Monday, August 11, 2014

Falsified AGW

Anti Persistency of Monthly Global Surface Temperature Series Falsifies Positive Feedback, which Falsifies AGW
Andre Bijkerk
Falsifying catastrophic antropogene global warming requires to demonstrate that the climate sensitivity for doubling CO2 is less than the lower value that is considered AGW. The IPCC generally defines this lower value to 1.5 degrees Kelvin per doubling CO2. However assuming the generally accepted 3.7-4 w/m^2 IR energy increase for doubling CO2, a value of less than one degree Kelvin can be calculated, when substituting this in the temperature differentiated Stefan -Boltzmann equation (F= σT4 ). However models generally assume a value of 1.2K for an equilibrium value. Hence to obtain any value higher than 1,2K the climate system must be governed bydistinct positive feedback mechanisms. Consequently, disproving AGW can be done by demonstrating a lack of positive feedback in the system, which is required to push the general climate sensitivity to more than 1-1.2K per doubling CO2
This has been done before, for instance: Kärner 2002, Kärner 2005, Kärner 2007. Here it will be demonstrated that for instance the HadCRUT4 global surface temperature series show characteristics of negative feedback.
Feedback is well known in system control engineering as well as its specific traits. Apart from the mentioned amplification effects another major element in feedback is persistence or anti-persistence.
If a system output deviates from the average/norm/ etc, then positive feedback tends to enlarge that deviation, while negative feedback tends to reduce it. Because it is feedback, essentially this happens after a certain time lag for the system to process the inputs, forcing and feedback signals. That is crucial. So the output of a system is what the system normally would have produced plus or minus the feedback factor of the previous output one time lag unit ago.
This is demonstrated in the excel sheet tab labeled "feedback demo" in this spreadsheet:
HadCRUT reversals.xlsx
HadCRUT reversals.xlsx
Shared with DropboxView on www.dropbox.comPreview by Yahoo
This demonstrates what basic proportional positive and negative feedback do to a simple random walk down in the C-column in a most elementary way. The strength of the feedback factor (typical range 0-1) can be adjusted in cell C2.
View on dl.dropboxuserconte...Preview by Yahoo
Fig 1 . Demonstration of positive (red) and negative feedback (blue) on a certain basic forcing signal (black) with a random monte carlo simulation (one dimensional random walk). Open circles denote reversals of the signal direction, demonstrating general persistency in direction with positive feedback and general anti-persistency in direction with negative feedback.
The graph shows the original random walk in black, and what would happen with positive feedback in red, and negative feedback in blue. Two things to note--first, the red graph is always fluctuating the highest as it has more gain and steers the system away from the center value . Alternately the blue negative feedback graph has the smallest absolute value all the time as it steers the signal back to the average value of zero here. The second point is that the red graph is smoother overall as it resists letting the signal return to lesser values, the feedback keeps pushing it away. Consequently, it has less reversals of direction, which are indicated here with the open circles. The blue line of negative feedback tries to reverse back to the center all the time and consequently it has clearly more reversals than the red line of positive feedback, steering away from the average value.
Obviously these reversals can be counted and compare to the original signal. If there are clearly more than 50% reversals the system would undergo negative feedback and with significant less than 50% reversals, it is positive feedback. Due to the chaotic nature, values close to 50% are hard to judge on weak positive or negative feedback, depending on the statistical significance. Note that using a feedback factor of 0.5 (positive) resp -0,5 (negative) in this demonstrator would lead to a positive gain of about 1,5 with about 38-40% reversals resp for the negative feedback, a gain of about 0,5 with about 60-62% reversals.
In exactly the same way, the data of a long series, for instance monthly global temperatures can be analyzed on persistency behavior, which would indicate the nature of feedback, if any.
To check that for monthly global temperature series, the longest series, HadCRUT04 monthly global temperature data from 1850 till now (June 2014) was analysed in the tab "HadCRUT reversals" of the spreadsheet. It counted 1192 reversals and 780 non-reversals as shown in cells D3 and D4..Hence a reversal rate of 60,4%. So the series demonstrates anti-persistency on a monthly basis. Consequently, the feedback on the global temperature system is predominantly negative under the sum of all negative feedback with approximately a monthly lag.
Then the question raises if there would be a difference in reversal rate from before and after the the start of the increase of CO2, which could indicate a difference in feedback once CO2 started to change significantly. To address this, the counts have been split to before (1850-1949) and after (1950-2014) the start of the CO2 enrichment.
The reversal rate with the monthly time constant is 59.8% before the CO2 and 61.5% afterwards. This suggests that the monthly feedback became more negative when the CO2 started to rise.
Next other time constants are analysed, ie, bi-monthly, quarterly, semi annual and annual, averaging any possible combination starting with each month. For instance for bimonthly, there are two datasets, for odd and even months, while for yearly results there are 12 sets, each beginning with a different month. It's all shown in the spreadsheet.
The result is in the box in the sheet labeled "HadCRUT reversals".
monthlybi-monthlyquarterlysemi annualannualReversals 1850-195059,8%57,3%58,9%56,2%57,8%Reversals 1951-201461,5%51,6%50,4%53,4%58,6%
This shows that all values are greater than 50%. Consequently there is no evidence of persistent behavior hence no evidence of distinct positive feedback on any of those time constants. It does show, however, that the persistency increased on a bi-monthly, quarterly and semi-annual basis during the CO2 enrichment period. Whereas on the annual basis the persistence acts again as on the monthly basis, becoming slightly stronger in the CO2 enrichment period.
Notice that this does not say anything about the nature of causes of positive or negative feedback, be it variations in the sun, variations in humidity, cloud cover, ice melting, ocean signals, volcanoes, polution, aircraft contrails, El Nino, anything. It merely shows if the reaction of the system is persistent or anti-persistent to (forcing) factor variation. It's not about what causes the variation, rather, it's about how the system reacts to those.
Since there is no reversal rate anywhere below 50% it is evident that the temperature series is not at any point driven by a sufficient strong positive feedback signal that would be required to create catastrophic global warming. Also it shows that intermediate time constants of months the persistence became less negative when the CO2 started to rise as af 1950. However both the monthly and annual time intervals negate this, showing a weaker persistence in the CO2 enrichment period. Consequently without evidence of significant persistence, the required strong positive feedback for catastrophic global warming does not exist. Therefore the actual climate sensitivity per doubling CO2 has got to be about the basic value ranging from 1K to 1,2K. So the minimum value of the IPCC, 1.5K, can be considered falsified.
Dataset HadCRUT4
Kärner O, 2002; On non-stationarity and anti-persistency in global temperature series. Journal of Geophysic Research Volume 107, Issue D20, pages ACL 1-1–ACL 1-11, 27 October 2002
Kärner O, 2005; Some examples of negative feedback in the Earth climate system. Central European Journal of Physics vol 3 No 2, pp 190-208
Kärner O, 2007 On a possibility of estimating the feedback sign of the earth climate system. Proc. Estonian Acad. Sci. Eng., 13, 260-268


