Its not warming: data vs. meme

>>Is the El Nino considered a global event or just a local USA event on temperature? Sounds like global in nature. Huge differences in US climate this winter.

>>Regarding the Paris "agreement" to try to hold to 1.5 deg C increase (stretch goal I know) how much is considered to have already happened?

>>As far as your analysis thank you but I do not personally have enough expertise to argue the point. I think you are showing the data in a different way than the "T-rise hiatus" analysts folks. If you have put your finger on a fatal scientific flaw of that on-going argument, congratulations are in order.

 

ENSO (el nino/La Nina) is an Pacific Ocean Osculation that has repercussions in global weather.

2015 will be around 1 degree C above pre-industrial levels
if you subscribe to the consensus sensitivity of 3 then there is already enough ghg to produce a 1.54 degree C temperature increase (3 C x log 2(400ppm/280ppm))
Which means either the sensitivity consensus from IPCC is wrong, or we are at maximum ghg in the atmosphere for that 1.5 C goal. The pledges so far promise that there will be over 2 degrees of warming.

The hockey stick graph in 2001 IPCC publication cherry picked the 1998 excepionally hot year for the time as the end point, and hottest year so far. It came from papers, but there really was no excuse for going from averaged data to individual years of data. That gave a very misleading expectation that temperatures would continue to rise as fast as that very fast blade.

Picking that same year from the chart's end point, there has been an hiatus, if we are looking just at individual years. IMHO this is equally misleading. If you use a moving average instead, there is a decrease warming rate, but the world has been warming, as expected. The problem comes in defense of the hockey stick. If you embrace the bad statistically methods used there, then you can use similar bad statistical methods to make the similar claim of a pause.

 

The difference between 1degC and 1.54degC is due to thermal lag. Think of it as thermal inertia. It is similar to the case where you set a burner on high on a stove but it still takes some time for the water in the tea kettle to boil.

Bill the Engineer

 

They're kissing cousins?
Sorry, couldn't resist!

 

...the main thing the "hiatus" observers are saying is at some point lets say 2020 after the El Nino impact is over that the rate of increase needs to not continue to decline, and still have the models still be correct.

That much I understand because I have used models all my life, and made some of my own models, so I am very sensitive to using models to try to predict God's natural order of behavior. Often times in my career I got obtained data that conflicted with the models. I am full of bullet holes, believe me. I am a data believer vs. model believer...there are 2 types of people..it turns out.

I do believe CO2 is going up ...that's data. I do believe CO2 is a greenhouse gas.

 

Well that is one possible hypothesis, but ... for it to be tested we need more detail. How many years is the delay? What is the transfer function? What parts are natural variation?

Since models are not nailing these down too well yet, we also must assume the null hypotesis is possible, sensitivity is not 3. Another decade or two should help tell the ghg climate signal from the weather noise. T Here may be other climate signals in addition too ghg.

 

Dealing with 'noisy data' is as much an art as applied statistical analysis. I deal with network data and the averages are less important than the saturation events that cause noticeable impacts on users. Averages have a place but they often suppress the peaks and valleys. The temperature records have a similar characteristics:

• too cold - crops can fail and critters die
  
• too hot - crops can fail and critters die

So I find it more enlightening to plot the local maximums and local minimums. Unlike the average, these are the extreme events that bound whatever is being measured and have effects:

Trends of peaks and valleys parallels the average giving a cleaner, trend.

The deniers put their eggs on 1998 and 2015 is blowing it away. I have no doubt that with a few cooler cycles, 2015 will be the new anchor for the next 'pause' claim. But deniers have strange understandings of trends:


I'm also not a great fan of using averages versus a Gausian weighted average that applies a stronger score to the middle values and weaker as the data points move away from the middle. It has the advantage of preserving the peaks and valleys. In contrast, a linear average that mutes the peaks and valleys.

In about 30 days, we'll have access to Berkeley temperature, sea level, and earth rotational data. I'll continue my earlier analysis and we'll see how another year has added to the planet changes.

