zyxwvutsrqzp Weather Vol. 55 December 2000 zyxwvu Meteorological effects of the solar eclipse of 11 August 1999 Edward Hanna Department of Meteorology, University of Reading A rare and spectacular total solar eclipse and cloud, and I received a tremendous crossed the south-west tip of Britain, sweeping response (>140 replies) in answer to my swiftly onwards into continental Europe, request for relevant observations (Hanna during the late morning of 11 August 1999 1999). These came from an array of (usually) (Fig. 1). (The last total solar eclipse in the UK well-equipped amateur and professional barely clipped Shetland in 1954, and the next weather stations, with reports not only from will not occur until 2081, in the Channel the UK but also several from abroad, and they Islands.) The rest of the country was subjected are summarised below. to a deep partial eclipse. For most of Britain, the eclipse lasted from about 1000 to 1235 BST, with maximum obscuration of the Sun by the Observational reports Moon between about 1112 and 1122BST, Some 81 observational reports were deemed exact timings being later further east (Bell good enough to be used in the analysis because 1997). Unfortunately the surface pressure they included frequently and carefully meas- analysis (Fig. 2) shows a weak low approaching ured readings of the above-mentioned ele- Ireland, with an associated cloudy trough hit- ments. An eclipse summary was compiled for ting the south-west peninsula just at the crucial each of these stations but only the more repre- hour when avid eclipse watchers could well sentative summaries for each geographical zyxwvutsrq have done without this nuisance! Ahead of the front, winds were light in a slack south or south-east airflow resulting from a weak transient ridge. Cloud conditions further east and north in Britain were still patchy but generally region, starting in and around the area of totality and generally working outwards, are presented here. I plan to make the full report available in certain specific libraries and archives and/or on the Web; please consult me clearer (Fig. 3 ) . While many of the south coast for the latest status. resorts east of the Isle of Wight had largely clear skies, with typically 7 or 8 hours of bright sunshine, those further west were mainly cloudy with only an hour or less of sunshine. Because of the extensive cloud - extending also over much of neighbouring Europe, and widely spoiling the view in France and Germany - for many the eclipse was probably of greater meteorological than astronomical sig- South-west England zy Neville Bailey at Denbury, Newton Abbot (Devon), measured a slight but nevertheless apparent temperature drop from 15.6'C at 1105 to 15.2 C between 1140 and 1145BST, using his Davis instrument package. The temperature sensor was in a Stevenson screen (SS) nificance. Moreover, it was exceptional in with standard exposure. There were 8 oktas of having a track crossing densely populated land stratocumulus (Sc) and low stratus (St) with masses (most eclipses occur over the sea), and altocumulus (Ac), with occasional partial was a near-midday, summer eclipse. Therefore breaks allowing the eclipse to be seen. Solar it provided a unique opportunity to assess the radiation peaked at 232 W m - * at ~ O ~ O B S T , influence on meteorological parameters such as then fell back to zero by 1110BST and for the surface air temperature (SAT),wind, radiation next two 5-minute readings, but was back up 430 zyxwvuzt Weather Vol. 55 December 2000 zyxwv Fig. 1 The circumstancesof the solar eclipseof 11August 1999for the British Isles. The near-horizontaldark lines mark the degree of obscuration of the Sun (shown as percentages on the right) at maximum eclipse. The nearly vertical curved lines mark the times of greatest eclipse (progressively later eastwards) at five-minute intervals. The zane of totality, over zyxwz CornwaU, south Devon and much of the Channel, is shown. (From Bell 1997.) 43 1 zyxwvutsrqzpyo Weather Vol. 55 December 2000 zyyxxwwzvvyuuxttsswrqrvqpupo Fig. 2 Surfacepressure arralysisfor 1300 Byron 1I August 1999 @om The Met. Ofice Daily Weather Summary) to 123 W m - by 1150BST. A 1 mph* zephyr with gusts to 3-4 mph fell calm or almost calm (Omph with the odd 1mph gust) for the hour centred on mid-eclipse. Graham Easterling took data from a relatively sheltered, south-east-facing site, about 1km from the sea and 60 m above sea-level, at Penzance (Cornwall). He used maximum and minimum thermometers in a standard SS in the centre of his back garden. The maximum temperature prior to the total eclipse was 15.7 -C at about 0945 BST, and the minimum recorded during the eclipse was 15.0 C at about 1140 BST. He wrote: “There was a sharp shower around 1000, lasting around 20 min. This suppressed the temperature prior to totality, and therefore made the further fall in temperature due to the eclipse appear small. Had the shower not occurred, the [pre-totality] temperature would certainly have risen more, probably peaking about a degree higher. The maximum temperature of 17.8 C [recorded during an evening clearance] was very low for August, and [it] was the coldest day since June 29. August 11 was the most consistently cloudy * 1 mph = 0.86kn 432 day for some time. After a clear sunrise, Ac and Sc quickly spread in from the west, and there was little blue sky left by 0900. From this time Cu [cumulus] developed quickly over the land, giving showers from around 0930, but it remained somewhat clearer over the sea. As totality approached the Cu tended to melt away, leaving medium-level cloud. It was noticeably clearer over the sea. After totality, the partial eclipse was zyx glimpsed at around 1145, as the Cu devel- opment reached its minimum. From 1200 the Cu started to re-form.” The wind was south-east light at ~OOOBST, veering light westerly - probably due to the encroaching weather front - at 1045 BST. Channel Islands At Guernsey Airport Met. Ofice (data courtesy of Tim Lillington), air temperature monitored by a Vaisala HMP35 sensor clearly dipped from 16.9‘C at 1031-1036 to 16.0’C at 1120- 1135 before rebounding to 16.9 ’C by 1220BST. The near-surface grass temperature fell from 22.9-C at 1023 to 15.5 C at 1124BST, and that over concrete from 22.1 C at 1021 to 16.8”C at 1 1 2 1 - 1 1 2 4 ~ s ~th, ese obviously being much larger changes (Fig. 4, zyxwvutzszyr Weather Vol. 55 December 2000 zyxzwyxvwu Fig. 3 AVHRR infrared satellite picture of the UKfor 0820 BSTon 11 August 1999. Note the colder (higherlthicker) zyxwv cloud spillingin across south-west England. (MP0Universityof Dundee.) see back cover). However, there was no noticeable change in the 15 cm depth soil temperature, which hovered at around 17.1 "C. No sunshine was recorded between 1000 and Pallot), an open-site air temperature, measured using a Rosemount platinum resistance thermometer, fell from 18.3 "C at 0936 to 16.4"C at 1124BST. There was extensive (6-7 oktas) 1300BST; solar radiation gauged by a Kipp & Sc/Ac at about 2100m and 1-3 oktas Cu at Zonen CM 11/14 pyranometer fell from 53.89 about 500m. "It was possible to obtain the zyxwv Wm-2 at 1001 to 1.13Wm-' at 1116, then rose to 58.89 Wm-2 at 1226BST. The light south-south-east wind, measured using a Vaisala WA21M anemometer at the end of the runway, slackened off from around 6 to 4kn from 1121 to 1143BST. odd glimpse of the sun through the Ac layers sufficient to observe the eclipse process." Wind was south-east veering south, about 4-9 kn. southern At Tersev Airport (data courtesy of Anthony Nigel Paice, observer at Middle Wallop Met. 433 zyxwvutsrqpo Weather Vol. 55 December 2000 Office, Stockbridge (Hampshire), took a com- ling dry (wet) temperatures fell from 16.1 prehensive set of standard manual readings at (15.0)”C at 1014 to 15.0 (14.3)“C at 1109 10 min intervals throughout the eclipse period. (1 114) BST and then followed a plateau until He found little effect on air temperature, after 1124BST before rising. Temperatures although there is a slight and suggestive time- were back up to their initial values by about lagged fall from 16.3”C at 1050 to 15.8 ‘C at 1200BST, and rose about a degree or so higher 1140BST. Grass temperature fell more substan- by 1215BST. Relative humidity (RH) rose from tially, from 20.4-C at 1040 to 14.7”C at 89% at 1009 and 1014 to 94% at 1109 and 1120 BST. The prevailing cloud was Sc, with 1119,thenfellto85%by 1215BST. Ac around mid-eclipse and a little Cu through- At l009BST the weak trough that gave out, giving a total cover of 7 oktas. The thick showers at about 09 15-0930 BST was moving Sc thinned and broke a little to reveal the away eastwards, visible as a line of Cu and eclipse from just after first contact to a few cumulonimbus (Cb) to the east and north-east. zyxwvzu minutes beyond the maximum obscuration of 98%. Mr Paice suggested that cloud cover might have muted the air temperature drop at this location. Wind was 4-6 kn, south-east. Malcolm Walker, Education Officer of the Royal Meteorological Society (RMS), and his wife were on the Isle of Purbeck, 1.5km west- Cloud at 1010BST was about 3 oktas Sc and 4 oktas Cu with Ac above. Over the next half an hour Cu decreased while Sc increased to around 6 oktas. There was Ac and cirrostratus (Cs) above, with only very occasional, fleeting glimpses of the Sun. These generally heavily overcast or gloomy conditions persisted until south-west of Kingston (Dorset), 150m above after 1200BST, after which the Sc decreased to sea-level. They took a comprehensive suite of 3 oktas, and two or three very tiny patches of observations on a bridle path beside a field of blue sky were finally seen at 1215BST. maize where the plants were mostly 1.O-1.5 m A sensitive Deuta Anemo hand-held cup high; there is a very slight slope upwards to the anemometer was used to measure wind speed. south, beyond which the land falls away There was a very light (2-4kn, gusting 6- quickly to the English Channel, 1.7 km distant. 10 kn) south-south-east wind. Mr Walker There is a fairly steep slope to the north, down a field to a tarmacked country lane and, on the other side of the lane, a narrow belt of woodland. Their site experienced a near-total (99.8%) eclipse at 1117BST. zyxwvut The Walkers measured ground (T,) and screen (T,) temperatures using a Maplins electronic digital thermometer that had been checked for accuracy against eight or nine mer- zyxwvu cury-in-glassthermometers owned by the RMS. T, was recorded 1 m above the ground by the external thermometer probe in a portable zyxwvutsr shield attached to a fence post, and represents noticed a sudden rustling of the maize leaves at 1103BST but the anemometer recorded no gusts; could this have been a downdraught? Occasional rustling of the maize leaves was noted again at 1119BST. Chris Watts at Romsey (Hampshire) used an Environmental Measurements Ltd (Sunderland) automatic weather station (AWS) in an open exposure, on a gentle slope facing north. Air temperature fell from 17.8 ’C at 1030 to 15.9 C at 1130 and 1140 BST: “Early morning cloud cover averaged 6 air temperature. Tg was recorded 7cm above oktas, mainly Ci [cirrus], Ac and Sc. The the ground by the sensor inside the thermo- clearest skies were to the north-east. In the meter. The instrument itself sat on an empty early stages of the eclipse cloud increased to leather case and a box lid was used to shield it 7 oktas, but the Sun could still be seen in from direct, but not indirect, solar radiation. A cracks between the clouds and also behind Casella whirling psychrometer was used to measure dry- and wet-bulb temperatures at about 1.75m height. T, (Tg)fell from 15.7 (17.3) C at 1024BST, values at which they had been fairly steady for 15 min beforehand, to c 14.7 (14.8) at 1119 (1124)BST. The whir- the lighter clouds, However, about 30 min before maximum coverage [at Romsey about 99%,1, St approached and soon the sky was completely overcast. The Sun was obliterated and did not reappear during the rest of the eclipse.” 