The Most Extra New Super Global Warming Thread

[xTYBALTx]

I. I believe I can isolate our disagreement:

Originally Posted by bkelly

What an average of the twelve studies offers is a better estimate of the average of the actual temperature.

A "better estimate"? Better than what? I find that statement quite presumptuous - it intrinsically presumes some studies to be more wrong than others, an assertion for which you provide no support.

II. On the consistency of the reconstructions:

Sequence of conversation thus far:

Originally Posted by bkelly

I think the data shows remarkable consistency in terms of the degree of the temperature anomaly throughout the time scale. In English, they generally trend up and down together.

Originally Posted by tybalt

You must be looking at a different graph, because the one you posted shows nothing of the sort.

Originally Posted by bkelly

No? You don't see each trial showing rising temperatures from 700 to 1000 and then a drop-off from 1000 to 1250? 1250 to 1500 is chaotic but 1500 to 1850 shows a slight cooling trend while 1850 to the present is quick warming.

a. "...each trial showing rising temperatures from 700 to 1000..."

Six sources go to before 1000. Five go to 700.

MJ2003 shows little/no rise from 700 to 1000.
HCA2006 shows very very very minor rise from 700 to 1000.
MSH2005 shows rising temps from 700 to 1125, and most of the rise by far is from ~950 to ~1125. As most sources cool, this one warms.
Severity of rise varies from +1C to +.1C. I believe the warming of the entire 20th century is estimated at +.4C; a differential of .9C between different sources is beyond significant.

b. "...and then a drop-off from 1000 to 1250"

MBH1999 is a meta analysis, I believe. Auto d/q in this discussion.
Timing of the drop varies by 100 years between some graphs, with severity of the fall varying from -1C to -.4C. I believe the warming of the entire 20th century is estimated at +.4C; a difference of .6C is beyond significant.

c. "...1500 to 1850 shows a slight cooling trend..."

PS2004 shows an increase over this period.
Severity varies widely, again, but there is less for this period than the other two. Perhaps data become more reliable as we get closer to the present?
MJ2003 shows very very little cooling for this period.

d. "...and the present is quick warming"

The severity of change in the period from 1900 to 1950 is no larger than can be observed in numerous other places on the graph. Even the total, if we accept literally the land monitors' data, is +.8C for ~100 years, which is large but not the largest from the available data.

III. My point in all of this:

Is that the debate should be other methodological differences, data source differentials, and the dozen other sources of inconsistency amongst our sources. This will help determine which of the dozen sources are better and worse, and will help us get a more accurate picture of past data. It should not be about the correct way to agglomerate "equally" disparate results of numerous studies. If all sources were equal, they'd all show the same result. Clearly they do not show the same results - some must necessarily be better than others.

[bkelly]

Originally Posted by Kerostasis

For reaching the 99.95%, we ask "given TRUE, how likely will all studies say TRUE", and "given TRUE, how likely will all studies say FALSE"...

Ah, I understand! You're saying that when twelve 51% chance trials agree, the odds that they agree upon a actual truth is equal to the odds of all twelve getting the right answer (.51^12) divided by the total number of ways all 12 trials might agree. When there are more ways to get the problem wrong, the odds of a false positive go down so an answer is more likely to be right.

You're right Kerostasis, with a T/F setup the probability that twelve 51% trials in agreement will be right equals .51^12 / ( .51^12 + .49^12) = 61.8%. My "difficulty to achieve" equals "probability of correctness" and 99.98% number is bogus.

Originally Posted by Kerostasis

What you need instead is a way to show that the studies are more reliable than 51%, rather than simply relying on the fact that there's lots of them.

But alas, I still disagree with this. Weak evidence becomes strong in large enough numbers. With twelve in agreement in our above example 51% becomes 61.8%. With 120 trials in agreement, the probability becomes 99.2%.

Hmm, I had not considered the two choice scenario for my example. I see this as a poor illustration now that I have that perspective.

