|
Alternative Climate Thinking |
|
Join the Fight Against Hyperbole and Propaganda in the Sciences |
|
proxies |
|
What is a Proxy? Most simply put, a proxy is something that represents something else indirectly. Some examples include: Proxy voting – where a designee is authorized to cast a vote on behalf of an absentee (sounds almost illegal). Proxy fights - which sound a lot more interesting than they really are…unless perhaps you’re a stockbroker. Proxy murder – hired assassins make quite a good living performing these. There’s also the double proxy murder where two people, who are strangers to each other, exchange murders…leaving each murder without means or opportunity (generally there is motive, however). Proxy war - where two powers of some kind use third parties as a substitute for fighting each other directly. Both China and Russia were famous for this kind of proxy fighting through the cold war era (Korea and Vietnam both come to mind). You’d think that a game of chess could proxy for real war. If only the big wars could be settled with a “bishop captures knight – mate!” Of course we have all heard of (or are part of) health care proxies - in which a legal document is used to specify an “agent” to make medical decisions on behalf of a incapacitated patient (Congress calls this The Health Care Reform Bill). There are even sexual proxies that we all have been guilty of using from time to time (feet for men and noses for women) with varying degrees of success. And finally, there is what we are here to elaborate on - proxy climate. Climate proxies are sources of climate information from (mostly) natural records which can be used to estimate climate conditions prior to the most recent period. For a proxy indicator to be useful, it should (must) be calibrated against modern instrumental (chemical, physical, and biological) data in order to provide a quantitative reconstruction of past climate. And this, to some extent, makes the assumption that the chemical, physical and biological world was exactly the same at a given point in space, then as now. We’re not saying that the physical laws were any different in the past; however, there is no way to know the exact state of the “system” at a given point in history without direct observation. There can only be assumptions based on present spatial and temporal information. In addition, since most of the experimental samples analyzed have been retrieved from some depth in buried strata (earth, rock, ice, etc.), and the biogeochemical processes change with both time and pressure. There is also the consideration of the sample itself and its preservation as it is brought to the surface. Bottom line; there is always uncertainty associated with proxies, so be ware. Before we get into the many different types of climate proxies available, let’s key in on two of the highlighted words in the previous paragraph – estimate and quantitative. No matter what anyone says, or what numbers and graphics come flying out of the data analyses – climate proxies ARE ONLY ESTIMATES – and can never be anything more than that! Now, they can be good estimates (or bad ones), but by their very nature they are, at best, only approximations of past conditions calibrated from the chemical, physical and biological processes of today (or other assumptions from relatively low resolution fossil/geologic data) . What makes all of this complex and interesting is that these estimates are consistently reported in a quantitative manner, which suggests objectivity instead of subjectivity – and which is not completely true. In the world of climate, quantitative implies precision (but not accuracy). So what if it’s reproducible, if it’s way off mark to begin with (or end with). We would feel much more confident in this type of data analysis if it was always presented in relativistic (trending) terms, but that’s not the nature of the beast. There are many proxies used for representing past climate. Some, like the flora and fauna of the past as represented in the geologic (fossil) record, are relatively easy to understand and offer a glimpse into the world as it was tens of thousands to millions of years ago. This is considered very low resolution data (relatively little variability) and generally only provides ranges of climate over relatively long periods of time (borehole data and glacier movements are good examples). However, it is useful when looking at the evolution of the Earth over its entire history. At the other end of the spectrum, there are higher resolution proxies that involve complex physical and chemical relationships in which changes in oxygen isotope ratios can give an indication of past (local) temperatures. Tree ring proxies are