Marshall Shepherd is former chair of the American Meteorological Society. He directs the Program for Atmospheric Sciences at the University of Georgia.
Marshall Shepherd in Forbes:
A tweet by Dr. Jason Patton at Oklahoma State Universityinspired this question,
How many of you will check “satellite-derived temperatures” today to determine what you are going to wear or whether you should cover your flower bed tonight?
Actually this statement is a bit tongue-and-cheek because satellite
derived estimates have a different objective. A public narrative has emerged that misrepresents how climate scientists record temperature. I spent the first half of my career at NASA and served as Deputy Project Scientist for the Global Precipitation Measurement (GPM) mission now in orbit. I also teach a Satellite Meteorology/Climatology class. Satellite datasets are important complementary datasets, but they are not typically used to validate ground measurements. With GPM, we meticulously established a ground validation program so that we have a robust set of “ground” measurements to tell us if measurements from 350-800 km in space are “accurate.”
At a recent Congressional hearing, Senator Ted Cruz stated that satellite data shows “no significant global warming for the past 18 years.” An article in Mother Jones summarizes an important exchange between the Senator and Admiral David Titley, a decorated Navy admiral and professor at Penn State University. Titley’s response has become an instant classic:
I’m just a simple sailor,but it’s hard for me to see the pause on that chart.
He was pointing to his chart of over 100 years of data. He also pointed out that Cruz was referencing a dataset that began just prior to a relative maximum in temperature associated with the 1998 El Nino.
2015 will be the warmest year on record and it’s not even close. If it were a track meet, 2015 would almost be lapping the field. A recent Washington Post Capital Weather Gang article headline sums it up, “2015 almost certain to be Earth’s warmest year on record by an enormous margin.” NOAA’s National Center for Environmental Information (NCEI) shows that no place in the continental United States had a mean temperature departure average that was relatively “ cool” for the September to November 2015. NOAA climatologist Deke Arndt tweeted:
Fall ’15 wall-to-wall warm unlike more regional patterns of past Ninos
I wanted to provide a quick overview of the science of satellite vs ground temperatures to inject some needed context into the discussion. I wrote to my colleague Dr. Roger Pielke, Sr. We have published papers together on the role of land cover changes on climate. He is a colleague that I respect even if we disagree at times. Roger emailed to me,
A major advantage of the satellites, as you, of course, know, is that the coverage is essentially global. It is also a layer (mass) weighted average.
There is no doubt that a disadvantage of thermometers is that it is difficult to cover the polar regions, vast areas of oceans, and other inaccessible regions. We have the same problem with rainfall measurement and is why we developed the GPM mission.
Satellites can measure temperatures in a different way using something called passive microwave radiometry. In passive microwave retrievals, temperatures can be “derived” from emissions if a certain set of constraints are applied. Different microwave wavelengths have different “weighting functions” that aid in extracting the layer of the atmosphere where the energy is being emitted. By piecing together the information from the different layers, you can build a vertical profile of temperature. It is important to note that we are also talking about “layers” of the atmosphere rather than specific levels that we get from instruments ascending on weather balloons. The Advanced Microwave Sounding Unit (AMSU) is one of the work horses of atmospheric temperature retrieval and is described here, but there are other instruments as well. The Eumetsat website notes:
AMSU-A uses oxygen absorption bands/lines for atmospheric temperature sounding. Window channels at 23.8, 31.4 and 89 GHz provide information on surface temperature….
Let me translate this to “English.” The satellite uses something called selective absorption. For example, ozone in the stratosphere is a selective absorber of the sun’s ultraviolet (UV) radiation but allows visible and heat (infrared) energy to pass through. Kirchoff’s Law tells us that good absorbers at certain frequencies will emit energy. Our bodies absorb and emit energy as infrared energy (heat). Oxygen is a good absorber and emitter at certain microwave frequencies that the AMSU instrument is sensitive to.
Like surface observations, satellites have their own challenges. It is remote sensing rather than a direct “in-situ” measurement. Clouds and precipitation influence (attenuate) the amount of emitted energy to the sensor (anyone have satellite TV?). For a precipitation scientist like me, the attenuated “signal” is actually useful and the basis for how we can measure rainfall from space. Other sources of error include variance in orbits, instrument degradation, and errors in conversion of data to temperature.
I often see criticism of the ground-based temperature record because adjustments are made. Trained scientists understand that generally “all” data must be quality controlled or adjusted. Satellite-based datasets are not different. Roy Spencer of the University of Alabama-Huntsville notes in this document (link) that data adjustments are required for a variety of reasons including calibration, satellite changes, and so on. So can we end this silly notion that only thermometer datasets are being adjusted? Even adjustments in different versions of the satellite dataset can shift the results from trend to no trend in one update.