Kerostasis, do you disagree that in theory enough weak evidence becomes strong or do you just think that global warming evidence is so weak that a boatload of it wouldn't be enough to change your mind?

bkelly

[bkelly]

Originally Posted by xTYBALTx

Is that the debate should be other methodological differences, data source differentials, and the dozen other sources of inconsistency amongst our sources. This will help determine which of the dozen sources are better and worse, and will help us get a more accurate picture of past data.

*chuckle* I am not surprised you would say this Tybalt. I admire you're spirit.

I think looking in detail at the results of one of the studies would be a good next step. I'll need some time to study up on one and make a good presentation of it. Perhaps I'll be ready before the end of the Thanksgiving weekend.

Is there one study you'd like to know more about Tybalt? Preferably one that includes tree rings for proxies.

bkelly

[xTYBALTx]

As outsiders, we are largely relegated to analyzing the analysis - we can look at the results and aide in drawing conclusions but can do little else. My PhD will be in Economics/Political Science, not in climactic science, so I don't think that it makes much sense for me to do my own analysis of studies; rather, as non-experts you and I can help understand the analyses performed by the experts on their own studies. That is assuming, of course, that the experts are not happy with the intellectually lazy state of affairs which exists right now, and therefore take it upon themselves to perform additional analysis.

[bkelly]

Originally Posted by bkelly

Is there one study you'd like to know more about Tybalt?

Originally Posted by xTYBALTx

...I don't think that it makes much sense for me to do my own analysis of studies; rather, as non-experts you and I can help understand the analyses performed by the experts on their own studies. That is assuming, of course, that the experts are not happy with the intellectually lazy state of affairs which exists right now, and therefore take it upon themselves to perform additional analysis.

*scratches his head* Um, is that a "no"?

bkelly

[Kerostasis]

Originally Posted by bkelly

Kerostasis, do you disagree that in theory enough weak evidence becomes strong or do you just think that global warming evidence is so weak that a boatload of it wouldn't be enough to change your mind?

bkelly

I agree that in theory enough weak evidence will become strong, but I had to disagree with you about the rate at which it becomes stronger. Individually weak evidence becomes strong as a group much more slowly than individually strong evidence becomes even stronger as a group.

As far as the particular evidences proposed here -- I havent looked at them closely enough to establish whether they are weak or strong to begin with. But the question becomes more complicated because we have to not only acheive the correct data but also acheive a correct interpretation of that data. For example, the question of Volatility that you were starting to examine is an important part of the interpretation.

[GlobalWarmin]

Drought Exacerbates Carbon Dioxide Problem

http://www.livescience.com/environme...ought-co2.html

[El_Gringo_Grande]

Funny story.

A couple of old farmers today where eating in the booth behind me. Him and his friend where discussing the weather. They talked about how warm this fall has been. One says it is probably global warming. The other says "Most of the scientist think it is just a natural fluctuation." The other says "Geary England has mentioned global warming a couple of times. There might be something to it." The first guy says "Well Rush says that almost all the real scientists think there is no such thing as global warming." The second guy says "Well that Rush is one smart guy."

What does that conversation say about Americas bread basket?

[Kerostasis]

Originally Posted by El_Gringo_Grande

What does that conversation say about Americas bread basket?

They don't listen real closely to either side?

[bkelly]

Hey! Hope you all had a good Thanksgiving.

Well, it seems that this thread isn't very active when I'm not putting out my thoughts. I'm happy to offer my perspective to people who want to hear a side closer to the scientists but I don't care to climb upon a soap box. I'll finish up this discussion about paeloclimatology and then talk about the land based temperature data since there have been questions.

But I'll still be around. If you'd like my perspective on something else, just ask.

bkelly

[bkelly]

Originally Posted by xTYBALTx

...the debate should be other methodological differences, data source differentials, and the dozen other sources of inconsistency amongst our sources. This will help determine which of the dozen sources are better and worse, and will help us get a more accurate picture of past data.

