Friday, October 24, 2008

The Case for Global Warming Skepticism

One of the hot-button environmental issues is Global Warming. While some people argue we must be more specific and refer to “anthropogenic Global Warming,” I do not do so. Not because I think humans are not affecting the environment in any way, but because I think it is scientifically impossible to accurately measure the temperature and compare it to historical trends in the first place. And if it is scientifically impossible to do so, then all Global Warming (anthropogenic or natural) is unscientific.

In order to demonstrate the scientific problems with Global Warming, we must first understand a bit of how scientific experiments work. The two key concepts will be our understanding of precision and accuracy. Without these two ideas firmly in place, we cannot even begin to weigh the evidence presented for Global Warming.

While precision and accuracy are often thought of as being identical concepts, in science there is a specific difference between the two. Precision refers to the level of certainty an instrument gives us. For example, if we measure time using an analogue clock with no second-hand, the clock is precise to the minute. That is, we can tell that it is 10:58. But we do not know if it is 10:58:03 or 10:58:57. On the other hand, we could have a clock with a second hand that would be more precise because it would illustrate the seconds. Furthermore, we could have a digital clock that would be able to give us fractions of a second as well.

However, at some point the precision ends. We might be able to use a stopwatch to calculate that something took 10.874 seconds, but we wouldn’t know if it was 10.8744 or 10.8740 or (possibly, depending on the specs of the stopwatch) if it was 10.8739 and rounded.

To tell how precise an instrument is we need to know how many decimal points the measurement goes to. The more numbers after a decimal point, the more precise the measurement is. As a result, 10 seconds is not as precise as 10.0 seconds (since 10 seconds could be rounded from 10.3 or 9.7, etc. while 10.0 could only be rounded from, say 10.03) and 10.000 would be even more precise.

In addition to precision, however, we must take into account the accuracy of the instrument. A stopwatch might give us a precise measurement down to the eighth decimal point (for example), but if the stopwatch is inaccurate then that precision is illusionary. For example, if two stopwatches measure the same amount of time and one says 10.000 seconds passed while another says 9.989 seconds passed, we would know that one (or both) are inaccurate instruments. There is, however, basic general agreement between the two instruments. We would be safe to say they were approximately similar, but we would need to give an error range in order to remain scientific. (And it should be noted that any scientific experiment that does not give you its error range is worthless.)

In addition to instrument accuracy we have to deal with human error. Sometimes, people misread instruments or write down the wrong number. Further, as we again use the stopwatch illustration, we have to deal with the lag between when the event being measured starts and when the experiment observer presses the button to begin the clock. This reaction delay must be figured into the experiments as well.

All that is fine and well, but what does it have to do with Global Warming? Well, first of all we know that Global Warming claims a specific temperature change over the course of the last century. For example, global temperatures are said to have risen about 0.6 degree Celsius since 1900. The problem with this is that our thermometers are far more precise today than they were in 1900 as the advancement of technology continues. Beyond that, we do not know how accurate the experimenters in 1900 were per say, or where exactly their thermometers were placed, or a host of other possible important factors to the experiment.

This is important for the simple fact that 0.6 degrees Celsius is a very small number. In fact, when you consider that most historical weather reports give temperatures in whole numbers (i.e., “On April 4, 1907, the high temperature was 68 degrees F”) this means the recorded temperature is not very precise at all.

This immediately brings us back to an important rule of precision. An experiment is only as precise as the least precise measurement used. For example, suppose you were trying to determine the volume of a cube. We know that volume is length x width x height. If we measure the length as 10.0 meters, the width as 5 meters, and the height as 14.973 meters, we multiply those numbers together to get 748.65 meters cubed. However, since the width is a whole number variable, the precision of the experiment can only be a whole number! We cannot have any decimal portion at all, so the true, scientific volume of the cube is 749 cubic meters. Due to this, the extra precision that we got measuring the height is irrelevant to the final answer. It can only be as precise as the least precise measurement.

Which means that if temperatures at any point in the data are in only whole number increments, we cannot have a temperature change of 0.6 degrees. The precision of the answer is more precise than the data given; it is invalid.

Now someone could argue that it doesn’t matter because the 0.6 degrees is in Celsius rather than in Fahrenheit, which is what virtually all of at least the earliest American data was measured in. However, this brings up another matter. The Fahrenheit scale is inherently more precise than the Celsius scale because the degrees are finer. That is, between freezing and boiling there are only 100 degrees on the Celsius scale, but there are 180 degrees on the Fahrenheit scale. This means that measuring in degrees F is 1.8x more precise than measuring in C. To show why this is a problem, both 87 and 88 degrees F round to 31 C. In fact, assuming infinite precision, 87F = 30.555…C and 88F = 31.111…C, a difference of 0.555…. Or, to put it another way, about 0.6 degrees C.