434 zyxwvutsrqp Weather Vol. 55 December 2000 There was only 72s of bright sunshine, as ally) taken higher up in or above the aunos- zyzxywxvwutvsur measured by an R&D Electronics sun recorder, between first contact and maximum coverage. The south-easterly wind, which had frequently been averaging around or over 3 (gusting 611)mph from 0820 to 1020BST, temporarily dropped back to about 1-13/* (gusting 36) mph at 1100-1 15OBST. South-east England pheric boundary layer. In confirmation, the weather station readings show a marked drop off in wind speed from about 1.5 to 0.2 m sat about 1135BST. At Beaufort Park, Bracknell (Berkshire) (data courtesy of Michael Collins, The Met. Office), air temperature fell from 17.5"C at 1023BST (still 17.3"C at 1043BST) to 14.5"C at 1 1 3 2 ~ sR~H. rose from 62.2% at 1027 to 88.5% at 1146BST. The ground temperature, Researchers at Alice Holt (Forestry Commis- measured by a freely exposed grass minimum sion), near Farnham (Surrey) (data courtesy of thermometer, fell more starkly from 28.0 "C at Mark Broadmeadow), used a Solent sonic 1024 to 12.5"C at 1 1 2 5 ~ s ~C.eilometer anemometer to measure wind speed and direc- (cloud-base height) data (of cloud directly tion 8 m above the 20 m oak canopy of a forest stand. Temperature and global radiation data were measured at a standard weather station 3km away on a fairly exposed sandy heath. Screened air temperature fell from 18.2"C at 1030 to 14.7"C at 1130BST. Cloud was said to zyxw be variable and the partial eclipse just about visible. Global radiation fell from 613 Wm-2 at 1000 to 13.2 Wm-2 at 112OBST. Wind zyxwv speed hovered at about 1.O-2.5 ms- I; although the wind was already light, perhaps it did not drop off perceptibly around mid- zyxzwy eclipse (unlike at other stations) because the above the station) suggest dissipation of convective cloud for an hour or so centred on mideclipse, with the lower cloud layer base rising from 968 m at 1047 to 2164 m at 1205BST (Fig. 5). Direct (diffuse) short-wave radiation fell from 766.3 (246.0) Wm-2 at 1009 (1039) to 0.0 (3.3) Wm-2 at 1052-1100 (1119)BST, and global radiation followed suit from 679.5 Wm-2 at 1015 to 14.0 Wm-2 at lll9BST (Fig. 6). Wind was variable, mainly south-easterly 1-6kn, falling entirely calm from 1049 to 1155BST. However, "The anemometer logger was a Munro Mk4 with a starting speed of Alice Holt measurements were (unconvention- about 4 kn. This was very disappointing ... 2500 - 2000 e - p 1500 0 al a0 0 - 0 4 8 loo0 - t 0830 0930 zyxwvutzsy 1030 1130 1230 1330 1430 Time (BST) Fig. 5 Lower cloud basefrom Mo&l C ceilometer at Beaufort Park (Berkshire) on 11 August 1999 435 zyxzywxvwuvztzsuyyrtxqsxrwzpwzqoypyvvnoxuxumnw 0830 0900 0930 lo00 1030 1100 1130 1200 1230 1300 1330 1400 1430 lime (BST) Fig. 6 Shon-wave radiation at Beaujort Park (Berkshire) during and around the solar eclipse of 11 August 1999. Direcr radiation indicated by open circles, diffuse by crosses and global by solid line. zyzxywxvuwtsvrquptosnmr 10 0100 0300 0500 0700 0900 1100 1300 Time (BST) 1500 1700 1900 2100 2300 Fig. 7 Visibility ufBeaufort I’urk (Berkshire) (solid line) and Corsock (Kirkcudbrightshire) (open circles) on 11 August 1999 The Met. Office is upgrading its wind systems from Munro to Vector precisely to obtain better low-speed performance.” Visibility dropped from 22.86km at 1015 to 17.25km at 1132Bsl but was back up to 22.08km by 1154 BST (Fig. 7). Martin Rowley at Bracknell (Berkshire), in a typical suburban garden situation, close to the Beaufort Park Met. Office site, used standard 436 mercury-in-glass thermometers calibrated by the supplier, Casella, in a home-built SS. The reading interval vaned from 10-15 min at the start of the period to 1 or 2 min either side of maximum eclipse, lengthening again on the other side. Screen temperature fell from 1 7 . 6 ” C at 1038 to 15.9‘C at 1143BST. T h e temperature indicated by a (non-standard) thermometer exposed to direct-sky radiation at zyxwvutsrqp Weather Vol. 55 December 2000 approximately 0.7 m, shielded from ground- eclipse the Cu, about 3 oktas, cleared comple- based (back-scattered earth) radiation and tely to reappear again after 1210BST. The wind insulated from its supports and stand, fell from was very light and did not appear to change in 254°C at 1018 to 1 5 5 ° C by 1125BST, and either strength or direction. climbed up to 27.4"C by 1223BsT. The Sun Bernard Burton at Wokingham (Berkshire) was directly visible for most of the time, with recorded a "very marked eclipse signature" zyxwvzuyt hard shadows and clearing skies after initially building Ac/Cu, and winds were light and variable falling to calm for several minutes around mid-eclipse. Thick, high Sc moved in from the west after 1300BST. Hugh Thomas sent data from Hurstpierpoint (West Sussex) - a Met. Office-approved, relatively sheltered, back garden station on a using his aspirated psychrometer and anemometer (10 m effective height) (Figs. 9 and 10): "The psychrometer is exposed in the front garden in a suburban environment. It is at a height of 1.5 m, and samples dry and wet bulb [temperature, screened in plastic tubes with aluminium foil] at 1s intervals, logging low sandstone plateau about 12km north of the averaged data over 1 min. Values of the coast and 2.6km north of the crest of the humidity mixing ratio are calculated using zyxwv South Downs. He used standard maximum, minimum and dry-bulb thermometers and a Casella thermograph: "At 1000 the screen max registered 20.0 "C an assumed air pressure of 1013mbar. The solar eclipse depressed the air temperature by >2degC [from 17.88"C at 1026 to 15.35 "C at 1128BSTl. The associated rise in dew point can be attributed to