Sunday, 8 November 2015

Monthly and seasonal EWP versus CET graphs: January to June plus fiscal year

In the previous post I had a look at CET versus EWP correlations for the first half the fiscal year (the second half of the calendar year). I will now look at the second half of the fiscal year (first half of the calendar year) to see how the patterns evolve, and as a last step I will see what the results are for the fiscal year as a whole.

January:

As can be seen, a general positive correlation between EWP and CET is evident in January, as in December and November. Outliers from the general pattern of warm, wet Januaries with maritime flow contrasting with cold, dry ones characterised by easterly winds from the continent do occur, but are less distinctive than those for December.

The most notable “cold-wet” outlier is January 1809, with CET of 2.0˚C, estimated Scotland temperature of -1.1˚C, estimated UK mean temperature of 0.9˚C, yet EWP of 134.3 millimetres. There exist other moderate “cold-wet” outliers (forming a semicircle between 100 and 110 millimetres and from 0˚ to 2˚C) in Januaries 1768, 1774, 1789, 1867, 1895, 1942, 1959, along with January 1979 – the coldest month over the contiguous US since before 1880, but very hot in southern Australia due to a super-monsoon and warm in the Far East:
If not to nearly the same extent as December 1886, Januaries 1895 and 1979 were both sunnier than average except in eastern coastal areas. In the southwest Torquay exceeded ninety hours in both months, and was only ten hours shy of the UK record for January in 1979. January 1942, however, was rather gloomy, with only 37 hours sunshine in England and Wales against a virgin mean of 47.7 hours – although 11 Januaries since 1929 have been gloomier. January 1959, however, resembles December 1886 very closely in setting sunshine records, and data from Durham and reports from elsewhere suggest January 1959 is very likely the sunniest since before 1881.

The “warm-dry” outliers of 1898 and 1916 are distinctly different. 1916 was the mildest January on record (though in between a very cold November and snowy, dull March) but was probably the UK’s windiest month since 1871 with gales on 25 days – some destructive even when no rain fell. 1898 was a classic anticyclonic gloom month with little sunshine, but the red diamonds in the upper left (1989, 2005) were ten to fifteen hours sunnier than the mean, and the “winterless winter” of 1948/1949 was extremely sunny.

February:

February shows the familiar pattern from the other winter months of warm, wet, westerly “maritime” months contrasting with cold, dry, easterly “continental” months. Outliers to this pattern are concentrated exclusively in the top left “warm-dry” section of the scatter plot:
  • record warm February 1779 with CET 7.9˚C and EWP of 13.5 millimetres
  • February 1998 with CET 7.3˚C and EWP of 20.4 millimetres
  • February 1790 with CET 6.6˚C and EWP of 20.9 millimetres
  • February 1903 with CET 7.1˚C and EWP of 38.8 millimetres
  • February 1846 with CET 6.4˚C and EWP of 37.1 millimetres
  • February 1815 with CET 6.5˚C and EWP of 44.0 millimetres
The lack of “cold-wet” outliers like January 1809 or December 1886 is highly notable. It is true that February 1900 (CET 2.6˚C; EWP 131 millimetres) was much colder in Scotland than in England, but February 1900 was 2.0˚C warmer than January 1809 at both Edinburgh and Gordon Castle, and not as cold in southern Britain.
Moreover, as can be seen from this precipitation map for the record cold February 1947, these “cold-wet” outliers are only so in the east. Often these supposed outliers are very dry on western slopes which, normally exposed to the westerly winds, are left in a rain shadow (more accurately a snow shadow) that is much stronger than the normall westerly rain shadow on the eatern slopes of temperate zone mountains. The then-record cold “Crimean Winter” February of 1855 was the driest – indeed the driest for any month – between 1845 and 1894 at notoriously wet Seathwaite in the Lakes District, with less than half an inch of water-equivalent precipitation or less than a third the average water equivalent precipitation for all of England and Wales.

Winter:

As can be seen, the winter graph shows a positive correlation between EWP and CET already explained for the individual months and not significantly altered by greenhouse pollution from Australian road transport and coal power. In complete contrast to the February plot, the most notable outliers are of the “cold-wet” type, most notably:
  • 1878/1879 (fourth coldest since 1766; CET 0.62˚C; EWP 250.0 millimetres)
  • 1978/1979 (red diamond; CET 1.58˚C; EWP 335.2 millimetres one of only four winters since 1910 drier in Scotland)
  • 1914/1915 (marginal outlier; CET 4.33˚C; EWP a then-record 423.0 millimetres)
“Warm-dry” outliers are not so pronounced, and include
  • the amazing winter of 1778/1779 (record warm February is earliest surviving record warm month and MSLP was over 1,030 millibars)
  • the second-driest winter in 1857/1858 (as noted earlier, December 1857 was warmer than 1934 or 1974 in Scotland and probably the UK as a whole, whilst its MSLP was comparable to February 1779)
  • the winter of 1988/1989 with CET of 6.52˚C and EWP only 185.5 millimetres. In contrast to December 1857, this winter was warm all though central and northern Eurasia due to a highly positive NAO index

March:

With March, I have noted with yellow diamonds the several very snowy Marches that occurred around a century ago during World War I. These were particularly disruptive in the emergency with cold delaying opening of the growing season in 1917, and causing human disruption during the snowy March 1916 – apart from 1947 the worst March of the twentieth century. That March 1916 and 1919 were exceptionally cold and wet is very clear from this graph.

It is extremely evident that the positive correlation seen for the previous four months has completely disappeared, and that the line looks exceptionally flat, with “outliers” being either warm (1938 and 1957), cold (1785) or wet (1947 and 1981).

