2. Department of Environment and Geography, Macquarie University, Sydney NSW 2109, Australia
Global warming is a hot topic. The fourth assessment report of the IPCC estimated a 0.74 ℃ warming from 1906 to 2005 (IPCC, 2007). For the last 6 years, this situation has even become worse (IPCC, 2013). Prior to the publication of the IPCC reports, however, very few scientists considered anthropogenic climate change possible. The fourth and fifth assessment reports of the IPCC issued in 2007 and 2013 strongly asserted that increasing CO2 emission from the combustion of fossil fuels and other human activities is the primary cause of the global warming.
The core theory of CO2-caused global warming proposed by the IPCC is based on three assumptions: (1) The Earth acts like a greenhouse, and the greenhouse effect of increasing CO2 is capable of raising temperature. (2) The available instrumental temperature records over the last century accurately reflect global temperature trends (Figure 1a). (3) The rising atmospheric CO2 is the result of the increasing consumption of fossil fuel. The conclusions by IPCC are logical deductions that should be tested and proven (or challenged) by facts.
Global temperature, CO2 emission, and consumption of fossil fuels have indeed risen. However, any of four possible causal relationships may exist between these three factors. (1) Rising CO2 causes global warming, as advanced by the IPCC. (2) CO2 and temperature are rising independently of each other. (3) CO2 and temperature both rise and fall as a consequence of a third factor. (4) CO2 and temperature are each regulated by independent factors. As early as 1958, scientists began measuring atmospheric CO2 at Mauna Loa, Hawaii. Since 1958, more and more observation stations have been built throughout the world. The causal relationship between the atmospheric CO2 rising and global warming can be clarified with the actual data from these stations (Figure 1b).
Global warming is believed to be an unequivocal process. The rising of atmospheric CO2 is strongly confirmed by observations. The amplitude of global warming over the last century is such that the IPCC estimated 0.85 ℃ rise in temperature in 2013 (IPCC, 2013), 0.74 ℃ in 2007 (IPCC, 2007), and 0.6 ℃ in 2002 (IPCC, 2002), Jones (1994) proposed a growth of 0.4 ℃ from 1851 to 1993. The correlation coefficients continue to rise over time. However, it must be noted that the impact of human activities and the natural environment on CO2and temperature is a large complex web of cause and effects that cannot be expressed as a simple equation. The IPCC's calculation of the global warming does not take into account the urban heat island effect (see subsection 3.2.2 for a detailed discussion).
Here, we will use the CO2 and temperature data from Mauna Loa, Hawaii, and other observation stations to explore the causal relationship between temperature and CO2, and the trend of climate change.2 Variations of CO2 and temperature 2.1 CO2 records and their seasonal variations
Early in 1958, the first atmospheric CO2 observation station was established at Mauna Loa, Hawaii (19.5°N, 55.6°W, altitude 3, 397 m). The atmospheric CO2 at that station has increased by 81.86 mg/kg from March 1958 to March 2013 (Pieter, 2013). Since 1958, more and more CO2 observation stations have been built (Figures 1b and 2a), such as the stations at Alert, Nunavut, Canada (82.5°N, 62.5°W, altitude 210 m), the northernmost permanently inhabited place in the world; Ascension Island, UK (8.0°S, 14.4°W, 85 m), an isolated volcanic island in the equatorial waters of the South Atlantic Ocean; Cape Grim, Australia (41.7°S, 144.7°E, altitude 94 m), the northwestern point of Tasmania; and Baring Head, New Zealand (41.6°S, 174.9°E, altitude 85 m), the southern end of the North Island of New Zealand. The four stations are far away from human activities. All CO2 records at these stations show seasonal rises and falls, as well as a rising long-term trend (Figure 2). There is a 6-month time lag in the seasonal rises and falls of CO2 between the Northern Hemisphere and the Southern. For example, the annual CO2 at Mauna Loa, Northern Hemisphere (green, Figure 1b), peaks when the CO2 at Baring Head, Southern Hemisphere (red, Figure 1b), troughs. The 6-month time lag in CO2 variations between hemispheres reflects the opposite seasons of the hemispheres. As we know, solar radiation regulates the seasons. Hence, solar radiation controls both the seasonal CO2 fluctuations and temperature fluctuations.2.2 Correlation analysis between CO2 and temperature 2.2.1 Variations of CO2 and temperature at Mauna Loa
