EMAIL EXCHANGE WITH DR. JUDITH CURRY

AN EMAIL EXCHANGE WITH DR. JUDITH CURRY

THE PROLIFERATON OF MODELS IN CLIMATE SCIENCE

NOVEMBER 5, 2010.  You don’t go to a spot in Tennessee, drill down into the ground, extract a core sample, analyze it and publish a graph that reports the evolution of the planet’s temperature for the last 1000 years.

There are measuring stations, on land, all over the globe—some stations far less reliable than others.  There are satellites overhead that have been recording radiance since 1978.  There are ocean measurements as well. There are tree cores that reveal information about temperature.

How all these records are interpreted and then coordinated is a matter of controversy.  Models are used.  These models are competitive with one another.  And they are strings of complex mathematical inferences which give more weight to certain data and less weight to other data.

After a year or so of sifting through scientific and (to me) arcane debate about the flaws in various models, I began to wonder whether all the models were so abstracted from plain observation that they were useless. 

On October 27, I emailed Dr. Judith Curry at the Georgia Institute of Technology.  Dr. Curry is a well-known climate researcher who, unlike many of her colleagues, has shown an interest in skeptics who reject the notion that manmade warming is a serious and imminent threat.

Here is that email:

veteran medical reporter query on climate science

InboxX

jon rappoport to curryja

show details 9:36 AM (3 hours ago)

Dr. Curry,

Jon Rappoport here.  I’ve worked as an investigative medical reporter for 30 years.

In uncovering various kinds of medical-research fraud, I’ve educated myself about how the use of sophisticated models can obscure and distort simple data and cover over the basic uncertainty of the research conclusions.

I’ve tried to apply the same understanding to debates about competing climate-science models. 

I fully realize that most scientists think models are absolutely necessary, and the proposal that important knowledge might be acquired without them appears ludicrous.

Nevertheless, I’ve tried to approach the warming question with an idiot’s eyes and, perhaps, a useful swipe or two of Occam’s Razor. 

So please remember—in the field of climate science, I AM a complete idiot.  I’m navigating by my own version of common sense, and I’m seeing problems with every step I take.  The only question is, are these problems less troublesome than those caused by building models that spawn other models?

Suppose, as regards land measurement of warming, we did the following:

Select the hundred most reliable land stations in the world—those that have a clear, continuous, daily record of temperature going back at least 75 years.

The environments around the stations haven’t changed so radically in 75 years that obvious warming factors were introduced (highways, factories, shopping malls, etc.).

The daily temp measurements, as far as we can tell, were carried out with integrity and accuracy.

We then perform one, and only one, slightly abstract arithmetical calculation: We take each day at each station, and average the temp measurements recorded at six in the morning, noon, and six in the evening—or as close as we can get to those times.

Now, for each of the 100 stations, we have a daily average temp.

We then graph that daily average for each station across those 75 years.

We then have 100 graphs in front of us.

For each graph, we do…nothing.  We just look at each graph.  We keep looking.  What are we seeing?  An obvious overall story of dramatic warming increase?  Decline?  Stability and sameness?  A mixed bag of uncertainty?

Then we draw a trend line for each graph. 

Then we ask, “How many degrees of warming for any given graph, across 75 years, would constitute a red flag?”

I would leave the answer to that question to you.

Let’s call the answer X.

Then we ask, “For how many of the 100 graphs do we see a trend of X or greater?”

Suppose the answer is 42. 

Well, is the geo-distribution of those stations clustered in one area?  Are the stations far apart?

What are the other 58 stations telling us?

In other words, we begin to construct a narrative.  A story told by what we see.  Others can examine our story and comment on it—without invoking complex models.

It seems to me this is a reasonable starting point.  And for those who want to jump off from it into the aether with their models, well, it should be apparent how valid or invalid their suppositions and reasoning are, right from the start. 

Does all this seem completely insane?

When I read comments on blogs in which people argue the flaws vs. merits of some extremely complex model, I get the uneasy sense they’re debating angels on the heads of pins—and they’ve wandered so far from what temperature is and means they’ve lost the thread. 

At the risk of losing you (if you’ve come this far), I can offer an analogy in the field of disease diagnosis and testing.  One can read papers in which the subtleties of antibody-test interpretation are argued: layers, bands, false-positives in non-risk populations, etc.  But one glaring fact, which is being ignored, stands out: Until 1985, the presence of antibodies specific to a particular germ was considered a good sign for the patient.  It meant his body had fought off and disposed of the invader.  Antibodies were certainly not a clue that the patient was ill or was going to get ill.  Then, without the slightest justification (except some clever circular reasoning), the whole business was turned on its head.  Antibodies were indicative of illness. 

I point this out to a number of well-known researchers, and they simply stare at me as if I’m mad.

So then I say, “You know, when you vaccinate a person, you’re producing antibodies.  And you’re saying the antibodies confer protection against illness.  But when those same antibodies are produced naturally, by the body, you say they’re a bad sign.”

Then they throw me out. 

Hope to hear from you.

Regards, Jon

Here is Dr. Curry’s brief reply.  I have deleted a sentence or two that contained private information Dr. Curry didn’t want published.    

Curry, Judith …10:29 AM (2 hours ago)

 Reply |Curry, Judith A to me

show details 10:29 AM (2 hours ago)

hi jon thanks for your message.  not at all insane, in fact a number of people are doing similar things…I will get to this general topic in december (at least according to my current plans).  Judy

I take Dr. Curry’s reply in a positive light.  Maybe there is another way. 

Skeptics have already shown the massive unreliability of many data that are widely accepted as useful.  So perhaps it’s time to say all the models are a sign of a bad habit; a useless and harmful addiction.

Throughout science, there are types of models that should never have been built in the first place.  They were interesting to the builders in the same way that advanced chess is interesting to those who can play at that level.  But by their very nature, they’re rubbish science.

They were doomed from the start to chart a flight path that was, a priori, in a universe vastly different from the observational data.  

JON RAPPOPORT

www.nomorefakenews.com

Jon Rappoport is the author of LOGIC AND ANALYSIS, a unique course for home schools and adults.  For inquiries: qjrconsulting@gmail.com