But there's a problem. The original graph I posted of twelve different climate recreations was a simplification. These studies are not able to pinpoint the average global temperature 400 years ago but instead offers a probability curve of any given year for a given temperature resolution. The lines on the graph are simply the maximum on those probability curves.

For example, one study's line might be at -.25 for the year 1400 but it isn't actually that definitive. If we looked at the data, we could find it's results only give a 30% chance that the temperature was between -.2 and -.3. The chances of the temperature being between -.24 and -.26 would probably be around 4%.

So your suggestion of finding one study in which to believe for a more accurate picture of past temperatures is backward. Any one of these studies alone will give you LESS certainty of past temperatures. Only when they're taken together can we say with 80% confidence that the average global temperature was -.3 to -.4 degrees colder in 1425-1450 than the mean global temperature from 1961-1990.

So, if you want to know exactly what the past millennium's temperatures were, you want an infinite number of these studies, not one.

Originally Posted by xTYBALTx

If all sources were equal, they'd all show the same result. Clearly they do not show the same results - some must necessarily be better than others.

You're speaking of objective truth and I agree with you in principle. Some of the studies must be closer to objective reality than others.

Now back to the real world where us flawed humans are prisoner to our unreliable senses and preconceptions. You are never going to see objective truth because there will always be unknowns and errors and therefore doubt. You're asking for the "right" answer to historical temperature recreations but there never will be one. Your request for more accuracy is a bottomless pit because no matter how deep we go, there will always be uncertainty.

As an example, consider the current global temperature estimates that have been thrown around this thread. These are created with thousands of direct samples (using a thermometer) instead of proxy samples (tree rings, ice cores). Earlier you said that these estimates "would be subject to the criticisms Lucky posted earlier, regarding the positioning of the instruments and how that positioning can affect collected temperatures."

So, in other words, you're skeptical of global temperature estimates made with direct readings today! How is any additional precision going to allay your skepticism of historical temperature recreations?

bkelly

[El_Gringo_Grande]

Originally Posted by Kerostasis

They don't listen real closely to either side?

Yes. Actually they don't listen at all to either side.

Geary England is a world famous Oklahoma meteorologist. (He had a role in the movie twister. Us Okies have to take our fame where we can get it! )

He is probably the most level headed meteorologist I have ever seen. During tornado times he will come on and point out what is going on and tell you to take your precautions. The other guys around here will claim the world is ending and go on forever about how bad it is.

Anyway he has mentioned global warming a couple of times in passing. He seems to be saying that there is probably something to it. But that is all he says.

Rush... Well I don't know what Rush says but I do listen to him occasionally and he is basically a shock jock of the conservative political set. I can't see anything he says being informative in any way. Entertaining, yes. Educational, no.

But I have a feeling that most people get their information from one of the above sources. The knowledgeable persons sound bite that imparts almost no information or the ramblings of a ideological demagogue.

Kinda like the whole evolution vs creationism/intelligent design. You have a whole lot of noise going around but in reality there is a bunch, and I mean a bunch, of sound scientific knowledge that never is consumed by the general public.

Unfortunately most people base their decisions on the noise.

[bkelly]

Originally Posted by xTYBALTx

And what's up with the land monitor based temperature data?

In case you're new to this thread or have a memory almost as bad as mine, Tybalt is asking about the century old networks of ground-based temperature monitoring stations like the ones managed by the United States Historical Climatology Network. These are important because past and present data from these stations are used to find global temperature anomalies in studies such as NASA's GISTEMP program. Before his departure, Lucky Shot questioned the accuracy of these temperature stations because many are in areas where they can be affected by an "urban heat island" effect or artificial heat sources. As evidence, he cited a power point presentation of surfacestations.org, a group working to audit these temperature stations, that showed 85% of a sample set of stations should be expected to give readings at least a degree high. He also gave links to videos of Bob Carter from James Cook University who mentions these stations and shows pictures from surfacestations.org.


So here's what's up Tybalt- I agree with Lucky's points. The numbers may not be spot on but in general he is correct that a lot of these stations read inflated temperatures because they're over asphalt and/or near artificial heat sources.