Figure 2 shows the variations of CO2 and temperature at the four stations, but it is hard to determine their causal relationship through visual observation. Correlation analysis is used to investigate their relationship. The results show that monthly CO2 and temperature records do not correlate with each other at the four observation stations; see Figures 3a, b, c, d. However, the monthly CO2 variation at Mauna Loa correlates well with the global monthly temperature change (Figure 3e, R2=0.7655). This finding indicates a significant difference between the global data and the individual records. The correlation between CO2 and the global mean temperature shows an obvious increase over that of the 188 individual records, suggesting that the global data process is defective and the produced global mean temperature data is unrepresentative and totally unreliable (see the discussion, section 3, for details).2.2.2 Annual growth-rate analysis of CO2 and temperature
Annual growth rate is defined as the difference in annual average value between this and last year. The annual growth rate of CO2 (GRC) and temperature (GRT) at the Mauna Loa station show rises and falls. Moreover, the GRC has a clear rising trend, while GRT both at the Mauna Loa station and globally remains flat (Figures 4a, b). This finding implies that mechanisms for GRC and GRT variations are quite different. This trend coincides with observations from Figures 2b and 3a, which suggest that CO2 and temperate are not correlated with each other. Thus, it is rather likely that the rise in atmospheric CO2 is not the cause of rising temperatures.3 Discussion 3.1 CO2 seasonal variation and greenhouse effect test
IPCC has suggested there was a simple relationship between CO2and temperature. One way to test this theory with a large data set is by studying seasonal variations in CO2 and temperature. Bacastow et al. (1985) analyzed CO2 variations between 1959 and 1982 and found that CO2 displayed seasonal rises and falls and daily fluctuations. The amplitude of its daily variations varied with the seasons. We studied the seasonal fluctuations in CO2 and temperature to determine the relationship between the two. For any given year, we used the annual ΔT=highest monthly average temperature for the year minus the lowest monthly average temperature for the same year; annual ΔCO2=highest monthly average CO2 for the year minus lowest monthly average CO2 for the same year. Using this approach, an average annual ΔT 5.10 ℃ corresponding to average annual ΔCO2 5.73 mg/kg at Mauna Loa, Hawaii, is obtained for the 55 years (1958-2013).
If there were a simple relationship between CO2 and temperature, then it would be the case that every 1 mg/kg rise in CO2 corresponds to a 0.89 ℃ rise in temperature. From March 1958 to March 2013, CO2 at Mauna Loa, Hawaii, rose by 81.86 mg/kg; using this formula, we should see a 72.86 ℃ rise in temperature. However the temperature at Mauna Loa actually rose by -0.62 ℃ during the 55 years of records, significantly less than predicted by this model. Again, this finding suggests that the relationship between CO2 and temperature as stated above fails to be verified.3.2 Is the Earth an effective greenhouse?
CO2-caused global warming critically assumes that the Earth is an effective greenhouse. If this is the case, CO2 variations will correlate well with temperature changes and even precede temperature changes.3.2.1 Relationship between CO2 and temperature
Figure 2b shows the variations of monthly CO2 and temperature at Mauna Loa station over the last 55 years. CO2 rises, while temperature remains flat. The correlation coefficient is R2=0.0335 at Mauna Loa; 0.03 at Alert, Nunavut; 0.0013 at Ascension Island; and 0.0001 at Cape Grim (Figures 3a-d). We can see that, at these four stations, CO2 and temperature are not correlated.
In Figures 3a-d, we can see that at Mauna Loa, Nunavut, Ascension Island, and Cape Grim, the correlation coefficients between monthly CO2 and temperature records show no correlation. Furthermore, in Figure 4, we can see that the annual growth rate of CO2 (GRC) rises, while the annual growth rate of temperature (GRT) remains flat, suggesting that different mechanisms drive their change. However, why is that in Figure 3e, monthly changes in Mauna Loa CO2 correlate with changes in global temperature (R2=0.7655)?
