Aurora catalogue
1 Dunstanburgh Castle, Northumberland, silhouetted by an auroral display on 16 February 2015. Downloaded from https://academic.oup.com/astrogeo/article/56/4/4.25/210105 by guest on 23 September 2021
An arch in the UK Mike Lockwood and Luke found 3090 newspaper reports of auroral Britain, the transition to the Gregorian cal- sightings, the most useful being regular endar was made on 25 March 1752 and so a Barnard report on a new monthly meteorological reports. The years correction of 11 days is needed for reports catalogue of auroral observations after 1923 have yet to be analysed, but are between 1700 and 1752, many of which made in the British Isles and Ireland. well reported in observatory yearbooks. were published in Philosophical Transactions. The yield from newspapers compares However, for other reports the correction with 5278 reports from the scientific lit- has already been made; care must be taken rches, streamers, polar lights, merry erature and 2034 from amateur observers. not to apply it twice. dancers… just a few of many names However, the largest contribution comes After removal of duplicates the dataset Aused to describe the aurora borealis from observatory yearbooks and reports, contains 17 761 independent reports from (figure 1) in historical documents in the which yielded 9239. Parish and county known locations between 1705 and 1950 UK. We have compiled a new catalogue archives have proved difficult to exploit (our on 7640 auroral nights (an average of 2.33 of 20 591 independent reports of auroral survey contains just 26 such per auroral night). For 1951 to sightings from the British Isles and Ireland reports); these are not widely “The dataset contains 1975 we do not, at the time of for 1700–1975 using observatory yearbooks, available in electronic form. 17 761 reports from writing, have the coordinates the diaries of amateur observers, news- We have added these to 2958 known locations from of individual sightings but paper reports and the scientific literature. reports of sightings from the 1705 to 1950” we do have summaries of the Our aim is to provide an independent data British Isles listed in previous totals as a function of latitude; series that can aid understanding of long- international surveys, in particular those by adding these allows us to identify 10 390 term solar variability, alongside cosmo- Fritz (1873, 1881) and Lovering (1868). auroral nights between 1705 and 1975. genic isotope data and historic records of This total of 20 591 reports contains dupli- geomagnetic activity and sunspots. cates of the same observation and so we Geomagnetic latitudes The most time-consuming part of this have stored details, wherever possible, such A feature of the British Isles is that they task has been sorting the newspaper as the observer’s name, the source and the cover a wide range of geomagnetic latitudes reports: a search of the British Newspaper location, to help identify these. The location – about 9° in all. The secular change in the Archive (compiled in partnership with the of a reported sighting tends to be general- geomagnetic field makes this spread of lati- British Library) for the term aurora borealis ized as it is copied from one catalogue to tudes a major advantage. To find the geo- revealed 13 673 articles. We have had to another. For example, many early sightings magnetic latitude of each location at a given sift out those relating to: racehorses and come from King’s Lynn (by Martin Folkes date we here use a spline of the Interna- ships of the same name; travelogues about and William Rastrick) but this is some- tional Geomagnetic Reference Field (IGRF) the polar regions; articles about scientific times generalized to “Norfolk” or even to and the gufm1 geomagnetic field models theories; articles and letters proposing false “England”: hence, one sighting can lead to (Jackson et al. 2000), with some smoothing associations with the weather; descriptions three or more, apparently different reports. to remove small discontinuities between of light shows, paintings and works of lit- Also, place names change: King’s Lynn the two at 1900. These models are used to erature; and advertisements for fireworks. has been referred to as Lynn, Lun, Lenn, compute the field inclination, i, which is We here define a “sighting” as a night on Lenne Regis, Lynn Regis, Len Episcopi and converted into a “dip” geomagnetic lati- which aurora was observed before dawn; Bishop’s Lynn. Metadata is also particularly tude, L, using the formula tan (L) = tan (i)/2. observations after midnight but before valuable in identifying observations that Values of L were evaluated at grid points dawn are given the date of the start of the have been duplicated because of the confu- with spacing 1° in latitude and 2° in lon- night before midnight. We have, so far, sion between Julian and Gregorian dates. In gitude and 2-dimensional interpolation
