What is the real state of the climate?

The Czar of Muscovy has a very thought provoking post up about what's going on in Climate Science.  You should go RTWT, as it has triggered the thoughts I'm jotting down here.  It's really quite moderate, and covers the high points exceptionally well.

It made me think that I should drop the snark for a moment and do a survey of my best guess as to what the climate is up to, and whether Anthropogenic Global Warming (AGW) - warming due to man's release of carbon dioxide - is real.  So as near as I can offer a straight up Climate Science 101, here it is.

Disclaimer: I'm not a Climate Scientist; nor do I play one on TV.  An Electrical Engineering/Economics/History background has given me some analytical tools that I rely on in my more serious posts, but I don't try to follow the climate statisticians.  Treat this as an educated layman's view.

1. Is the climate changing?

It sure is.  Nobody (pro-AGW or skeptic) disputes that we've recently emerged from the Little Ice Age, which ended maybe 150 years ago.  The historical record is indisputable that it was much colder 200 years ago - stories of frozen Dutch canals and winter carnivals held on the Thames ice are easy to find.  It simply doesn't get that cold these days.

Going back a little further, nobody really disputes the Medieval Warm Period, which seems to have been as warm or warmer than things are today.  The MWP was entirely uncontroversial in the 1990s until Michael Mann's 1998 "Hockey Stick" graph seemed to say that it wasn't all that warm.  Lots of subsequent research shows that Mann et al were wrong on that bit, and even Mann's new 2009 paper has the MWP back.

So on a scale of two or three centuries, there's no dispute at all that things are significantly warmer than they used to be.  There are also shorter term changes, most notably the El Nino/Southern Oscillation (ENSO), which switches between warmer and cooler on a roughly 30 year interval.  Note that there are other, less significant cycles as well, which makes climate reconstruction pretty chaotic and non-linear.

2. What about the last century?

Looking at things over 200 years is not really very interesting.  So what about the last 100 years?  After all, the temperature data that we have is pretty spotty if you go back 150 years; we're on much firmer ground if we look at the period starting around 1880 (or even better, 1910).  This is more of a mixed picture.

2a. What data do we have?

There are three major historical temperature databases:

1. The Global Historical Data Network (GHCN), maintained by the US National Oceanic and Atmospheric Association (NOAA).  GHCN contains records from many hundred (1600 or more) US weather stations and more from other countries.

2. GISS Surface Temperature Analysis (GISTEMP), maintained by NASA's Goddard Institute for Space Studies.  It has been providing climate analysis since the 1970s.

3. The (UK) University of East Anglia's Climate Research Unit (CRU) provides both land surface temperatures(CRUTEM) and combined land and sea surface temperatures (HadCRUT). The CRU data is the data set relied on by the UN IPCC.

There seems to be considerable data sharing between these data sets.

2b. How complete is the data?

This is where we enter our first controversy.  Everyone agrees that the data is incomplete.  Weather station data has gaps (perhaps someone was sick one day, and didn't read the thermometer), stations are sometimes relocated (is the new site the same as the old one vis a vis temperature?), equipment gets upgraded (does the new equipment read the same as the old?).  Nobody disputes this.

What is controversial is how these gaps and discontinuities are treated.  Each of the major data sets has a method to "homogenize" the data, where the data is adjusted to (hopefully) address problems.  Some skeptics (full disclosure: including me) think that adjustments are excessively aggressive, and have led to a notable mismatch between "raw" data and "adjusted" data.  The CRU announced in 2009 that they have lost the raw data, so it's not possible to check whether adjustments were made properly or not.  It is claimed that the data could be reconstructed, but as far as I know, nobody has done this yet.

Another problem area that doesn't get much attention (but should) is that each of the data sets strives to provide a global view of temperature.  We simply don't have very good global coverage by weather stations.  Europe and North America in particular are over-represented in the data, and the oceans are very under-represented.  To deal with this, the globe is divided into areas called "grids" roughly 1200 km square.  Weather stations within a grid are averaged ("homogenized"); grids without any station data are interpolated from nearby grids that have stations.

