Denier’s Favorite Go-to Chart? Teensy Weensy problems..

February 19, 2016

Science sharpshooter Dana Nuccitelli takes deadly aim at climate denialdom’s current favorite graphic.
The visual, created by Dr. John Christy and Roy Spencer,  heavily referenced in Senator Ted Cruz’s recent bogus show-hearing in December,  has not been submitted to any publication, where it might be subjected to review.

Turns out there are reasons for that.

Dana Nuccitelli in the Guardian:

There’s a chart that’s become the favorite, go-to among those who distort climate science in order to justify opposition to climate policies. It’s shown up in many conservative media outlets like The Wall Street Journal and The Australian. It’s been presented by congressional Republicans’ favorite climate witness, John Christy, in recent hearings called by Ted Cruz and Lamar Smith.

The chart, which tries to compare estimates of the temperature of the Earth’s atmosphere to model simulations, is full of problems. Its purpose is clear – to make it appear that climate models are unreliable, and thus cast doubt on all of climate science and the tremendous risks posed by human-caused global warming. In other words, its purpose is to undermine climate policy efforts to reduce those risks by slowing global warming.

But as with all charts of such singular, unscientific purposes, it’s simply another example of cherry picked data being presented in a multiply misleading way, as the annotated version below illustrates (for an animated version, click here).

1) The data are misleadingly misaligned
To start, the observational and simulated model data are aligned at a single point at the start of the graph, in 1979. That choice serves to visually exaggerate any difference between the models and data. When doing this sort of comparison properly, climate scientists will average the data over a period called the “baseline,” usually between 10 and 30 years, so that any abnormally hot or cold single years in the data or model runs don’t throw off the alignment.

In fact, there was recently a paper published by Hawkins & Sutton showing that the choice of baseline can significantly impact how accurate the models appear in this sort of comparison, and thus it’s important to carefully choose an appropriate baseline period and perhaps test the comparison with multiple different periods. Christy didn’t even use a baseline period to begin with! NASA Goddard director Gavin Schmidt did the comparison using a baseline period of 1979–1988, and as you can see, the result looks quite different.

bogusgraph2

Temperature of the atmosphere to the mid-troposphere in the tropics, comparing model simulations (gray envelope) to observational estimates from satellites (colored lines). Illustration: Gavin Schmidt

2) The uncertainty is ignored!

Normally, climate contrarians love to harp on uncertainties. In fact, the whole argument against implementing climate policies to slow global warming hinges upon the backwards logic that we don’t know enough to take action.

However, in Christy’s chart, no uncertainty ranges are shown whatsoever. The difference this makes is clear in Schmidt’s version above, which shows the spread of individual climate model simulation runs (in grey), although doesn’t show the significant uncertainty in the observational data. When both are taken into account, the observations are consistent with the range of model projections.

3) Observational data disagreements are hidden

Christy’s graph also averages together multiple different observational data sets, which aren’t in terribly close agreement. Gavin Schmidt shows some of these in his version of the chart. Christy’s version also somehow glosses over the fact that atmospheric temperature estimates from satellites and weather balloons diverge in recent years.

4) The chart isn’t peer-reviewed or easily reproducible

Another problem is that Christy doesn’t say which observational data sets he’s averaging together, but there are more than three groups that use satellite data to estimate the temperature of the atmosphere (Schmidt shows four of them in his chart). So he could be omitting some of the data that shows larger warming trends.

We don’t know because this is “blog science.” The analysis hasn’t been vetted by other experts, or published in a scientific journal after being subjected to peer-review. We don’t even know if Christy is accurately representing the climate models, because they don’t produce a simple “bulk atmosphere temperature” output. As such, Christy’s chart should be taken with a grain of salt.

5) We don’t live on Mount Everest

The average elevation of the “bulk atmosphere” shown in Christy’s graph is 25,000 feet, which is just below the peak of Mount Everest, and not far below the elevation at which commercial aircraft generally fly. The temperature at such high elevations isn’t very relevant to humans.

