Review of an article on climate change
I remember being introduced to the term 'global warming' way back when, probably sometime in the Eighties. Back then, it wasn't kosher to say that anthropogenic sources of carbon dioxide were exacerbating the 'greenhouse effect', just as cigarette smoking may not definitely cause lung cancer. It was with some degree of interest that I started going through the various chapters of the IPCC's Fourth Assessment Report in class, aided by my lecturer and classmates in sifting through the dense scientific language and physical concepts of climate research.
The class assignment seemed interesting enough: review a paper by a team of climate change skeptics. On first impression, the paper seemed decent enough- it survived a peer-review process after all, even if it was published in some medical journal. I set aside my doubts about what physicians and surgeons know regarding climate change and delved deeper. Red flags started popping out when I found out that the journal wasn't even listed on MedLine, and is apparently a propaganda rag for the right-wing with dubious peer-review practices.
Then I read the paper. Initially I was disappointed that the authors weren't putting up much of a (intellectual) fight against anthropogenic causes of climate change, but when the errors started becoming more blatant and egregious, I started getting angry. How can this even be considered a scientific work? Even with my limited knowledge in climate issues, some things were just plain wrong. In any case, I've decided to post my review here for you to judge, and hopefully to provide some measure of reason in this madness. A small voice perhaps, but surely it's better to light a small candle than to curse the dark?
Here's the paper I'm reviewing, which you should read before proceeding:
Robinson AB, Robinson NE, Soon W (2007) Environmental effects of increased atmospheric carbon dioxide. Journal of American Physicians and Surgeons 12: 79-90
And my review:
Be warned! It's pretty long!
In science, it is always good to question the prevailing dogma, as this leads to progress. However, refutations should be accompanied by clear and compelling data. The former are certainly abundant throughout the paper, but unfortunately, the latter appear to be lacking.
The scope of the article is a general statement about the effects of carbon dioxide on global climate, but instead of producing globally averaged data, the authors tended to showcase only regional data. While the Sargasso Sea is a large body of water, it cannot solely represent the entire Earth’s temperature changes, which do not occur uniformly over the world’s oceans and continents (Christensen et al 2007). No references were given for the statement on page 1, paragraph 3, that “The average temperature of the Earth has varied within a range of about 3oC during the past 3000 years”, as well as two statements on the Sargasso Sea, that “Figure 1 (temperatures in the Sargasso Sea) is illustrative of most geographical locations…” on page 2, paragraph 2, and “The 3000-year range of temperatures in the Sargasso Sea is typical of most places” on page 4, paragraph 6. Confusingly, the last two unverified statements are immediately followed by the contradictory assertions that “…there is great variability of temperature records with location and regional climate” and “Temperature records vary widely with geographical location…”. The author of the original paper from which the Sargasso Sea data arose also concluded based on his data that the changes described for the “surface waters over the Bermuda Rise…most likely reflect climate change on the basin or hemispheric scale” (Keigwin 1996).
The use of regional data to approximate global trends continues in Figure 3, where Arctic air temperatures are compared with solar activity, looking at surface temperatures in the United States, and comparing Antarctic ice core temperatures to global emissions of carbon dioxide. According to the unreferenced data in Figure 6, variations in global temperatures can be extreme (more than 100oC), and thus it seems contradictory and unrealistic to extrapolate regional information to the global scale. In light of this, as the geographic range of the localities used in the query in Table 1 are unreported, one is unable to place the results presented in the table in the global perspective.
Underreporting of data is not restricted to spatial scale, and extends to other figures as well. Graphs attempting to show trends between hurricane phenomena and time only reported the occurrence of hurricanes that made landfall (Figure 9), or ‘violent’ hurricanes (Figure 10). Similarly, only category F3 to F5 tornados are included in Figure 8. These strong to violent tornados are relatively rare in the United States (about 2% according to http://www.ncdc.noaa.gov/oa/climate/severeweather/tornadoes.html), no waterspouts were included in the figure, and it is not immediately apparent why other types of hurricanes and tornados were excluded from analysis. In another example, no species data available for the plant experiment in Figure 23, which was used to extrapolate the global plant response to increased atmospheric carbon dioxide.