Now, here's three reasons why Lucky's arguments don't matter for the global warming debate.


1) Many of these stations are reading hot and this is a serious problem if we want to know today's average temperature. However evidence for global warming doesn't need a quantitative measurement of the temperature, only an accurate measurement of the change in temperature over the years.

For example, consider a station that was installed 100 years ago over asphalt causing it to read 5 degrees hotter than it actually is. After a year of readings and some math, it's determined that the average temperature of 1907 at that location was 24 degrees. Now, we know that number is inaccurate. Regardless, we take readings from the same station over the same asphalt today and calculate an average reading of 25 degrees we can state with confidence that 2007 is a degree warmer than 1907 despite the measurements being 5 degrees off.

It's consistency that's important for global warming estimates, not absolute temperature measurements. This is why one station's readings are never directly compared to another, only their deviation from the mean temperature of a specific period. So even though some of these stations are in artificially warm environments, they can still yield excellent data on long-term climate change provided the skew in their readings are stable.


2) You see what's coming next. "But the skew in the readings aren't stable! Cities sprawl into these station's environment. Asphalt is put down right next to these boxes so the readings go up. Bob Carter showed a picture of air conditioners that were feet from these stations and we know those weren't there for a hundred years!"

Again, you're right. Climatologists know that readings from these stations depend not only on the environment but also on the instruments, the shelters, the time of observations and the means of which the data is averaged. Things change and since any change can affect the results, a great deal of thought has gone into the best ways to handle these situations. In general the data is published "as is" with the onus on the consumer to account for these discrepancies so he or she may decides what is best. I'll describe the method used by the National Climatic Data Center since the graphs from GISTEMP that I provided use their information.

To start, a reference series is created for each year of each station that is assumed to be an accurate model of the climate for that area. The reference series is created from three sources, a highly correlated neighboring station, that station's previous years readings tempered by the five most correlated neighboring stations and an average of the readings form the central three the closest five stations. The actual data is then compared to the reference series and if a potential discontinuity is found, more analysis is done. If there is 95% certainty that a "change point" was found, temperatures from that point on are adjusted on a monthly basis which is then smoothed over the year. An overview of this process is detailed here.

So when Bob Carter's slides show a jump in temperature data when a station or its environment changes, it's not the forehead smacking, "my goodness, what were we thinking?" moment that he implies it to be. It's known, it's expected and it's well accounted for.


3) In the end, none of this actually matters. When confronted a global warming skeptic will often point out the problems with these stations, but it's only FUD and he or she will dance around the topic without saying anything concrete. It's no wonder because the argument can be quickly identified as bullpatties when in plain sight. So let me say what Lucky Shot didn't:

The temperature increase of the last few decades are illusion created by increased readings from these monitoring stations due to modifications of the land around the sensors including the installment of new heat sources.

Wow, that makes a lot of sense. Of course asphalt and air conditioners are going to raise the temperatures of these stations so it's going to be hotter in areas that cities are expanding. And this would be easy to prove! All we have to do is show that the temperatures are going up in and around cities. So find a global map showing temperature anomalies and...


Click for a larger version
source: http://data.giss.nasa.gov/gistemp/gr...+2005+1998.pdf

Oh crap! We're not seeing the most warming in populated areas but in the north and south poles and Russia, a country with a low population density. If global warming is only due to thermometers being close to roads and air conditioners, why are we not seeing it in areas with people, roads or air conditioners?

And by the way, if you're skeptical about the adjustments I described in step 2, notice what they say in the above maps, that the US was about the same or cooler in 2007 as compared to 1998 and cooler still in 2005. Our cities and air conditioner usage has grown over those nine years, yet the adjusted numbers don't show a temperature increase.


Whew, I need some lunch.

bkelly

[GlobalWarmin]

Ominous Arctic Melt Worries Experts

http://www.livescience.com/environme...ctic-melt.html

[CingularDuality]

Ominous. That's quite a descriptive word.