It’s possible that the true shape of this curve is a triangle – one sees much more CET ranges in dry Marches (both March 1785 and March 1938 had EWP of under 20 millimetres and March-April EWP under 30 millimetres) – than in most wet Marches, though the two EWP outliers in 1947 and 1981 make claims of a “triangular”-shaped scatter plot look dubious and we can assume that in March EWP and CET show little correlation.

If we look at the red diamonds controlled by Australian greenhouse gas emissions, the conclusion is not really different. Despite the clear presence of several Marches (1990, 1997, 2011) that were both warm and very dry, the silver diamonds in the top left corner (1779, 1938 and 1961) and several cool, wet Marches during the 1970s and early 1980s suggest no fundamental change. However, before 1884 very cold, dry, easterly Marches occurred as not observed since in 1785, 1786, 1807, 1808, 1845 and 1883. Nevertheless, warm, wet, westerly Marches (1903, 1912, 1981) were not observed to oppose them as would be expected if the EWP/CET relationship had changed.

April:

As we can see, for April the EWP versus CET relationship is again negative – as it was from July to September. The line of best fit is less steep than for July and August, but nonetheless not flat like for March.

There exit numerous red diamonds in outlying parts of this graph – both hot and wet. Nevertheless, because the extreme hot outliers of 2007 and 2011 were both exceedingly dry, it is not likely that man-made global warming had altered the shape of the graph at all.

The “cold-dry” outliers of 1837 (coldest April on record) and 1771 (EWP 31.3 millimetes, CET 5.5˚C) are more striking than the “warm-wet” ones of 1792 and 1961 (CETs both 10.0˚C, EWPs 97.7 millimetres in 1792 and 98.1 millimetres in 1961) – the latter being the wettest April on record in southwestern Australia, which has seen huge rainfall declines due to its own greenhouse emissions.

May:

In May we see a clear negative correlation between EWP and CET, as observed in the summer months in our first set of scatter plots. The most striking outlier is the “hot-wet”, non-anthropogenic May 1811 with CET of 12.8˚C but a very wet EWP of 121.9 millimetres – the eighth-wettest on record and one of only four cases where May was the wettest month of the calendar year (the others being 1773, 1820 and 1983). The striking character of 1811 – a cool summer, two cold winters but very hot spring and autumn – is noted in Kirkpatrick Sale’s Rebels Against the Future: The Luddites and Their War on the Industrial Revolution – Lessons for the Computer Age.

The prime “cool-dry” outlier of 1876 (EWP 23.6 millimetres; CET 9.6˚C) featured two very cold days at the beginning but was the beginning of the summer when W.G. Grace hit the first two first-class triple centuries. From the context of a warming world, May 1991 – the third-driest on record but with CET 0.4˚C below the virgin mean of approximately 11.2˚C – is also notable although it was obviously a similar but more eastward Atlantic block to September 1986 which I discussed before.

Spring:

As we would expect, in the spring season over the UK, hotter seasons tend to be drier. It’s notable that here one sees the shift brought about by Australian greenhouse gas emissions much more as the diamonds on the upper right (hot and wet) are almost all red.

In contrast to the numerous moderate “warm-wet” ouliers brought about by man-made global warming, there are two exceedingly marked “cold-dry” outliers from before 1974: the record-cold spring of 1837, which was 0.92˚C colder than the winter of 1833/1834, and the record-dry spring of 1785, the core of easily the driest fiscal year since 1750.
Spring seasonMarch EWPAnomalyApril EWPAnomalyMay EWPAnomalyMarch CETMarch CET AnomalyApril CETApril CET anomalyMay CETMay CET anomaly
178518.8 mm-40.0 mm10.1 mm-48.3 mm25.9 mm-38.4 mm1.2˚C-4.1˚C8.4˚C+0.4˚C12.3˚C+1.1˚C
183730.4 mm-28.4 mm50.4 mm-8.0 mm36.7 mm-27.6 mm2.3˚C-3.0˚C4.7˚C-3.3˚C9.9˚C-1.3˚C

June:

For June, the slope of the EWP versus CET graph is strongly negative, as for all the summer months. Despite two major “hot-wet” outliers since 1974 in 1982 and 2007, there is not the concentration of red diamonds in the upper right that we saw for the spring season. The major “cool-dry” outlier is the foggy, obviously northerly and blocked June 1923, which was very cool over Europe and uniquely wet in Australia, where the Mallee received up to five times its normal rainfall. June 1923 clearly had a large Arctic block causing hot weather in Canada as well as the cool in Europe:
It’s notable that in both May 1991 and June 1923 Scotland was less cool than southern England, no doubt because it was nearer the centre of the block and less exposed to Arctic airflows.

Fiscal Year (July to June):

Although I have not yet calculated the Spearman and Pearson correlation coefficients, the full fiscal year EWP versus CET scatter plot suggest that the positive correlations between November and February outweigh the negative correlations we observe from April to September.

It is possible, though I have not checked, that the choice of the fiscal year over other possible twelve-month ranges affects the result by dividing over two years extreme hot and dry summers like 1826, 1868, 1911, 1921 and 1976. Such dry years as 1826 and 1921 were in fact hot as a whole, despite the association of dryness with cold during the winter, the quintessential “continental” year of 1780 was only 0.11˚ cooler than the 1766 to 1974 average despite an extremely cold January, and 1947 with its long, hot summer was 0.61˚C hotter despite its record cold February.

Compared to the individual months, outliers are more numerous for the whole fiscal year. The very cold year of 1813/1814 and the very dry year of 1784/1785, as well as the very wet anthropogenic years of 2000/2001, 2006/2007 and 2013/2014, especially stand out. The hot, dry anthropogenic years of 1975/1976 and 1991/1992 are also marked outliers, as is the extreme “cold-wet” outlier of 1878/1879, where a CET of 7.30˚C was recorded for the twelve months ending October 1879.

No comments: