The Netherlands has one of the longest instrumental temperature records in the world, and probably the most complete record covering the last 300 years. The record from De Bilt (Berkeley Earth ID: 175554) had nearly 3700 months of data in 2013 that stretched back to 1706 (see Fig. 41.1 below). Only Berlin Tempelhof (Berkeley Earth ID: 155194) has earlier data that extends to 1701, but it has fewer months overall and significant gaps in its record before 1756.
Fig. 41.1: The temperature trend for De Bilt since 1706. The best fit is applied to the interval 1731-2005 and has a positive gradient of +0.29 ± 0.04 °C per century. The monthly temperature changes are defined relative to the 1976-2005 monthly averages.
As I showed in the last post, Belgium also has one long record that stretches back to the 18th century, but it has virtually no other data before 1973. The Netherlands is much better in this respect. There is one other dataset with some sporadic 19th century data, and overall there are five long station records with more than 1200 months of data. In addition, there are another 25 medium records with more than 480 months of data. Details of all these 30 stations (and other shorter records) are listed here, while their geographical locations are shown on the map in Fig. 41.2 below.
Fig. 41.2: The locations of long stations (large squares) and medium stations (small diamonds) in the Netherlands. Those stations with a high warming trend are marked in red.
It can be seen from the map above that the stations in the Netherlands are fairly randomly distributed across the country, but that their number appears to be significantly less than the 30 stations stated previously. This is because in nearly a dozen cases two or more stations are located within 10 km of each other. I intend to look at this is more detail in a later post, where I will look at what this says about data reliability.
The other impact of this clustering is the effect it could have on the station weightings in the regional average. Normally if a cluster of records is found the weighting of each record should probably be reduced as they will tend to repeat each other's data and geographical coverage. However, as most of the station records appear in effect to be paired up, they will almost all have the same reduced weighting, so the weighting reduction should largely cancel. This is largely confirmed by the results I will show later in this post. The other point to note, is that the clustering really only impacts the medium stations, most of which have data after 1970 only. So the weighting problem will only have a slight effect on the overall trends after 1970.
Fig. 41.3: The temperature trend for the Netherlands since 1706. The best fit is applied to the interval 1731-2005 and has a positive gradient of +0.31 ± 0.04 °C per century. The monthly temperature changes are defined relative to the 1976-2005 monthly averages.
If we average the anomaly data for all the long and medium stations we get the trends shown above in Fig. 41.3. The overall trend indicates that the region has warmed by about 0.31 °C per century since 1700. This equates to an overall warming of about 0.97 °C. But as I explained in Post 14, the current human industrial and domestic energy consumption in the country suggests that the region should have warmed by at least 1.0 °C over the same period simply as a consequence of all the heat that is produced each year by human activity. So, just as for Belgium, we see little need to call on the effects of carbon dioxide emissions and the Greenhouse Effect to explain the observed temperature rise.
The other interesting feature of the data in Fig. 41.3 is the shape of the temperature trend between 1800 and 1950. There is clearly a peak around 1860 that is seen not just in the De Bilt record in Fig. 41.1, but also in the Zuid-Limburg station data. This suggests that temperatures in the mid-19th century in the Netherlands were actually higher than they are today. This is a phenomenon that I have identified and highlighted previously in other countries and regions such as New Zealand (see Post 8), Australia (see Post 26) and South America (see Post 35). In fact it appears to occur over most of the Southern Hemisphere, or at least in those parts that have sufficient data before 1900.
The anomaly data used to construct the trend in Fig. 41.3 was derived by first calculating the monthly reference temperatures (MRT) for the period 1976-2005 for each record, and then subtracting these from the raw data. These were then averaged. Temperature records were only included in the trend in Fig. 41.3 if they had at least 480 months of data, and at least 320 months of this data was within the MRT interval of 1976-2005. This was to ensure that all temperature anomaly records were measured relative to identical reference points. The result was that three medium stations were excluded because they had insufficient data after 1975. These were the stations at Den Helder, Maastricht and Groningen.