A&G • August 2015 • Vol. 56 • www.astrongeo.com 4.25 Aurora catalogue was used to derive L for the coordinates 68 of each observation. The left-hand panel of figure 2 showsL as a function of date 66 for various sites important in the history of auroral observations in the British Isles. 64 )
The right-hand panel shows the distri- o ( 62 bution in magnetic latitudes of auroral Λ observations from known sites (computed 60 in the same way) in seven catalogues from the northern hemisphere from 1650 to 1950. 58 The catalogues used are by Fritz (1873, 1881) Haroldswick and Lovering (1868) (both global), Silver- 56 Aberdeen man (New England), Silverman (all USA; see Silverman 1992), Rubenson (1879, 1882; 54 Edinburgh Sweden), Nevanlinna (1995; Finland) and Kendal
52 Downloaded from https://academic.oup.com/astrogeo/article/56/4/4.25/210105 by guest on 23 September 2021
the new catalogue described here (British geomagnetic dip latitude, Kings Lynn Isles and Ireland). The distribution shows a 50 marked peak at L between 57° and 63° (the London two dashed lines in the left-hand panel). 48 Plymouth The decay at L < 57° with decreasing L is probably accurate because throughout the 46 1650 1700 1750 1800 1850 1900 1950 2000 0 5 10 period there were many available observ- year N(Λ)dΛ/1000 ers at these latitudes. However, the decay at higher L may be influenced by a decrease 2 (Left): The geomagnetic dip latitude L of various sites in the British Isles as a function of date. in the available observers with increasing (Right): The distribution of auroral sightings with L (from well-specified locations taken from a basket of L. We here refer to the band of peak occur- seven catalogues from various longitudes in the northern hemisphere and including the new catalogue rence at 57° ≤ L ≤ 63° as the “auroral oval”. presented here): the number of sightings during 1650–1950 at magnetic latitudes between L and The large changes in L with time for a L + dL, N(L)dL, is shown as a function of L for dL = 1°. All magnetic latitudes are computed in the same given site are caused by the secular changes way from a spline of the IGRF and gufm1 models. in the geomagnetic field associated with the motions of the magnetic pole. The red observing locations moved poleward Scotland included: Rev. A C Henderson, line shows L for Haroldswick, on the Island to Edinburgh, Aberdeen and eventu- who made observations from Whalsay of Unst in Shetland, which was poleward of ally the Orkney, Hebrides and Shetland in Shetland, then from Buckie and Forfar, the auroral oval at the start of the interval islands. Given that these areas are sparsely between 1912 and 1918; and Rev. W M but has been within it since 1796. Aberdeen populated, one might have expected the Dunbar of Applegarth Manse, Dumfries was also poleward of the oval initially, but rate of auroral recordings to fall. That they (observations 1838–1848). But clergymen “moved” to lower L such that it was within did not has much to do with an engineer looking to the heavens (in more than one it for 1783–1853. Edinburgh was initially called Robert Stevenson and his three sons sense) was far from a new phenomenon; close to the poleward edge of the oval but David, Alan and Thomas who designed there are many prior examples from further moved equatorward of it in 1841. Ken- and built 55 lighthouses around Scotland south including: nonconformist preacher dal, King’s Lynn, London and Plymouth between 1811 and 1886. Lighthouse keepers William Rastrick (King’s Lynn, 1722–1726); were all within the oval at the start of the regularly reported aurorae until operations and Rev. William Derham (Upminster, interval, but moved equatorward of it (leav- became mechanised in the second half of Redbridge and Windsor, 1707–1728). ing the oval in 1824, 1796, 1788 and 1789, the 20th century. Between 1920 and 1950, There were also a great many secular respectively). As sites moved equatorward there were 813 auroral observations from amateurs such as physician Thomas of the oval in the geomagnetic reference lighthouses constructed by the Stevensons Hughes from Stroud, Gloucestershire, frame, the occurrence frequency of auroral and 1229 from all Scottish lighthouses. who recorded aurorae and meteorologi- sightings there fell. At the end of the Maun- cal information in his diaries from 1771 to der minimum (1715), centres of population Amateurs and gentlemen scientists 1813 (Harrison 2005). Those with inherited such as London, Plymouth and King’s Lynn Amateur observers also made vital con- wealth were able to devote themselves were ideally placed to observe aurorae. tributions. Anthony Johnson reported 300 to science. One such was Edward Joseph So, for example, John Huxham, a physi- auroral nights between 1913 and 1930 from Lowe, who began his observations of natu- cian with a mixed reputation (detailed in Haroldswick on Unst, Shetland and the ral phenomena in 1840 at the age of 15 at his the intriguing biography by Schupbach Rev. Charles Clouston made an astonishing family home of Highfield House, Not- 1981), observed and recorded aurorae on 875 reports of auroral nights between 1833 tinghamshire (now part of the University 122 nights between 1728 and 1748 from and 1882 from the manse in Sandwick, Ork- of Nottingham). A co-founder of the Royal Plymouth (Herrick 1838). Between 1786 ney. Clouston was one of many clergymen Meteorological Society, Lowe gained the and 1793, John Dalton in Kendal (with his who regularly observed and recorded natu- nickname The Big Snowflake on account of collaborator, Peter Crosthwaite in Keswick) ral phenomena. He was related by marriage his impressive bushy white beard. Auroral recorded 240 auroral nights (Dalton 1834) to Balfour Stewart, who first postulated a sightings were part of his book Natural Phe- and figure 2 shows that by this time the conducting layer in the atmosphere (the nomena and Chronology of the Seasons (1870) auroral oval had moved north such that ionosphere) as a cause of geomagnetic vari- and have been included in several auroral Kendal was close to the magnetic latitude ations and after whom the Balfour Stewart catalogues. Lowe’s own copy of his book at which Huxham had made his observa- Auroral Laboratory (BSAL) at the Univer- was recently discovered in the Museum of tions half a century earlier. Kendal became sity of Edinburgh was named. Rural Life at the University of Reading and equatorward of the oval in 1824; optimum Other notable clergy observers in has proved to be valuable for its marginalia.
4.26 A&G • August 2015 • Vol. 56 • www.astrongeo.com Aurora catalogue
The first national network of amateur academics of the day began to recognize any such sight either from home or abroad.” observers was assembled by George Dalton’s genius. In 1810 he refused Sir This analysis did omit some isolated J Symons (1838–1900), who contacted, Humphry Davey’s offer to propose him for sightings in 1661 from London (reported in befriended and encouraged amateur fellowship of the Royal Society (probably newspapers and the Leipzig University the- observers, founding and managing the because he could not afford the annual ses by Starck and Früauff) and Derham’s in British Rainfall Organisation. As well as subscription); nevertheless, he was pro- Upminster. A similar reappearance of auro- recording rainfall data, he compiled lists posed and elected without his knowledge rae was reported between 1716 and 1720 in of phenological observations from his in 1822, and elected as a member of the Italy and in New England (Siscoe 1980). network of observers, and of aurorae. From French Académie des Sciences in 1816. He The almost complete absence of auroral 1860, he published these observations in made his last meteorological recording the reports from the British Isles during the annual booklets entitled The Distribution day before he died in 1844. Dalton’s careful Maunder minimum is remarkable. As of Rain over the British Isles, continued after catalogue of observations clearly showed figure 2 shows, all the major centres of his death by the BRO and taken over by the the decline in auroral activity and so ended population in southern England were ideal Met. Office in 1919. These also appeared any serious debate about the reality of the locations to see aurora under normal condi- from 1866 onwards, in Symons’s Monthly Maunder minimum and the notion that tions. There were a few isolated auroral