Bob Wilson

 

Heres the correct global temp from UAH

 

Could you show your work. I'm not sure what you are doing. Are you doing a least mean square fit to individual years? Averaged years? Something else?

Correlation should be shown to S x log2([curr ghg]/[past ghg])
The graphs in question used 1902 - 1998. They drew a linear regression line to smoothed data to 1902 then actual yearly data un smoothed.

For ghg levels this should be a good source.
http://data.giss.nasa.gov/modelforce/ghgases/Fig1A.ext.txt

Data should have been smoothed to include past data. Outliers thrown out if they distort the data. Of course the given charts exagerated the impact, and show a much higher slope at the end compared to what I'm guessing you are doing.

Fifth IPCC assesment says that slope smaller for the recent decade, which agrees with my data. I believe you are using some kind of naive smoothing, which exagerates pre 1900 temperatures. More data will improve this, which is why I think you are seeing an increase. A least mean square single sloped line is also not very good.

One poster suggested a time delay. ghg concentration in 1850 - 285.2, 1910 - 300.1, 1998 - 366.27, 2015 - 400

If we were 100% correlated with ghg concentration in the given year we should see 15% of the ghg warming in the period 1850-1910, 59% between 1910 and 1998, and 26% from 1998-today. Feel free to add ENSO changes and time delays for your ghg function.

 

Follow the red shaft of the linear regression, then the large shaft. If this is a good graph (I would say its misleading without a lot of explanation, and they did explain, but it mostly is glossed over in the summary) then the 1998 to 2014 graph is also appropriate, and IMHO misleading. We have smoothed data in this graph then in 1902-1998 yearly data plotted. Why isn't it smoothed? Why a shaft in red to the instrument record. It is meant to focus on the bottom of the slaft (the drawn line) to the top of the blade (1998). The blade did not continue though, as many at the time of publication argued was likely.

I think that is what this 2001 graph was meant to show.

I was asking what graph you are using. Is it individual years, which I believe are misleading when added to paleo. If smoothed by what method. Yes I have seen numberous. I haven't seen one that claimed warming was faster including 1999-2014 than leaving it out. That is why I asked.

I suppose if you start in 1850 and only draw a single line, you may be there.

 

Sure, easy to answer your questions. The graph I posted, which is the one I remembered, and the first one that came up in google was figure 2.20 from the IPCC in 2001. If you right click to just open the figure, or google yourself you will see some features. There is a dashed line drawn in to form a shaft of a "hockey stick". Its labeled linear trend 1000-1900. There is a dashed line not on the legend but labeled 1998 instrument record.

The paleo record is naturally smoothed, as each of the proxies is smoothed or moved in time. I believe in AR5, the IPCC with a better paleo record did not like talking about periods of less than 30 years. Why go further in the smoothing by adding a linear dashed line to pretend noisy data was not noisy? Why not smooth the instrument record by 30 years to work with the paleo record, or 10 years, or at least 3 years which was common with hadcrut at the time? Was it to call out 1998 as they did with the extra graph line? Of course it was. This is not my private beef, these were criticisms at the time of the chart.

Which means, if you like this chart figure 2.20 from ipcc 2001. Then those charts with the 1998-2013 with a line showing no warming are also valid. My point is not to defend these later charts, its to denounce all these misleading charts.

The warmists are pushing to include the single year 2015 even before all the data is in. The skeptics seem to want to cherry pick the same date that the IPCC cherry picked in 2001, 1998. That does not seem skeptical but cherry picking and misleading. Now my memory is not perfect, that fig 2.20 did include 1999, you simply can not see it, but its in the legend, nor is it called out by its own line that draws extra attention to it.

Why not smooth at least 10 years in these charts of temperatures if we are using them to look at ghg warming. The noise in a 10 year period is highly susceptible to things like ENSO according to IPCC ar5. Now if we are using the temperature record to model ENSO and not green house gas warming, then sure 1998 is a very important year.

