This was an eye opener
- Thread starter goldenfamaliy
- Start date
- Status
greenbean36191
Premium Member
:rolleye1: Rehashing points ALREADY addressed in this very (short) thread-
In any event, this point is largely irrelevant to the discussion since it does not address the imbalance in C cycling, which is the concern. The relative size of emitters does not matter. What matters is the increase in emissions vs. the available capacity of C sinks to take up that increase. There has been no observed increasing trend in natural CO2 sources over the recent past, while anthropogenic sources have increased dramatically. Again, isotopic signatures confirm that the increased CO2 is from anthropogenic sources, not natural.
Carbon is often talked about as a "budget" because it's an analogy that makes sense. If we have a $100 budget and natural sources account for $98.50 of that budget and we spend another $3 it doesn't matter who spent more- we still went over budget (exceeded the capacity of natural carbon sinks).
Statistics dictate that in order to say anything about a trend you have to have a long enough time period that the signal overcomes the natural variability of the system for that time period. For annual temperature anomalies, that means periods close to 30 years in most cases.
Then PLEASE, by all means bypass the media and go straight to the scientific literature. Show us that scientific data supporting the opposition from the scientific literature. It should be easy to find. I certainly haven't come across it, and it doesn't sound like Bill has either. Authors like Oreskes have done reviews of the literature and not found it.
Wrong. Every global climate model includes variations from the sun. Changes in the sun's total irradiance, sun spots, and magnetic variations are tracked regularly. We also have measurements of oxygen and beryllium isotopes that give us ways to measure solar variation in the past. There has been no change in solar output that can account for current temperature trends by any known mechanism. The current warming pattern does not match the pattern that solar warming would cause anyway. With solar warming the upper parts of the atmosphere would warm faster than the lower parts. Instead what is observed is that the lower atmosphere is warming while the upper atmosphere cools. Again, there is no known mechanism by which solar variation can cause that pattern, yet it is exactly the pattern expected with greenhouse warming.
There have been multiple periods in Earth's history when CO2 has been one of, if not the major cause of climate change prior to the current event. It happened during the Carboniferous and during periods of major basalt floods like the Deccan and Siberian traps to name a few.
No one is saying differently. The presence of natural cycles does not draw into question the possibility of anthropogenic effects as well. Rain has occurred naturally for billions of years as well, but in the last 60 we have figured out how to seed clouds to make it rain when and where we want it to. Does the fact that it has rained naturally for billions of years make you question the proven technology of cloud seeding can in fact work? Do you question that a lit match can start a forest fire because forest fires have occurred for hundreds of millions of years and matches have only been around for about 150 (coincidentally, about the same length of time as concern about global warming). Of course you don't because the logic is silly. We know that different causes can have common effects.
Again, there is no reason why anyone who takes the time to research this topic should be repeating this talking point as it's very easy to check by looking at actual numbers. The total CO2 output of ALL volcanoes worldwide is orders of magnitude less than from anthropogenic sources in a single year. The specific numbers have already been posted in this thread. We know the numbers from accounting estimates as well as from analysis of the isotopic signatures of the emitted CO2. You can also look at the atmospheric record of CO2 and you will notice that even exceptionally large volcanoes like Pinatubo don't even emit enough CO2 to register as spikes in the record.
In any event, this point is largely irrelevant to the discussion since it does not address the imbalance in C cycling, which is the concern. The relative size of emitters does not matter. What matters is the increase in emissions vs. the available capacity of C sinks to take up that increase. There has been no observed increasing trend in natural CO2 sources over the recent past, while anthropogenic sources have increased dramatically. Again, isotopic signatures confirm that the increased CO2 is from anthropogenic sources, not natural.
Carbon is often talked about as a "budget" because it's an analogy that makes sense. If we have a $100 budget and natural sources account for $98.50 of that budget and we spend another $3 it doesn't matter who spent more- we still went over budget (exceeded the capacity of natural carbon sinks).
That's nice, but the discussion has been about science, which the article did not address. Everyone, including scientists are entitled to their opinion, but it is not science without evidence. Starting off an article with the claim that "Global warming (i.e, the warming since 1977) is over." puts the author on shaky ground to begin with because such a claim is statistically untenable.
No. All of the major temperature records are available online, so you can analyze the data yourself. None of the data show a statistically significant cooling trend and in fact, the trend from the most recent decade is similar to that of the past 4. The "cooling trend" is produced by selecting short time periods that are dominated by noise and often high start points (e.g. 1998). These are not valid statistical methods and do not allow for any determination of a real trend whatsoever. For example, the you calculate the trend for the most recent decade, the margin of error is about 1.5 times the value of the trend. That margin of error means that the actual trend lies somewhere between cooling and record breaking warming. This is true for any single decade you look at for the last 40 years at least.
Statistics dictate that in order to say anything about a trend you have to have a long enough time period that the signal overcomes the natural variability of the system for that time period. For annual temperature anomalies, that means periods close to 30 years in most cases.
Joshsmit56001
Member
<a href=showthread.php?s=&postid=15311249#post15311249 target=_blank>Originally posted</a> by greenbean36191
[B
"There has been no change in solar output that can account for current temperature trends by any known mechanism."
If this is true please explain the graph below
stugray
Premium Member
Here's another "truth":
The sun behaves "Well" to our model so far as the data fed into the model has historical significance.
The sun has been present for ~4.5 billion years.
We have less than 20 years of "calibrated data" to feed into the model.
What does this mean....?
It means that we have a decent solar model over the time we have accurate solar "behavior" data.
I concede to greenbean36191 for the details of the current climate model, but I still question the validity of ALL the data fed into the model.
Here are a couple of links to stories that are relevant to this conversation:
http://www.space.com/scienceastronomy/090706-sunspot-activity.html
http://www.spaceflightnow.com/news/n0906/29ulysses/
http://sohowww.nascom.nasa.gov/
These are more spacecraft that demonstrate how we are now ( within the last 20 years ) getting data that we never had before.
Stu
The sun behaves "Well" to our model so far as the data fed into the model has historical significance.
The sun has been present for ~4.5 billion years.
We have less than 20 years of "calibrated data" to feed into the model.
What does this mean....?
It means that we have a decent solar model over the time we have accurate solar "behavior" data.
I concede to greenbean36191 for the details of the current climate model, but I still question the validity of ALL the data fed into the model.
Here are a couple of links to stories that are relevant to this conversation:
http://www.space.com/scienceastronomy/090706-sunspot-activity.html
http://www.spaceflightnow.com/news/n0906/29ulysses/
http://sohowww.nascom.nasa.gov/
These are more spacecraft that demonstrate how we are now ( within the last 20 years ) getting data that we never had before.
Stu
ScooterTDI
New member
Since I have some background in atmospheric modeling (not climate modeling), I think I need to point out that modeling is not science. Modeling is a tool used to get a grasp on something you already know that you don't understand well.
All models are built on a house of cards of assumptions. Hopefully the assumptions on which a model is founded are sound, but often if even a single assumption is flawed, the models estimates can be in drastic error.
I really dislike the term "prediction" in reference to the results of a modeling exercise. Models are not intended to and should not be used to attempt to predict the future. The proper use of modeling is to explore the possible interations that could occur in complex systems given a set of assumptions to simplify the system enough to afford calculation. Nothing else.
Some modelers get carried away with the idea that their model is SOOO good that its results must reflect reality. The good modelers understand that the model is a much simplified version of reality and the worst assumption you can make is the assumption that you've included ALL relevant factors in the model.
Scott
All models are built on a house of cards of assumptions. Hopefully the assumptions on which a model is founded are sound, but often if even a single assumption is flawed, the models estimates can be in drastic error.
I really dislike the term "prediction" in reference to the results of a modeling exercise. Models are not intended to and should not be used to attempt to predict the future. The proper use of modeling is to explore the possible interations that could occur in complex systems given a set of assumptions to simplify the system enough to afford calculation. Nothing else.