Meteorological Magazine (which became the there was no secular change in solar activ- reports from mainland Europe during Downloaded from https://academic.oup.com/astrogeo/article/56/4/4.25/210105 by guest on 23 September 2021 Meteorological Magazine in 1920 and was ity. This debate still resurfaces today – see the Maunder minimum and sailors and published until 1993). the rebuttal by Usoskin et al. (2015). fishermen who visited the far north were It now seems strange that these reports told of continued sightings during this time appear in weather bulletins; however, The Maunder minimum by locals. Thus it seems likely that aurora when Symons started, many people were On the night of Tuesday 17 March 1716 continued during the Maunder minimum, convinced that the aurora was a harbinger (Gregorian or “new” date), auroral displays possibly with less intensity and a lower of bad weather. The name “meteorology” were seen across much of Europe, includ- latitudinal span, and usually at very high comes from the Greek meaning “the study ing in London. Very few people in the Brit- latitudes. The few lower-latitude aurorae of things in the air” and the term “meteors” ish Isles had seen aurora before (Fara 1996). that were seen may well have been stable in the 18th century was used to refer to all The event was so rare that it provoked the sub-auroral red (SAR) arcs generated from luminous appearances in the sky, including commissioning of reviews by the Royal ring current particles or, as noted by Zhang aurora. After 1892, sightings from amateur Society (carried out by Edmond Halley) (1985) for ancient Korean observations, pos- observers were also published in the jour- and by l’Académie des Sciences of Paris sibly even intense airglow or zodiacal light. nals of the British Astronomical Associa- (by Giacomo Filippo Maraldi, also known After 1716, reports of active auroral arcs, tion (BAA), compiled by G J Burns and after as Jacques Philippe Miraldi). It generated arches and streamers in the UK became 1925 by W B Housman. These data are in much interest in the Royal Prussian Acad- much more common. the present catalogue. emy of Sciences in Berlin, particularly from The remarkable John Dalton also con- the husband-and-wife team of Gottfried Bombing of London tributed greatly to our catalogue. Reading and Maria Kirch, and later their daugh- On the night of 7–8 March 1918, aurora was the two editions of Dalton’s Meteorological ter Christine. All these reviews found observed throughout south and east Eng- Observations and Essays (published in 1793 evidence of earlier aurora, but very few land (see Nature editorial, 1918). There are and 1834), one is struck by during what we now call the Met. Office reports, newspaper articles and the precise and methodical “Dalton’s observations Maunder minimum. BAA reports and the associated magnetic way that all his recordings ended serious debate Halley himself had disturbance was monitored at Kew obser- (including of aurora) were about the reality of the observed the 1716 event, and vatory. This night became significant in classified and tabulated – Maunder minimum” correctly noted that the auro- the history of British auroral observations with none of the flowery ral forms were aligned with because of its role in early air warfare. language and unnecessary detail that the geomagnetic field. Halley was an avid Aerial bombardments in the first world characterized contemporary science watcher of the skies who had wanted to see war began in 1915. Many of these early air publications. With his ground-breaking aurora and actively searched for it: “… [of] raids hit small villages, largely because research into atomic physics, gas laws, all the several sorts of meteors [atmospheric navigation was so crude. London and other meteorology and colour blindness, Dalton phenomenon] I have hitherto heard or read cities were bombed in 23 airship raids in was very much a forerunner of the modern of, this [aurora] was the only one I had not 1916. Both sides noted that the raids were professional scientist. Yet for most of his as yet seen, and of which I began to despair, more effective when moonlight overcame life he was an amateur, because he was a since it is certain it hath not happen’d to the navigation problem. This became part Quaker and so classed as a “dissenter”. any remarkable degree in this part of of the strategy for the British air defences This truly brilliant man was barred from England since I was born [1656]; nor is the and for the German aviators who increas- attending or holding a post in a university. like recorded in the English Annales since ingly used aircraft rather than Zeppelins. From the age of 15 (in 1786) he and his the Year of our Lord 1574.” In his paper to On the night of 7–8 March 1918, seven brother ran a Quaker school in Kendal, the Royal Society, Halley lists reports of planes approached London along the Dalton having taught Latin from the age of the phenomenon, both from the UK and Thames estuary, navigating using the 12. He made his first record of an aurora in abroad, in the years 1560, 1564, 1575, 1580, reflection of the bright aurora from the the same year. 1581 (many reported by Brahe in Denmark), water; five turned back but two got In 1893 he moved to Manchester to 1607 (Kepler in Prague) and 1621 (Galileo in through. There was much speculation in become a teacher of mathematics and Venice and Gassendi in Aix, France). Strik- the press that the German air forces had natural philosophy at the New College, a ingly, thereafter Halley found no credible predicted the aurora that had so greatly dissenting academy. After seven years he reports until 1707 (Rømer in Copenhagen helped them – an idea that amuses present- resigned to try to ease the college’s deepen- and various in Berlin) and 1708 (Neve in day scientists who are still working hard ing financial woes, and made a living as a Ireland). He states: “And since then [1621] to gain such a capability with advanced private tutor. The better and more generous for above 80 years we have no account of satellites such as the STEREO mission
A&G • August 2015 • Vol. 56 • www.astrongeo.com 4.27 Aurora catalogue
19 Oct. 1726 24 May 1788 9 No v. 1871 29 Mar. 1943 (a) (b) (c) (d)
Downloaded from https://academic.oup.com/astrogeo/article/56/4/4.25/210105 by guest on 23 September 2021
3 The locations of auroral sightings on a given night (yellow spots). The red lines are contours of constant geomagnetic latitude (L). Note that the contour shown in full is for L = 60° in (a) and (b), whereas it is for L = 55° in (c) and (d). Beneath each map the panels show: the number of auroral nights in the British Isles and
Ireland per year, NA; the group sunspot number, RG; and the open solar flux, FS. The coloured sections show to where each parameter has evolved at that date. These data have been compiled into an animation that can be viewed on our website (http://www.met.reading.ac.uk/~spate).
(Davis et al. 2012) and projects such as Solar of population. On this night aurora was Science implications Stormwatch (Barnard et al. 2014; see also recorded from Plymouth to King’s Lynn. Having compiled the dataset, we are now page 4.20 of this issue). Map 3(b) is for 24 May 1788, when the beginning to study its implications. The Before this air raid, aurora had been L = 60° contour has moved to the north most obvious problem with the data comes reported in a somewhat ad-hoc fashion in of England. Aurora was seen at just four from long-term trends in the availability of the yearbooks of observatories such as Kew, locations on this night. This is typical observers at a given geomagnetic latitude and in the remarks from observers in the of this period, when the data sequence and their ability, and willingness, to record BRO rainfall bulletins. In 1911 these were relies heavily on a very few observers and, what they saw. Cloud cover means that the organized into The British Meteorological and in particular, John Dalton and Thomas data are more robust when the number Magnetic Year Book published by the Met. Hughes. The northward shift of the auroral of observers is high and increasing street Office and the Air Ministry. Initially, aurora oval may have contributed to the drop lighting may have reduced the number of reports were somewhat hidden away, often in regular observers; this is offset by the potential observers. It is remarkable that only denoted by small symbols in tables quality, commitment and rigour of the few the variation in the number of auroral of atmospheric data. The report published active observers. But with fewer observers, nights, NA, in our survey (see figure 4) is in 1919 (covering 1917) introduced a table no matter how committed and skilful, it very similar to that from corresponding of auroral data, probably in response to becomes more likely that cloudy skies will surveys at other longitudes – from the USA, the air raid. This valuable table contin- cause auroral nights to go undetected. Map from Finland and from Sweden, covering ued in exactly the same format in these 3(c) is for 9 November 1871 when the L = 60° the same range of L. The small differences yearbooks (renamed The Observatories’ contour has moved even further north and are all consistent with the changes in L Yearbook in 1922) until 1954. For 1955–1967, the L = 55° contour lies across southern with time at the different longitudes and data from the Lerwick observatory only Scotland. This is a highly active day and the variation with longitude of the length were included, with summaries of the rest aurora was seen at 34 locations between of the observing season (set by