118: Arctic Circle - temperature trends STABLE before 2000

If the Antarctic is the region of our planet that is most often associated with climate change, then the Arctic is the region that is often considered to be experiencing the most extreme climate change. Yet the reality is rather different. Since 1970 the mean temperature has indeed risen by about 2.5°C, yet between 1920 and 2000 the mean temperature (as indicated by the 5-year average) fluctuated by over 2°C (see Fig. 118.1 below). The largest temperature change since 1920 is +2.8°C between 1964 and 2010, yet the peak in 2010 is only 1°C higher than the previous peak in 1938. That increase may still sound like a lot, but when put in context against the size of the natural variability it is relatively modest, and is not as statistically significant as a 0.5°C or 1°C temperature rise seen in regions where the historic temperatures were previously stable.

The other point to bear in mind is the greater variability in temperatures that is seen in cold climates. It is not uncommon for the monthly anomalies for stations in both the Arctic and Antarctic to exceed ±10°C, or even ±20°C in some cases. In more temperate climes ±5°C would be considered extreme. The fact is that when the climate is more extreme, it is less stable. This is also seen around much of the equator as well.

Fig. 118.1: The mean temperature change since 1920 for the Arctic above 68.5°N (excluding Scandinavia) relative to the 1981-2010 monthly averages. The best fit is applied to the monthly mean data from 1921 to 2000 and has a negative gradient of -0.57 ± 0.18 °C per century.

The trend for the mean temperature anomaly (MTA) shown in Fig. 118.1 was determined by averaging the monthly anomalies from all the station records in the Arctic within 2,400 km of the North Pole that had at least 480 months of data before 2014 (for a list see here). This amounted to 88 datasets in total, of which five were long stations with 1200 months of data and the remaining 83 were medium stations with over 480 months of data. Of these medium stations 30 had over 900 months of data. It should be noted that an additional 34 stations from Scandinavia (Norway, Sweden and Finland) were excluded as their number and density would (in my opinion) have distorted the analysis. These countries will be analysed separately in the near future, and the discussion at the end of this post will compare the results shown in Fig. 118.1 with the equivalent result if the Scandinavian stations were included (see Fig. 118.7).

Fig. 118.2: The number of station records included each month in the mean temperature anomaly (MTA) trend for the Arctic in Fig. 118.1.

The process of determining the MTA in Fig. 118.1 involved first determining the monthly reference temperatures (MRTs) for each station using a set reference period, in this case from 1981 to 2010, and then subtracting the MRTs from the raw temperature data to deliver the anomalies. If a station had at least twelve valid temperatures per month within the MRT interval then its anomalies were included in the calculation of the mean temperature anomaly (MTA). 

The total number of stations included in the MTA in Fig. 118.1 each month is indicated in Fig. 118.2 above. This indicates that after 1950 there were at least fifty active stations, but before 1920 there were less than about ten. As ten is generally too low a number to produce a reliable trend, particularly over a large region like the Arctic, the data in Fig. 118.1 was truncated with only data post-1920 being shown. However, if all the data is considered the MTA trend will have data extending back to 1866 as shown in Fig. 118.3 below. Note also that the low number of stations before 1920 results in a much higher variance of points about the mean (yellow line). This is more evidence of the greater unreliability of this earlier data, which is why the plot shown in Fig. 118.1 is a more statistically reliable.

Fig. 118.3: The mean temperature change since 1860 for the Arctic above 68.5°N (excluding Scandinavia) relative to the 1981-2010 monthly averages. The best fit is applied to the monthly mean data from 1921 to 2000 and has a negative gradient of -0.57 ± 0.18 °C per century.

The locations of the 88 stations used to determine the MTA in Fig. 118.3 are shown in the map in Fig. 118.4 below. This appears to show that the geographical spread is fairly uniform, although there does appear to be more stations in the eastern hemisphere than in the west. The distribution of those 35 stations with over 900 months of data is fairly even as well, which suggests that the simple average of the anomalies from all stations used to construct the MTA in Fig. 118.1 should still yield a fairly accurate temperature trend for the region as a whole. The one noticeable deficiency is the absence of stations within the 80th parallel. In fact the closest medium station to the North Pole is Alert in Canada (Berkeley Earth ID: 153879) which is over 840 km from the North Pole. The result is that we do not know what is happening to temperatures at the North Pole, and we have never known.

Fig. 118.4: The (approximate) locations of the 88 longest weather station records in the Arctic within 2400 km of the North Pole (excluding those in Norway, Sweden and Finland). Those stations with a high warming trend between 1911 and 2010 are marked in red while those with a cooling or stable trend are marked in blue. Those denoted with squares are long stations with over 1200 months of data, while diamonds denote medium stations with more than 480 months of data.