Some modelers get carried away with the idea that their model is SOOO good that its results must reflect reality. The good modelers understand that the model is a much simplified version of reality and the worst assumption you can make is the assumption that you've included ALL relevant factors in the model.
Scott
Joshsmit56001
Member
Very well said!
ScooterTDI
New member
A phrase used by many modelers I know:
"All models are wrong!"
Scott
"All models are wrong!"
Scott
HippieSmell
Occupy Reef Central
So, Scott, what is your take on climate modeling? I don't know what your experience is (sounds like a little in models for daily weather forecasts), but are you saying that climate models aren't sophisticated enough to be allowed as evidence in climate policy decision making? If not, what is the alternative?
ScooterTDI
New member
My take on climate modeling is that it is a valid method to explore the complex interations that may occur and examine how perturbations may affect different aspects of the system, but it is not a method that can predict the future. If modeling results are included in policy decision-making, I fear that the modeling results will be misinterpreted as a window into the future.
The idea that a model can be used to predict the future is a very common misperception/misunderstanding of the role of modeling and the information that it can actually provide. I can't even count how many times someone has asked me "How accurate are these predictions?". This sort of question implies a fundamental misunderstanding of the purpose of modeling. People look towards modeling as if it has an answer for the future, but it doesn't.
My view has nothing to do with the level of sophistication of the models. I've seen plenty of models that are more sophisticated that the available data actually allows. These models are often "worse" than the simpler models that have fewer (but better) assumptions. All models are based upon assumptions, otherwise it wouldn't be a model, but rather a direct calculation.
The results produced by modeling are generally highly dependent on the assumptions that are built into a model. A good modeler can manipulate these assumptions to produce any result desired while still maintaining valid logic for the choice of assumptions.
The alternative is to accept that our understanding of climate change is limited and consider that limitation while making policy decisions. Hysterical changes based upon chiefly on modeling results is probably not the best approach. I'd advocate well thought out gradual changes to reduce emmision and wean us from oil.
Policy decisions that seek to reduce the chance of adverse climate change should be enacted for more reasons than just the environment. Besides the many political/international benefits, the fact is that the Earths oil reserves are not infinite. Even if climate change were not a concern, we would still need to find alternative energy sources. In my mind, that really is the most important reason to move towards renewable energy sources.
Scott
The idea that a model can be used to predict the future is a very common misperception/misunderstanding of the role of modeling and the information that it can actually provide. I can't even count how many times someone has asked me "How accurate are these predictions?". This sort of question implies a fundamental misunderstanding of the purpose of modeling. People look towards modeling as if it has an answer for the future, but it doesn't.
My view has nothing to do with the level of sophistication of the models. I've seen plenty of models that are more sophisticated that the available data actually allows. These models are often "worse" than the simpler models that have fewer (but better) assumptions. All models are based upon assumptions, otherwise it wouldn't be a model, but rather a direct calculation.
The results produced by modeling are generally highly dependent on the assumptions that are built into a model. A good modeler can manipulate these assumptions to produce any result desired while still maintaining valid logic for the choice of assumptions.
The alternative is to accept that our understanding of climate change is limited and consider that limitation while making policy decisions. Hysterical changes based upon chiefly on modeling results is probably not the best approach. I'd advocate well thought out gradual changes to reduce emmision and wean us from oil.
Policy decisions that seek to reduce the chance of adverse climate change should be enacted for more reasons than just the environment. Besides the many political/international benefits, the fact is that the Earths oil reserves are not infinite. Even if climate change were not a concern, we would still need to find alternative energy sources. In my mind, that really is the most important reason to move towards renewable energy sources.
Scott
stugray
Premium Member
ScooterTDI,
Thank you you made my point far more eloquently than I did.
IMO we cannot make drastic decisions based on the current model(s) because:
1 - Even the long term data we are feeding into the models is imperfect ( It is based on many of these assumptions & other models ).
2 - The Models themselves have only been 'experimented with' since we have had computers ( and I mean GOOD ones ).
3 - We only have 'data continuity' & 'calibrated data' for the very recent past.
When the modeled trends are on the order of hundreds of thousands of years to Millions of years, how can we trust applying current data when we have so little?
I am not saying that the satellite data is perfect ( it relies on other models ).
I am just saying that the validity of ALL of the 'Climate' & ' Atmospheric' & ' Solar' Models will converge given enough time & data & computing power.
Stu
Thank you you made my point far more eloquently than I did.
IMO we cannot make drastic decisions based on the current model(s) because:
1 - Even the long term data we are feeding into the models is imperfect ( It is based on many of these assumptions & other models ).
2 - The Models themselves have only been 'experimented with' since we have had computers ( and I mean GOOD ones ).
3 - We only have 'data continuity' & 'calibrated data' for the very recent past.
When the modeled trends are on the order of hundreds of thousands of years to Millions of years, how can we trust applying current data when we have so little?
I am not saying that the satellite data is perfect ( it relies on other models ).
I am just saying that the validity of ALL of the 'Climate' & ' Atmospheric' & ' Solar' Models will converge given enough time & data & computing power.
Stu
HippieSmell
Occupy Reef Central
Well, I'll ask you this then. Do you think that there is a link between CO2 and temperature that warrants a reduction in emissions? Why or why not? How does modeling impact your opinion?
ctenophors rule
New member
thats not what he is saying, he is saying that a model is based on a bunch of assumptions, then the data is fit in, so to speak, to hopefullly support those assuptions.
the fact the AAC02 effects the temperature would be an assumption that no scientist above the age of 10 would dispute.
all that the model does is show how much of an effect it will have.
whether is it accurate or not, depends on the other assumptions, like, how much AAC02 will be pruduced in the next 20 years. what natural spikes in co2 levels will occur. will the increased water vapor form enough clouds to counter act its temperature trappinf protperties by deflecting enough heat, or might it even reerse the heating trend slightly.
stuff like that.
.........(i think)
HippieSmell
Occupy Reef Central
I didn't say he was saying anything. I'm asking what he thinks.
ScooterTDI
New member
I don't doubt that CO2 concentrations can be correlated to temperature. However, correlations to not demonstrate causality. In fact, it is generally much more difficult to demonstrate causality than to demonstrate a simple correlation.
Here is a good example of what I mean:
Illicit drug use and clinical depression have been correlated. That is easy to demonstrate. It is not so easy to discern whether drug use causes depression or whether individuals already experiencing depression begin using drugs to self-medicate. The reason I mention this as an example is that I have seen peer-reviewed articles in scientific journals that claim that drug use causes depression based solely on a correlation.
As I said before, I think there are many reasons to reduce emissions and move towards alternative energy sources. One of the benefits of emmision reduction is to reduce the possibility that we will negatively impact the environment, though personally this would probably not be my primary motivation.
Modeling results can be considered, but only with the caveat that they shouldn't be treated as a prediction. Modeling may help us understand how certain aspects (i.e. temperature) of the environment might respond to specific changes (i.e. CO2 concentrations), but it must be understood that we do not know 1) if our models address all of the relevant factors involved and 2) if the underlying assumptions in our models are valid.
Scott
ScooterTDI
New member
This brings up an important aspect of modeling.
Let us imagine we have a data set that seems to demonstrate some kind of trend over time. We want to know the future, so we decide to make a model even though this is a misguided use of modeling.
You start exploring all the different ways of modeling the existing data. Some possible models you may discard because one or more assumptions would have to be made that don't really seem reasonable to you. Often, this is a subjective judgement.
Even though you through out tons of models, you still have many models that will fit the existing data well, but they all have different predictions about the future. What do you do? How do you pick the model to use to make predictions about the future? Again, this is often a subjective decision. Some modelers would opt for simplicity. Others, would opt for the most detailed model. Some would find one that made the most "sense" to them.
No matter what choice is made, it is not possible to be certain that the particular model selected will have any accurate predictions about the future. The only conclusion that can be made is that the selected model does a good job of describing the historical data.