sunlight) at (compiled by the BSAL). This ceased in 1975 Guernsey in the Channel Islands and a given L. This gives us real hope that there when the BSAL was closed. These year- Perthshire in Scotland. These observations is important geophysical information avail- books include the lighthouse observations; were mainly made by Symons’ network of able from the series, especially in conjunc- they have been digitized and placed online amateurs (by then extending to Orkney) tion with other surveys. by the NERC’s British Geological Survey. and other people recorded in newspapers. The colour histogram in figure 4(a) shows Map 3(d) is for 29 March 1943 when aurora the number of auroral nights per year in
Maps of auroral sightings was seen from 29 sites between Leeds and the British Isles and Ireland, NA. The black Figure 3 shows four daily maps of auroral Baltasound on the island of Unst in Shet- line shows the results of the global survey sightings. Map (a) is for 19 October 1726, land. These data were mainly supplied by by Krivský & Pejml (1988) who combined early in the data sequence when the L = 60° observatory staff and lighthouse keepers several catalogues. This covers the interval contour lies across the south of England and compiled into the joint Air Ministry/ up to 1900 and is extended here using the and aurora were visible from major centres Met. Office reports. global annual values of Legrand & Simon
4.28 A&G • August 2015 • Vol. 56 • www.astrongeo.com Aurora catalogue
activity. Hence, we are employing statisti- (a) cal methods to use the different datasets to quantify any effects of variability in the efficiency of auroral detection. Figure 4 demonstrates the well-known relationship between auroral occurrence (b) and sunspot number, but also reveals that this relationship is not as simple as is often assumed. First, the relationship to sunspot
numbers is highly nonlinear, with NA fall- ing to very low values during the Dalton (c) minimum (around solar cycle 6) and even during the less-pronounced Gleissberg minimum (around cycle 14). Secondly, the variation in the amplitudes of the cycle Downloaded from https://academic.oup.com/astrogeo/article/56/4/4.25/210105 by guest on 23 September 2021 (d) peaks in NA and RG shows some similarities but also considerable differences. Thirdly, some solar cycles are very different in
waveform with very large peaks in NA during the declining phase. In particular, cycle 18 shows a maximum almost at the end of that cycle, a feature also seen in cycle
4 Time series plots of: (a) the number of reported auroral nights in the British Isles and Ireland, NA; 10. Figure 4(b) shows that the OSF usually
(b) the open solar flux, FS; (c) the group sunspot number, RG; (d) the average number of distinct locations peaks after the sunspot number but this from which aurora is observed on each auroral night, nS. The annual histogram elements are coloured is not enough to explain the extremely according to the magnitude of the parameter plotted. The grey and white areas highlight odd- and delayed peaks in the number of aurorae. even-numbered sunspot cycles, respectively, which are numbered along the top of panel (a). The black Figure 5 shows why we are confident line in (a) shows the results for NA for the global combination of catalogues compiled by Krivský & Pejml that we can overcome the limitations of the (1988) which has been extended to after 1900 using the annual global values by Legrand & Simon (1987). auroral sighting data and derive mean- In (b) the blue line shows 11-year running means of the annual values and the cyan line is derived from ingful geophysical insights. The left plot
NA using a polynomial fit derived from a fit to the reconstructed OSF values after 1845, see figure 5(b). shows the mean NA value as a function of the mean FS for averaging over FS bins that 5 (1987). The UK catalogue has very similar more common. The regression slope of nS are 0.06 × 10 Wb wide. Here we have only long-term variation. Figure 4(b) shows the against NA over 20th-century solar cycles used data after 1845, for which the OSF was open solar flux (OSF), FS, and 4(c) the group is very similar to that for the long-term reconstructed using a number of pairings sunspot number, RG. The OSF is as derived variation over the whole period, which of geomagnetic data series (Lockwood et from geomagnetic activity data for 1835 suggests that the long-term variation of al. 2014). It can be seen that, on average, NA onwards by Lockwood et al. (2014) using the both is real and not strongly influenced increases monotonically with FS but the