What Is the Temperature of the Earth?

The news are all reporting 2014 as the hottest year on record, but no one is asking how such a measurement can be made without bias.

“The heat is on: NOAA, NASA say 2014 warmest year on record,” Seth Borenstein states on PhysOrg.  The data videos in the article look pretty conclusive.  But for anyone who thinks scientific conclusions should rest on evidence, not on authority (even prestigious specialty organizations like NOAA and NASA), it might bear examining the question of earth’s temperature from a philosophical viewpoint: How does one measure it?  We can’t put a thermometer under Mother Earth’s tongue.

Even when we use a thermometer for a human, we are arbitrarily taking a measurement from one part of the body to the exclusion of others, because we consider it “useful” as a proxy for a person’s health. The reading might be different in the rectum, or the ears, or at the abdomen compared to the extremities. If mouth temperature under the tongue is useful, there is a bit of natural fluctuation around the mean; your temperature is rarely exactly 98.6° F. Scientists know that all measurements involve some degree of error. They use different methods of averaging, such as mean, mode, and median, to select a measurement that is useful for their purposes.

The error bars and biases are greatly exacerbated when trying to come up with “a number” that represents a global measurement. For earth’s temperature, consider just a few of the sources of error and bias:

• Every square mile of the earth’s surface has its own temperature at a given time.  The poles are colder than the equatorial regions.
• Each point’s temperature is constantly changing.  It’s colder at night than during the day.
• The temperature varies with height above the surface.  The atmosphere has a temperature profile from ground to stratosphere that, too, is constantly changing.
• The temperature varies with depth beneath the surface; it’s hotter underground, and colder with depth in the oceans.
• The weather is constantly changing; air speed and humidity can influence measurements.
• Surfaces near a thermometer, such as black asphalt or grass, can influence measurements.
• Each environment has its own thermal inertia.  Some rocks cool down more slowly than other surfaces. The measurement might include some residual heat from the past instead of the current temperature.
• Humans or machines that read the instruments can make mistakes.
• There are multiple methods of taking temperature: thermoelectric, mercury and other liquids, or bimetallic strips, for instance. Who decides which type of device to use?
• If a change in device type is made at a station, or is rolled out at all stations over time, how do measurements with the new device correlate with measurements from the previous device?
• All recording devices must be calibrated. It becomes difficult to ensure equal calibration for monitoring stations around the world. Some countries could be more careless. Some stations might drop out from damage, neglect, or vandalism.
• It’s impossible to gather data from every point on the globe, so decisions have to be made about where to put monitoring stations.
• It’s impossible to gather data continuously from a given monitoring station, so decisions have to be made about what times of day to collect data points.  Should it be daily highs and lows? Or should it be temperatures at noon and midnight? Who decides?
• Earth’s climate has natural cyclic variations over multiple time scales, some of them poorly known or unknown.

This is not a complete list. So what is the “temperature of the earth”? There is no such thing.  There is “a” temperature at “a” place, at “a” time, under certain environmental conditions, at a given humidity and wind speed, as measured by an artificial device that may or may not be calibrated properly and working properly.  By itself, a temperature reading signifies nothing about climate, because it’s trying to measure a moving target in a very small location on a huge planet.

It should be obvious that human bias enters every factor. Someone has to decide where to put the monitoring stations, what device to use, and what data points to collect at what times. When the data are in, the numbers must be crunched and the error mitigated according to some model or method. A scientist or a program based on a model may decide to toss out data that appear anomalous according to someone’s criteria, but what if those data points are meaningful? Models and methods are human constructs, devised for their “usefulness” – but even so, models are only simulations of reality, not reality itself. It’s impossible to know all the factors that could influence the results.

Much more troublesome is trying to link temperature trends (if they are meaningful at all) to causes.  Seth Borenstein’s article is adamant that humans have caused 2014 to be the hottest year on record. Who could possibly know that? Pictures of smokestacks at power plants can’t prove it; that titillates the emotions by visualization, perhaps, but isolated pictures are not global evidence. Recently, for instance, it was reported that more methane (a potent greenhouse gas) was emitted by earthquakes than previously thought (Science Magazine).  Hardly a month goes by without some new factor being reported that could alter the models’ conclusions.  Here’s one from a few days ago on PhysOrg: “Mountain system artificially inflates temperature increases at higher elevations,” suggesting that “warming in the mountains of the western U.S. likely is not as large as previously supposed.” Here’s another one announced on PhysOrg 11 days ago: “A new NASA-led study shows that tropical forests may be absorbing far more carbon dioxide than many scientists thought, in response to rising atmospheric levels of the greenhouse gas.” Examples like this could be multiplied.

