India has some of the longest temperature records in southern Asia. The longest of these is for Chennai (Berkeley Earth ID: 155503) and dates back to 1796, although the data is only continuous from 1875. Overall, there are 40 long stations with over 1200 months of data, and another 59 medium stations with over 480 months of data (see here for a complete list). Unfortunately what complicates the analysis of temperature data for this region is an unusual split in the distribution of the station time series.
Generally when we look at temperature data for a specific region we find that there are more stations after 1970 than before, and that the number of stations decreases the further you go back in time. For India the picture is slightly different. The largest number of long and medium stations were in the 1970s, but decreased slightly up to 2010. But there are also about a dozen long or medium stations with no data after 1957 as well as over thirty with no data before 1970. This means that no single choice for the 30-year period used to define the monthly reference temperatures (MRTs) can enable all the available stations to be incorporated into a single temperature trend for the region. For this reason I have used two different time intervals to determine the MRTs and calculated a separate trend for each. Then I have amalgamated the two trends into a single trend (see Fig. 71.6).
Fig. 71.1: The (approximate) locations of the long and medium temperature records in India. Those stations with a high warming trend are marked in red while those with cooling or stable trends are marked in blue. Those denoted with squares are long stations with over 1200 months of data.
Overall, 98 station temperature records met the twin requirement of having over 480 months of data and also having at least 12 years of data within the two MRT intervals (for an explanation of how the MRTs are calculated see Post 47). The geographical distribution of these stations is shown in Fig. 71.1 above. This clearly shows a very even spread of stations across the country, which effectively negates any need for the use of homogenization or gridding in order to determine the mean temperature trend for the whole country.
It is also apparent from the colours of the station markers in Fig. 71.1 that there is no uniform temperature trend across the country. Indeed there are more stations with stable or cooling trends (for the period 1876-1985) than there are stations with warming trends. A warming trend is defined as being one where the total temperature change over the length of the station time series exceeds 0.25°C and the trend gradient is more than twice the uncertainty in that gradient.
Fig. 71.2: The temperature trend for India based on an average of anomalies from all long and medium stations. The best fit is applied to the monthly mean data from 1876 to 1985 and has a positive gradient of +0.11 ± 0.05 °C per century. The monthly temperature changes are defined relative to the 1921-1950 monthly averages.
In order to maximize the number of stations in the trend before 1950 a 30-year MRT interval of 1921-1950 was selected. The anomalies of all the monthly data in each qualifying temperature record were calculated relative to the monthly averages for this interval, and then the anomaly time series were averaged. The resulting trend is shown in Fig. 71.2 above. This indicates that there was virtually no warming in India before 1985, with a small rise of about 0.6°C happening thereafter. The trend before 1875 is probably unreliable as it is based on insufficient data as indicated in Fig. 71.3 below.
Fig. 71.3: The number of station records included each month in the mean temperature trend for India in Fig. 71.2.
However, the trend after 1970 does not include all the available data, and so is also potentially less accurate for the years after 1970 than it could be. As Fig. 71.3 above shows, less than 40 stations were included in the mean trend after 1975 yet over 70 were available. Including those missing stations changes the trend after 1975, although not by a huge amount, as Fig. 71.4 below shows.
Fig. 71.4: The temperature trend for India based on an average of anomalies from all long and medium stations. The best fit is applied to the monthly mean data from 1876 to 1985 and has a positive gradient of +0.12 ± 0.05 °C per century. The monthly temperature changes are defined relative to the 1981-2010 monthly averages.
In order to incorporate all the stations with data that is mainly after 1970 it is necessary to change the MRT interval to a later period. Choosing 1981-2010 for this period results in a slightly different trend as Fig. 71.4 above illustrates. While the trend before 1985 is very similar to that in Fig. 71.2, afterwards the temperature rise is less pronounced. This change in outcome is a consequence of having more data for this later period (but less for the period before 1960) as Fig. 71.5 below shows.
Fig. 71.5: The number of station records included each month in the mean temperature trend for India in Fig. 71.4.
Ideally we would like to combine the two sets of data shown above into a single trend that utilizes all the available sets of station data. There are several ways to do this, but the simplest is to splice the data before 1957 in Fig. 71.2 to the data after 1957 in Fig. 71.4. The rationale for this is that Fig. 72.2 is more accurate in determining the trend before 1957 because it incorporates more stations into the average than does the trend in Fig. 71.4. For data after 1957 the reverse is true.
Fig. 71.6: The temperature trend for India created by combining the trend in Fig. 71.2 before 1957 and the trend in Fig. 71.4 after 1957. The best fit is applied to the monthly mean data from 1876 to 1975 and has a positive gradient of +0.09 ± 0.06 °C per century. The monthly temperature changes are defined relative to the 1957-1972 monthly averages.