also an example of high resolution proxies that are able to resolve year-to-year changes; however they also involve the use of statistical techniques that lose much of the longer-term temperature detail in parts of the record. Nevertheless, it must be kept in mind that no class of proxy (low, medium or high resolution) is explicitly correlated to temperature (and/or CO2) under all conditions. The List – Proxies in alphabetical order: Borehole Proxy – This is one of the more interesting proxies in that the data is found from direct measurement of temperature inside boreholes drilled into the Earth’s crust. The correlation with the past is found from the difference in the expected increase in temperature with depth (the geothermal gradient) which can then be interpreted in terms of changes in temperature at the surface, in the past. (The idea is that this heat energy has slowly diffused downward over the years, warming (or cooling) the soil layers below the surface.) Data is available for the past 150 years or so. Find more information at: http://www.geo.lsa.umich.edu/climate/. Coral Proxy - Consist of stable isotope (dO18 to dO16 ratios) and trace metal (Strontium to Calcium and Magnesium to Calcium ratios) analyses from corals located around the globe. Corals are proxies for sea surface temperature and salinity in the upper ocean environment. Data for this proxy is available for about the last 200 to 300 years. Not the best proxy for atmospheric temperatures, however. More information can be found at: http://www.ncdc.noaa.gov/paleo/reports/trieste2008/corals.pdf. Floral and Faunal Data Proxies – Is past climate reconstruction based on plant and animal fossil information. Most of this data set is made from preserved remains large enough to be visible without a microscope. They include plant and animal life (including insects) formerly living and growing in the region of concern. Together, this information along with pollen data can be used to reconstruct a terrestrial environment of the past. For more information, see: http://www.ncdc.noaa.gov/paleo/fauna.html http://www.ncdc.noaa.gov/paleo/plantmacros.html http://www.ncdc.noaa.gov/paleo/insect.html Historical Proxies - Includes records of paleoclimatic markers obtained from human historical references and documentary evidence, such as hunting and fishing records, harvest dates and records, and oceanic ice on (free) dates. (This would include much information concerning the Mideval Warm Persiod (MWP) and the Little Ice Age (LIA) Trieste 2008 Workshop Report on Documentary Data . Ice Core Proxies – One of our favorites and one which we will expand on later. The climate indicators that come from ice cores include oxygen isotopes, methane concentrations, and dust content to name a few. Trieste 2008 Workshop Report on Ice Cores Lake Levels (Sediments) Proxies - As water levels in lakes fluctuate with changes in moisture balance (precipitation minus evaporation) over time, so do the fossil shoreline deposits and other features that are indicators of past moisture balance as well as climate within the lake’s basin. As with both coral and ice proxies, stable oxygen isotope (dO18 todO16) and trace metal analysis are the two primary means of reconstructing past climate histories. This is often considered a subset of paleolimnology which will be described in more detail below. Loess and dust proxies – This data provides vivid examples of how changes in climate can affect of processes that shape the landscape. Some useful chemistry analyses can also be performed. Adjacent layers are analyzed to show the evolutional history of the winds as well as other processes dealing with terrestrial surface deposition over time. Not the best temperature proxy source, but useful for background information. Paleofire Proxies – Are proxies constructed from fire history data derived from other proxy sources, such as tree scars (tree ring analysis) and charcoal (in sediment) records. For more information, go here: http://www.ltrr.arizona.edu/sngc/studies/pftrd.htm Paleolimnology proxies – Consist of multiple data sets from inland fresh, saline and brackish water bodies. A wide variety of measurements from lake and bog sediments are used to designate past water temperature, physical properties, biology, and chemistry. http://www.indiana.edu/~diatom/paleo.html Paleoceanography proxies – Are a superset of proxies produced from data derived from many different proxies found in deep sea sediments including trace metal and isotopic composition of fossil plankton, species composition, and lithology (rock strata, composition and physical makeup). http://www.tos.org/oceanography/issues/issue_archive/issue_pdfs/19_4/19.4_dybas.pdf Pollen