Let's take a look at the global surface temperature average since 1850:

Temperature since 1850.gif
Source: Met Office via Corporate Responsibility

Now, take a look at the CO2 emissions for the same period:

Graph of Global Carbon dioxide Emissions, 1850 to 2009.
Source: NASA Earth Observatory
We can see many things here, but one of them is that there was a prolonged period of warming in the 1800s (about 1860 - 1880) that correlated to minimum amount of CO2 emissions. In other words, there was a natural process going on during this period that caused the climate to warm. This is not controversial or new, although the contrarians would like to make the case that it is. There is natural variability in the climate. We know this because the climate scientists have done the hard work to discover it.

What you have done is to compare the period of low-CO2 emissions to the period with high-CO2 emissions to see if there was a relationship between temperature change and CO2 emissions. However, your procedures were invalid.

We can see by comparing the two graphs above that, in fact, there was a considerable amount of climate change between 1860 and 1880, even though CO2 emissions were very low during this period. What that means is that there was something else causing climate change during this period. Then, we can see there was a great deal of climate change between 1950 and 2006 (end of data on graph) and there was also a great deal of CO2 emissions. So, what you thought you were doing is changing one variable versus another (CO2 levels versus temperature), but in fact, you were changing at least three variables simultaneously - CO2 levels, temperature and what ever it was that caused the climate change between 1910 and 1940. You wanted a dependent variable (temperature) against an independent variable (time) and looking at two different periods where you assumed only one thing changed (CO2 levels). But, this is not the case.

Obviously, something was going on in the 1800s and you did not do anything to isolate or identify that cause and simply assumed that what ever it was that caused that change back then is not relevant to the calculations. You have, at least, a second dependent variable in the data that is not accounted for. If climate scientists did this contrarians would be all over them, as they should. Having a second dependent variable invalidates the experiment and this is a pretty typical exercise in most introductory labs.

Let's look at the price of gasoline as an example. You can plot the cost of gas at the pump versus time and compare it to the regulatory and tax situation in some period in the past and some period more recently. By doing this you can conclude that the cost of gas has been rising because of an increase in regulatory oversight and greater taxes. Both of which are valid observations. But, there is also the vastly increased cost involved with having to go to greater and greater extremes to find the crude oil. So, if we were to compare the cost of gas to regulatory periods we would certainly get a plot. But, that plot would not include the fact that it costs a lot more to get crude oil today than it use to. Plotting data always gives a plot, but that doesn't mean it means anything.

Until such a time as you identify the second dependent variable and take it into account, your calculations make no sense and prove absolutely nothing.

You did not prove man made global warming is not real.

No comments:

Post a Comment