55. Austria - temperature trends STABLE to 1980

The temperature trend for Austria from 1780 to 1950 is qualitatively very similar to that of its neighbour Hungary over the same time period. In both cases the temperature declined from around 1800 to about 1880, and then rose again afterwards. Thus temperatures in the middle part of the 20th century were broadly similar to those in 1800. This can be seen most clearly by comparing the 5-year moving average for Austria in Fig. 55.1 below (yellow curve) with the equivalent curve for Hungary in Fig. 54.1 in Post 54. For this period the temperature trend is therefore stable but with considerable natural variation. This also suggests that the trend in Fig. 55.1 is accurate as it is in effect corroborated by the data from its neighbour, Hungary.

 

From 1951 to 1980 the mean temperature in Austria increases slightly. Then around 1988 there is a sudden jump in temperature of about 1 °C similar to that seen in Hungary, Czechoslovakia, the Baltic States, Germany and Denmark. In fact comparing the mean temperature for the period 1991-2010 with that for 1961-1980 indicates a sudden rise of 0.99 °C, while the average for 1991-2010 is 1.37 °C above the overall 1781-1950 trend (see the red best fit line in Fig. 55.1).

Fig. 55.1: The temperature trend for Austria since 1767. The best fit is applied to the interval 1781-1950 and has a positive gradient of +0.05 ± 0.08 °C per century. The monthly temperature changes are defined relative to the 1981-2010 monthly averages

The best fit line in Fig. 55.1 is calculated for the period 1781-1950. The reasons for this choice are statistical accuracy and impartiality. I could have chosen the period 1781-1980. This would have represented a full two centuries of data, but it would give a misleading value for the trend (of 0.20 ± 0.06 °C per century) because it would be calculated over a non-integer number of cycles in the natural variability. This variability peaks around 1780 and 1950. That is why the trend in Fig. 55.1 is calculated between those two dates. 

Fitting to the full length of the data also poses similar drawbacks. In addition, the anomaly data clearly shows a different form of behaviour after 1980 compared to before. It would therefore be inappropriate to analyse both time-frames with a single best fit line. For that reason I have restricted the linear regression analysis to the 1781-1950 interval. For more explanation of the rationale I have employed here I suggest referring to the discussion of Fig. 4.7 in Post 4, the discussion of Fig. 18.3 in Post 18, and the discussion of Fig. 30.3 in Post 30.

Fig. 55.2: The number of station records included each month in the mean temperature trend for Austria when the MRT interval is 1981-2010.

The anomalies used to calculate the trend in Fig. 55.1 were determined relative to the mean temperatures for the interval 1981-2010. This interval corresponds to a maximum in the number of available station records that had over 480 months of data (see Fig. 55.2 above) and therefore should lead to more accurate results for the trend. In all, 27 stations in Austria have over 480 months of data (see here), of which 26 have sufficient data in the MRT interval to qualify for inclusion in the overall trend. The exception is Obir (Berkeley Earth ID: 5111) which although having 1153 months of data has none after 1944. It is therefore excluded. For a detailed explanation of MRTs and their use in determining the temperature anomalies please refer to Post 47.

Of the 26 station records included in the trend in Fig. 55.1 fifteen had over 1200 months of data. The geographical locations of these long stations are shown on the map in Fig. 55.3 below. The remaining medium stations (with over 480 months of data) are also shown as small diamonds.

Fig. 55.3: The locations of long stations (large squares) and medium stations (small diamonds) in Austria. Those stations with a high warming trend are marked in red. Those with cooling or stable trends are marked in blue.

The map in Fig. 55.3 illustrates how evenly the long and medium stations in Austria are distributed across the country. This allows their anomalies to be averaged without any need for different weightings for different stations to be employed. This equal weighting approach was used to construct the trend in Fig. 55.1. I have also used it to construct a Berkeley Earth version based on their adjusted data. This is shown in Fig. 55.4 below.

Fig. 55.4: Temperature trend in Austria since 1767 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 1831-1980 and has a gradient of +0.36 ± 0.03 °C/century.

There are two things that are striking about the trend for the Berkeley Earth adjusted data in Fig. 55.4. Firstly, it agrees almost exactly with the trend published by Berkeley Earth and shown in Fig. 55.5 below even though the trend in Fig. 55.4 uses an equal station weighting approach while the trend in Fig, 55.5 does not. This suggests that adjusting the station weightings has a minimal effect on the result, and so validates the approach taken to calculate the trends in both Fig. 55.4, and more importantly Fig. 55.1. But secondly, much of the cooling between 1820 and 1900 is erased. The result is that the temperature trend for 1831-1980 is reduced from +0.73 ± 0.10 °C per century to a more modest +0.36 ± 0.03 °C per century. In other words, the trend looks more like the IPCC hockey stick.

Fig. 55.5: The temperature trend for Austria since 1750 according to Berkeley Earth.

The difference in the trends from 1831-1980 for the data in Fig. 55.1 and Fig. 55.4 is the result of adjustments made to the data by Berkeley Earth. These adjustments are commonplace in the Berkeley Earth data and have been documented in many of my previous posts, but usually they tend to increase the trend compared to that seen for the raw data. In this case, though, these adjustments actually reduce the trend between 1831 and 1980. However, before 1831 the adjustments effectively add warming to the trend by reducing temperatures before 1830. The net effect of these two sets of adjustments is to flatten the curve between 1770 and 1980 and make it appear more like a hockey stick. These adjustments are shown in Fig. 55.6 below.

Fig. 55.6: The contribution of Berkeley Earth (BE) adjustments to the anomaly data in Fig. 55.4 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 negative gradient of -0.394 ± 0.004 °C per century. The orange curve shows the contribution just from breakpoint adjustments.

Conclusions

1) The temperature trend for Austria is qualitatively very similar to that of its neighbour Hungary. This effectively allows the trends from these two adjacent countries to corroborate each other's results. 

2) The temperature trend for Austria is stable up until 1980. The net upward trend of 0.4 °C is comparable to the natural variation. 

3) After 1980 there is a sudden increase in temperature of about 1 °C that occurs around 1988. The reason for this is unknown.