procedures discussed by Lockwood (2013). by the availability of observers. Note that spread in the data (given by the standard This is here extended to before 1835 using there are years early in the series when the deviation and quartile ranges shown) is the model of Owens & Lockwood (2012), as average number of sites from which an very large at larger FS. The right-hand plot applied by Lockwood & Owens (2014), and aurora night is observed approaches four, shows the result of the same procedure using RG as model input. There is an anom- which is in the upper decile of values for applied to 11-year running means of the alously high value of observed NA in the recent decades. This indicates that larger data. The nonlinearity of the relationship UK in 1941, not seen in the other surveys; events were detected and recorded soon is now clear and the scatter in the data is this may be caused by the glow from fires after the Maunder minimum as effectively greatly reduced, especially away from the generated by second world war air raids as they were in the 20th century. How- change in slope in the middle of the plot. (which peaked in 1941) being mistaken ever, this alone does not mean that we can The dashed line shows a 6th-order polyno- for red aurora. If so, it is somewhat ironic treat the sequence shown in figure 4(a) as mial fit to these smoothed data, constrained because our survey has shown that the homogeneous. For example, it may be that to pass through the origin. inverse story, of citizens or firemen rushing for early years only active, dynamic auroral Using the fit shown in figure 5(b), the to a fire that turned out to be aurora, is as displays (“streamers”) were recorded, 11-year smoothed NA data,
A&G • August 2015 • Vol. 56 • www.astrongeo.com 4.29 Aurora catalogue closer when the UK auroral data are com- bined with the data from elsewhere. At this point it is worth asking the ques- tion “what causes low-latitude aurora?”. Combined analysis of global ultraviolet auroral images from the Swedish Viking satellite and the Tsyganenko model of the magnetospheric magnetic field have indi- cated that the main, lower-latitude auroral oval maps to the earthward edge of the cross-tail current in the plasmasheet of the geomagnetic tail (Elphinstone et al. 1993, 1995). This current sheet thins and moves earthward under the additional magnetic pressure exerted by the magnetic fields in the tail lobe when the open magnetospheric Downloaded from https://academic.oup.com/astrogeo/article/56/4/4.25/210105 by guest on 23 September 2021 flux is increased by enhanced magneto- pause reconnection. Thus the main auroral oval migrates equatorward during the growth phase of substorms (Milan et al. 2012). The expanding–contracting polar cap paradigm predicts that this growth-phase increase in open flux is later balanced by an increase in the rate at which open flux 5 Mean values for 1845–1975 of the number of auroral nights, NA, as a function of mean values of the 5 5 is lost by reconnection in the cross-tail cur- open solar flux, FS, for bins of FS that are 0.06 × 10 Wb wide and centred on [0.03:0.06:0.51] × 10 Wb, left) rent sheet (Lockwood et al. 2009). The lag for annual values and (right) 11-year running means of annual values. Error bars are ± one standard devia- time, t, in the tail response means that the tion in both dimensions and grey areas show the quartile ranges for each bin in both dimensions. The long-term average of the magnetospheric dashed line is a 6th-order polynomial fit to the smoothed data, constrained to pass through the origin. open flux is enhanced when the dayside reconnection rate is enhanced and this is variations in the energetic particle popula- the series, with locations, up to the present reflected in average positions of the auroral tion in the near-Earth plasma sheet. day. This will enable us to study in more oval that are further equatorward (Milan detail how the data depend on near-Earth et al. 2009). On annual timescales, the Future work interplanetary and magnetospheric condi- magnetopause reconnection rate depends The catalogue of locations of auroral sight- tions, as measured by spacecraft. on the magnitude of the interplanetary ing presented here covers the years from One reason why this work is important magnetic field (IMF) and hence on the OSF; 1700 to 1951. The series of data on auroral is that we have very few observations that the north–south IMF orientation factor nights has been extended up to 1975 using allow us to study centennial variations in averages to a near-constant value on these the summaries