Proxies – Derived from pollen grains washed or blown into lakes and accumulated in sediments. These can provide a record of past vegetation. Obviously, different types of pollen in lake sediments reflect the vegetation that was present around the lake at a given point in time, and thus the climate conditions can be determined (within a range) that would favor given vegetation. Considered relatively low resolution data. http://www.ucar.edu/learn/1_2_2_10t.htm Speleothems (aka Cave deposits) - Deposits formed from groundwater within underground caverns which include, but are not limited to, stalagmites and stalactites. Many are stratified and contain certain compounds which can be radiometrically dated. From the thickness of depositional layers as well as the isotopic records, climate proxies can be made. While speleothem δ 18O of calcite data is considered good data for temperature proxies, speleothem δ13C data is not (important later on). See, http://www.nps.gov/archive/wica/Speleothems.htm and http://www.gsf.fi/esf_holivar/baker.pdf . Tree ring proxies - Last, but certainly not least, is the most (in)famous of proxies. Tree ring proxy correlations are drawn from ring width or wood density measurements. On the plus side, tree ring analysis is relatively easy to go about…at least in the beginning. There’s no digging, drilling or deep sea diving to perform, although getting to suitable trees may involve hiking into some pretty remote areas. Tree slices or cores are removed and then taken back to the lab for analysis. On the minus side, trees only grow on land and since a little over 70% of Earth is covered by oceans, seas and lakes, tree rings, at best, can only tell us about a small part of it. In tree ring analysis (dendroclimatology), it is vital that every physical and geographical detail of the site be documented very carefully, such as: slope orientation, soil/bedrock composition and even proximity to other trees (all will make a huge difference in tree ring growth). The general and most simple proxy correlation regarding tree rings and temperature stems from the assumption that the warmer the environment, the greater (wider) growth of the tree. Again, this is grossly oversimplified as tree ring growth is influenced by a great many other factors other than temperature itself, such as : precipitation amount and timing, solar availability (sunlight and clouds), pests, competition, forest fires, soil nutrients, and even the annual duration of snow on the ground. In the end, all a tree ring can really tell us is whether or not the biogeochemical and physical conditions during the growing season were favorable to tree (ring) growth. It is this great complexity of the real physical and biogeochemical science behind tree (ring) growth which brings in a certain amount of “quantitative subjectivity” with regard to corrections to the raw data in order to produce a proxy that works. Other problems are both spatial and temporal in nature. Trees that are good candidates for this type of proxy (undisturbed longevity for one) are relatively few and far between, so the density of the data network is rather sparse. Trees are also picky in as far as when they like to grow – during the day and…well… over the growing season (duh). This seems to eliminate a goodly portion of the year inasmuch as both nocturnal and wintertime temperatures make up a good portion of the global mean temperature (on any annual or greater timescale). Even on the surface, tree rings do not seem the best source for temperature proxies and as we will see a little bit latter, other proxies such as isotopes in coral, ice, minerals and sediments are a much better choice. For general information on tree rings - http://www.icogitate.com/~tree/treerings.ac04.htm The Hockey Stick No discussion concerning the subject of tree rings would be complete without addressing the recent use and abuse of tree ring data to further the global warming agenda and even set world policy. This is, known as the global warming “hockey stick.” The so-called “hockey stick” is a graphical representation of a proxy reconstruction produced by Michael Mann, Raymond Bradley and Malcolm Hughes (MBH) in a paper titled, “Global-scale temperature patterns and climate forcing over the past six centuries,” published in Nature (volume 392, 23 April 1998, pages 779-787). It has since been used by the United Nations' Intergovernmental Panel on Climate Change (IPCC) in its past two (third and fourth) assessment reports, published in 2001 and 2007 respectfully, to influence world policy regarding the issue of global warming (most notably the Kyoto Protocol and most regrettably Al Gore’s horribly flawed motion picture titled “The