If we next consider the change in temperature based on Berkeley Earth (BE) adjusted data we get the MTA data shown in Fig. 118.5 below. This again was determined by averaging each monthly anomaly from the 88 longest stations in the Arctic. The mean temperature follows a similar trajectory to that of the unadjusted data in Fig. 118.3 with temperatures fluctuating by over 1°C throughout the 20th century but rising to record highs after 2000.

Fig. 118.5: Temperature trends for the Arctic above 68.5°N based on Berkeley Earth adjusted data. The best fit linear trend line (in red) is for the period 1876-2010 and has a positive gradient of +1.80 ± 0.05°C/century.

Any differences between the MTA in Fig. 118.3 and the BE versions using adjusted data in Fig. 118.5  are mainly due to the data processing procedures used by Berkeley Earth. These include homogenization, gridding, Kriging and most significantly breakpoint adjustments. 

The magnitude of these adjustments can be determined by calculating the difference in the MTA values seen in Fig. 118.3 and Fig. 118.5. The result is shown graphically in Fig. 118.6 below. The blue curve is the difference in MTA values between adjusted (Fig. 118.5) and unadjusted data (Fig. 118.3), while the orange curve is the contribution to those adjustments arising solely from breakpoint adjustments. Both are relatively small for most of the 20th century, varying by about 0.2°C about their midpoints. The large offset between the blue and orange curves is mainly due to a difference in MRT interval used. I used 1981-2010 for the data in Fig. 118.3 whereas Berkeley Earth (BE) tend to use 1961-1990. 

The main conclusion to be drawn from the data in Fig. 118.6 is that the BE adjustments have two effects. Firstly, they reduce the amplitude of the data oscillations before 2000, and secondly, they reduce slightly the temperature rise after 2000.

Fig. 118.6: The contribution of Berkeley Earth (BE) adjustments to the anomaly data in Fig. 116.4 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 1921-2000 has a positive gradient of +0.179 ± 0.013 °C per century. The orange curve shows the contribution just from breakpoint adjustments.

Finally, there is the issue of the omitted Scandinavia data that was mentioned at the start of this post. This amounts to five long stations (all in Norway) and a further 29 medium stations (one in Sweden, six in Finland, and the rest in Norway). Including these stations in the averaging process yields the MTA time-series shown in Fig. 118.7 below.

Fig. 118.7: The mean temperature change since 1860 for the Arctic above 68.5°N including Scandinavia relative to the 1981-2010 monthly averages. The best fit is applied to the monthly mean data from 1936 to 2005 and has a positive gradient of +0.09 ± 0.21 °C per century.

Comparing Fig. 118.7 with Fig. 118.3 indicates that the inclusion of the Scandinavia data reduces the temperature variability between 1920 and 2000, and also reduces the temperature rise from 1964 to 2010 by about 0.5°C (from 2.8°C to 2.3°C). What it certainly does not do is make the climate change worse. Nor does the BE adjusted data. In fact the most extreme temperature changes are seen in Fig. 118.1.

Summary

According to the raw unadjusted temperature data, the climate of the Arctic experienced large temperature variations of up to 1.7°C (or ±0.85°C about the midpoint) throughout the 20th century (see Fig. 118.1). 

Since 1964 temperatures have increased by 2.8°C from their minimum value.

Since 2000 the Arctic has warmed by up to 1.5°C but it is unclear how much of this is permanent climate change and how much is just more natural variability. Comparing the data in 2010 with the previous peak in 1938 suggests that 1.0°C of the rise may be permanent.

Over the same period adjusted temperature data from Berkeley Earth appears to show similar results (see Fig. 118.5) but with about 0.5°C less warming since 1970 and a variation of ±0.5°C before.

 


Acronyms

BE = Berkeley Earth.

MRT = monthly reference temperature (see Post 47).

MTA = mean temperature anomaly.

Link to list of all stations within 2,400 km of the North Pole and their raw data files.

51. The Baltic States - temperature trends STABLE to 1980

The Baltic States are the countries of Lithuania, Latvia and Estonia that used to be part of the USSR and are now part of the EU. For the purpose of geographical convenience I will also include the enclave of Kaliningrad in this analysis, for while it is actually a part of Russia, it is not contiguous with Russia, but is instead bordered by Poland, Lithuania and the Baltic Sea.