The global monthly mean temperature comes from calculation of global individual data after corrections for landscape, latitude and other factors. It shows good correlation with the CO2 monthly record (R2 > 0.7655), but it is not a natural simple record. A test is thus required to prove if this calculation is appropriate. In theory, for the individual temperature records, their individual correlation coefficients R2 with CO2 should be possibly divided into two groups almost equally according to their R2: one group should have their R2 > 0.7655 (for big cities with strong heat island effects); another group with R2 < 0.7655 (for cities with less heat island effect). Therefore, across the globe, 188 individual records of monthly mean temperature (see supplementary material) are selected for correlation with the four CO2 monthly records (CO2 in Figure 3), respectively. The mean correlation coefficients of the four methods are obtained: 0.016, 0.058, 0.024 and 0.020. The maximum correlation coefficients of the four methods are: 0.293, 0.319, 0.417 and 0.366 (see supplementary material). This test that indicates, R2=0.7655 obtained by the global monthly temperature (GISS, 2013) is not proved by 188 selected records (mean R2=0.024, for correlating them with monthly CO2 at Mauna Loa). These very different correlation coefficients R2 (0.7655 and 0.024) indicate remarkable deviation between global individual monthly mean temperature and the global monthly mean temperature by calculation. As the R2=0.024 calculated from 188 records selected from over the world is rather low, it implies no relation at all between the temperature and CO2, However the R2=0.7655 obtained by GISS (2013) shows a strong relation between the global monthly mean temperature and CO2. This test therefore makes it clear that a big difference of correlation coefficients R2 (=0.7655-0.024=0.7415) was produced during the data process for obtaining the global monthly mean temperature by GISS (2013). This inappropriate calculation thus creates an untrue picture: CO2 emission by humans drives global warming.
We suggest that it is the urban heat island effect on data of the global temperature (rising population density, rising industrial activity, rising CO2) that is causing this rising temperature (see the next section for further discussion of the urban heat island effect). Indeed, when we look at data from Mauna Loa, Nunavut, Ascension, and Cape Grim, four stations far from urban areas, least affected by the urban heat island effect, we see there is no obvious correlation between CO2 and temperature change. Furthermore, temperature from 188 stations both close and far away from urban areas also shows poor correlation with CO18.104.22.168 Urban heat island (UHI) effect
IPCC has suggested that global temperature has increased by 0.74 ℃ over the last century and 1.28 ℃ over the last five decades (Figure 1). However, the 55-year temperature record at Mauna Loa, Hawaii, shows a clearly flat trend (Figure 4), conflicting with the sharp temperature rise in Marcott's hockey-stick graph (Marcott et al., 2013). At the very least, this finding indicates that temperature records are highly sensitive to the location of the observation. The global temperature stack comprises both urban and rural temperature records. As urban areas have expanded dramatically in the last 50 years, meteorological stations have become closer to urban areas or surrounded by multiple urban areas. This process is how the urban heat island of cities affects temperature records. Although many local studies have demonstrated that the microclimate within cities is on average warmer than if the city were not there (IPCC, 2007), however, some investigators believed that, on hemispheric and global scales, urban-related trend is an order of magnitude smaller than decadal and longer time-scale trends (see IPCC, 2007). Hansen et al. (2010) suggested that the impact of the urban heat island effect is too weak to affect the global temperature stack. However, many local measuring studies on the urban heat island effect disagree with this conclusion. In most climates, maximum UHIs occur a few hours after sunset; maximum intensities increase with city size and may commonly reach 10 ℃, depending on the nature of the rural reference (Heisler and Brazel, 2010). Similar studies on various metropolises in North America, Australia, and Europe have been reported (Figure 5a)-particularly for Columbia, Maryland, where the population increased from 1, 000 to 25, 000, and the urban heat island effect caused a rise in recorded temperature of 5 ℃ (Bryant, 1997). Studies from various major cities in China, e.g., Beijing (Zheng and Liu, 2008), Shanghai (Zhu et al., 2006), Xian (Zhu et al., 2006), and Lanzhou (Ma et al., 2009) demonstrate a 0.70~1.07 ℃ rise to the annual mean temperature from the urban heat island effect. Most studies indicate that the urban heat island effect has increased recorded annual mean temperatures by > 1 ℃. We suggest that the IPCC's 0.74 ℃ temperature rise over the last century could be sufficiently explained by the urban heat island effect.3.2.3 Trend of climate change: cooling or warming?
