Wednesday, October 22, 2014

Climate Change Modeling Has Some Problems. History Might Help!

On September 19, Steven E. Koonin wrote an article for the Wall Street Journal titled "Climate Science Is Not Settled" .  Although there are many interesting observations in the article, what caught my attention was the topic of Climate Models. http://online.wsj.com/articles/climate-science-is-not-settled-1411143565?mod=WSJ_hp_RightTopStories   The basic message is that all of the modeling that has been done cannot accurately predict our current climate results.  Here are a couple of quotes from the article:

"As a result, the models give widely varying descriptions of the climate's inner workings. Since they disagree so markedly, no more than one of them can be right."

"Although the Earth's average surface temperature rose sharply by 0.9 degree Fahrenheit during the last quarter of the 20th century, it has increased much more slowly for the past 16 years, even as the human contribution to atmospheric carbon dioxide has risen by some 25%. This surprising fact demonstrates directly that natural influences and variability are powerful enough to counteract the present warming influence exerted by human activity."


Models are ways of predicting a future result which in this case is the Average Annual Temperature of Earth.  For most of us and in our experience, the range of highest to lowest temperatures seems to occur within a year, which captures all of the seasons.  So no surprise, that modelers are focused on the Average Annual Temperature and all the "influencers" that occur within a year.  By studying many years, say 200 or so, you could get a good idea if this Annual Average Temperature predicted by the models, is accurate.  (Useful worldwide temperatures date back only to 1880, hence the 200 year history.)

If you wanted to predict (or even measure) the Average Annual Temperature on Earth, clearly you would not gather climate data only for March and September!  You would be missing some important information.  Likewise, as we take the Average Annual Temperature and find that it is increasing recently and setting records, as compared to 1880, MIGHT WE BE MISSING SOMETHING BY NOT LOOKING AT OTHER CENTURIES OR MILLENNIUMS??  Might there be other temperature cycles we need to understand before we can declare that "these are the highest temperatures we have ever seen, and, therefore, caused by human influences?"  The operative word here is "we".  Could there be some other temperature cycles beyond those that occur within a year which could be affecting the climate models?  Should we instead be looking at the Average Century Temperature for instance?

As many of my posts suggest, there is a great deal to be learned from studying the history of any problem, so I will try that approach again.  The history of worldwide climate data has all been within the last 200 years.  However, there is some data that goes back 800,000 years and this amount of history might give us some insight to the "cycles" of both temperature and the corresponding CO2 levels, which are often in the news.  This data comes from ice core samples, 2 sites in Antarctica are titled EPICA and Vostok (temperature data is listed a degrees centigrade variation from present); and one Arctic site called GISP2 in Greenland (temperature is degrees centigrade).  The process and science of these measurements can be read at  http://www.climatedata.info/Proxy/Proxy/icecores.html but I will focus on the data itself which can also be downloaded from this site.  But to be clear, this is very localized data and does not reflect the Global Temperature.  However, it is the relative changes over time at this location which can help us understand if there are any longer term "temperature cycles" that are affecting the globe such as the Milankovitch Cycles.

Here are the key points that are supported by the graphs and analysis which follow:

  1. Earth is getting warmer!  It is supposed to be getting warmer since we have been in an interglacial warming period that began about 17,000 years ago.  This warming began at the end of a long glacial period and when combined with the interglacial warming period, lasts about 100,000 years.  These cycles have been repeated for at least 800,000 years.  There is a longer term temperature affect in play beyond the 4 seasons in a year.
  2. If you look at the previous interglacial warming period 127,000 years ago, we have not yet reached the previous high of 4.84 degrees.  100 years ago we were 1.8 degrees hotter than we are today, because we are at "0", the base line for all these temperature measurements.  We are not yet as warm as we have been in previous temperature cycles.
  3. The most recent interglacial temperature rise has taken longer than the previous two, but not yet reached their highs.  The pattern and length of our recent interglacial rise, however, looks very similar to that of 422,000 years ago.  This also holds true for CO2 levels.  Using this as the benchmark, we could have 11,000 more years of warming before hitting the historical highs.  This might yet be another even longer temperature cycle we need to understand.
  4. Looking at the last 50,000 years of temperatures, which begins in the middle of the last glacial period, you see a very rapid rise in temperature about 10,000 years ago.  Since this time, the temperature has been fairly stable, until the last 600 years, when the temperatures dropped to a new, lower stable level.  This is the period of highest human activity.
  5. Using only the last 50 or 200 years is not nearly enough data to understand the amount and sources of the earth's changing temperature.  Without this understanding, we cannot model or assess the impact of the human contribution to Global Warming.  We need utilize and model at least 422,000 years of history.  If we have not yet hit the historical highs from all these years ago, how do we conclude that humans are to blame?  Are we the cause of global warming, or rather are WE to adapt to the predictable warming yet to come as our ancient forefathers have done???


ANALYSIS AND GRAPHS

The EPICA ice is the deepest, so it yields the greatest amount of history going back 800,000 years.  Below you will find a control chart (X, MR) of the Temperature data.  The upper chart is the actual temperatures (tracked as plus/minus from the current temperature) and the lower chart is a measure of the variation in these numbers.