decreased and the thermograph 19.5"C. By 1045 the mixing as the lower boundary layer became dry bulb was down to 19 "C. At the height of decoupled [there was a marked effect of the the partial eclipse at 1118, the dry bulb was eclipse on the 1 min variation in humidity down to 16.5"C and the thermograph mixing ratio, which temporarily became 16.6"C. I had put a max/min thermometer much less variable as convectiodturbulence on the clothes line. At 1010 it registered decreased]. The subsequent gradual 20.5 "C and by 1118 it had fallen to 13.5 "C. increase in dew-point evident after 1300 is At the height of the eclipse, in the still air due to background synoptic advection. with very weak sunshine, it felt distinctly During the eclipse the light wind decreased. chilly and the clothes line madmin seemed to give the more realistic readings. A large area of Sc cloud cleared away c. 0950 and during the period of the partial eclipse there was unbroken sunshine [Fig.8(a)]. From about 1057 to 1127 the sunshine card was During the period 1000 to 1045, Cu cloud (Cu humilis and Cu mediocris) continued to increase in amount and development, with the bases gradually rising in line with normal boundary layer diurnal heating and mixing. Between 1045 and 1130 it was not burnt. There was a slight increase in noted the Cu development became sup- zzyyxwxvwuvtusrtq windforamomentaround 1115but, over- all, there was virtually no wind during the period of the eclipse." Maurice Brockman at Watford (Hertfordshire) had telemetered probes: one for air tem- perature (T,) inside an s ~ a;nother for Sun pressed, firstly with the Cu tops becoming less 'rounded' and more ragged, and then with the amount gradually decreasing. Between 1130 and 1215 there was virtually no Cu. Fresh Cu began to develop at 1215, but the bases seemed a little higher than temperature (Tsunfr)om a probe inside a black before. During the whole period, there was bulb mounted on the roof. T , (Tsunfe)ll from 2 4 oktas of Sc near 2000m, all thin (Sc 17.4 (24.6)"C at 1024 (1000) to 15.2 stratiformis translucidus perlucidus), not (14.6) "C at 1127-1 136 (1121-1 125)BST, enough to prevent the R&D sunshine highlighting the approximate 10min lag of T , recorder recording sunshine, and 3-6 oktas changes in response to changing insolation. thin Ci. Cloud in the Sc layer was gradually The Sun was partially obscured for a period about 1125-1135BST by thin Ac. During the increasing due to the synoptic-scale advection of moister air, and this had thickened to 437 zyxwvutsrqzpy Weather Vol. 55 December 2000 zzyyxxwwzvyuxtwssrvrqquppt Fig.8 Sunshine recordsfrom a Campbell-Stokes recorderfor (a) Hurstpierpoint (WestSussex), (b) UniversityofRead- ing (Berkshire), (c) Northamptonshire Grammar pltsford Hall, and (d) Moel y Crio (Flintshire) for I I August 1999 (tinies art‘ G‘ICIT) give 8 oktas cover in the early afternoon.” measured temperature rose from 13.O‘C at 0900 to 15.4 C at 1050, then dropped back to The air minimum temperature in the screen 14.0 C by 113OBST. Between 0900 and during the eclipse was 15.6 ’C. The grass mini- 1000 BST there was about 7 oktas of sc, but the zyxwvu mum thermometer, which had been reading about 18 C after resetting at 1000BST, probably fell to a minimum of around 11 C during the same period. An open-scale barograph, even using a scaled magnifier which has a resolution of 0.05 mbar, showed no marked varia- sky then soon cleared from the west, leaving scattered Ci and continuous sunshine. Around mid-eclipse, Mr Selfe noted the chill in the air at 14 C, when he reckoned that under normal circumstances for that day the temperature should have been about 19“C (the eventual tion in pressure during the eclipse. maximum temperature was 19°C on a sunny afternoon). The wind remained flat calm, East Anglideastern England except for a very slight breeze for a few seconds just before 1100 BST. R. W. Selfe and family at Benfleet (Essex) used Dr Charles Briscoe at Buxton (Norfolk) French-style maximum and minimum thermo- used a Davis Groweather AWS. The tempera- meters in an ss at the lower (east) end of their turelhumidity sensor - within a white, louvred garden, about 3 m from a tall (approximately radiation shield - and solar recorder were on a 12 m) tree hedge running north-south. The standard Davis mounting in a properly exposed 438 zzyyxwxvwuvtsurqztp zyxwvutsrqponm Weather Vol. 55 December 2000 zyxwvu 0.4 - h cn Y 0) -02 C '0 e a 'E 9 0 EP 0 .0c-) *Xc z.gQ.2 r 3 a.4 0 2 4 6 8 zyxwv 10 12 14 16 18 20 22 24 Time (GMT) Fig. 9 One-minute variation in humidity mixing ratio from the Wokingham (Berkshire) psychrometerfor 11 August 1999 =NN.11 RCRlDlRlKNOTSl24 . Fig. 10 Part of the Wokingham (Berkshire) anemograph tracefor 1 1 August 1999 (time is GMT) (river authority-approved) park site with trees, but none closer than twice their height, and an anemometer was mounted atop his house. Data were recorded every 10 min. Air temperature fell from 15.4"C at 1040 to 13.1"C at 1 150BST.At the same time, RH peaked sharply at 82%, having risen by about 12%, and the already light (3mph) north wind distinctly fell calm. The temperature recovered more rapidly than it had fallen, to 16.2"C at 1250 peaking at 16.9"Cat 1340BST. Conditions were prob- ably mainly sunny, as nearby Cromer recorded the hirrhest sunshine in the country (10.5 I hours). The solar radiation data show a morning peak of 479 Wm-2 at 1020,followed by a fall to 62 W m P 2 at 113OBST (just after mideclipse), then a rise to a much greater peak of 765 W m P 2 at 1330 - already 745 Wm-2 by 1250BST. Hourly measured evapotranspiration suffered a lull at 1200BST when it was just 0.108mm, compared with 0.204mm at 1100 and 0.407mm at 1300BST. Reporting from Greetham (Rutland), former Met. Office professional Paul Bartlett took temperature readings from two garden U- type thermometers, both north-facing, well exposed over grass, right in the middle of nearly open countryside, and checked for accuracy (estimated k0.5 degC) against values from nearby RAF stations. After rising to an initial peak of 17.0"C at 1005 and again at 1025BST, the temperature fell dramatically to 1 1.O "C at 1132,but was back up to 17.0"Cat 1249BST. It was pretty overcast with Sc around 0945BST, despite just being able to glimpse the Sun, but the Sc thinned and cleared from the west during the height of the eclipse, allowing an increasing view of the Sun; shadows eventually emerged after the main event. Winds were light with initially a north-north-east, then a southeast drift, or occasionally calm. Mr Barlett did n_ ot. _ : . 