The mean temperature trend for the region is shown in Fig. 51.1 below. This was achieved by averaging the temperature anomalies for all the weather station temperature records in the region, where the temperature anomalies were measured relative to the monthly reference temperature (MRT) in each case. The MRTs were calculated for the interval 1991-2010. This is rather later and shorter (only 20 years rather than 30) than usual due to the need to maximize the available data and avoid the jump in temperature in 1988. For a more detailed explanation of the MRT calculation process, see Post 47.

 

Fig. 51.1: The temperature trend for the Baltic States since 1775. The best fit is applied to the interval 1781-1980 and has a negative gradient of -0.08 ± 0.08 °C per century. The monthly temperature changes are defined relative to the 1991-2010 monthly averages.

 

For 200 years up to 1980 there was no anthropogenic global warming (AGW) occurring in the Baltic States. In fact the mean temperature for the region fell by about 0.15 °C. Then around 1988 it suddenly jumped by about 1.1 C (see Fig. 51.1 above). Even then the temperature is less than it was in the 1820s, although the data for that period needs to be treated with some caution. That is because it is based on less than five station temperature records (see Fig. 51.2 below). 

However, the more significant factor in explaining the caution over the temperature peak around 1824 in Fig. 51.1 is probably the fragmentation of some of the temperature records in that era, particularly for Dorpat, Tallinn and Riga. This, when combined with the low number of stations overall, can lead to discontinuities in the temperature trend. 

Having said that, data from Vilnius, Sovetsk and Mitau all appear to show similar peaks in the temperature trend around 1824, and their data are continuous. So maybe the peak around 1824 is real. In which case temperatures in the 1820s really were higher than today.

Fig. 51.2: The number of station records included each month in the mean temperature trend for the Baltic States when the MRT interval is 1991-2010.

The temperature trend shown in Fig. 51.1 is the average of just 23 medium and long station records with over 480 months of data. Of these, seven are long stations with more than 1200 months of data. In fact four have over 1800 months (or 150 years) of data. The 23 stations are also distributed evenly over the region as shown in Fig. 51.3 below, with each of the four regions (Kaliningrad, Lithuania, Latvia and Estonia) also containing one of the four longest records. The HTML links above link to a list of stations for each region.

Fig. 51.3: The locations of long stations (large squares) and medium stations (small diamonds) in the Baltic States. Those stations with a high warming trend are marked in red.

What is interesting is comparing the trend based on the original true temperature data in Fig. 51.1 with the equivalent trend based on the data used by Berkeley Earth after they have adjusted the data. The Berkeley Earth version is shown in Fig. 51.4 below.

Fig. 51.4: Temperature trend in the Baltic States since 1775 derived by aggregating and averaging the Berkeley Earth adjusted data for all long and medium stations. The best fit linear trend line (in red) is for the period 1841-2010 and has a gradient of +0.45 ± 0.04 °C/century.

Unlike the original data which has a slight negative trend before 1980, the Berkeley Earth adjusted data has a strong positive trend of 0.45 °C per century. In total this equates to a warming of over 0.8 °C before 1980. When the temperature jump after 1980 is included, the total temperature rise since 1800 is over 2 °C. This may be consistent with IPCC briefings, but it is not consistent with the actual real data in Fig. 51.1.

Fig. 51.5: The contribution of Berkeley Earth (BE) adjustments to the anomaly data in Fig. 51.4 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 1841-2010 has a gradient of +0.351 ± 0.007 °C per century. The orange curve shows the contribution from breakpoint adjustments.

Overall, the Berkeley Earth adjustments appear to add between 0.6 °C and 1.0 °C to the warming, depending on how you view it. If we consider the net adjustments made to the data (the blue curve in Fig. 51.5 above) which are the difference between the mean anomalies in Fig. 51.1 and Fig. 51.4, these appear to add about 0.6 °C of warming. The difference in the gradients, however, results in over 0.9 °C of warming being added. Either way, these are significant modifications to the real data that completely change its properties.

Summary

1) In the 200 years before 1980 the mean temperature of the region decreased by 0.15 °C (see Fig. 51.1).

2) Once again we see a sudden rise in temperature in 1988 of about 1 °C that is difficult to explain (see Fig. 51.1). Similar rises were seen in Poland (see Post 50), Germany (see Post 49) and Denmark (see Post 48).

3) Even after the 1988 temperature rise, temperatures post-2000 are still below those pre-1830 (see Fig. 51.1).

4) The temperature trend based on Berkeley Earth adjusted data has a warming of over 0.8 °C before 1980 and over 1 °C of additional warming after 1980 (see Fig. 51.4).

5) Adjustments made to the temperature data by Berkeley Earth via breakpoint adjustments and homogenization have profoundly changed both the magnitude of the warming since 1800 and its significance (see Fig. 51.4 and Fig. 51.5).