This means that virtually all of the touted Global Warming temperature difference could possibly be nothing more than just the imprecision of conversion between C and F.

But there is another problem with the methodology used to calculate Global Warming. It’s based on average data. Unfortunately, I’ve yet to see a report that indicates how the averages are determined. I’ve even emailed specific people who have written on the topic and gotten no response. Granted, I am pretty much unknown; still, this information is necessary for us to be able to make an informed decision as to the veracity of Global Warming.

Let me give an example of what that is the case. If the average is simply the average between the highest temperature of the day and the lowest temperature of the day, two radically different days can give the same average result. For instance, if the high was 80 degrees and the low was 40 degrees, the average would be 60. But the average would also be 60 degrees if the high was 120 and the low was 0. Granted, that is a rather extreme (and unlikely) example; but more realistically, 82H and 38L also average out to 60.

But beyond that, there are even more problems. Two days with identical highs and lows can themselves be radically different once you factor in the temperatures throughout the day. For example, suppose the high and the low occur within a 6 hour range and the two days look like this.

Day 1: hour temperatures from 6 – noon = 40, 50, 55, 60, 70, 80
Day 2: hour temperatures from 6 – noon = 40, 40, 50, 60, 70, 80

(The second day had cloud cover that kept the cooler temperatures in the morning, but once the clouds burned off the heat increased.)

The average (keeping the precision of the “experiment” above) of all the numbers for Day 1 = 59 degrees. The average for Day 2 = 57 degrees. That’s 2 degrees F different, more than 1 whole degree C too…and neither of those matches the 60 degree average between the high and the low alone.

While those numbers are arbitrary, they are not unlikely numbers at all. Indeed, it is very probable that the cloud cover effect could happen in the morning while the afternoon temperatures remain similar.

It is therefore critically important that we know how the averages are calculated. Indeed, another possible way that averages are collected is by simply taking the high temperature for the date and averaging it out for every other year for that same date. I.e., saying “The average high temperature on June 7 is 87 degrees.” In addition to not accurately representing how hot a day actually is (given the above, since two days with the same high can have radically different average temperatures when you break the day down hour-by-hour), we are also left with the fact that averaging on a daily basis ignores an important calendar phenomenon.

Leap year.

Yup, that pesky leap year thing throws off our precision because comparing June 7 of this year to June 7 of last year is not a precise comparison. The Earth is not in exactly the same place as it was that time last year (of course this also ignores the rotation of the solar system, etc. which probably would affect temperature well below the precision our instruments can detect anyway). In fact, it is possible that June 7 of this year is more likely correlated to June 6 of three years ago than June 7 of last year. As a result, “record highs for this date” are also pretty much pointless. They’re okay for giving a general idea of the weather, but they play havoc with trying to maintain any kind of precision on temperatures.

Unfortunately, I do not know which method scientists actually use to try to determine the average temperatures and to come up with their number of 0.6 degrees C. As I stated earlier, no one that I’ve e-mailed about this topic has ever bothered to answer my question. In order to make an informed decision, we must know this.

But even not knowing the actual method used, the methods I’ve shown above would be unable to provide any precise data for the past 100 years. And I do not see how any other method of determining this number could work. As a result, I have no reason to believe in Global Warming at all, let alone anthropogenic Global Warming. Scientists must provide the details of their experiments, the details of how they determined these averages, the error bars for the temperatures collected at the beginning of the 20th Century, etc. before we can even hope to accept it as a theory. Anything less than this disclosure renders Global Warming as unscientific.


  1. Though I do agree with you that Anthropogenic Global Warming is a myth, I still wonder how you would account for the melting of the ice cap(s) (not sure if both are melting.) It would seem something would have to be getting hotter to melt them.

  2. Lucas,
    I may be wrong, but it's my understanding that while the surface cover of the poles has gotten smaller, the poles have gotten thicker. I'm not sure that this isn't some normal oscillation between surface coverage and thickness where the mass of ice remains relatively similar, but deposition makes thicker and eventually displaces to create more surface area, which floats off as ice burgs and melts. I don't know that we have enough historical data to know for sure.

    Great observations! I've always said that if you want to know how certain science really is, just watch the weatherman.