Inconvenient Truth” that ironically led him and the IPCC in sharing the Nobel Peace Prize in 2007). The premise of the Mann, et al. study showed relatively stable temperatures from A.D. 1000 (and in more recent versions, from 0 A.D.) to about A.D. 1900 with a subsequent, and dramatic, increase in temperature between A.D. 1900 and 2000. The conclusion drawn by authors of their reconstruction is found most obviously and directly from Mann’s own quote, “Our results suggest that the latter 20th century is anomalous in the context of at least the past millennium. The 1990s was the warmest decade, and 1998 the warmest year, at moderately high levels of confidence." And, this conclusion is being directly attributed to the human technological presence and the burning of fossil fuels over the past 100 to 150 years. Please take note that this climate reconstruction was only done for the northern hemisphere – not the entire globe as we have been led to believe. Although the MBH research has been found faulty in all of its various incarnations over the last dozen or so years by a number of individuals, the idea behind its powerful visualization has never been completely dropped and still has many political, scientific and media circles who continue the propaganda behind this hypothesis even to this day. So just what went wrong? A treatise could be written on what NOT to do in science, but you’ll have to settle for the highlights in this case instead. Readers, however, are encouraged to follow this story in more depth - to gain more insight into this grossly misleading (and often unscientific) process. Back in the mid 1990s, scientists thought the world was warmer in medieval times – aka Medieval Warm Period (MWP) – which it was. However, it just so happened that CO2 was on the low side at that time and didn’t fit with the climate models. In 1998 Michael Mann and company (MBH) decided to ignore this past climate history and produced the “graph that scared the world” – a proxy reconstruction (using choice tree ring data) that showed the “1990s was the hottest decade for a thousand years.” Now the warming temperatures (more or less) over the past century and a half “fit” the current rise in carbon! The rest, shall we say, is history: The IPCC used the information in their 2001 Third Assessment Report; world governments copied it and distributed it; and the media went wild and told everyone. A star was born! In this particular case, however, it was to be a slowly falling star. (It should also be noted that the Little Ice Age (LIA) that followed the MWP was also not captured in the Mann et al. analysis.) Everything was just fine and dandy for the next few years. Lots of back patting, handshaking, and “atta boys.” Grant money would now flow in virtually unimpeded and bring much needed cash flow to the entire field of climatology in general. Would the accolades ever end? They would…or at least they would slow down a bit. A couple math/statistics wizards from north of the border (Canada, for those of you from elsewhere) decided to take a closer look at the analysis behind the graph and test it…see if it was really what it appeared to be. What they found was a real shocker of the time and also put them on a path of vindication (and sometimes retribution) that continues to this day. It all started with a publication McIntyre and McKitrick produced in 2003 titled, “Corrections to the Mann et. al. (1998) Proxy Data Base and Northern Hemispheric Average Temperature Series” (McIntyre S & McKitrick R, Energy & Environment, v 14, num. 6, pp 751-771 (2003)) that showed a wide variety of statistical and mathematical errors along with poor data handling, the use of old/poor proxies, mixing proxy and non-proxy data, and terrible quality control. McIntyre and McKitrick also found that the actual values for the MWP were higher than for any period in the 20th century and that the hockey stick shape of the MBH 1998 (proxy) reconstruction was caused by cherry picking data and the incorrect use of principle component analysis (PCA) which Mann had purported to use. See the McIntyre and McKitrick paper in its entirety here: http://www.uoguelph.ca/~rmckitri/research/MM03.pdf It’s the incorrect use of PCA that actually “forces” the hockey stick shape to occur. PCA is used to find prominent features of a data set, but what Mann and company did was to use wrong (irregular) data normalization procedures. What happened is this. Standard normalization calls for each data set to be normalized over the entire data period. This is important, since the original Mann data period should have been A.D. 1400 to A.D. 1980. To do this, each data set that they were to use would have their averages subtracted so