The result of this splicing is the trend shown in Fig. 71.6 above. It clearly exhibits little significant warming from 1876 to 1976. Any slope in the trend is comparable to the uncertainty in its value, and is also much less than the natural fluctuations in the 5-year averaged data. After 1975 the mean temperature rises by about 0.6°C over the next 37 years. Overall it should be noted that the total temperature rise of about 0.5-0.7°C since 1900 is much less than the 1.5°C that is generally claimed by climate science for global warming.
It should be noted that in order to combine the two trends into one it is necessary to offset one or both in order to avoid a discontinuity at the junction. This is achieved by comparing the mean values for each trend in Fig. 71.2 and Fig. 71.4 over the same time interval. In this case the interval chosen was from 1957 to 1972, with the data from Fig. 71.2 then being adjusted downwards by 0.0276°C and the data from Fig. 71.4 being adjusted upwards by 0.3778°C.
Fig. 71.7: The temperature trend for India since 1800 according to Berkeley Earth.
The results shown in Fig. 71.6 clearly differ from the global temperature trend since 1900 as advanced by most mainstream climate science. They also disagree with the trend for India published by Berkeley Earth using adjusted data and shown above in Fig. 71.7. This adjusted data appears to suggest that the climate in India has warmed by 1.5°C since the early 1800s. So which data trend is correct?
Well, if we average the Berkeley Earth adjusted data from all 98 stations used in this analysis (the adjusted data is listed on the Berkeley Earth site along with the raw data) we obtain trends that are very similar to those in Fig. 71.7 as the data in Fig. 71.8 below shows. This demonstrates two points. Firstly, it provides strong evidence that the averaging method we use to combine anomalies from the different stations is sufficiently accurate to avoid the need to use gridding and homogenization techniques. If this were not true then the data in Fig. 71.7 and Fig. 71.8 would not agree so well, but the uniform geographical distribution of stations across India facilitates this. But secondly, it shows that most of the warming presented in Fig. 71.7 comes from adjustments made to the data, and not from the raw data itself, otherwise the data in Fig. 71.7 would agree with the data in Fig. 71.6.
Fig. 71.8: Temperature trends for India based on the average of anomalies for all long and medium stations using Berkeley Earth adjusted data. The best fit linear trend line (in red) is for the period 1814-2013 and has a gradient of +0.63 ± 0.01°C/century.
Finally, comparing the adjusted data in Fig. 71.8 with the unadjusted data in Fig. 71.6 allows us to quantify the contribution made to the trends in Fig. 71.8 by the Berkeley Earth data adjustments. These net adjustments are shown below in Fig. 71.9.
Fig. 71.9: The contribution of Berkeley Earth (BE) adjustments to the anomaly data in Fig. 71.8 after smoothing with a 12-month moving average. The blue curve represents the total BE adjustments including those from homogenization. The linear best fit (red line) to these adjustments for the period 1931-2010 has a positive gradient of +0.244 ± 0.004 °C per century. The orange curve shows the contribution just from breakpoint adjustments.
The data in Fig. 71.9 indicate that the adjustments made to the data by Berkeley Earth add at least 0.5°C of warming to the overall temperature trend between 1870 and 2010, with some additional warming being added via adjustments to the data between 1820 and 1860. This explains why Fig. 71.7 and Fig. 71.6 appear so different. Yet it is Fig. 71.6 that, in my opinion, is the most reliable of the two plots simply because the statistical methods employed to create it are less complicated and the temperature data used is free from adjustments. It is also clear, though, that the data in Fig. 71.6 before 1875 is much less reliable as an indicator of the true mean temperature in the region due to insufficicient stations in the average. If we ignore that data the trend for India will be as shown in Fig. 71.10 below.
Fig. 71.10: A rescaled version of Fig. 71.6. The best fit is applied to the monthly mean data from 1876 to 1975 and has a positive gradient of +0.09 ± 0.06 °C per century. The monthly temperature changes are defined relative to the 1957-1972 monthly averages.
Conclusions
It is clear that there was no climate change in India before 1975. Temperatures were actually stable for over 100 years prior to 1975 (see Fig. 71.10).
Since 1975 there has been a modest temperature rise of about 0.5°C (see Fig. 71.10), but this is a long way short of the 1.5°C claimed by climate science as being the mean temperature change for the entire Northern Hemisphere.
The temperature rise of 0.5°C since 1975 could be due to increased greenhouse gas emissions. Carbon dioxide levels in the atmosphere have increased from 330 ppm in 1975 to nearly 420 ppm now. But there could be other reasons.
Since 1975 the population of India has more than doubled. It has also industrialized, although not as much as China. Both of these factors could have caused an increase in urban heating, and consequently an increase in average temperatures. It therefore remains a matter of conjecture as to how much of the 0.5°C warming in India is directly due to carbon dioxide emissions.
No comments:
Post a Comment