  3. Lucas
    From what I've heard, the north pole ice is shrinking, but the south pole ice is expanding (both, of course, due to Global Warming...which is another problem with the theory in that everything is claimed to be caused by GW, even direct contradictions).

    In any case, I have no problem supposing that the North Pole ice sheet is getting smaller. That's something that could actually be demonstrated with a small degree of instrumental error, etc.

    But all it would prove is local warming. That is, it would say it's warming at the North Pole. But that says nothing about anywhere beyond the Arctic.

    That does actually tie into a bit with the averages again too. Since we don't know how the averages are collected, then it is difficult to see how a radical increase in temperatures at the North Pole (one site I read said they're about 15-20 degrees warmer), even if true, could translate to Global Warming if the world as a whole is still only 0.6 degrees warmer. (Also, note the ambiguity of scientists who mix the F and C scales; it's 15-20 degrees warmer in F, but it translates to 0.6 average difference in C. Again, this plays to the fact that F is more sensitive than C because it is more precise (almost double) and saying there's a 15-20 degree shift looks more impressive than saying there's a 7-10 degree shift.)

    However, it's also not certain that the northern ice sheets are actually dwindling either, as we've seen Alaskan glaciers increased last year and many places are recording record (or near-record) lows currently. Not that I put much stock in those "records" for the reasons I stated in my original post.

    And I suppose one final point could be made, and that's that it actually doesn't necessarily take extra heat (in terms of temperature) to melt ice. You do need extra heat in the form of calories (not to be confused with food calories). I don't remember the exact formula, but while it takes only 1 Calorie to raise 1 gram of room temperature water 1 degree Celsius (that's it's definition, BTW), it takes something like 70 Calories to melt ice, and those 70 calories do NOT increase the temperature of water. That is, a gram of water at 0 degrees C is frozen; add 70 (or whatever the real amount is) calories of heat and the ice melts, but is still 0 degrees C. This is the amount of heat needed for water to go through phase transitions, and the same thing happens for water to convert from liquid to steam (except, if I recall, at that point it takes something like 300 calories).

    Again, don't take those numbers as 100% accurate as I'm going from memory, but the concept is the same. In any case, I'm not sure how this would actaully affect GW models; but then I'm not sure if any of the Algoreites have even considered that in the first place :-)

  4. precision rules in project management

    Measure with micrometer
    Mark with chalk
    Cut with chainsaw.

    How hot is it today ??

  5. Fred: hilarious!

    ...and true, at least as far as project management goes in my industry.

    "How hot is it today ??"

    Good question. It's 54 degrees Fahrenheit here right now...
    Oh, wait! That's Accuweather. says it's 55F.
    I just checked the thermometer outside and it reads about 48 degrees. Hmmm: What elevation is best? I'm sure is cooler uphill. It's about 70F in my home office. It's probably over 80 just over the stove. It's also below freezing in my freezer. I'm feeling pretty well so I would expect my core temperature to be 98-point-something. How big of a "here" would it need to be? How could we determine an average temperature wherever "here" is supposed to be? Globally? Good luck.

  6. I love the comment about science being as reliable as the weatherman.

    If your car started as often as the weatherman was correct, we'd all lose our jobs and go hungry.

    Science is actually very good at predicting outcomes, that is why your microwave oven cooks food and not you.

    But you keep your healthy skepticism about science, and just know for sure God will take you to heaven when you die.

    Hot-button environmental issues...

    You think it is scientifically impossible to measure temperature.

    Sorry I didn't get past the first paragraph, it felt like at any moment you were going to talk about how God created the Heaven and Earth in 6 days...

    Your first comment by Lucas wonders why the ice caps are melting... and that is what usually causes humans to start looking for answers, you know, questions about what we see.

    Your lack of belief in science I suppose is balanced by a belief in God.

    Your car starts, your computer boots, yet we can't measure temperature.

    Pretty silly stuff... certainly a case for ignoring your blog.

  7. My comment about the weatherman has more to do with realizing that we have a false sense that science is an accurate source for detecting truth. Science has been hijacked by politics. If it can be argued that the Bible can be interpreted more than one way because we observe that different people do indeed interpret it more than one way, and that some ministers have abused the trust people have placed in them as interpreters of biblical truth for personal gain, then it can also be argued that science as a sociological entity can be misused by some scientists and patrons of science, people who have been entrusted to develop accurate scientific conclusions, for their own political purposes.