We live on the Earth’s surface, so we care about the global surface temperature. And because we live at the surface, we have an easier time measuring the temperature here, so we know average global surface temperatures with more certainty than the atmospheric temperatures that Christy focuses on.
6) The rest of the global warming data show climate models are accurate

Atmospheric temperatures are just one of the many climate changes simulated by climate models. For example, climate models have done an excellent job predicting how much temperatures at the Earth’s surface would warm.

globalmeancmip

Global mean surface temperature series (red) along with five different Monte Carlo surrogates based on forced signal + ARMA noise realizations (gray) using CMIP5 all-forcing experiments. Illustration: Mann et al. (2015)

They’ve also done well predicting the rate of sea level rise.

slrgraph

Estimated changes in the observed global annual mean sea level (GMSL) since 1950 relative to 1961–1990. Estimated changes in global annual sea level anomalies are presented based on tide gauge data and based on sea surface altimetry. The shading shows the largest model projected range of global annual sea level rise from 1950 to 2035 for the IPCC First Assessment Report (FAR), Second (SAR), Third (TAR), and for Church et al. (2011) based on the CMIP3 model results. Illustration: IPCC AR5

Arctic sea ice is actually declining a bit faster than models predicted.

arcticicestroev12

Projected vs. observed September Arctic sea ice extent. Illustration: Stroeve et al. (2012)

And then there’s ocean heating. Only about 1–2% of the energy imbalance causing global warming goes into warming the atmosphere, and only about 3% is used up melting ice. Over 90% of that energy goes into warming the oceans. And climate models are doing a very good job predicting the rate at which the oceans are heating up.

oceanheat

Observed (red) vs. CMIP5 model simulated (gray and black) ocean heat content. Illustration: Cheng et al. (2015)

Proven reliable: climate models and contrarian cherry picking

As the statistician George Box said, “all models are wrong, but some are useful.” Climate models are certainly useful, and are doing a pretty darn good job predicting global warming. Their predictions have been far more accurate than those made by climate contrarians, who keep telling us that the Earth will start to cool any day now, as we keep breaking heat records.

The only reason to focus on the temperature of the atmosphere, most of which is far above where humans actually live, which we can’t measure as accurately, and which only accounts for about 1–2% of overall global warming, is because it’s the only data that seems to indicate that climate models might not be accurate. That’s called cherry picking. And even then, to make that argument requires presenting the data in a multiply flawed, misleading way.

Climate models are doing a good job because we understand basic physics. We know that as we keep increasing the greenhouse effect, more heat will be trapped, and the planet will keep warming. Ice will keep melting, sea levels will keep rising, and various bad things will happen, because we’re warming the planet much faster than it does naturally. That’s a dangerous physical reality we can’t escape just by creating cherry picked, misleading charts.

Advertisements

6 Responses to “Denier’s Favorite Go-to Chart? Teensy Weensy problems..”

  1. firstdano Says:

    HotWhopper did a good analysis of the original Christy fib as well.

    Best,

    D

  2. indy222 Says:

    Aghhh! You forgot the BIGGEST reason Christy’s chart is bogus – and that is the failure to include the PDO negative cycle from the exact start of the chart to the end of the chart – perfect alignment there! – and that climate models are not able to guess the star/end of PDO cycles. ie they effectively simply average over PDO as if it’s not really an oscillation.

  3. grindupbaker Says:

    Yes. The CMIP5/UAH comparison graph at the end of (1) would have been nicely augmented by imposing that chart of the ENSO & PDO on it showing the ENSO indices in red (higher) and blue (lower) colorations if I’m recalling that correctly.

  4. Sir Charles Says:

    Graph by Berkeley Earth:

  5. grindupbaker Says:

    I want to study a bit this MSU method if I can find some hobby time this year. Can anybody point me to the graph or table of the temperature weighting functions by altitude that UAH uses in order to obtain an average temperature of atmospheric molecules ?

    For RSS according to Fig. 3 in “Construction of the RSS V3.2 Lower-Tropospheric Temperature Dataset from the MSU and AMSU Microwave Sounders” by CARL A. MEARS AND FRANK J. WENTZ published in American Meteorological Society on 11 February 2009 temperature weighting functions as a function of
    altitude are approximately (just my eye-ballling of the plot so perhaps I misunderstand its meaning):
    14-20 km -0.022 to -0.032
    13 km 0.000
    12 km 0.030
    11 km 0.030
    10 km 0.14
    9 km 0.23
    8 km 0.30
    7 km 0.54
    6 km 0.45
    5 km 0.48
    4 km land = 0.43 ocean = 0.45
    3 km land = 0.33 ocean = 0.38
    2 km land = 0.24 ocean = 0.27
    1 km land = 0.12 ocean = 0.14
    land surface 0.195
    ocean surface 0.105


Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: