Temperature Data Adjustments

[blizzard1024]

All,

 

Many of you know I am a skeptical when it comes to all science, especially extraordinary claims like a world wide disaster from climate warming. However, I honestly am trying to understand these data adjustments because EVERY single adjustment I see(except the satellite data which is being attacked by the way), have warmed the present and cooled 100 years ago and in some cases significantly. I am a meteorologist with a strong background in fluid dynamics and using models to forecast the weather. I really am not much of a statistics guy like many of you may be. I have read this TOB stuff and I can see how double mins from a 7 am to 7 am ob site would skew the data however, what about the maxes?? Why would this have anything to do with warming the maxes as someone has stated? Again, this looks very fishy to the skeptical community that the very folks who are getting all the grant money are also adjusting data making it warmer now and cooler in the past. Even if these adjustments are scientifically valid as many of you point out it still is a tough sell to the general populous because it is very complicated. I have read some of the papers and they are poorly written and frankly I get lost because I am not a statistics guru. I would like to have a civil discussion and I would like for many of you to find a way to describe this stuff in a way that is easy to understand. Plus many papers are behind paywalls UGH.  

thanks.

[ORH_wxman]

Without getting overly into the nitty gritty of the statistics, this is a good overview of the idea behind temp adjustments:

http://judithcurry.com/2014/07/07/understanding-adjustments-to-temperature-data/

Basically our two largest changes in the way we measure temperature (TOBS and MMT) has led to a warm bias in older temps and a cold bias in more recent temps if they are not adjusted. It's a coincidence that it worked out that way but it's necessary if we want to be accurate.

[skierinvermont]

All,

 

Many of you know I am a skeptical when it comes to all science, especially extraordinary claims like a world wide disaster from climate warming. However, I honestly am trying to understand these data adjustments because EVERY single adjustment I see(except the satellite data which is being attacked by the way), have warmed the present and cooled 100 years ago and in some cases significantly. I am a meteorologist with a strong background in fluid dynamics and using models to forecast the weather. I really am not much of a statistics guy like many of you may be. I have read this TOB stuff and I can see how double mins from a 7 am to 7 am ob site would skew the data however, what about the maxes?? Why would this have anything to do with warming the maxes as someone has stated? Again, this looks very fishy to the skeptical community that the very folks who are getting all the grant money are also adjusting data making it warmer now and cooler in the past. Even if these adjustments are scientifically valid as many of you point out it still is a tough sell to the general populous because it is very complicated. I have read some of the papers and they are poorly written and frankly I get lost because I am not a statistics guru. I would like to have a civil discussion and I would like for many of you to find a way to describe this stuff in a way that is easy to understand. Plus many papers are behind paywalls UGH.  

thanks.

 

Actually, the adjustments to global temperature are net-negative. The raw data shows more warming than the adjusted data. This is because the adjustments to ocean data are more negative than the adjustments to land are positive. Your conspiracy theory of scientists receiving grant money and then adjusting temp data upward is unsupported by the facts. The net adjustments are negative. The "skeptics" focus on the positive but necessary land adjustments but have no problem with the negative ocean adjustments. It's pretty clear who has an agenda.

 

Also the revisions to UAH have been more positive than the adjustments to surface land data.

[chubbs]

As discussed in the post above temperature homogenization adjustments have reduced global warming due to predominantly upward adjustment to early ocean data.

 

http://variable-variability.blogspot.com/2015/02/homogenization-adjustments-reduce-global-warming.html

 

[Polar Vortex 2014]

All,

 

Many of you know I am a skeptical when it comes to all science, especially extraordinary claims like a world wide disaster from climate warming. However, I honestly am trying to understand these data adjustments because EVERY single adjustment I see(except the satellite data which is being attacked by the way), have warmed the present and cooled 100 years ago and in some cases significantly. I am a meteorologist with a strong background in fluid dynamics and using models to forecast the weather. I really am not much of a statistics guy like many of you may be. I have read this TOB stuff and I can see how double mins from a 7 am to 7 am ob site would skew the data however, what about the maxes?? Why would this have anything to do with warming the maxes as someone has stated? Again, this looks very fishy to the skeptical community that the very folks who are getting all the grant money are also adjusting data making it warmer now and cooler in the past. Even if these adjustments are scientifically valid as many of you point out it still is a tough sell to the general populous because it is very complicated. I have read some of the papers and they are poorly written and frankly I get lost because I am not a statistics guru. I would like to have a civil discussion and I would like for many of you to find a way to describe this stuff in a way that is easy to understand. Plus many papers are behind paywalls UGH.  

thanks.

 

The observation (set max) used to occur in the afternoon from 5 to 7. This caused a lot of artificially high maxes. Think of how many days its warmer from 5 to 7 than the following day maximum temperature. Go back through the older records, and notice how many high temperatures are in fact the set max from the previous day's observation, and then get back to us.

[blizzard1024]

Without getting overly into the nitty gritty of the statistics, this is a good overview of the idea behind temp adjustments:

http://judithcurry.com/2014/07/07/understanding-adjustments-to-temperature-data/

Basically our two largest changes in the way we measure temperature (TOBS and MMT) has led to a warm bias in older temps and a cold bias in more recent temps if they are not adjusted. It's a coincidence that it worked out that way but it's necessary if we want to be accurate.

Thanks. Just what I was looking for. I know Dr. Curry and she has common-sense about this issue. Unfortunately, it is still too complicated for the general public or policy makers(they need things almost impossibly dumbed down). But for my purposes  this is perfect. take care.

[blizzard1024]

What is everyone's opinion on the CFSV2 ,  MSU and RSS. I know a lot of people are skeptical of these datasets because they don't show as much warming. But the surface record with all the needed adjustments, station moves plus trying to get the temperatures over the vast oceans 100 years ago...it would seem the satellite record is the way to go. Also what about the radiosonde data? why isn't that shown each month? I haven't seen radiosonde data in a long time for climate records??

[skierinvermont]

What is everyone's opinion on the CFSV2 ,  MSU and RSS. I know a lot of people are skeptical of these datasets because they don't show as much warming. But the surface record with all the needed adjustments, station moves plus trying to get the temperatures over the vast oceans 100 years ago...it would seem the satellite record is the way to go. Also what about the radiosonde data? why isn't that shown each month? I haven't seen radiosonde data in a long time for climate records??

 

The radiosonde data has much larger uncertainty bars and I would characterize it as a rough estimate of temperature trends. The MSU data/RSS is somewhat more accurate than radiosonde but is estimated to have uncertainty on the order of .1C/decade for long-term trends. There also exist major discrepancies between MSU products and versions and the methodology has required major revisions. In 1997 the scientists at UAH were insisting the earth was not warming. A year later they were forced to revise their methodology. Satellite data is not the golden egg of temperature measurement. There are many calibration and merging issues to deal with that have no exact solutions. 

 

The adjustments to surface data can be tested quantifiably for accuracy. How much does changing TOBs time change measurement? This can be tested empirically.

 

 

CFSv2 is not considered nor designed for long-term estimate of temperature. There is an obvious step down in temperature that occurred in 2010. I think the methodology changed then. 

[nflwxman]

The radiosonde data has much larger uncertainty bars and I would characterize it as a rough estimate of temperature trends. The MSU data/RSS is somewhat more accurate than radiosonde but is estimated to have uncertainty on the order of .1C/decade for long-term trends. There also exist major discrepancies between MSU products and versions and the methodology has required major revisions. In 1997 the scientists at UAH were insisting the earth was not warming. A year later they were forced to revise their methodology. Satellite data is not the golden egg of temperature measurement. There are many calibration and merging issues to deal with that have no exact solutions. 

 

The adjustments to surface data can be tested quantifiably for accuracy. How much does changing TOBs time change measurement? This can be tested empirically.

 

 

CFSv2 is not considered nor designed for long-term estimate of temperature. There is an obvious step down in temperature that occurred in 2010. I think the methodology changed then. 

 

Yep, and it's equally funny how Joe Bastardi and Co. completely ignore the fact that the CFSv2 adjusted it's baseline in 2010.  The dataset actually has 2008 as warmer and 2010, but of course that doesn't raise any red flags with the hardcore deniers.  

[FloridaJohn]

What is everyone's opinion on the CFSV2 ,  MSU and RSS. I know a lot of people are skeptical of these datasets because they don't show as much warming.

 

They are skeptical of those datasets not because of how much warming they show, but how much uncertainty there is in their results. As mentioned above, compare the size of the error bars of the various temperature record sets and that will show you how much uncertainty exists in each set. There is no conspiracy here, just science.

[tamarack]

The radiosonde data has much larger uncertainty bars and I would characterize it as a rough estimate of temperature trends. The MSU data/RSS is somewhat more accurate than radiosonde but is estimated to have uncertainty on the order of .1C/decade for long-term trends. There also exist major discrepancies between MSU products and versions and the methodology has required major revisions. In 1997 the scientists at UAH were insisting the earth was not warming. A year later they were forced to revise their methodology. Satellite data is not the golden egg of temperature measurement. There are many calibration and merging issues to deal with that have no exact solutions. 

 

The adjustments to surface data can be tested quantifiably for accuracy. How much does changing TOBs time change measurement? This can be tested empirically.

 

 

CFSv2 is not considered nor designed for long-term estimate of temperature. There is an obvious step down in temperature that occurred in 2010. I think the methodology changed then. 

 

There must be some long-term datasets which include hourly obs and max/min, that could be mined for these variations (by someone who could do this as part of their day job.)  As purely a guess, I'd estimate a move from midnight TOBS to 7A would lower avg temp by 0.5F, and from midnight to 4P would raise the avg by about the same.  The change ought to be directly proportional to the avg diurnal range for a particular location.

[blizzard1024]

 

 

The adjustments to surface data can be tested quantifiably for accuracy. How much does changing TOBs time change measurement? This can be tested empirically.

 

 

Maybe for the TOB but what about all the station moves and non-climatic trends associated with these, also instrumentation changes? Plus what about oceanic temperatures? These would seem to be rife with large error bars...especially 100 years ago data vs today's.

[FloridaJohn]

Maybe for the TOB but what about all the station moves and non-climatic trends associated with these, also instrumentation changes? These would seem to be rife with large error bars...especially 100 years ago data vs today's.

What if they kept the old station in place for a time after the new station was up and running? That would give them a way to compare the data from the two stations. What if they did the same thing for instrument changes? What if they compared the changed station to a nearby unchanged station?

 

Error bars for older data are indeed larger than newer data. That shows improvements in instrumentation and methods. But even the old error bars are definable and the limitations are well understood.

[chubbs]

Don't ignore the factors that cause warming to be underestimated like movement out of city centers, increased irrigation, and improved temperature screens and shelter paints

 

http://variable-variability.blogspot.com/2015/04/raw-temperatures-show-too-little-global-warming.html

[Jonger]

The Earth is warming... 

 

The rate is debatable though. I put far more stock in Satellite measurements than ground measurements.

 

1998 still easily takes the award for the warmest year on record.

 

 

[skierinvermont]

The Earth is warming... 

 

The rate is debatable though. I put far more stock in Satellite measurements than ground measurements.

 

1998 still easily takes the award for the warmest year on record.

 

 

 

Yes, because you read denier blogs and aren't familiar with the literature on satellite measurement which is rife with merging issues, revisions, correction estimates, and calibration. 

 

You would have been right along with Spencer and Christy screaming the earth wasn't warming in 1997 - then a year later they are forced to revise and suddenly their results show large decadal warming. 

[skierinvermont]

Maybe for the TOB but what about all the station moves and non-climatic trends associated with these, also instrumentation changes? Plus what about oceanic temperatures? These would seem to be rife with large error bars...especially 10

0 years ago data vs today's.

 

Station move can be dealt with mathematically by essentially treating them as separate stations. They do get merged IIRC but in such a way that there is no change in the global trend compared to if they were treated as completely separate stations. 

 

Instrumentation - that one is obvious and is dealt with in the Curry piece if you read it. If you're going to ask questions you have to read the answers. You should know the answer to this question already since it was dealt with very explicitly in the Curry piece.

Instrumentation changes can be tested empirically by testing the equipment. That one is probably the most obvious and simple. Test the thermometers in all sorts of conditions and see what the bias is.

 

 

Oceanic temperatures show large adjustments down. I'm not as familiar with these. You could do more research - and I'm sure you'll find equally justifiable answers for the adjustments as you do for land temps.

[blizzard1024]

 

 

 

Oceanic temperatures show large adjustments down. I'm not as familiar with these. You could do more research - and I'm sure you'll find equally justifiable answers for the adjustments as you do for land temps.

 

and oceans make up around 70% of the Earth's surface. That makes the surface temperature record useless. The satellite record was corraborated by an independant datatset- the radisondes. The atmosphere is warming less as you go up which is counter to the entire AGW theory. That is very inconvienent. So we find a way to trash all these datasets because they don't fit our beliefs. Also I am amazed at the ISCCP dataset and how well it explains the temperature trends between 1983 and 2009. Oh but wait I think you personally called the folks on this dataset and they told you it isn't accurate enough for a long term climate record. Of course  it isn't!!! It explains the global temperature changes very well.... see below  

 

 

As a scientist you should question and be skeptical instead of labeling people and being nasty on this forum. Wait are you even a scientist??

[blizzard1024]

The radiosonde data has much larger uncertainty bars and I would characterize it as a rough estimate of temperature trends. The MSU data/RSS is somewhat more accurate than radiosonde but is estimated to have uncertainty on the order of .1C/decade for long-term trends. There also exist major discrepancies between MSU products and versions and the methodology has required major revisions. In 1997 the scientists at UAH were insisting the earth was not warming. A year later they were forced to revise their methodology. Satellite data is not the golden egg of temperature measurement. There are many calibration and merging issues to deal with that have no exact solutions. 

 

The adjustments to surface data can be tested quantifiably for accuracy. How much does changing TOBs time change measurement? This can be tested empirically.

It is interesting that the RSS UAH and radiosondes all show a peak or near peak 2007 2008 then a slow fall much like the CFS. So the CFS is not much dufferent that these independent datasets.

 

 

CFSv2 is not considered nor designed for long-term estimate of temperature. There is an obvious step down in temperature that occurred in 2010. I think the methodology changed then.

[StudentOfClimatology]

Just FYI, the ISCCP dataset has very large bars..over 1.5 standard deviations on either side of the trend line. We just don't have the ability to measure long term changes in global cloud cover yet, which is the largest source of uncertainty in regards to estimated climate sensitivity to GHG forcing.

[skierinvermont]

Just FYI, the ISCCP dataset has very large bars..over 1.5 standard deviations on either side of the trend line. We just don't have the ability to measure long term changes in global cloud cover yet, which is the largest source of uncertainty in regards to estimated climate sensitivity to GHG forcing.

 

No, it's a conspiracy theory SOC . ISCCP, stitched together from a million different satellites, is perfect and has zero issues.

[blizzard1024]

No, it's a conspiracy theory SOC . ISCCP, stitched together from a million different satellites, is perfect and has zero issues.

 

How does the surface station data not have large error bars??? Come on. This is crazy especially over the ocean 100 years ago. Also sea ice extent and volume 100 years too.   The earth is warming. I think we all know that but it is being overestimated by the surface stations. That's my opinion. The satellite record is the best and it is corroborated by the radiosonde data. There is no conspiracy here...its call tribalism and politics which has ruined climate science IMO. what needs to be done is to slash funding by 90% to modeling long term climate change and put that money into seasonal forecasting. That would by far be more beneficial to society and the world instead of this "problem" which seems to be occurring much more slowly than the climate models project. anyway, I am not arguing with you skier dude. It seems that you folks are hellbent on believing the datasets that show your view and throw out all the datatsets that don't. That is not good science. Why does the cloud data show such a nice correlation to the temp record?? Is that really a coincidence???  I think it is a smoking gun that cloud cover variations for whatever reason modulate the Earth's short term climate. Increasing CO2 levels has some effect...but we just don't know if the feedbacks are positive enough in today's climate system to cause alarm. Plus the oceans can absorb a lot of heat too which can mute the external forcing by drawing it our over 100s of years....long TCR.

[skierinvermont]

How does the surface station data not have large error bars??? Come on. This is crazy especially over the ocean 100 years ago. Also sea ice extent and volume 100 years too.   The earth is warming. I think we all know that but it is being overestimated by the surface stations. That's my opinion. The satellite record is the best and it is corroborated by the radiosonde data. There is no conspiracy here...its call tribalism and politics which has ruined climate science IMO. what needs to be done is to slash funding by 90% to modeling long term climate change and put that money into seasonal forecasting. That would by far be more beneficial to society and the world instead of this "problem" which seems to be occurring much more slowly than the climate models project. anyway, I am not arguing with you skier dude. It seems that you folks are hellbent on believing the datasets that show your view and throw out all the datatsets that don't. That is not good science. Why does the cloud data show such a nice correlation to the temp record?? Is that really a coincidence???  I think it is a smoking gun that cloud cover variations for whatever reason modulate the Earth's short term climate. Increasing CO2 levels has some effect...but we just don't know if the feedbacks are positive enough in today's climate system to cause alarm. Plus the oceans can absorb a lot of heat too which can mute the external forcing by drawing it our over 100s of years....long TCR.

 

See you pretend to be all open and inquiring, but in reality you have your mind made up.

 

ORH gave you some excellent information in the first post that answers all your questions, but you don't actually read it or process it because it doesn't compute with your pre-determined position. 