A significant number of figures in the paper attempt to demonstrate trends in climate variables over time. It is disappointing to observe that no statistics are presented for most of them (e.g. Figures 2-5, 7-15). While it is possible to fit a straight line through any scatterplot and measure its gradient, knowing the standard errors, sample size, and regression and/or correlation statistics would allow the reader to judge the statistical significance of the trend. Additionally, sampling artifacts were not accounted for. For example, the variable of average rainfall over the contiguous United States (Figure 7) subsumes many geographically distinct areas within this region, and it is unknown if sampling effort was even across space and time, which would impact the variance within the data.
Aside from the examples brought up earlier, there are other statements in the article that do not have references. The authors assert on page 3, paragraph 4 that carbon dioxide acts as a fertilizer for plants, improving plant growth, and leading to positive cascading effects down the food chain. Most astounding is the statement that “The extent and diversity of plant and animal life have both increased substantially during the past half-century”, which is clearly at odds with the realities of widespread habitat destruction, decline in fisheries, and catastrophic extinctions in recent times (e.g. Brook et al 2003, Sala and Knowlton 2006). Also unverified is the argument on page 9, paragraph 1 that “Trees respond to CO2 fertilization more strongly than do most other plants, but all plants respond to some extent”, which is used to show the benefits of increasing atmospheric carbon dioxide concentration.
The significance of the trendlines asides, while it is true that correlation does not prove causation, non correlation does not prove non-causation, as the authors imply for the relationship between sea level rise and hydrocarbon use. The relationship may not be directly linear to begin with, and cannot be summarily dismissed. The authors’ stand would be clearer if they had provided a viable alternative hypothesis as to how ‘solar irradiance’ generates the largest radiative forcing effect on global climate. The consideration of only two factors in the control of global temperature (methane was brought up, and dismissed as being unimportant) is overly simplistic. The IPCC’s 4th Assessment Report on Radiative Forcing (Forster et al 2007) lists 16 factors that may impact climate on local and global scales. As a result of the multitude of factors that could lead to temperature changes, the comparison in Figure 16 is misleading. Any effect that carbon dioxide has on forcing global climate would be in concert with the other factors, which were unlikely to remain constant throughout the long duration of the comparison.
The authors reject predictions from computer-based climate models as incorrect, and lacking no empirical support. However, they offer no details on the kinds of errors arising from models, whether it is the assumptions made or overly simplistic cycles and feedback loops, or even which models they are specifically referring to. The computer model uncertainties listed in Figure 19 came from a paper written by one of the co-authors of this article and a personal communication from 1995, which probably were not original sources of the models’ data, and the latter reference is highly outdated. The statement that computer models have been “invalidated by numerous observations” was not accompanied by any specific references. No mention was made of the improvement in modeling expertise over time- current climate models are verified using hindcasts on data collected previously, before being used for future predictions (Christensen et al 2007).
The incomplete picture painted by the authors extends to the choice of data, which could bias the analyses reported in the article. The selection of gas, oil and coal use as a proxy for atmospheric carbon dioxide concentration in the correlation figures, and disregarding other inputs from biomass burning, cement production and deforestation, was not explained in the text. In Figures 3, 5 and 13, it is not clear how the various metrics of solar activity (5 listed by the authors), were summed into the single variable ‘solar irradiance’. It is also not apparent how glacier length is a good representation of global warming, seeing as how the authors themselves note that precipitation is a factor in glacier length. Furthermore, the response variable is described as ‘normalized glacier length’, without explaining what it was normalized to. A closer examination of the scale in Figure 2 shows that glacier length was measured at the precision of centimeters. Considering the massive sizes of glaciers, the magnitude of their seasonal variation in length, and the precision of the methods used to measure them, without any indication of the standard errors, we cannot ascertain if the differences in length shown in the figure are significant at all.
While standard error terms were missing from many of the figures, it is puzzling why the authors chose to use standard deviation of tree ring width as the metric to represent tree growth in Figure 21. The mean would give a much clearer picture of differences in growth, whereas standard deviation can change due to sampling artifacts. No units are given for the standard deviations as well. Other analyses are plagued by confounding variables. Although the graph purports to show trends in United States forests, the metric used is volume of timber, presumably with commercial value. One does not know if the increase in timber volume is due to an increase in timber plantations, a reduction in logging, or a response to increased atmospheric carbon dioxide as suggested by the authors (no references). The comparison of warming trends on other planets to Earth’s (page 5, paragraph 11) is unrealistic, given the vastly different atmosphere present on each planet. The attribution of some warming effect to the ‘urban heat island’ effect in Figure 15 is also confounded, as population changes over time, and the effects were not separated in the analysis.
Lastly, subjective terms that are not clearly defined crop up in the article, such as the “Qualitative Greenhouse Effect” presented in Figure 18. With no qualification of the term, or any measurement units, it remains unknown how the authors are able to come up with a scale and compare among the four categories presented. Carbon dioxide and methane are labeled as “minor greenhouse gases” without qualification as to whether the term refers to the gases’ relative atmospheric volume or radiative forcing effects. Statements like “The climate of the Earth is now benign” do not add to the scientific argument in the paper and show a tendency of the authors to editorialize.
Due to numerous omissions, inappropriate generalizations, a lack of clarity in the analyses and a dearth of statistical analyses of data, it is hard to conclude anything about the hypothesis that the authors have put forth. I find that the authors are unable to support their argument that increasing carbon dioxide has no deleterious effects on global climate.
References:
Brook BW, Sodhi NS, Ng PKL (2003) Catastrophic extinctions follow deforestation in Singapore. Nature 424: 420-426
Christensen, J.H., B. Hewitson, A. Busuioc, A. Chen, X. Gao, I. Held, R. Jones, R.K. Kolli, W.-T. Kwon, R. Laprise, V. Magana Rueda, L. Mearns, C.G. Menendez, J. Raisanen, A. Rinke, A. Sarr and P. Whetton, 2007: Regional Climate Projections. In: Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.
Forster, P., V. Ramaswamy, P. Artaxo, T. Berntsen, R. Betts, D.W. Fahey, J. Haywood, J. Lean, D.C. Lowe, G. Myhre, J. Nganga, R. Prinn, G. Raga, M. Schulz and R. Van Dorland, 2007: Changes in Atmospheric Constituents and in Radiative Forcing. In: Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M.Tignor and H.L. Miller (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.
Keigwin LD (1996) The Little Ice Age and Medieval Warm Period in the Sargasso Sea. Science 274: 1504-1508
Sala E & N Knowlton. 2006. Global Marine Biodiversity Trends. Annual Review of Environment and Resources 31:93-122
The class assignment seemed interesting enough: review a paper by a team of climate change skeptics. On first impression, the paper seemed decent enough- it survived a peer-review process after all, even if it was published in some medical journal. I set aside my doubts about what physicians and surgeons know regarding climate change and delved deeper. Red flags started popping out when I found out that the journal wasn't even listed on MedLine, and is apparently a propaganda rag for the right-wing with dubious peer-review practices.
Then I read the paper. Initially I was disappointed that the authors weren't putting up much of a (intellectual) fight against anthropogenic causes of climate change, but when the errors started becoming more blatant and egregious, I started getting angry. How can this even be considered a scientific work? Even with my limited knowledge in climate issues, some things were just plain wrong. In any case, I've decided to post my review here for you to judge, and hopefully to provide some measure of reason in this madness. A small voice perhaps, but surely it's better to light a small candle than to curse the dark?
Here's the paper I'm reviewing, which you should read before proceeding:
Robinson AB, Robinson NE, Soon W (2007) Environmental effects of increased atmospheric carbon dioxide. Journal of American Physicians and Surgeons 12: 79-90
And my review:
Be warned! It's pretty long!
In science, it is always good to question the prevailing dogma, as this leads to progress. However, refutations should be accompanied by clear and compelling data. The former are certainly abundant throughout the paper, but unfortunately, the latter appear to be lacking.
The scope of the article is a general statement about the effects of carbon dioxide on global climate, but instead of producing globally averaged data, the authors tended to showcase only regional data. While the Sargasso Sea is a large body of water, it cannot solely represent the entire Earth’s temperature changes, which do not occur uniformly over the world’s oceans and continents (Christensen et al 2007). No references were given for the statement on page 1, paragraph 3, that “The average temperature of the Earth has varied within a range of about 3oC during the past 3000 years”, as well as two statements on the Sargasso Sea, that “Figure 1 (temperatures in the Sargasso Sea) is illustrative of most geographical locations…” on page 2, paragraph 2, and “The 3000-year range of temperatures in the Sargasso Sea is typical of most places” on page 4, paragraph 6. Confusingly, the last two unverified statements are immediately followed by the contradictory assertions that “…there is great variability of temperature records with location and regional climate” and “Temperature records vary widely with geographical location…”. The author of the original paper from which the Sargasso Sea data arose also concluded based on his data that the changes described for the “surface waters over the Bermuda Rise…most likely reflect climate change on the basin or hemispheric scale” (Keigwin 1996).
The use of regional data to approximate global trends continues in Figure 3, where Arctic air temperatures are compared with solar activity, looking at surface temperatures in the United States, and comparing Antarctic ice core temperatures to global emissions of carbon dioxide. According to the unreferenced data in Figure 6, variations in global temperatures can be extreme (more than 100oC), and thus it seems contradictory and unrealistic to extrapolate regional information to the global scale. In light of this, as the geographic range of the localities used in the query in Table 1 are unreported, one is unable to place the results presented in the table in the global perspective.
Underreporting of data is not restricted to spatial scale, and extends to other figures as well. Graphs attempting to show trends between hurricane phenomena and time only reported the occurrence of hurricanes that made landfall (Figure 9), or ‘violent’ hurricanes (Figure 10). Similarly, only category F3 to F5 tornados are included in Figure 8. These strong to violent tornados are relatively rare in the United States (about 2% according to http://www.ncdc.noaa.gov/oa/climate/severeweather/tornadoes.html), no waterspouts were included in the figure, and it is not immediately apparent why other types of hurricanes and tornados were excluded from analysis. In another example, no species data available for the plant experiment in Figure 23, which was used to extrapolate the global plant response to increased atmospheric carbon dioxide.
A significant number of figures in the paper attempt to demonstrate trends in climate variables over time. It is disappointing to observe that no statistics are presented for most of them (e.g. Figures 2-5, 7-15). While it is possible to fit a straight line through any scatterplot and measure its gradient, knowing the standard errors, sample size, and regression and/or correlation statistics would allow the reader to judge the statistical significance of the trend. Additionally, sampling artifacts were not accounted for. For example, the variable of average rainfall over the contiguous United States (Figure 7) subsumes many geographically distinct areas within this region, and it is unknown if sampling effort was even across space and time, which would impact the variance within the data.
Aside from the examples brought up earlier, there are other statements in the article that do not have references. The authors assert on page 3, paragraph 4 that carbon dioxide acts as a fertilizer for plants, improving plant growth, and leading to positive cascading effects down the food chain. Most astounding is the statement that “The extent and diversity of plant and animal life have both increased substantially during the past half-century”, which is clearly at odds with the realities of widespread habitat destruction, decline in fisheries, and catastrophic extinctions in recent times (e.g. Brook et al 2003, Sala and Knowlton 2006). Also unverified is the argument on page 9, paragraph 1 that “Trees respond to CO2 fertilization more strongly than do most other plants, but all plants respond to some extent”, which is used to show the benefits of increasing atmospheric carbon dioxide concentration.
The significance of the trendlines asides, while it is true that correlation does not prove causation, non correlation does not prove non-causation, as the authors imply for the relationship between sea level rise and hydrocarbon use. The relationship may not be directly linear to begin with, and cannot be summarily dismissed. The authors’ stand would be clearer if they had provided a viable alternative hypothesis as to how ‘solar irradiance’ generates the largest radiative forcing effect on global climate. The consideration of only two factors in the control of global temperature (methane was brought up, and dismissed as being unimportant) is overly simplistic. The IPCC’s 4th Assessment Report on Radiative Forcing (Forster et al 2007) lists 16 factors that may impact climate on local and global scales. As a result of the multitude of factors that could lead to temperature changes, the comparison in Figure 16 is misleading. Any effect that carbon dioxide has on forcing global climate would be in concert with the other factors, which were unlikely to remain constant throughout the long duration of the comparison.