Sociologists of science probably find it intriguing that the “climate of opinion” on anthropogenic climate change has divided sharply between the “consensus” and the “deniers”.  Both have well-qualified experts to point fingers across the aisle. The point of this analysis is not to take sides, but to point out, from a philosophical perspective, that such answers cannot be known by the methods of science.  There are too many variables.  The error bars are too large.  There are too many unknowns and unknowables. Scientists don’t fully understand all the factors and feedback mechanisms, like clouds, ocean currents and forests, that could alter the models significantly.  Human biases are unavoidable.

For example, consensus scientists were caught red-faced when having to admit a 15-year “warming hiatus” since about 1995. They rushed to rescue their conclusions (i.e., about warming being man’s fault) by looking for ad hoc explanations. Here’s one that appeared a few days ago on PhysOrg: “The ‘warming hiatus’ that has occurred over the last 15 years has been partly caused by small volcanic eruptions.”  The article claims, “a series of small 21st century volcanic eruptions deflected substantially more solar radiation than previously estimated.” But this explanation, while giving the consensus a story to tell the press, raises more questions than it answers. What other factors are substantially greater “than previously estimated”? If small volcanoes have this much power to influence the climate in short order, why aren’t the alarmists blaming the volcanoes instead of the power plants?  And if smoke and dust lower the temperature, why isn’t the solution to throw even more soot from power plants into the atmosphere?

Consensus scientists know and believe that climate swings greater than anything observed today have occurred naturally in the past. So the current debate reduces to identifying what factors out of large number of possibilities, including unknowns and unknown unknowns, tend to indict human beings for a trend that might be purely natural. Experts in scientific ethics might well ask, also, why warming is such a worrisome thing, if animals and plants thrived in the past under even greater climate swings.

When scientific institutions (or consensus deniers) take leave of their empirical modesty and become emotional advocates for causes that cannot be rationalized by scientific methods, even in principle, it’s a good time to ask whether ideology or politics is influencing their behavior.  This goes for Clarke and Lawler, who passionately argued on The Conversation that people need to trust experts, otherwise they are being anti-intellectual. Some questions, though, are not questions of science; they are questions of philosophy about science. Many scientists are not trained to think critically about the limitations of science.

Again, this is not to take sides in a “heated” debate, but to step back and look at the debate philosophically. Professor Jeffrey Kasser, in his Teaching Company series on Philosophy of Science, tells a somewhat humorous story about the difficulty of objectively measuring a property of a material, namely “fragility.” It seems simple at first; you hit something, and if it breaks, it’s fragile, right? But what if it breaks only when hit hard, or with a certain kind of object? At the end of a long train of factors to consider, he ends with having to define fragility with a long list of arbitrary methods: you have to hit the object with a standard hammer with a standard whack at a given angle, etc. etc.

Even taking the temperature of a room could require a long list of directions that some human had to decide: use a certain kind of thermometer, at a certain height off the floor, holding it with gloves instead of bare hands, and so forth. But then, what kind of gloves, and how thick? Does the measurer have to wear a white lab coat? There are an infinite number of conditions that might change the measurement. We know some are silly and unlikely to affect the outcome, such as what the measurer had for breakfast. But those criteria cannot be defined scientifically; they are arbitrary, based on what the people who define the method consider useful. No human can know all the factors that come into play. And that’s just for measuring a small room. How much more defining the “temperature of the earth.” We hope you see that such a measurement is meaningless!

Interestingly, temperature itself is a vexed question. What is it? There are several definitions; motion of molecules, that which feels hot to the touch, that which raises mercury a certain number of millimeters in a tube, etc. But what is temperature? The operationalist Bridgman said, “Temperature is what thermometers measure” (Stanford Encyclopedia of Philosophy). The mathematician Fourier forswore the debate about heat, opting instead to offer an equation that describes what heat does, not what it is.  Often in science, the deeper the questions you ask about seemingly simple phenomena, the more puzzling they become.  Even Newton did not deign to feign what gravity is.  He only described what it does, and what calculations you can make with an equation.

Finally, we should think about scientific ethics. Ostensibly, it is the job of a scientist to describe and explain a phenomenon, not to advocate for what politicians should do about it.  So why are all these climate scientists like Michael Mann and his friends warming up to U.N. climate summits and telling them what must be done?  You may agree with him, and it is his right, like any other person’s, to have political beliefs. But to claim a belief about climate change is scientific goes far beyond the ability of science to know. The take-home lesson from this entry is that science is not objective; it is profoundly human, and humans are often driven by ideology.  Don’t be influenced by majorities and pictures and graphs, when the underlying data cannot be conclusive. If it’s consensus, it’s not science; if it’s science, it is not consensus.