 

And to the bolded: yes you are arguing with me (just read your very next sentence), you just don't have any evidence

[blizzard1024]

See you pretend to be all open and inquiring, but in reality you have your mind made up.

 

ORH gave you some excellent information in the first post that answers all your questions, but you don't actually read it or process it because it doesn't compute with your pre-determined position. 

 

And to the bolded: yes you are arguing with me (just read your very next sentence), you just don't have any evidence

 

I read Judy Curry's piece from ORH and it makes sense for land readings. Still I don't see how there can't be large error bars. I especially have trouble when stations are moved. The station that is moved then senses a new microclimate which has non-climatic trends. These microclimates typically have large variations in temperature over short distances on clear calm mornings when the atmosphere is not mixed. Just look at a hi res IR satellite loop when arctic air is over the U.S or southern canada. Moving a station will no doubt change the temperature averages and they can try to account for these but there has to be unknowns. I used to ride my bike to work years ago when I lived close enough and I went through  a rural area then a town and university setting. The changes in temperature over a 2 mile distance on those clear calm mornings is incredible. The official HCN station is at the university where it is much warmer on clear mornings vs rural locations. 120 years ago when they started records at this station it was mainly farm land there...yet there is an adjustment downward at that station too 120 years ago. How can this be? It over compensates for the UHI at this relatively small town. The arctic regions have a long winter where the boundary layer is stable so UHI effects have to be large. At least large enough for error bars to be large. This explains why minimum temperatures have gone up more and why the Arctic winter temperatures have shown the most increase.  And the oceans??? I won't even go there because the error bars are so large.

And yet 2014 was the hottest year ever by something like .03 degrees??? That is laughable.

 

I really am trying to get a handle on the magnitude of human induced climate change. I believe there is evidence for modest warming of 1 to 1.5 C at best and we have already seen about .5C or so. Plus the world's oceans are buffering the response of the climate system to the small external forcing of increased GHGs from humans and spreading it out over a longer period of time than the climate models show.  There is nothing that is alarming enough to wipe out the world's economy by abandoning fossil fuel use.

[skierinvermont]

The station move question was already answered. They are essentially treated as completely separate stations. This eliminates all possible bias.

[FloridaJohn]

I read Judy Curry's piece from ORH and it makes sense for land readings. Still I don't see how there can't be large error bars.

Those two sentences contradict each other, at least to me. If you read it and understand it (and it makes sense to you), then you should be able to see why there are not large error bars.

[FloridaJohn]

There is nothing that is alarming enough to wipe out the world's economy by abandoning fossil fuel use.

 

No one has proposed such a thing. Moving away from fossil fuels will not "wipe out the world's economy." Multiple studies have shown that doing nothing will be more expensive for humanity, and markedly worse for the world's economy.

[blizzard1024]

No one has proposed such a thing. Moving away from fossil fuels will not "wipe out the world's economy." Multiple studies have shown that doing nothing will be more expensive for humanity, and markedly worse for the world's economy.

This belongs on another thread   but energy = prosperity. You are dead wrong here. Solar, wind and renewables are not even close to replacing fossil fuels. We are going to have to live with fossil fuels for at least another several decades unless there is a major break through in renewable energy efficiency which is possible I suppose but not likely(I hope I am wrong here)....

[skierinvermont]

This belongs on another thread   but energy = prosperity. You are dead wrong here. Solar, wind and renewables are not even close to replacing fossil fuels. We are going to have to live with fossil fuels for at least another several decades unless there is a major break through in renewable energy efficiency which is possible I suppose but not likely(I hope I am wrong here)....

 

Wind energy is less expensive than coal in most markets. Solar is getting close.

 

You're about 15 years out of date with your facts. Welcome to 2015. 

 

There was more new wind power than any other power source created in the last 5 years. It is the fastest growing source of energy, not just in relative terms, but in absolute terms. Why? Because it is the most economical in the current market.

 

And with Tesla's foray into the affordable electric car market, it becomes affordable to drastically reduce transportation as a source of fossil fuels. 

[nflwxman]

Wind energy is less expensive than coal in most markets. Solar is getting close.

 

You're about 15 years out of date with your facts. Welcome to 2015. 

 

There was more new wind power than any other power source created in the last 5 years. It is the fastest growing source of energy, not just in relative terms, but in absolute terms. Why? Because it is the most economical in the current market.

 

And with Tesla's foray into the affordable electric car market, it becomes affordable to drastically reduce transportation as a source of fossil fuels. 

 

We can produce wind at about 3 cents a kWh in many markets.  This is essentially equal to Natural Gas at this point.