Scott
greenbean36191
Premium Member
This is an excellent example of why scientists are required to test for statistical significance rather than just looking at apparent trends. The data in the graphs is correct and the smoothed curve is correct for the data that is plotted. The problem is that there is a major endpoint artifact. Because the graphs stop at a high point purely by chance, the end of the curve is drawn upwards towards it. However, the time series does continue after the endpoint of the graph, so we know that the apparent upward trend shown in that version is spurious. The high point was actually about 1950 and there has been no trend since.
The same is true of total solar irradiance, galactic cosmic rays, and solar flares (this graph continues from where the NOAA one left off).
While the trend prior to 1950 is real and can account for a great deal of the warming prior to that time period, there is no known mechanism by which the warming would continue at the rate it has while there has been no trend in solar output. Even by Stott et al.'s estimate (Stott is generally considered to be a contrarian) changes in solar output can only account for about 1/3 of the observed warming, even if an unproven GCR/ cloudcover feedback is assumed. Besides, as I've already mentioned before the pattern of warming is consistent with greenhouse warming, but inconsistent with warming due to solar variation.
To blame it all on the sun you would have to demonstrate some unknown mechanism and feedbacks for solar warming as well as an unknown mechanism by which the influence of greenhouse gases is nullified.
Again, you're misusing "calibrated", presumably to mean satellite data rather than surface observation. There are two problems with that.
First, all scientific data is calibrated. Every official surface station out there uses calibrated instruments. They have been doing so continuously for about 160 years and less systematically for almost 400. We also have calibrated proxies that go back millions of years.
Second, besides your job description, why do you believe that the satellite data is more accurate than surface readings. For one, if your chief complain is a lack of long-term data, why do you insist on the shortest available dataset (which is 30 years for temp and 31 years for solar activity- not 20)? Second, using the same satellite data, different processing produces different results. RSS and UAH use the same data and get different values and different trends for example.
Also, I'm sure you realize that satellite measurements are based on physical models rather than direct measurements, right? So why do the satellite models produce the best data available, but physical models in general are not to be trusted?
Joshsmit56001
Member
We obviously have two different sets of graphs here so we will have to agree to disagree.
greenbean36191
Premium Member
As an undergrad in a population modeling class who would fail you for an assignment on the spot if you used "prediction" rather than "projection" when referring to model output.
Huh? The two chief uses of models are to forecast and hindcast. You hindcast to test your understanding of the fundamental factors at play and to test the skill of the model and you forecast to project what to expect in the future if those factors continue to interact in the way they have in the past.
Yes, but there is a big difference between assumptions built into the model itself and assumptions built into the input scenarios.
When your assumptions built into the model are based on physics or on relationships between variable that have been constrained within certain ranges then assuming continuation is pretty sound.
Your input scenarios are always going to contain questionable assumptions. However, you can violate assumptions of the input scenario leading to incorrect projections, but the model itself is still sound.
I've never met a modeler in any field that makes that assumption. It's understood that models are simplifications of reality. However, rather than making any assumptions about whether you've included all of the most important factors you can TEST it by applying the model to hindcasts.
The rest of the saying is "All models are wrong, but some models are useful."
Well they are our best estimates of what to expect. Would you prefer policy based on naive assumptions or best estimates?
Yes, but I'm not sure how the limitation of understanding gives any comfort. It's a two way street. It means we're just as likely to be UNDERestimating the impacts as we are to be overestimating. Judging by the performance of forecasts, we've been erring more to the side of underestimation.
greenbean36191
Premium Member
No we don't. Your graph ends in the 1980s and the moving average appears to be going up because of an endpoint artifact. When you plot the rest of the data after the end of your graph that apparent upward trend goes away because the artifact is eliminated. It's a common problem that occurs with plotting moving averages. There's no hocus pocus going on there. You can plot the two different time periods and apply identical smoothing on your own to check for yourself. The data is available here: ftp://ftp.ngdc.noaa.gov/STP/SOLAR_DATA/SUNSPOT_NUMBERS/MONTHLY.PLT
- Status
Similar threads
