National Geographic "The War On Science"

From the link

This is called back testing. The ability to back test your model with old data to calibrate perimeters is a great way to make your model look good on all the old data. In order to be selected as an IPCC these models had to have parameters set by back testing. Which makes the historic test worthless when evaluating models, they pass by criteria of how they were built. Image a stock market model built in 1928 that back tested well in 1928, would it predict the market well? That sites statement is non-sense. The models have not done well.

Past results of back tested models do NOT ensure future results!

What makes a model a good model is if it accurately predicts the future. The models selected in IPCC IV did a poor job of accurately predicting temperatures and precipitation in the period between their back testing period and actual data. What IPCC V judged was that they poorly modeled natural variation in these things, so there is no way to tell if they accurately modeled temperature sensitivity to green house gasses. Specifically the IPCC thought that ocean osculations like ENSO had more of impact than models have predicted.

When models fail its important to create a better model. Perhaps some models rejected by the IPCC for too low greenhouse gas sensitivity better modeled the results.

There has been a great deal more research on osculations and oscean currents they relate to natural variation of temperature, percipitation, and melting antartic ice since IPCC IV. This data can be used to make better models. Natural variation could have made the temperature gap higher or lower.

IPCC increased the range of climate sensitivities in IPCC V based on this research to 1-6 degrees C per doubling of Carbon dioxide. 1 represents no feedback, 6 represents 6x feedback from the physics of ghg warming. Until natural variation is better understood in the models it is hard to nail down a good number. Historically Alley has found a 2.8 degree correlation, but his data includes tipping points, and it is not well understood how the sensitivity reacts without tipping points.

Gavin Schmidt is the most senior modeler at NASA and has stepped into James Hansen's role. He is also one of the founders and big contributor to the RealClimate blog. Schmidt has said that human caused ghg is responsible for 80%-120% of warming. 120% would mean that the world would be cooling, perhaps to another ice age if man was not burning fossil fuel. Until Schmidt's or anyone eases model is believed though, I would like to see it accurately predict the future over a 20 year period, and NASA's model isn't close to accurate. Curry says 50% or more, and she and Schmidt have done a flame war about her conservatism (literally agreeing with IPCC's lower bound) and his hand waving.

In the period between 1960 and 1990, if only ghg caused the change in temperature than sensitivity is 3. It is hard to tell if natural variation was responsible though without better models.

 

It's called "hindcasting" in meteorology.

 

+1

Thank you for the correct term for climate and weather. I was using the generic modeling term.

In reality if we talk about climate as over 30 year periods, I would hindcast only older data. If I had the data and were to model it, I would hindcast or back test until 30 years ago (instrument record 1880-1984) to set parameters, then test it on the recent period (1985-2014). If the model was close, I would use the best theories available to tweak (methodology uncertain to make better) parameters. Unfortunately the data is sparse going back, with first weather satellites not launched until around 1960. That means that modeling something like ENSO may have less than 50 years of good data, and I would want to use all of it.

The good news is we are collecting more and better data which will lead to more accurate models. Computational power and software for meteorological and climate modeling keeps getting better and less expensive. Modeling should continue to improve. Past IPCC leadership perhaps has constrained IPCC models to over estimate ghg and underestimate natural variation.

We should not though listen to blogs like skeptical science, and believe that we can devine climate sensitivity from inaccurate models. We do have other methods to give this sensitivity bounds though.

 

I'm not sure what you are disagreeing with me over. I agree with this statement. You need to test against historic conditions. A model that fails to work in the past is unlikely to work in the future.

What I said though was that the models that the IPCC uses have had their perameters set by hindcasting, so this is a poor test of future performance. The way we test future performance is how well a model does with data that it was not trained with. IPCC V said these models didn't do very well. Having built some back tested models myself, I can assure you that if your model assumptions are incorrect, you can still get a model that still looks good with historic data.

That means that the models need to be improved. I don't think B.E.S.T. or IPCC or even Gavin schmidt of real climate disagrees with me there.

Quick google of realclimate for a more recent gavin schmidt post brings of up this. I don't like his graphic as I find it misleading, but the text and references are excellent for understanding sensitivity and the models. After reading I will correct a misstatement. No feedback gives a sensitivity of 1.2 degrees C for a doubling of ghg, not 1 degree as I posted.
RealClimate: On sensitivity: Part I

 

Sorry about that bob. I knew if you were disagreeing with your link, you weren't understanding my point. I'm glad that the link was helpful.

IMHO skeptical science was using an argument that meant they clearly did not understand how the models were selected, or was trying to fool people.

 

Do we really care exactly what the catastrophe is going to be? Or are people of the opinion that you can increase the energy in the system and somehow get benign results?

Thank You Kindly.

 

Corwyn, I think we do care. I think there are number of areas with possibly negative outcomes. Not all of equal sizes. Also if responses require money, or lots of money, some type of ranking is required.

It seems to me that we get too bogged down focusing only on air temperature patterns. The others are as I have proposed before, though perhaps not comprehensively.

If the models were better they could be very helpful in sorting out risks of climate change. We'd need to look elsewhere to assess risks of CO2 on marine food sources and making crops lower in protein and yet more appealing to pests.

Mining and burning fossil C poses human risks both obvious and others not yet generally agreed upon. Finally, about half the world's human population would certainly be much better off with more access to energy (not to mention potable water).

When the global banking industry...um...mis-stepped, Big buckets of money appeared to aid in soft landing/recovery. That being an industry 'too big to fail'. It is not impossible that the fossil-C industry might need assistance to redirect its activity towards sustainability (I know that word has been declared illegal in Florida, but I'm not there )

All of that constitutes the matter at hand. It can't be waved away by assertions of climate scientists' malfeasance, or by unmet demands for particular 'debating' structures.

If by any chance there is a war on science, I suggest an immediate armistice. We are in need of the science toolkit.

 

The increase in temp from 1850 until today has been hugely beneficial to the human race as well as other species.
That rise was mostly natural as co2 wasnt increasing until after the mid of the 20th cent.
Prior to 1850 there was crop failure starvation and disease caused by cold temps.
Many millions died .
Witches were the cause then,now its co2.
Today we have the best climate weve had on Earth for centuries.
The past 10,000 years were warmer than today and the planet prospered.
We have the added benefit of increased plant population and growth of 20% due to higher co2 levels.
Higher temps of a few degrees will most likely improve our environment even more.
But it doesnt appear that added co2 has had any warming effect that isnt compensated for by lessening water vapor.

 

When IPCC models fail in the first few years of a 100 year prediction, its hard not to use a broad brush to say that they are worthless.
BTW the model which predicts a 1 degree warming and is correct so far is IMO by Richard Lindzen. LOL! The AGW skeptic who was once the IPCCs top climate scientist.
Also super computers cant fix garbage in garbage out, from climate scientists who dont understand how climate or weather works.
Super computers will only give mistaken climate scientists, the wrong answer faster.
If you want to improve climate models weaknesses,try first learning by studying and knowing how water vapor controls climate.Learn how oceans control climate.Then learn how the Sun controls climate.
Then model that.

 

Mojo@131
"That rise was mostly natural as co2 wasnt increasing until after the mid of the 20th cent."
Not all have missed my posts on this matter.

"Prior to 1850 there was crop failure starvation and disease caused by cold temps."
Most biological processes (including disease) are enhanced by high temperature and water availability. One would certainly agree that crop production per se is less in cold years - with everything else held constant. In the future it remains unclear whether everything else will be constant.

"Today we have the best climate weve had on Earth for centuries."
This statement is true unless one has been whacked by flood, drought, fire, or typhoon. I'm still happy that mojo has not been whacked.

"The past 10,000 years were warmer than today"
Not all have missed my posts on this matter.

"lessening water vapor."
It is increasing, or presents no significant trend. Not all have missed my posts on this matter.

mojo @132
"Richard Lindzen. LOL! The AGW skeptic who was once the IPCCs top climate scientist."
Top or not, we may not here be able to decide. But why not just put the question to Dr. Lindzen - how much does he expect global surface air T to change in the next 1 and 2 decades? I am not sure he has put money down on that.

"Learn how oceans control climate."
Yes yes yes! You threw a lot of pasta at the refrigerator and this one sticks. It is almost as if you have been reading my PC posts. We need this. More CO2, more outgoing energy trapped, but the oceans still decide whether that trapped energy will spoil your enterprise.

Referring to my post @130, the past (glorious as it often was) might not dictate the future. It is clear that mojo disagrees. So the point is in which direction we ought to look. Does looking back favor fossil-C burners? You betcha.

How shall we look forward instead? No reason to suppose that it will be simple.

 

One ought not suppose that mojo's renewed stridency in these issues is related to latest global air T, Arctic sea-ice (low) maximum extent, nor the latest level-5 cyclone through Vanautu. Anything like that would point towards 'motivated thinking'.

Instead, what we all want to talk about is how to chart a good course in the 21t century. Would be hard to to that while tossing aside science, am I right?

 

Bob Wilson and other engineers may appreciate a study recently published regarding the hydrologic cycle constraining the atmosphere's role as a heat engine in a warming climate....


Constrained work output of the moist atmospheric heat engine in a warming climate


There have been dueling studies of whether California's drought since 2011 is natural variability or AGW-induced. NOAA researchers conclude that the drought is natural variability...


http://cpo.noaa.gov/sites/cpo/MAPP/Task%20Forces/DTF/californiadrought/California_Drought_2-pager.pdf


...and that the drought is almost opposite to what GCMs have predicted....


Causes of Calif. drought natural, not man-made: NOAA


Another study concludes that the drought is AGW-induced....


Climate change identified as the root of California's hot, dry winters | Inhabitat - Sustainable Design Innovation, Eco Architecture, Green Building


There appears to be a correlation between California's drought and the tornado "drought" that has occurred since 2011 (including near-record low number so far this year)....




Storm Prediction Center WCM Page


The persistent ridge along the west coast induces troughing in the eastern CONUS and northwest flow over the central plains, which is not conducive for spring-time severe weather outbreaks. So if AGW is responsible for the drought in California (a bad thing), it also is likely to be responsible for record-low number of tornadoes (a good thing).