A skeptic argument that refuses to go away is that arctic areas are under-represented, and data is interpolated from grids further south.  This is a problem for trend analysis when you realize that the number of stations used in the data sets has changed.  For example, GHCN dropped from around 5000 stations to 1500 or so.  If stations at higher latitudes or higher elevations were removed, then a spurious warming trend would appear.

2c. What can be done about the data problems?

Automated data collection has been implemented in the last two decades or so, removing the need for manual thermometer reading by weather station personnel.  Some (perhaps many) thermometers were relocated during this upgrade process, which seems to have introduced some discontinuities.  Long term trending data from these new automated stations, however, should be more accurate.  Whether you can easily compare data from station XYZ before and after the move is an area of some controversy.

New data collection types have come online as well.  Satellites (UAH) now collect air temperature (technically the temperature of the tops of low level clouds) for the entire globe.  These records go back to 1979, and frankly are an improvement if only due to the real global coverage.

Sea temperatures are now measured by the ARGO network of floating robot sensors.  This is exciting because the oceans can store much, much more heat than the air, and ARGO now can comprehensively measure this worldwide.  The system is still brand new, only reaching full operational capability in 2007.

3. What about older data?

Proxies are use to get an idea of temperatures in pre-thermometer eras.  Tree rings, ice cores drilled from glaciers, and pollen counts from bogs are all used.  The action, though, is really around the questions "what's been happening since 1910?" and "why?".  If we can resolve questions about the existing data, we don't need proxies; if we can't, proxies might not help.

4. How much confidence do we have in the data?

This is probably the question.  One of the good things that we've seen in the climate science debates during 2010 is a renewed focus on the uncertainties of the calculations.  Dr. Judith Curry from Georgia Tech is perhaps the most prominent non-skeptic scientists on this topic.  There isn't yet a consensus on how to approach this, or how it should be reported in the press.

5. What about Carbon Dioxide?

It is indeed a greenhouse gas, and the planet would likely be frozen without the greenhouse effect (maybe 15°C).  90% of this effect comes from water vapor, about 8% from carbon dioxide, and the rest from other trace gases.

6. Why the focus on carbon dioxide, if 90% of the effect comes from water vapor?

The hypothesis is that "positive forcings" (feedback loops) mean that small increases in CO2 will result in large increases in temperature.  This is extremely controversial, and is largely based on computer model simulations rather than on experimental observation.  Different values for the feedbacks give wildly different values of temperature increase from a doubling of CO2, from less that 1°C to over 5°C.

7. Does all the increased CO2 come from us?

No.  Most comes from natural sources; the biggest of these is likely out-gasing from the oceans as their temperature rises.  Ice core data suggests that CO2 increases several hundred years after temperature rises (approximately 800 years later).

That is a quick survey of the key climate science issues, presented as neutrally as I can.  I've left a huge amount out, but this is probably what an educated layman absolutely has to know to follow the debate closely. 

The Czar is correct that the politics of the debate have gone beyond crazy:

The second part of our opinion falls on the politicization of climatology. This is pretty bad; in fact, it is as bad as we have ever seen it. Some scientist in Jakarta reports that yesterday’s noon temperature was a degree hotter than it was a year ago, and within days some governmental agency is demanding that all profitable American industries shut down, and billions of American dollars be dumped on a Nigerian ex-secretary of the treasury who is certain he can get his gold back and split it with you. The paranoia and hysteria is incredible. Be worried about if A then Z logic. Ask to see the letters in between.

Worse, if you express any doubts about the logic here, liberal noise makers start screaming at you that you fail to understand the science! And, frankly, the Czar has a pretty good background in climatology and meteorology (the Mandarin, too, by the way), and finds the scientific arguments pretty unconvincing, or extrapolated past sanity. Especially on the anthropogenic part.

My personal feeling here is that a major politicization 10 years ago seized on the then warmer temperatures (1998 was a "super El Nino" event) to get press and public attention.  This seems to be cutting both ways now, as a series of three cold winters in a row have undermined public confidence in the climate predictions.  But that's just opinion.