Fig. 41.4: The number of sets of station data included each month in the temperature trend for the Netherlands.
The actual number of stations used to construct each monthly point in
the trend in Fig. 41.3 is illustrated above in Fig. 41.4. This shows
that the trend before 1900 is almost entirely due to the data from De
Bilt in Fig. 41.1, while the data from 1900 to 1950 comes from the five
long stations. After 1950 as many as 27 station records were used for each monthly average.
Fig. 41.5: Temperature trends for all long and medium stations in the Netherlands since 1750 derived by aggregating and averaging the Berkeley Earth adjusted data. The best fit linear trend line (in red) is for the period 1801-1980 and has a gradient of +0.29 ± 0.03 °C/century.
So the question is, how significant are these results? And also how reliable are they?
Well, one way to test this is to compare these results against those produced by climate science groups like Berkeley Earth. The first thing to remember, though, is that the Berkeley Earth anomaly data for each station record is different from that which I have calculated here because it uses homogenization and breakpoint alignment to adjust the data, techniques that I have profound misgivings about because they could introduce warming to the overall trend that is not actually there. That is why I restrict my analysis to the raw data with all its imperfections.
However, if we apply the same averaging process to the Berkeley Earth adjusted data as I have employed to the raw data, we see that the trends we get (as illustrated in Fig. 41.5 above) agree very well with those published by Berkeley Earth and shown in Fig. 41.6 below. In fact the size and positions of most of the peaks in the two figures are virtually identical. This suggests that the two processes (mine and Berkeley Earth's) are broadly consistent, even if the anomaly data for each station that is used in the averaging is different. What it also shows, though, is that the Berkeley Earth trend that incorporates homogenization and breakpoint adjustments is somewhat different from the trend I have presented in Fig. 41.3 that avoids using such controversial techniques. For example, according to Berkeley Earth, the warming in the Netherlands since 1900 is at least 1.5 °C, and there was no warm period in the mid-19th century. It is these disagreements over data and methodology, and the effects they have on the resulting temperature trends, that partly fuels the climate scepticism debate.
Fig. 41.6: The temperature trend for the Netherlands since 1750 according to Berkeley Earth.
If we try to quantify the difference between the Berkeley Earth temperature trend and the raw trend I have constructed in Fig. 41.3 we find that the adjustments made by Berkeley Earth have two main effects. The first is to flatten the trend before 1900. The second is to exaggerate the temperature rise after 1900 by about 0.3 °C. These adjustments are illustrated in Fig. 41.7 below.
Fig. 41.7: The contribution of Berkeley Earth (BE) adjustments to the anomaly data after smoothing with a 12-month moving average. The linear best fit to the data is for the period 1901-2010 (red line) and the gradient is 0.266 ± 0.009 °C per century. The orange curve represents the contribution made to the BE adjustment curve by breakpoint adjustments only.
Conclusions
It is clear from Fig. 41.3 that there has been a large degree of warming in the Netherlands over the last 300 years, but that this is probably less than than the 1.5 °C we are being led to expect for anthropogenic global warming (AGW) in the Northern Hemisphere as claimed by the IPCC and the HadCRUT4 data.
The magnitude of this warming is probably only about 1 °C. However, this temperature rise is only what one would expect from the growth of industrial energy use over this period (for the Netherlands I have previously calculated that it should be about 1.0 °C) as I explained in Post 14.
However, there is also evidence of significant natural variation in the temperature record (such as the warming in the mid-19th century) that is inconsistent with current IPCC claims.
Consequently, the data presented here does not really add support to the theory that carbon dioxide is the primary driver of warming, otherwise the warming seen in the Netherlands should be much larger, and there would be no anomalous fluctuations in temperature before 1900.
Finally, there is the issue of historical perspective. If temperatures in the recent past were both higher than now and at times lower than now, why are we worried about current temperatures when they appear to be fluctuating between normal bounds?
No comments:
Post a Comment