439 z Weather Vol. 55 “think the [6 degC] temperature drop was that surprising in retrospect. Convection at 17 “ C had just begun so there was probably zyx a super-adiabatic lapse rate adnear the surface. Then convection died as the temperature dropped, and resumed at the same zyxwvu temperature after the eclipse ...I admit the deficits of my thermometers, but in such zyxwvut light winds would an ss have altered them? And if they were [indicating too] high or low zyxwvut it wouldn’t change the observed [tempera- December 2000 showed a somewhat lesser drop of about 1degC*]. The recovery of temperature was then quite rapid with pre-eclipse values reached at 1200. RH remained largely constant until 1100, increasing during and immediately after maximum eclipse and then falling after 1145. The rise in temperature up until 1045 was clearly seen in the presence of (convective) Cu as the dominant cloud type, with Ac developing from 5 to 6 oktas between 1000 and 1030. From 1100 ture] drop.” and through maximum eclipse, Cu had dis- appeared completely leaving the higher-level Ac, which had formed an extensive blanket Midlands of 7 oktas. By 1145, cloud types had chang- ed dramatically; Ac having lifted to Cc [cir- At Halesowen Climatological Station in the rocumulus], and convective Cu having West Midlands, part of the Met. Office Co- returned under the marked increase in tem- operating Network (data supplied by Cedric Roberts), dry- (wet-) bulb temperature decreased from 16.5 (13.6) C at 1020 (1013) to 14.1 (12.1) C at 1124 (1121)BST. Grass (slab) temperature fell from 28.6 (18.5) C at 1013 (1026) to 11.6 (14.0) C at 1122 zyxwvzuytx (1123)BST, and surface soil temperature from 25.8 C at 1016 to 14.6 C at 1124BST. The decline in soil temperature was largely muted at even a few centimetres depth. Global radiation fell from 661 W m - 2 at 1016 to 20 Wm-’ at 1115BST. At 1000BST there was 6 oktas Ac, moving from the north-west. Cloud varied quite a lot, being around 7 oktas Ac during the zyxwvuts eclipse; at no time was the Sun completely perature which marked the post-eclipse phase. The pressure readings obtained from the mercurial barometer (uncorrected for gravity or sea-level) reveal a marked increase in pressure from 1000 to 1045, i.e. in the phase immediately following first contact, then a gradual reduction in pressure from maximum eclipse until 1145 when pressure stabilised and rose again. There was a clear lull in wind speed between 1119 and 1145, but a dramatic increase from 1200 to 1215 [and another short peak in wind strength at 1030, 49 min before maximum eclipse]”. obscured. An R&D Electronics Sun sensor, calibrated against a Campbell-Stokes, recorded intermittent sunshine, especially before mid-eclipse. The south-east wind, measured by a Vector certified anemometer, briefly died down from around 4 to 2% kn (Fig. 11). Mike Lewis is the Station Manager at Northamptonshire Grammar Pitsford Hall Weather Station. The station uses official Met. Ofice standard equipment and observation procedures. To quote from Mr Lewis’s thorough report: The sunshine card tellingly shows the diminution of bright sunshine at approximately 1045 and its re-emergence at 1130 (Fig. 8(c)). Matthew Roe at Riverside, Redditch (Worcestershire), used R&D Electronics tempera- ture sensors housed in a standard small ss in a sheltered back garden. Air temperature fell from 16.4”C at 1045 to 15.1‘C at 1130, and rose to 16.8”C at 1215BST. These data agree with those from sheathed mercury thermometers housed in the same screen. The temperature drop of just over 1degC also shows up “. . . up to 1045 a steady increase in tem- perature was experienced to a maximum of 16.2 “C, From 1045 the decrease in temperature was noticeable, reaching a minimum of 14.6 C at 1130 - 11 min after maximum eclipse. [The thermograph ~~ * Thermographs, traces of which were available from several stations, generally showed smaller temperature drops during the eclipse than ordinary (mercury or electronic) thermometers, perhaps due to their greater thermal inertia and slower response to the relatively rapid temperature changes. 440 4- zyzxywzxvywuxtvwsruvqutpstorzsnqyrm 0100 0300 0500 0700 O900 1100 1300 1500 1700 1900 2100 2300 Time (BST) zzyyxxwwvvuzut Fig. 11 Wind speed at Cranfteld (Bedfordshire) (dashed line), Halesowen (West Midlands) (solid line), Normanby (North Yorkshire) (crosses) and Klindre (Powys) (open circles) on 11 August 1999 well in the Casella thermograph trace. There was no discernible change in RH, which stayed fairly steady at about 74%. Cloud was 7 oktas Cu, Sc, Ac and Ci from 1015 to 1100BST, then 6 oktas Ac and Ci. Shallow Cu dispersed by 1115BST, reforming about an hour later; thin Ac covered the Sun throughout. The wind, measured by an R&D Electronics anemometer, was light (2-6mph) and from the south-east; there was a noticeable 8mph gust ofwindat 1 1 1 0 ~ ~ ~ . Northern England John Goulding took readings with a Davis AWS at Normanby, Middlesbrough (North Yorkshire). The temperature probe resided in an SS in a suburban garden with some shelter from surrounding trees, and the anemometer and wind vane were mounted 10m above ground. The temperature levelled off at 15.5"C at 1054-1055, then dropped back to 14.3"C at 1152-1 154BST, before recommencing its rise. The weekly trace from the Casella thermohygrograph mounted within the screen also shows this effect. Cloud was thin and steadily increased from 3 oktas (Ci, Cc and Cs with small Cu) at 0930 to 8 oktas (uniform Cs) at 1200BST, before dispersing somewhat again; the early Cu had vanished by 1130BST. The eclipse was fully visible throughout. The light (1-3mphY with gusts to about 5mph) and veering westerly wind fell entirely calm during the period 1112-1135BST (Fig. 1l), then resumed from a more northerly point. Wales Dr Donald Perkins of Llansadwrn (Anglesey) used standard instruments in a garden weather station. He reports that: "partially cloudy conditions prevailed including Ci and intermittent Cu of varying thickness. The screen temperature drop was 2.1 degC (maximum to minimum) over a period of 85 min. There was an increase in RH peaking at 90% before the temperature minimum. Larger and somewhat faster changes were evident in the probe and grass temperatures. The [digital] probe [with a small white remote sensor] was small and should give a reasonable indication of air 44 1 zyxwvutsrqp Weather Vol. 55 temperature at 2.4m [over grass]. This measured a drop of 4.2degC over a period zyxw of 80 min. The drop in the temperature at December 2000 wind was initially light (1-3kn) and from a northerly point, and sometimes calm, but fell consistently calm (save at 1122) from 1052 to the grass surface was 7.5 degC over a similar 1133BsT(Fig. 11). period. There seems also to have been a small effect on the soil temperatures at 5cm. Though small (only 0.3degC), this is probably significant in view of the number of observations and comparison with data of 12 August. There is even a faint suggestion of lulls in the rise of the 20 and 30cm depth temperatures. There seemed to be a dispersion of [Cu] cloud during the eclipse and an increase in haze. Wind on the day was fairly light but there were some changes which could have been induced by the temperature drop and cloud dispersion.” Scotland Anthony Bowles submitted data from a sophisticated Vaisala AWS at Corsock, a fairly wellexposed moorland site in the upper Urr Valley (Kirkcudbrightshire). Screened temperature fell from 15.0‘C at 1038 to 12.2’C at 1139 and 1145BST. Meanwhile, RH rose from about z 75 to 88%. His wife reported scattered Ac at about 3300m at about the time of the eclipse. The wind was very light (about 1-3kn) and variable, perhaps with a south-easterly bias. Visibility was reduced from about 30-5Okm generally during daylight hours to about 25 km - not far off night-time values - from about Solar radiation, measured using a solar thermopile connected to a millivoltmeter (readings given as a percentage of the normal maximum unobscured by clouds at noon), initially peaked at 74.4% at 1000, then fell back to just 2.3% at 1115, climbing again to 90.7% at zyx 1230 BST. John Goodger at Velindre, Brecon (Powys), a rural Met. Office climatological station, took zyxwvutsr AwS readings, with the temperature probe 1045 to 1130 BST (Fig. 7). J. K. Blackshaw of Elderslie, Johnstone (Renfrewshire), took readings from individual sheathed maximum/minimum thermometers in an ss in a sheltered garden position. He noted only a slight temperature drop, from 13.7 C at 1100 to 13.3 C at 1130 and 1145BST, while the humidity rose slightly from 70 to 76%. The sky was covered with Sc and altostratus (As). A few Cu clouds formed over inside an SS. The temperature profile shows a the hills to the south prior to the eclipse; they short, sharp jump of about 0.5degC just appeared to lose strength and decay slightly before 1000BST when the screen was opened around and just after the eclipse, before grow- for 0900 GMT readings and replenishing of the ing again as the temperature began to climb. wet bulb. Shortly after and for the next hour, The wind fell from force 1 to calm from 1115 temperature levelled out at about 15.2 C, but to 1145 BST (during and just after mid-eclipse) fell slightly and slowly from 15.4 C at 1029 to “but this may just be coincidence as the wind 14.4 C at 1138-1 139, thereafter rapidly rising was so light anyway”. to 18.6 c by 1 2 4 5 ~ ~Fr~om. 0900 to 0950BsT it was bright with 7 oktas of thin Sc and 1-2 oktas of Cu mediocris and fractus Northern Ireland below. From 0950 to 1050 BST cloud thickened and increased to 8 oktas; it remained dull for the next hour. From 1150 to 1200 BST the cloud rapidly thinned (about 7 oktas), and from then until at least 1 3 0 0 t~her~e ~was broken cloud with sunny intervals. The last stages of the eclipse were visible at times. It was already overcast at the time of the main (albeit modest) temperature decrease, so most of this can arguably be attributed to the eclipse. The zyx Trevor Boyd at Helen’s Bay, Bangor (Co. Down), a Met. Office climate station, noted a screened dry- (wet-) bulb temperature drop from 15.8 (12.9)‘C at 1030-1045 (1030) to 13.7 (11.4)-C at 1130 (1115)BST. Meanwhile grass temperature fell from 18.2 to 13.0”C. Ground temperatures at 2 10cm depth were not discernibly affected. It was generally cloudy (7-8 oktas) with small breaks, with the Sun often visible through cloud, but com- 442 zyxwvutsrqp Weather Vol. 55 pletely cloudy briefly at 1115; there were a few zyzxyxwwvuvut drops of rain shortly after 1145BST. There was December 2000 was reduced to Cu humilis between 1130 and 1230 LST (during the eclipse), then it became a light (2-4 kn) south-east drift. Ireland At Dun Laoghaire (Co. Dublin) “the slowly thinning cloud sheet (Ac stratiformis transluci- zyx mediocris again, and at 1530 LST congestus. The eclipse delayed subsequent thundery activity during the afternoon, and also prevented the valley breeze for two hours. The Campbell-Stokes sunshine recorder showed no trace from 1100 to 1120 LST. dus) permitted all phases of the eclipse to be seen. The