The authors reject predictions from computer-based climate models as incorrect, and lacking no empirical support. However, they offer no details on the kinds of errors arising from models, whether it is the assumptions made or overly simplistic cycles and feedback loops, or even which models they are specifically referring to. The computer model uncertainties listed in Figure 19 came from a paper written by one of the co-authors of this article and a personal communication from 1995, which probably were not original sources of the models’ data, and the latter reference is highly outdated. The statement that computer models have been “invalidated by numerous observations” was not accompanied by any specific references. No mention was made of the improvement in modeling expertise over time- current climate models are verified using hindcasts on data collected previously, before being used for future predictions (Christensen et al 2007).
The incomplete picture painted by the authors extends to the choice of data, which could bias the analyses reported in the article. The selection of gas, oil and coal use as a proxy for atmospheric carbon dioxide concentration in the correlation figures, and disregarding other inputs from biomass burning, cement production and deforestation, was not explained in the text. In Figures 3, 5 and 13, it is not clear how the various metrics of solar activity (5 listed by the authors), were summed into the single variable ‘solar irradiance’. It is also not apparent how glacier length is a good representation of global warming, seeing as how the authors themselves note that precipitation is a factor in glacier length. Furthermore, the response variable is described as ‘normalized glacier length’, without explaining what it was normalized to. A closer examination of the scale in Figure 2 shows that glacier length was measured at the precision of centimeters. Considering the massive sizes of glaciers, the magnitude of their seasonal variation in length, and the precision of the methods used to measure them, without any indication of the standard errors, we cannot ascertain if the differences in length shown in the figure are significant at all.
While standard error terms were missing from many of the figures, it is puzzling why the authors chose to use standard deviation of tree ring width as the metric to represent tree growth in Figure 21. The mean would give a much clearer picture of differences in growth, whereas standard deviation can change due to sampling artifacts. No units are given for the standard deviations as well. Other analyses are plagued by confounding variables. Although the graph purports to show trends in United States forests, the metric used is volume of timber, presumably with commercial value. One does not know if the increase in timber volume is due to an increase in timber plantations, a reduction in logging, or a response to increased atmospheric carbon dioxide as suggested by the authors (no references). The comparison of warming trends on other planets to Earth’s (page 5, paragraph 11) is unrealistic, given the vastly different atmosphere present on each planet. The attribution of some warming effect to the ‘urban heat island’ effect in Figure 15 is also confounded, as population changes over time, and the effects were not separated in the analysis.
Lastly, subjective terms that are not clearly defined crop up in the article, such as the “Qualitative Greenhouse Effect” presented in Figure 18. With no qualification of the term, or any measurement units, it remains unknown how the authors are able to come up with a scale and compare among the four categories presented. Carbon dioxide and methane are labeled as “minor greenhouse gases” without qualification as to whether the term refers to the gases’ relative atmospheric volume or radiative forcing effects. Statements like “The climate of the Earth is now benign” do not add to the scientific argument in the paper and show a tendency of the authors to editorialize.
Due to numerous omissions, inappropriate generalizations, a lack of clarity in the analyses and a dearth of statistical analyses of data, it is hard to conclude anything about the hypothesis that the authors have put forth. I find that the authors are unable to support their argument that increasing carbon dioxide has no deleterious effects on global climate.
References:
Brook BW, Sodhi NS, Ng PKL (2003) Catastrophic extinctions follow deforestation in Singapore. Nature 424: 420-426
Christensen, J.H., B. Hewitson, A. Busuioc, A. Chen, X. Gao, I. Held, R. Jones, R.K. Kolli, W.-T. Kwon, R. Laprise, V. Magana Rueda, L. Mearns, C.G. Menendez, J. Raisanen, A. Rinke, A. Sarr and P. Whetton, 2007: Regional Climate Projections. In: Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.
Forster, P., V. Ramaswamy, P. Artaxo, T. Berntsen, R. Betts, D.W. Fahey, J. Haywood, J. Lean, D.C. Lowe, G. Myhre, J. Nganga, R. Prinn, G. Raga, M. Schulz and R. Van Dorland, 2007: Changes in Atmospheric Constituents and in Radiative Forcing. In: Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M.Tignor and H.L. Miller (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.
Keigwin LD (1996) The Little Ice Age and Medieval Warm Period in the Sargasso Sea. Science 274: 1504-1508
Sala E & N Knowlton. 2006. Global Marine Biodiversity Trends. Annual Review of Environment and Resources 31:93-122
Comments