    Therefore, inasmuch as any scientific conclusion is of political value, it must be taken with a grain of salt and investigated carefully. If a conclusion is easily debatable and those who dissent are politically attacked, there stands a high likelihood that the scientific conclusion being debated is influenced by a bias of unrelated presuppositions. It’s one thing to presuppose the philosophical foundations of logic and another to desire an outcome because one prefers one system of government over another and desires to change it.

    With regard to likelihoods, scientific discovery utilizes bivalent logic. However, given simple syllogistic limitations, one realizes that scientific discovery cannot demonstrate a conclusion to be deductively true. It can only inductively demonstrate a conclusion to have a likelihood of being true. Different conclusions have different likelihood for different reasons. You can't equivocate the accuracy of science in general from any particulars, and that was my point. To many of the ignorant among us look at technology and are told that science is a great thing for telling us what is true, and the media has been effective at implying viscerally that science can also tell us what is essentially true. However, science is not so monolithic in its discovery of testable truth because there are many foundational things we either cannot test accurately or cannot test at all.

    As I have said, science is bivalent. It involves measuring the truth-value of explicitly defined subjects with what should be explicitly defined givens and presuppositions. However, science is not univalent. That is, there is no subject it can test that is only true in every case. Where a univalent subject affects bivalent subjects, the influence of the univalent subject can be tested. For example, the law of gravity is observationally univalent. We cannot test the law of gravity directly, but can observe its influence on other subjects. God is univalent. We cannot test Him. However, we can test His influence on other subjects. In this regard, Christian apologetics has a plethora of exceptional likelihoods demonstrating God's influence. It is right to say that these are not deductive proof. As I have said, science cannot demonstrate a conclusion to be deductively true. But the exceptional likelihoods are notable.

    Rather, since God is univalent, He is as self-evident as such as the law of gravity. AS it is, the law of gravity is observationally sound, but is yet mysterious in the apparent observation of some things to defy it. God, being greater than the law of gravity, is much more mysterious. Would it seem that many defy Him, yet only for a time? Rather, are those things that seem to defy Him rather playing a role in His glory?

    Finally, we discuss six days as though time is a constant; and the big bang is promoted without fully considering temporal dynamics. We don't understand the universe as a mechanism much less the non-mechanistic substance on which its existence rests. Yet would we observe the revelation of a Creator if He chose to reveal Himself? For there are some to whom he has revealed Himself. Can those to whom He has not speak of which they have not observed? Would that He reveal Himself to all, but that He glorifies Himself in all ways.

  8. TitanicExplorer said:
    Sorry I didn't get past the first paragraph

    Trust us, we could tell.

    TitanicExplorer said:
    Science is actually very good at predicting outcomes, that is why your microwave oven cooks food and not you.

    A) That's not a prediction of an "outcome" per se (although what you actually meant by that is anyone's guess at this point).

    B) Not all that is under the label of "science" is equally "scientific." After all, phlogiston was scientific and patently false.

    C) If we go by the "if it works it's true" method, then tell me what exactly science has given us with the Global Warming hysteria.

    You said:
    Your first comment by Lucas wonders why the ice caps are melting... and that is what usually causes humans to start looking for answers, you know, questions about what we see.

    Yes, and I see reports of "record cold" weather around the US, growing glaciers in Alaska, a thickening ice sheet in Antarctica. I look out my window and see winters pretty much exactly like they've always been through my entire life. And I'm supposed to ignore all this because some ex-politician produces a movie? Because some Socialistic goverments have redistribution concepts and want scientists to provide cover for it?

    You said:
    Your lack of belief in science I suppose is balanced by a belief in God.

    Science is not something that ought to be "believed in." It is supposed to be something that is demonstrated.

    I couldn't care less if you do or don't believe in God; it is irrelevant to this issue. The only thing that's relevant is whether Global Warming is scientific. That's the only issue here.

    Which brings us back to:
    it felt like at any moment you were going to talk about how God created the Heaven and Earth in 6 days...

    Funny how you're the only one bringing religion into this.

  9. Great article! I'm going to keep a copy to quote from it in the future, you have some good examples. There are few in the sciences (other than in the applied sciences where mistakes are costly or hard sciences like Physics and Chemistry where discipline runs strong) that pay attention to decimal precision in their data or their conclusions.

    As for the earth warming, it's definitely happening, and I don't even need to take a single measurement because I trust the core Laws of Thermodynamics (Newton, Carnot, Gibbs, Boyle, Brown, et al.). However, it's not happening because of anything mankind is doing or not doing, it's the effect of a radiative sun acting on a body in space.