 

Because it removes the peak and valleys:

• < 1 year - shows seasonal variations, why we don't set our hair on fire for 'hottest month' nonsense
  
• 1 year - moderates the seasonal effects
• > 1 year - excellent approach to hide peaks and valleys
• > 10 years - hide them better
• > 100 years - might as well not even discuss it
• > 1000 years - perfect for orbital mechanics
• > 10000 years - Ok, we might see the ice ages

It is called 'framing the question' and I'm OK if others choose to leave the discussion by using a math trick to make it go away.

In about a month, the middle of February, I can resume add data through June of 2015 to my earlier plots. This will barely include any of the current El Nino data. For that, popcorn is advised.

Bob Wilson

 

Phil Jones refered to this grafting of unlike time frequency and extention as Mike's Nature trick, Mike is Micheal Mann. He first wrote against it, but decided to include it in the IPCC 2001 report.

The statistical trick is to not average when the other parts of the graph are averaged. This represents unlike things to make the data seem stronger than it is. Mann also said in now disclosed conversations that he thought his results were portrayed as more firm than they were. The paleo record diverged from the temperature record, so temperature record is substituted. Biaffra's data was also transformed using this mathematical technique by jones in ipcc 2001, so that it would look more like the mann data.

Certainly in marketing such statistical methods are used to exagerate actual differences. These are quite legal, but do we want our science treated like corporations treat claims.

Certainly if we are being unscientific and doing these tricks to exagerate, then the pause is fair game. It too uses statistical techniques of individual years and variation to show no statistically significant warming. If we force both to use proper techniques there is warming, but if we allow this foolish single year religious discussion you can't come to any scientific clarity, only competing advertising claims.

Sure, but individual years are pretty meaningless when it comes to ghg warming, as has been repeated by scientists that actually care about the science. For those with a political bent, yes 2015 is likely to score points for the warmists, as they have ignored the slower warming data.

To me what is the point of these individual years when we know there is multidecadal oscilations that have greater variance than ghg warming in these single years? If you are doing some proper multi year averaging, then yes adding 2015 will continue the trend, as 2014 did.

Are we so starved for news stories that we should forget the science. Why not look at 11 year averages or at leat 5 year averages. Lets go small. When 2015 comes in you can add the data point for 2013 5 year average. Better yet for looking at the trend would be 11 year (half solar cycle, which is key in natural variation) which will add the 2010 data point. Do we really need to guess and add false detail and add 2015 individual year?

 

Your points are valid if one is trying to stitch together modern and paleorecords which also includes ice cores. The accuracy and coverage significantly degrades as one uses non-instrument, proxies. Just I prefer modern instrument era data. The story here is quite compelling.

I enjoy the modern temperature records but am fascinated by sea level and earth rotation. We're air breathing critters and tend to notice air temperature. In contrast, the sea and earth rotation involve masses whose behavior as the surface warms is interesting.

BTW, I had never really looked at the Antarctic ice cover but this year promises to be entertaining. I never realized the extent the sea ice disappears in the Southern 'summer.'

Bob Wilson

 

That is the graph that I have a problem with in that old report. The good news is in AR5 the current IPCC report, they finally appear to have taken the criticism to heart and have corrected the misleading nature of it all. That is how science should work, where mistakes are corrected when data comes out. The process simply was slowed because of politics.

As to strictly temperature record, it is all about what you are trying to do with it. If you simply want yearly data, well that is what you should take. If you are trying to reflect ghg to temerature, in a white box type model, you should either account for the solar cycle and oscilations, or do some sort of smoothing to remove them. Definitely if you are trying to figure out solar cycle, seasonal graphs may be the best expression.

Here seasonal records probably are best. What are the hottest 3 months of the year for the region, what are the coldest 3 months, and use those, instead of global yearly.

Antartic ice seems to reflect not just land and sea surface temperatures, but underwater geothermal heating seem to have a big impact.