air temperature fell 0.9degC between 1030 and 1130”, while the tempera- Analysis of UK regional SAT drop ture taken from the grass minimum thermometer, exposed in direct sunshine over the grass surface, fell 3.4degC. The thermograph also shows the fall. Screened (or properly aspirated, e.g. psychrometer) SAT drops and cloud cover recorded during the eclipse (usually during the hour or so centred on 1 1 3 0 s ~ ~are) sum- “The mean depression of the air temperature due to the eclipse can be shown to be around 2degC, as the normal trend in temperature would have been upwards on a relatively sunny morning. The winds stayed constant south-east force 2 or 3. Therefore the factors of wind strength/direction and cloud amount [6 oktas] remained constant zyxwvuts during the event itself, i.e. the temperature changes are due to the eclipse, not some other complicating factor.” The contemporaneous 0.3 degC decline in wet-bulb temperature was less marked than the zyx fall in the dry-bulb temperature. Air tempera- marised by region in Tables 1 and 2. Most acceptable reports were submitted from southeast England (21), followed by south-west England (12) and the Midlands (12), although three other regions each have at least 6. Despite inevitable differences in station exposure, instrumentation and observation method, south-east England clearly had the largest mean SAT drop (2.3degC). This was nearly twice that in cloudier south-west England (1.2 degC) - despite the eclipse being total or more nearly so in the latter region. Indeed, East Anglideast England and the Midlands (where, again, cloud was much more broken) recorded markedly higher SAT drops. ture readings were taken from a standard SAT profiles from seven UK AWSs are plotted screen exposure; the moderately sheltered sta- in Fig. 12 (see back cover). Each station is tion, which lies about l l km south-east of representative of a different region: Ched- Dublin city centre and 1km south-west of the burgh, East Anglideast England; Corsock, Irish coast, reports daily to the Irish Met. Ser- Scotland; Guernsey Airport, Channel Islands; vice (this report thanks to Rob Weatherill). Halesowen, Midlands; Plymouth, south-west England; Portchester, central southern Eng- Continental Europe land; and Reading, south-east England. They were chosen purely on the basis of being com- At Pontremoli, north Tuscany (Italy), Maurizio prehensive, reliable records, already compu- Ratti reports that the air temperature reached a terised, and giving a broad geographical maximum of 26.3 “C at midday, followed by a spread. They all show a clear drop depicting gradual decrease. The maximum eclipse of the eclipse. This is clearly most marked in the 89% occurred at 1237 local summer time (LST, case of Chedburgh (2.2 degC), Corsock 1037 GMT). The temperature dropped to (2.8 degC), Halesowen (2.4 degC) and Read- 24.8”C 25 min later (a drop of 1.5degC). ing (2.1 degC), at none of which skies were Another screen at Pontremoli gauged a overcast; the progress of the eclipse was report- 3.0 degC decrease. The former is in a suburban edly at least partly visible from all these four site near his house, while the latter is an open stations, and indeed it was comparatively clear countryside site along the River Magra, which at Reading. All these stations are in the east or may- ex-ulain the difference. Cu mediocris cloud north of the UK. Portchester (central southern 443 zyxwvzuytsxrwqp Weather Vol. 55 December 2000 v Table 1 Screened (or properly aspirated) surface air temperature (SAT) drops and cloud cover recorded at various sites during the eclipse. Maximum fractional obscuration of the surface of the solar disc by the Moon is also shown. zyx Site Obscuration* SAT drop Cloud cover (degc) (oktas) South-west England Culdrose (Cornwall) Denbury (Devon) Gloucester Lapford (Devon) Liskeard (Cornwall) Norman Lockyer Observatory (Devon) Northay (Somerset) Penzance (Cornwall) Plymouth University (Devon) Salcombe (Devon) Shillingford Abbot (Devon) Westonbirt Arboretum (Gloucestershire) Channel Islands Guernsey Airport Jersey Airport Central southern England Chandlers Ford (Hampshire) Middle Wallop (Hampshire) Portchester (Hampshire) Purbeck (Dorset) Romsey (Hampshire) Shide (Isle of Wight) South-east England Alice Holt (Surrey) Beaufort Park (Berkshire) Birdham (West Sussex) Bracknell (Berkshire) Cranfield (Bedfordshire) Dorking (Surrey) Dover (Kent) Epsom Downs (Surrey) Guildford (Surrey) Hurstpierpoint (West Sussex) Motspur Park (Surrey) Northwood (Middlesex) Reading University (Berkshire) Reigate (Surrey) River (Kent) Royston (Hertfordshire) Steyning (West Sussex) Warlingham (Surrey) Watford (Hertfordshirc) Wokingham (Berkshire) Worcester Park (Surrey) East Anglialeastern England Benfleet (Essex) Buxton (Norfolk) Cavendish (Suffolk) Chedburgh (Suffolk) Morley (Norfolk) Stratford St. Mary (Essex) Winterton (Lincolnshire) 1. O O T 1. O O T 0.96 0.99 1. O O T 1.oo 0.99 1 .OOT 1. O O T 1.OOT 1 .00 0.97 1 .oo 0.99 0.99 0.98 0.99 1.00 0.99 0.99 0.98 0.97 0.99 0.97 0.95 0.97 0.98 0.97 0.97 0.98 0.97 0.97 0.97 0.97 0.97 0.95 0.99 0.97 0.96 0.97 0.97 0.96 0.92 0.94 0.94 0.92 0.95 0.89 0.4 0.4 0.3 (dry bulb) 0.8 (AWS) 2.9 0.8 1.2 2.0 0.7 0.5 2.4 1.9 0.9 cloudy 8 718 7 718 8 (inferred) 718 8 8 8 (inferred) 718 718 zzyyxxwwvuv 0.9 8 1.9 8 1.6 6-8 0.5 7 1.2 8'15h 1.O (electronic) 8 1.1 (whirling psychrometer) 1.9 718 2.2 8 3.3 variable 3.0 part 2.0 scattered 1.7 3-718 2.2 patchy 2.0 not much 2.4 617 2.7 5'15h 3.2 0-2 3.2 not much 3.6 0-2 1.6 patchylwell broken 2.1 not much 1.7 2-4 1.1 7 1.2 not much 3 not much 1% 1-7+ zyxwv 2.2 patchy 2.5 patchy 2.7 3-6 1.4 scattered 2.3 not much 1.8 7'15h 2.2 7 1.5 7 2.4 7 0.7 6-8 Midlands Churchover (Wanvickshire) 0.94 1% 718 444 zzyyxwxwvuvtsurqztzpsy Weather Vol. 55 December 2000 Table I Continued Site Obscuration* SAT drop Cloud cover (degC) (oktas) Halesowen (West Midlands) 0.94 2.4 7 Harborne (West Midlands) 0.94 2.0 6 Harriseahead (Staffordshire) 0.92 1.7 47 zyxwvuts Hockley Heath (West Midlands) 0.94 1.4 Market Harborough (Leicestershire) 0.93 1.7 NorthamptonshireGrammar Pitsford Hall 0.94 1.6 Riverside (Worcestershire) 0.94 1.3 Sutton Coldfield (West Midlands) 0.93 1.5 Tipton (West Midlands) 0.93 1.5 Warstock (West Midlands) 0.93 1% 4-6 7 5-7+ 617 3-6 2-5 broken-scattered Weston Coyney (Staffordshire) 0.92 1.4 3-6 Northern England Beverley (East Yorkshire) 0.89 1.o Carlon-in-Cleveland (North Yorkshire) 0.86 1% Chester (Cheshire) 0.91 1 Drumburgh (Cumbria) 0.85 0.7 High Bradfield (South Yorkshire) 0.90 1.8 Normanby (North Yorkshire) 0.86 1.2 Pateley Bridge (North Yorkshire) Sheffield (South Yorkshire) 0.88 0.90 11.