published in the Observa- the Sun and near-Earth space. The recent timescales, see Lockwood (2013). From this tory magazine by the BSAL, work that was decline in solar activity raises the real pos- we would expect NA to vary in a very simi- carried out mainly by James Paton and his sibility of a return to Maunder minimum lar way to OSF. Figure 4 shows that this is successor, Douglas H MacIntosh. These conditions within about 50 years (Barnard et only partly true. In particular, we note that data, giving the locations of all sightings, al. 2011), with implications for the operation there seems to be an “accumulative” effect are now held on paper at Aberdeen Univer- and design of all systems that are influenced whereby NA grows after several cycles of sity. Work has continued since then under by space weather. In addition, recent model- large OSF. Two possibilities that we will the auspices of the BAA and these data are ling (Macock et al. 2015) shows implications investigate are if this is related to long-term held, again as paper records, at the RAS’s for the stratosphere and even tropospheric variations in the time constant t a nd/or i f Burlington House in London. We hope in climate in some regions but little effect glob- there is any effect of potential cycle-to-cycle the near future to add these data to extend ally; see review by Lockwood (2012). ●
Authors Jezebel O’Reilly as part of her References Sohn, Vienna) ics 119 (7) 5193 Nevanlinna H 1995 J. Geomag. Mike Lockwood and Luke student project. We also thank Barnard L et al. 2011 Geophys. Fritz H 1881 Das Polarlicht Leip- Lockwood M et al. 2009 J. Geoelectr. 47 953 Barnard, Department of University of Reading colleagues Res. Lett. 38 L16103 zig (F A Brockhaus, Germany) Geophys. Res. 114 A01210 Owens M J & M Lockwood Meteorology, University of Clare Watt, Matthew Owens and Barnard L et al. 2014 Space Harrison R G 2005 Astron. & Lockwood M et al. 2014 2012 J. Geophys. Res. 117 A04102 Reading, UK. Chris Scott for many valuable Weather 12 657 Geophys. 46 4.31 Annales. Geophys. 32 383 Rubenson R 1879 Akad. discussions, the staff of Reading’s Barnard L et al. 2015 Astron. & Herrick E C 1838 Amer. J. Arts & Lovering J 1868 Memoirs Am. Handl. 15 Acknowledgments library and Museum of Rural Life, Geophys. 56 4.20 Sci. 38 January (8) 297 Acad. Arts. & Sci. New series Rubenson R 1882 Akad. The authors are grateful to the RAS Library at Burlington Dalton J 1834 Meteorological Jackson A et al. 2000 Phil. 10 (1) 9 Handl. 18 many individuals who aided the House, and NERC’s British Observations and Essays second Trans. Roy. Soc. A 358 957 Maycock A C 2015 Possible Schupbach W 1981 Medical compilation of this catalogue, Geological Survey in Edinburgh edition (Harrison and Crossfield, Krivský L & Pejml K 1988 impacts of a future Grand Solar History 24 415 directly or indirectly. In particular for digitizing the observatory Manchester) Publications of the Astronomical Minimum on climate: strato- Silverman S M 1992 Rev. we thank Roger Brugge of yearbooks. This work has been Davis C J et al. 2012 Space Institute of the Czechoslovak spheric and global circulation Geophys. 30 (4) 333 the University of Reading for funded by STFC consolidated Weather 10 S02003 Academy of Sciences 75 changes J. Geophys. Res. Atmos. Siscoe G 1980 Rev. Geophys. supplying a complete set of the grant number ST/M000885/1. Elphinstone R D et al. 1993 J. Legrand J P & Simon P A 1987 in press Space Phys. 18 647 BRO rainfall bulletins, and Keith Atmos. Terr. Phys. 55 14 1741 Annales Geophys. 5 161 Milan S E et al. 2009 Annales Usoskin I G et al. 2015 The Johnson of Loch of Hundland, Catalogue/Website Elphinstone R D et al. 1995 J. Lockwood M 2012 Surveys in Geophys. 27 659 Grand Maunder minimum in Orkney, for supplying previously A listing of the full catalogue Geophys. Res. 100 A7 12 075 Geophysics 33 (3) 503 Milan S E et al. 2012 J. Geophys. 1645–1715: a reassessment from uncollected observations by the and the animation of the Fara P 1996 J. Hist. Astron. Lockwood M 2013 Living Res. 117 A03226 multiple datasets Ap. J. in press Rev. C Clouston. The auroral data auroral sightings will be posted 27 239 Reviews in Solar Physics 10 4 Nature 1918 Editorial: Aurora Zhang Z W 1985 J. British were extracted from E J Lowe’s shortly at http://www.met. Fritz H 1873 Verzeichniss Beo- Lockwood M & M J Owens and Magnetic Storm of March Astronomical Assoc. 95 205 personal copy of his book by reading.ac.uk/~spate bachteter Polarlichter (C Gerold’s 2014 J. Geophys. Res. Space Phys- 7–8 Nature 101 32
4.30 A&G • August 2015 • Vol. 56 • www.astrongeo.com