Several cooling events have occurred over the last six years. The worst snowstorm of the past 50 years swept across vast areas of southern China in early 2008. Midwest and Northeast areas of the United States were in a snowstorm at the end of 2008. In early 2010, a snowstorm hit the Xinjiang area in China and some areas in Canada; in December 2010, almost all of Europe was in a snowstorm. Thirty states of the United States were hit by a snowstorm in February 2011; Japan was in a snowstorm in February 2012. On October 4, 2013, a “ridiculous“ autumn storm covered the Midwest and the Plains of the United States (Figure 5b). At the end of 2013, a once-in-a-century snow event occurred in Kunming in south China and in Egypt. A Polar vortex brought dangerously cold weather to Canada and large parts of the United States on January 6, 2014. The Polar vortex brought the coldest temperatures in more than 20 years to many areas. The temperature of the stratosphere (pressure 50~100 hPa) shows a cooling trend (Angell, 1988). Some scientist believe that the snow disasters are the result of global warming, it is just based on logical deduction. In January 2005, frost and snow passed through Europe and extended farther into the Sahara Desert in North Africa. Meteorology professor Leroux (2005) suggested that this event was the result of global cooling rather than global warming. On the satellite images of January 25, 2005, he found three moving Polar highs (MPHs), that covered most of the northern Atlantic and northern Europe. He emphasized that temperature was determined by climatic dynamics other than the hypothetical simulation equations. Hence, the frequent and widespread snowstorms over the last six years intuitively and empirically suggest that the Earth is cooling.
Global warming is propagated by the fourth and fifth reports of the IPCC. The Hollywood movie The Day After Tomorrow and the documentary An Inconvenient truth, featuring Al Gore, the former vice-president of the United States, raised international public awareness of global warming. Scientists know that climate change is a complicated process, and we do not yet understand all the details. All the logical deductions should be tested by facts before turning it into a widely accepted theory. The urban heat island effect in North American and Europe is able to raise temperature about 10 ℃ or higher, while that in China causes a 0.7 ℃ annual temperature increase. Both facts suggest that the urban heat island has a great impact on instrumental records. If we correct the global temperature stack by removing the urban heat island effect, the real global temperature increase over the last century will be far less than the 0.74 ℃ estimated by (IPCC, 2007), or perhaps the climate is even undergoing a cooling trend.4 Conclusions
1) The monthly variations in CO2 and temperature at Mauna Loa, Hawaii, do not correlate with each other (R2=0.0355). From 1958 to 2013, CO2 rose while temperature remained flat. Hence, we seriously question whether CO2 is the driving force behind temperature variation.
2) Both the instrumental CO2 and temperature records at the Mauna Loa, Hawaii, station show seasonal rises and falls. But there is 6-moth difference in seasonal CO2 and temperature fluctuations between the records in the Northern Hemisphere and the Southern Hemisphere. As we know, the reversal of seasons is determined by the changes in solar radiation. Thus, it is most likely that these seasonal rises and falls of both CO2 and temperature are driven by changes in solar radiation.
3) By studying the monthly relationship between CO2 and temperature over several decades, we established a theoretical transfer function between CO2 and temperature. Using this function, the rise of 81.86 mg/kg in CO2at Mauna Loa, Hawaii, between 1958 and 2012 should have resulted in a 72.86 ℃ rise in temperature, when in fact the temperature only rose -0.62 ℃. Thus, we submit that changes in atmospheric CO2 may not be the cause of global temperature changes.
4) In contrast to IPCC's suggestion that global temperatures rose 0.85 ℃ over the last century, from 1958 to 2012, temperatures at Mauna Loa, Hawaii, did not rise over this period. Mauna Loa is far from cities, where temperature variations are most affected by the urban heat island effect. Independent studies in North America, Europe, Australia, and China have shown that the urban heat island effect could lead to recorded temperature rises > 1 ℃. Thus, we suggest that the IPCC's 0.85 ℃ temperature rise over the last century could be sufficiently explained by the urban island heat effect.
5) The global monthly mean temperature produced by GISS (2013) shows a high correlation with Hawaii CO2(R2=0.7655). However R2=0.024 is obtained by 188 selected records from individual stations around the world. This test indicates global monthly mean temperature showing high correlation with Hawaii CO2(R2=0.7655) was inappropriately corrected and calculated during data process. An untrue picture is therefore created, that CO2emission by human activity drives global warming.Acknowledgments:
The authors gratefully acknowledge support for this research from the National Natural Science Foundation of China (Grant Nos. 41210002, 41602190 & U1405231). This manuscript benefited from Dr. Byrnes's help in writing and from the comments of an anonymous reviewer.
