Weather and Climate
In-class exercise #1
In this exercise, we will relate clouds observed by satellites at night to temperatures at the
Infrared radiation is heat. Three things control how cold it will get overnight in winter:
cloud cover, snow cover, and winds. (During the rest of the year, snow cover doesn’t
control low temperatures, but the other two factors do). If clouds exist, outgoing
radiation (at nearly all wavelengths) is absorbed and re-emitted back down towards Earth,
inhibiting cooling. If skies are clear, outgoing radiation can escape to space, allowing
cooling. Light (or no) winds also promote cooling. Where winds are strong, any cool air
that settles into low regions near the ground (remember that cold air sinks) can be
scoured out as warmer air aloft is mixed down. Snow cover also promotes cold
temperatures. Snow is an excellent and efficient emitter of infrared radiation. If there is
snow on the ground, then heat is very quickly emitted and temperatures can fall rapidly.
The table below summarizes things.
Snow Cover Clouds Wind
Exists Expect colder Expect Warmer Expect warmer
temperatures temperatures temperatures
Missing Expect warmer Expect colder Expect colder
temperatures temperatures temperatures
How can you detect clouds at night with satellite data? The Earth emits radiation – all
things with a finite temperature do – and clouds emit radiation too so if you have a
radiation detector on your satellite and it detects radiation at the appropriate wavelength,
you should be able to detect the difference between warm temperatures (the ground) and
cold temperatures (cloud tops). Radiation detectors on satellites are called radiometers.
Radiance – emitted radiation – at a particular wavelength is detected by the radiometer,
and the temperature of the surface emitting that radiation can then be determined
theoretically (using a couple of different equations and some assumptions).
The difficulty arises when a low cloud (fog, for example) has a temperature that is very
similar to the surrounding ground. In such a case, you can’t use temperature to
discriminate between cloud and ground. But even then, there are techniques to find
clouds. For example, an animation might show clouds moving while the features on the
ground are stationary. A second method to find clouds requires observations of emitted
radiation at 11 microns (the wavelength where radiation emitted by the Earth peaks – that
is, the Earth emits more 11-micron radiation than any other wavelength) and observations
of emitted radiation at 3.9 microns.
The Stefan-Boltzmann equation relates the amount of radiation emitted to a temperature.
From the powerpoints, you might recall that the radiation R = T4. The amount of
radiation is a constant times the absolute temperature to the fourth power. This equation
assumes that the object emitting the radiation is emitting the maximum possible at that
temperature – in other words, that it is emitting as a blackbody. In reality, all objects can
be characterized by an emissivity; if it’s 1, then the body emits as a blackbody. Many
bodies however have emissivities < 1, which means they don’t emit quite so much
radiation as theoretically possible. Even more important – at least from a meteorological
observations point of view – objects can have emissivities that are a function of
wavelength. In other words, at one wavelength they may emit as a blackbody, but at
another they do not. Liquid cloud droplets have that characteristic. At 11 microns, cloud
droplets emit just about as much radiation as they should given their temperature. At 3.9
microns they do not. So, for a given temperature, the amount of 11-micron radiation
emitted by a cloud full of water droplets is the maximum; the amount of 3.9-micron
radiation emitted by a cloud full of water droplets is not the maximum – it’s a little less.
The detector on the satellite assumes that the object emitting the radiation that the sensor
is detecting is the maximum possible – it assumes that the amount emitted is the
maximum amount possible for a given temperature and the sensed radiance is used to
compute a temperature based on the blackbody assumption. For 11 microns in this case,
that’s a good assumption, and the temperature inferred from the intercepted radiation is
close to the temperature of the radiating surface. For 3.9 microns, the assumption that the
emitter is a blackbody is not so good; not as much radiation therefore makes it to the
detector, and the decreased amount of radiation is interpreted as a cooler surface that is
emitting as a blackbody. So if you compare the temperatures inferred from the observed
radiances at 11 and 3.9 microns, there’s a difference due to the different emissivities at 11
and 3.9 microns where liquid water droplets occur. You can use this difference to define
where low clouds are at night.
In this exercise, you will be looking at station data that includes temperature (upper left),
dewpoint (lower left), pressure (upper right) and wind information. In addition, there is
cloud data to consider, but the cloud data are presented as a difference field between the
11 and 3.9 micron observations (note that 1 micron is 10-6 m, or a millionth of a meter.
The wavelength of visible light is about 0.5 microns).
The data are shown at this website:
http://www.ssec.wisc.edu/~scottl/satmet/LowCloud.html (It’s LOTS of data – you might
not want to do this over a slow line, where slow is dsl or slower)
In the presentation, the low clouds are represented by darker shading and relatively clear
skies are grey – so all of Minnesota is clear, for example. Note the southward
progression of low clouds over Wisconsin as time passes.
A note about times: times in meteorology are most frequently referenced to Greenwich
England – at the Prime Meridian (0o E); these times are called GMT (for Greenwich
Mean Time) or Z time. You’ll notice the Z after the time in the plots. Central Standard
Time (CST) is 6 hours behind GMT/Z time. The first satellite image, 02:01Z on 15
January 2008 is 8:01 PM CST on 14 January 2008.
Locate station KPBH in north central Wisconsin (Phillips, WI). How does the
temperature there change with time during the course of the night? Is the station clear or
cloudy? What are the winds like? I’ll tell you that there’s snow on the ground there.
Which of the factors promotes cooling and which does not?
Locate station KMSN in south central Wisconsin (Madison) and KRPD (Rice Lake) in
northwest Wisconsin. How do the temperatures at these stations change with time? Can
you relate the changes to skycover or windspeed? Fill out the table below.
Station Time (Z) (CST) Clouds Wind Temperature
KMSN 0600Z midnight 9F
KRPD 0600Z 3F
KPBH 0600Z 16 F
If you look at the loop on-line, you’ll notice something dramatic happens starting around
13Z. This is what time in central standard time? It’s 7:00 AM…what normally happens
around 7 AM? After 7, instead of dark low clouds (which correspond to a 3.9 micron
temperature that is cooler than the 11 micron temperature), the low clouds become very
bright because the temperature computed from the 3.9 micron radiance becomes much
larger than the temperature computed from the 11 micron radiance. Suddenly, when the
Sun rises, the amount of 3.9 micron radiance becomes very large compared to the amount
of 11 micron radiance? Can you think of a reason why? The sun does emit a lot of
radiance at 3.9 microns – much more than the amount at 11 microns. That radiation is
reflected from the clouds when the Sun is up. Thus, the character of the 11-3.9 difference
changes from night time (when the emissivity of clouds differs at 11 and at 3.9) to
daytime (when the radiance reflected differs at 11 and 3.9)