Arctic too. I’m only talking at the extreme points, but if you see temperature data for the Baltic which is > 3C throughout, it is slightly surprising to see ice registered in the exit between Germany, Denmark and Sweden. Ditto at the mouth of the St Lawrence river.
It’s the usual cycle of technology needing to be optimised before it is totally reliable, but it points to a need to decide when satellite-only is reliable……….
I would hope that the “Ground Observers” were the Airport hourly Meteorological report record which should be RATHER DETAILED, and extensive, though I have no idea how long/far back the data has been retained. Cloud cover reports are rather accurate and detailed.
As a weather observer I used to wear the red goggles before going out to do the observation. Amazing how the Cirrus shows up, even the Disappearus.
Regarding Marks comments about #of cloud layers & Cirrus Appearus I’d like to put in my two cents worth from my many years as a weather observer (certified). In the tropics there is ALWAYS a cirrus layer about 30Kft which is only visible at sunrise/sunset. The satellite might catch that but with the exception of sunrise/set obs it never gets recorded. METAR code which replaced AIRWAYS code in the 80’s or 90’s (I can’t remember which) allows for a MAX of 3 layers of clouds to be reported unless a cumulonimbus cloud was present and then it is four. In the tropics around thunderstorm activity there were many times in which I recorded six or more layers (I think my record was nine). In my opinion METAR is *garbage* when it comes to accuracy of sky cover observations. Knowing the garbage METAR code requires the observer to report I would put my money on the satellites being right before on the observations.
When I became a forecaster I put in ALL the layers I thought the pilots might consider significant and ignored what the code said I was supposed to put in.
There! I got that off my chest. I’ve been waiting many years to get that rant out! 🙂
Some more two cents worth. When it comes to thin cirrus it is very difficult for an observer to see more than one or two layers. There were times when my first clue I had it wrong was the shadows of contrails on the cirrus below the contrast.
As a Photo Interpreter for the Air Force, I’ve seen a few miles of very-high-level imagery. There are some cloud layers you can easily see through from one direction, but are totally opaque from another direction. Most of this is cirrus cloud cover. Sun angle plays a huge role in this. I’m not sure how that would affect satellite detection. The altitude of the clouds also plays a part in how transparent they are. Low clouds are usually not transparent, while high clouds usually are.