|AHCCD, 2013. Homogenized Surface Air Temperature Data Access. Environment Canada. http://www.ec.gc.ca/dccha-ahccd/default.asp?lang=en&n=1EEECD01-1. Cited in November 26, 2013..|
|Angell JK, 1988. Variations and trends in tropospheric and stratospheric global temperatures, 1958-1987. Journal of Climate, 1:1296-1313. DOI:10.1175/1520-0442(1988)001 &1296:VATITA>2.0.CO; 2..|
|Bacastow R, Keeling C, Whorf T, 1985. Seasonal amplitude increase in atmospheric CO2 concentration at Mauna Loa, Hawaii. 1959-1982. Journal of Geophysical Research:Atmospheres (1984-2012), 90: 10529–10540. DOI: 10.1029/JD090iD06p10529|
|Bryant E, 1997. Climate Process and Change. Cambridge University Press, pp. 1-228.|
|CDIAC, 2013. Carbon Dioxide (CO2). ORNL. http://cdiac.ornl.gov/trends/co2/#. Cited in November 26, 2013.|
|CMDL, 2013. Monthly Average Temperature. MAUNA LOA SLOPE OBS, HAWAII. ftp://aftp.cmdl.noaa.gov/met/mlo/. Cited in November 26, 2013.|
|CSIRO, 2013. Cape Grim Greenhouse Gas Data. CSIRO. http://www.csiro.au/greenhouse-gases/. Cited in November 26, 2013.|
|ESRL, 2013. NCEP-DOE AMIP-II Reanalysis (R-2). NOAA. http://www.esrl.noaa.gov/psd/data/gridded/data.ncep.reanalysis2.gaussian.html. Cited in November 26, 2013.|
|GISS, 2013. GISS Surface Temperature Analysis (GISTEMP). National Aeronautics and Space Administration. http://data.giss.nasa.gov/gistemp/. Cited in November 26, 2013.|
|Hansen J, Ruedy R, Sato M, et al., 2010. Global surface temperature change. Reviews of Geophysics, 48:RG4004. DOI:10.1029/2010RG000345.|
|Heisler GM, Brazel AJ, 2010. The Urban Physical Environment:Temperature and Urban Heat Islands. In:Aitkenhead-Peterson J, Volder A (eds.). Agronomy Monograph 55. Urban Ecosystem Ecology, pp. 29-56. DOI:10.2134/agronmonogr55.c2.|
|Houghton J, 2009. Global Warming:The Complete Briefing. Cambridge University Press, pp. 1-438.|
|IPCC, 2002. Report of the Nineteenth Session of the Intergovernmental Panel on Climate Change (IPCC) Geneva, April 17-20 (am only), 2002.|
|IPCC, 2007. Climate Change 2007--the Physical Science Basis:Working Group I Contribution to the Fourth Assessment Report of the IPCC. Cambridge University Press.|
|IPCC, 2013. Fifth Assessment Report (AR5)--Climate Change 2013:The Physical Science Basis. http://www.ipcc.ch/report/ar5/wg1/.|
|Jones PD, 1994. Hemispheric surface air temperature variations:a reanalysis and an update to 1993. Journal of Climate, 7:1794-1802. DOI:http://dx.doi.org/10.1175/1520-0442(1994)007 < 1794:HSATVA>2.0.CO; 2.|
|Leroux M, 2005. Global warming-myth or reality? The Erring Ways of Climatology. Chichester, UK:Springer, pp. 1-509.|
|Ma Y, Wang S, Wei H, 2009. Characteristics of urban heat island over Lanzhou in recent 50 years. Meteorological Science and Technology, 37: 660–664.|
|Marcott SA, Shakun JD, Clark PU, et al, 2013. A reconstruction of regional and global temperature for the past 11, 300 years. Science, 339: 1198–1201. DOI: 10.1126/science.1228026|
|Pieter T, 2013. Full Mauna Loa CO2 record. NOAA/ESRL. http://www.esrl.noaa.gov/gmd/ccgg/trends/. Accessed on November 26, 2013.|
|WRCC, 2013. Monthly Average Temperature. MAUNA LOA SLOPE OBS, HAWAII. http://www.wrcc.dri.edu/cgi-bin/cliMONtavt.pl?himaun. Cited in November 26, 2013.|
|Zheng S, Liu S, 2008. Urbanization effect on climate in Beijing. Climatic and Environmental Research, 13: 123–133.|
|Zhu JQ, Tang X, Jiang H, 2006. The air temperature variation and heat island effect in urban area of Shanghai City. Plateau Meteorology, 25: 1154–1160.|