The first 400,000 years show a different pattern with a range of temperatures from 3.15 to -9.63 than the last 400,000 years, with a temperature range of 4.84 to -10.58.  You will also see a change in the patterns in the second half of more distinct and rapid rise in temperatures (interglacial periods) and longer lower temperatures (glacial periods).  This cycle of glacial and interglacial periods has also been increasing from 74,000 to 113,000 over past 3 cycles or about 20% increase each cycle.  The most recent cycle is not yet over and is already 130,000 years.  Also notice that the variation in the temperatures is also increasing, which for the modeler, creates issues in creating accurate forecasts.  Below you will see the similarities to the corresponding CO2 levels.  Notice however, that there is not a large increase in the CO2 variation.



Now, comparing the first 400,000 years to the more recent 400,000 years can help us determine if the change in cycle patterns and ranges results in a statistical difference for temperature or CO2.



In the last 400,000 years, temperature dropped .49 degrees or 9.7% while CO2 actually rose 0.9%.  Here is another anomaly for modeling CO2 and temperature.  The pattern of the most recent temperature rise is quite different from the previous 2 interglacial rises which are quite steep and short.  However, the interglacial rise 422,000 years ago looks very similar to the most recent rise, and needs a closer look.

The first two interglacial periods below have the same average temperature and CO2's but the second one is less than a third the length of the first, 28,600 years vs 8,000.


For the third and fourth interglacial periods, the average temperature has risen, the CO2 levels are the same, but the length of time is about the same, 8,200 years and 8,000.

Finally, comparing the most recent interglacial to the similar one 422,000 years earlier, the average temperatures and CO2 levels are the same, but has not yet reached the highest temperature or CO2's of the past.  Also, the length of the most recent period is 17,000 years.  Could we be repeating the interglacial warming pattern of 422,000 years ago?  Could it take 11,000 more years to reach our interglacial high and only then know if "we" caused it?


The next several graphs of EPICA data will focus on the last 422,000 years, which actually cover the same time frame as the Vostok data.  To show how similar both the data sources are, the next 4 control charts (X,MR) will show the EPICA and Vostok temperatures followed by the CO2 levels.

EPICA and Vostok Temperature
Patterns very similar

EPICA and Vostok CO2
Patterns very similar

The next analysis will show the changes in average temperatures over  each of these 4 glacial/interglacial cycles using EPICA data.  I have done the same analysis on the EPICA CO2 data as well as the Vostok Temperature and CO2 data.  To prevent graphic overload, I will summarize all of this in a table at the end of this section.

The cycles here are defined as the maximum interglacial temperature to the next maximum interglacial temperature.  Each figure contains the X Chart, a histogram, the F and t statistic and a summary of the two cycle's data ("before" and "after").  The blue highlight in the lower left corner signifies a statistical difference either in the averages or the variations between the two cycles.  For the first two cycles, the average temperature dropped .46 degrees or 9%.


From the second to third cycles, the average temperature dropped .48 degrees or 8.8%.

From the third cycle to the fourth cycle (most recent), the average temperature rose .6 degrees or 10%.

However, comparing the first cycle to the fourth, the average temperatures are NOT statistically different!  This coupled with the 17,000 year interglacial temperature rise pattern studied earlier, leads me to conclude that any modeling needs to consider at least 422,000 years of history to establish all of the independent variables that could impact temperature changes.  Since good climate databases go back only a couple of hundred years, this could partially explain the problem with the current climate models.

Note below that the Vostok temperature for the first and last cycles, shows a difference when the EPICA data shows the temperature to be the same.  This is further evidence that the current interglacial warming may not be over.


Looking at the most recent interglacial rise of the last 17,000 years using another ice core database might help reinforce the point of needing more history to get accurate models.  The GISP2 data includes 50,000 years of temperature data at a much more granular level, about every year.  This control chart (X, MR) shows that there has been a dramatic shift up in temperature about 10,000 years ago and has been pretty steady since.  So, we need to focus on the significant event of 10,000 years ago!  It sure looks like this is just a part of the interglacial warming that began 17,000 years ago.  But, lets look closer at these last 10,000 years to determine if the human affect of global warming can be seen.


By zooming in on these last 10,000 years, which appeared to be fairly stable, shows even more information.  In the last 900 years the temperature dropped 1.6 degrees from earliest average.  But, over the last 300 years, the temperature is rising again, but has not yet returned to the earlier highs.  So again, our issue is not what happened in the last 300 years, but what happened 10,000 years ago.  Our current models will just not help with this question!  To confirm that recent human industrialization is the cause of "global warming", we would need to see global temperatures statistically higher than we have seen in the last 422,000 years!

In conclusion, to show the human affect of global warming, we need to see CO2 and temperature levels that have not been seen in human history.  That has not yet happened!  Modeling needs to include more of our earth's history in order to forecast when, and how high, our current interglacial warming period will go.  Only then can we know if, or how much, humans have impacted this interglacial cycle.