that they would have a mean of zero. This new data would then be multiplied by a number such that the average variation around the mean would equal one. This is called normalizing the data set to zero mean and unit variance. What Mann, Bradley and Hughes (MBH) did instead was to normalize each data set to have a zero mean for the period 1902 to 1980 to more or less match the instrumented records over this time. This was the “trick” in the normalization procedure that forced the hockey stick shape. McIntyre and McKitrick figured this out by putting in some arbitrary red noise (random walk noise) test data, which had no trends, into the MBH procedure…and lo and behold what should appear? The dreaded hockey stick shape! Note: MBH didn’t use only tree ring data/proxies, however, it must be pointed out that much of the tree ring data was spurious and that many of the kinds of trees they chose were actually very well correlated to CO2 (increase CO2, increase growth) and NOT temperature! Bottom line, it was the persistent properties of the tree ring data used that primarily forced the shape of the analysis, not the cumulative properties of all the proxies. Two more papers by McIntyre and McKitrick followed in 2005 that reasserted their position on the Mann study and added more damage. The first was “Hockey sticks, principal components, and spurious significance” (Geophysical Research Letters, Vol. 32, L03710, 2005) (http://climateaudit.files.wordpress.com/2009/12/mcintyre-grl-2005.pdf ) which showed that the evaluation of statistical skill for the Mann et al. analysis was grossly overestimated using the reduction of error (RE) statistic and that validation should include statistics, such as the R2 (coefficient of determination) and CE (coefficient of efficiency) statistics (for cross validation) which would have pointed to flaws in their method much earlier. The second 2005 paper by the two Canadians, “The M&M Critique of the MBH98 Northern Hemisphere Climate index: Update and Implications” (Energy & Environment, v 16, no.1, pp. 69-100, 2005 – found here: http://climateaudit.files.wordpress.com/2009/12/mcintyre-ee-2005.pdf ) addressed some issues initially raised in the 2003 paper that were unresolved. Some of the more important points: Actual methodologies used in the Mann et al. analysis were different from what was stated in their paper. Modification of the PCA for North American Tree Ring Data caused this PC to be “dominated by a subset of bristlecone pine ring width series which are widely doubted to be reliable temperature proxies.” (As stated earlier, this comes partly from the fact that bristlecone and foxtail pines are much better correlated to atmospheric CO2 – fertilization - than to temperature.) For the Gaspé cedar ring width series, another tree ring proxy used, MBH did not use archived data. Instead they (MBH) did an extrapolation and essentially changed the start date of the series. M&M also pointed out the lack of independence in supposed “independent” studies that were said to support the MBH work – including two papers that Mann himself was co-author on. Finally, they end the paper with examples of the failure of the peer review process. Others who have refuted MBH 1998 Dr. Edward Wegman, (George Mason University) Chairperson of the National Academy of Sciences (Applied and Theoretical Statistics) who with Doctors David W. Scott (Rice University), and Yasmin H. Said (The Johns Hopkins University) were asked by congress to review the work of MBH. Some of the more important findings are as follows: The MBH method creates a hockey-stick shape even when supplied with random input data; that the MBH method creates a PC1 statistic dominated by bristlecone and foxtail pine tree ring series (and which there are relatively few of); there is also evidence in the literature, that the use of the bristlecone pine series as a temperature proxy may not be valid (i.e. suppresses "Medieval Warm Period" in the hockey stick handle); that bristlecone pines exhibit CO2 fertilized growth, thus falsely enhancing warming in the hockey stick blade; that there is no evidence that Mann or any of the other co-authors have had significant interactions with mainstream statisticians; that the sharing of research materials, data, and results was done haphazardly and begrudgingly by Mann and Co. Bottom line: That Mann’s conclusion, that the decade of the 1990s was the hottest decade of the millennium and that 1998 was the hottest year of the millennium, cannot be supported by their analysis. This review also strongly supports the findings of McIntyre and McKitrick. Dr. Wegman has also been quoted since, with regard the MBH 1998, saying that “Method Wrong + Answer Correct = Bad Science.” The full report can be found here: http://republicans.energycommerce.house.gov/108/home/07142006_Wegman_Report.pdf Additionally, the shortcomings of the MBH 1998 methodology has also been (independently) confirmed in other publications by Lee, Zwiers and Tsao in the August 2008 issue of the Journal of Climate Dynamics titled “Evaluation of proxy-based millennial reconstruction methods” (http://www.springerlink.com/content/wj52683504977744/ ) and by Christiansen’s 2005, “The Shortcomings of Nonlinear Principal Component Analysis in Identifying Circulation Regimes” in the Journal of Climate (http://ams.allenpress.com/perlserv/?request=get-abstract&doi=10.1175%2FJCLI3569.1 ). Indirectly, Craig Loehle’s paper “A 2000-year global temperature reconstruction based on non-treering proxies” and “Correction to: A 2000-year global temperature reconstruction based on non-tree ring proxies” in Energy & Environment (http://www.freesundayschoollessons.org/pdfs/climate-history.pdf ) shows that the nonlinear response of trees to temperature explain a (mathematical) divergence problem, and why some trees are not good proxy “thermometers.” “Proxy climatic and environmental changes of the past 1000 years,” by Willie Soon and Sallie Baliunas, explain that due to differing natures that proxy indicators cannot be combined into a hemispheric or global quantitative composite and that many records reveal that the 20th century is probably not the warmest nor a uniquely extreme climatic period of the last millennium. http://www.int-res.com/articles/cr2003/23/c023p089.pdf Even Hans von Storch, a German born professor of Meteorology and a member of the advisory boards of the journal Journal of Climate (and pro global warming advocate) has claimed that the method used by Mann et al. probably underestimates the temperature fluctuations in the past by a factor of two or more…effectively eliminating large temperature variance (like during the MWP) that would normally show that the 20th century is not abnormally warm and Mann's so-called "hockey stick" would therefore be a much weaker argument for recent extraordinary climate change. It should be noted that due to the large uncertainty surrounding the MBH 1998 work, their method has NEVER been used by others for similar studies. Yet, the results are still being sold (politically and even psudoscientifically) as gold. The new and Improved Hockey Stick The newest Mann offering came out in the late summer of 2008 and is titled “Proxy-based reconstructions of hemispheric and global surface temperature variations over the past two millennia” (PNAS September 9, 2008, vol. 105, no. 36, pp. 13252 – 13257)…followed up by yet another gem, “Global Signatures and Dynamical Origins of the Little Ice Age and Medieval Climate Anomaly,” (Science, 27th November 2009). This is still a story in progress, yet ten years in the making. Of course, the underlying reasoning behind this “new” study was to find the same data/proxy correlation(s) that MBH tried to do back in 1998. This time, however, they were to do it with better statistics and with less reliance on tree rings. Have they succeeded? Well, yes and no. Yes, they essentially achieved finding a way to mimic their earlier work, but have one again sacrificed good science in doing so. McIntyre and McKitrick (who else) have since addressed this new work in a letter to the PNAS titled “Proxy inconsistency and other problems in millennial paleoclimate reconstructions” (PNAS February 10, 2009, vol. 106, no. 6) which points out some serious new issues, including: Statistical confidence interval problems; calibration/correlation problems, similar to their previous work, which auto-generates the pre-determined “hockey stick” shape; and, most interestingly, that the new study has used some sediment proxies with their axes flipped literally upside-down along with some data that is not even qualified to use as a temperature proxy (e.g., speleothem δ13C). As this discussion/debate is still underway, we will add more as news develops. You can probably be sure of a few things, however. This new Mann et al. method will not be used in any other new studies and I would seriously doubt at this point that the IPCC will include any of these new graphs as reference in their upcoming Fifth Assessment Report (2014)…if they still have any hope of being taken seriously by anyone. To keep up to date on the new aspects of this case from McIntyre and McKitrick’s point of view, go to climateaudit.org (http://climateaudit.org/ ) and start looking and reading.
|
|
To contact us: |
|
E-mail: contact@climatethoughts.org |
|
-What is the shortest book in the world? "The Environmentalists' Book of Jokes", of course. |