%o 718 415 516 617 6-8’ish 3-8 718 broken Wales Llansadwrn (Anglesey) Moel y Crio (Flintshire) Penmaen (Swansea) Presteigne (Powys) Velindre (Powys) 0.91 2.1 0.91 1.9 0.97 % 0.94 0.9 0.95 1.o 5-7 5-6+ 8, then 87 718 Scotland Corsock (Kirkcudbrightshire) Elderslie (Renfrewshire) Pittrodie (Perth) Westruther (Berwickshire) 0.85 2.8 0.82 0.4 0.80 0.82 11..oo scattered 8 7 718 Northern Ireland Helen’s Bay (Co. Down) 0.87 2.1 718 Ireland Dun Laoghaire (Co.Dublin) 0.91 0.9 6 Continental Europe Pontremoli (Italy) Schliichtern-Elm (Germany) 0.89 1.5 & 3.0 0.97 1.3 patchy thought cloudy *Values inferred to the nearest hundredth from Bell (1997). T indicates a total eclipse. Table 2 Mean screened surface air temperature (SAT) drops during the eclipse for different regions Region Number of sites Mean SAT drop (deg C) South-west England Channel Islands Central southern England South-east England East AnglideasternEngland Midlands Northern England Wales Scotland Northern Ireland Ireland Continental Europe 12 1.2 2 1.4 6 1.4 21 2.3 7 1.8 12 1.6 8 1.2 5 1.3 4 1.3 1 2.1 1 0.9 3 1.9 445 zyxzwyvxuwtvsruq Weather Vol. 55 December 2000 England) had an intermediate SAT drop of well as direct radiation at mid-eclipse. There is 1.2 degC. Much smaller (flatter) SAT declines a widespread misconception amongst the are evident for Guernsey Airport (0.9 degC) and Plymouth (0.5degC) in the south-west quadrant of the UK. The coincident timings in the dips of these SAT profiles from such scattered locations is impressive and leaves no doubt that they can, in large part, be attributed zyxwvuts to the eclipse. Conclusions zyxw Despite the encroaching front introducing zy synoptic changes to the relatively static weather public of shade compared with ‘sun’ temperatures. Since major solar eclipses so clearly affect local and regional weather patterns, it is hoped that this paper will stimulate similar meteorological observations and further research at future such events. The next is the totality scheduled to cross southern Africa on 21 June 200 1, and here weather conditions may well be clearer, although logistics ( i e . a dearth of suitable in situ stations) may preclude widespread useful measurements. However, interesting pattern over the UK, the above reports (mainly results may be obtained through the RMS’S from the UK) together show quite convincingly schools’ MetLink project (Walker, personal that the eclipse caused a widespread decline in communication). Through these spectacular SAT of typically 1-3 degC at screen height. As astronomical phenomena, we gain an insight the boundary (surface) layer temporarily into the workings not only of the solar, but also became decoupled from the higher atmos- of our own atmosphere. zyxwvutsr phere, ground temperature (where measured) fell considerably more than this. The ‘effective’ decline in SAT would be expected to have often been greater than suggested, because the eclipse occurred at the time of day (late morning) when temperatures were evidently rising Acknowledgements I am deeply indebted to all those kind observers who sent me their reports; without their extensive efforts this study would have been most steeply - and convection or mixing was quite impossible. As far as possible, credits and strong where thick cloud layers were absent - acknowledgements have been integrated into as part of the natural diurnal cycle. Some of the text but it has not been possible to use all the largest temperature drops attributed to the the reports (and only a sample are shown here) eclipse (locally up to about 6degC in SAT) so please allow me to express my collective gra- were measured in south-east England and East titude. I also thank Roger Brugge for forward- Anglia, where skies were relatively clear and ing several reports and helping me unravel the the (approximately 97%) eclipse was very near synoptic code. Graham Bartlett, Library Infor- to total. However, the screened SAT drop in mation Manager at the National Meteorologi- these regions was more generally 2-3 degC. In cal Library and Archive, supplied a copy of the south-west England (across a large part of Daily Weather Summary for 11 August, and the which totality occurred) generally thick cloud University of Dundee Satellite Receiving Sta- cover typically muted the SAT response to a tion (Neil Lonie) provided high-resolution UK modest 0.5-2.0 degC drop. Solar radiation satellite images. (obviously) fell to zero or near zero at mid- eclipse. There is considerable evidence from quite a large number of stations that low-level, References convective Cu cloud dispersed during and Bell, S. (1997) The RGO guide to the 1999 total eclipse immediately following mid-eclipse, and already light winds (any changes in which were unlikely to have caused the widespread SAT drop) widely fell calm. Practically, people felt colder during the eclipse than the above declines in SAT would ofthe SUN. Royal Greenwich Observatory zyx Hanna, E. (1999) Solar eclipse weather effects - can you help? Weather, 54,p. 223 ~~ Correspondence to: D r E Hanna, 23 Donnington suggest because of the masking of diffuse as Drive, Shanklin, Isle of Wight PO37 7JA. 446