July 30th, 2009

July 22 Great Solar Eclipse PFS Bulletin No. 14 [FINAL]

Published at 23:33. Filed under Precision Forecast Service; 1 Comment

OVERVIEW

This is July 22 Great Solar Eclipse PFS Bulletin No. 14 [FINAL], issued in both English and Chinese. Observations and analysis of the actual weather on July 22 as well as review and conclusion of this compaign are presented.

OBSERVATIONS

Satellite observations

Images from geostationary meteorological satellite Fengyun 2-D of totality zone over mainland China on July 22 0045Z, 0115Z and 0115Z are shown as Figure 1, 2 and 3, with the totality eclipse path manually marked in red line.

Images from geostationary meteorological satellite Fengyun 2-C of totality zone over East Asia and western Pacific Ocean on July 22 from 0000Z to 0330Z with 30min interval are shown as Figure 4-11 (listed at the buttom). An animation of these images is also shown below (click for raw size file).

Surface observations

• Surat, India: SYNOP (42840) data depicts an overcast weather with 10/10 cloud cover and moderate rain at 0000Z. Sky condition was about the same at 0300Z.
• Bhopal, India: SYNOP (42667) data depicts an overcast weather with 10/10 cloud cover and light rain at 0000Z. Sky condition was about the same at 0300Z. The data also depicts a temperture drop of 0.4C from 21/1800Z to 22/0000Z and a difference of 1.8C between 21/0000Z and 22/0000Z under similar weather situations.
• Patna, India: SYNOP (42492) data depicts a partly cloudy weather with 9/10 cloud cover at 0000Z and 8/10 cloud cover at 0300Z. The clouds were mainly stratocumulus cumulogenitus at 00Z (5/10), becoming cumulostratus and alto cumulus translucidus at 03Z (each 4/10).
• Litang, China: SYNOP (56257) data depicts a mostly clear weather with 4/10 cloud cover at 0000Z, the clouds were stratocumulus cumulogenitus and cirrus densus (each 2/10), with the buttom at a height of 600-1,000 meters. The weather had become more cloudiness at 0300Z, with 9/10 cloud cover, but most of the clouds were cirrus densus (8/10). Observers around this area (Haizikou Pass, Dinggong Grassland, Derong) reported successful observing and similar weather situation. No significant drop of temperature was observed at 0000Z according to the data.
• Kangding, China: SYNOP (56374) data depicts an overcast weather with 10/10 cloud cover (all were stratocumulus cumulogenitus) at 0000Z, becoming 9/10 with mainly cumulus (7/10) at 0300Z. Observers in this area reported observing failures. However, observers at Zimei Pass (approximately 70 kilometers southwest) reported a "mostly clear" weather during the totality. No significant drop of temperature was observed at 0000Z according to the data.
• Chengdu, China: METAR (ZUUU) data depicts a mainly cloudy weather with a few showers around the time of eclipse.
• Chongqing, China: SYNOP (57516) data depicts a fine weather with 7-8/10 cloud cover (all were cirrus densus) at 0000Z and 0300Z. Eclipse observations around this area were successful. No significant drop of temperature was observed around the totality according to the SYNOP data and METAR data from ZUCK.
• Yichang, China: SYNOP (57461) data depicts a mostly clear weather with 5/10 cloud cover at 0000Z, mostly alto cumulus translucidus and cirrus densus; at 0300Z the sky was even better, with 2-3/10 cloud cover. Eclipse observations around this area (such as Yichang, Jingzhou, Jingmen) were highly successful. The data also depicts a temperture difference of 0.8C between 21/0000Z and 22/0000Z and 2.3C between 21/0300Z and 22/0300Z under similar weather situation of the two days.
• Wuhan, China: SYNOP (57494) data depicts a mostly cloudy weather with 9/10 cloud cover at 0000Z and 0300Z with cirrus densus and growing alto cumulus translucidus. Cloud blocked the totality in some areas; in other areas the totality were not satisfying due to cirrus. The temperture at 0200Z was approximately 1C lower than 0100Z according to METAR data from ZHHH.
• Anqing, China: SYNOP (58424) data depicts a mostly cloudy weather with 9/10 cloud cover (all were alto cumulus opacus) at 0000Z and 0300Z. Generally speaking, observations around this area (including Tongcheng, Tongling, Xuancheng) were unsatisfying, with several sites (generally close to the Yangtze River) failed to see the totality. But observers about a hundred kilometers to the south (Huangshan, Jiujiang) enjoyed a much better view of the totality.
• Hangzhou, China: SYNOP (58457) data depicts a mostly cloudy weather with 9/10 cloud cover at 0000Z and 0300Z with mostly stratocumulus cumulogenitus and alto cumulus translucidus. Observers over the north (such as Shanghai, Suzhou, Wuxi) experienced rain during the eclipse. However, observers around Hangzhou and Ningbo reported that the sky had clear up at the time of eclipse.
• Yaku-shima, Japan: METAR data from RJFG depicts the weather to be cloudy around the time of eclipse with showers at first.
• Iwo Jima, Japan: METAR data from RJAW depicts the weather to be cloudy around the time of eclipse with showers at first. However, reports from Iwo Jima indicating that the totality was visible thru clouds.

ANALYSIS

Surface analysis chart by Hong Kong Observatory at July 22 00Z is shown as Fig. 12. Combining with a 500hPa geopotential height from NOAA/ESRL PSD reanalysis data (Fig. 13), we can see the remain of western Pacific subtropical high over Southeast China and a shearline at around 30N over East China.

Combining with products from KMA, JMA, and analysis products of NOGAPS and GFS (graphs not presented), the following features at July 22 00Z can be depicted:

• A deep low pressure was lifting over India Peninsula, bringing cloudy and rainy weather to totality area over India.
• A weak ridge was persisted over 95E on the backside of a low level trough at 105E (over the western Sichuan Basin), the totality area had generally less cloud despite the complex terrain.
• A moderate ridge, possibly the remain of subtropical high, persisted over 110E (eastern Sichuan Basin and western Hubei Province), enabled observers around this area to get an expectional view of the eclipse.
• A large stationary front associated with a mid-latitude cyclone over Hokkaido from eastern Yangtze Valley and East China Sea to Japan was persisted, with a possible weak low pressure area at the buttom (around Wuhan). The front had prevented many sites from the totality; however, a number of observers stayed away from the front were lucky enough to get the eclipse -- although not exceptional -- thru clouds.
• A splitted subtropical high was centred near Taiwan.

REVIEW

PFS for the event was started on July 6. However, it was not until July 9 (bulletin No. 4) did the forecast manage to made a "decent" guess of the weather on July 22. The following paragraph was copied from bulletin No. 4:

...we can expect large areas over Eastern China -- generally east of 115E 30N (i.e. east of the trough) -- are being covered by heavy cloud. Rains are likely, since a clear shear line is presented on the 850hPa wind chart. On the other hand, the weak ridge over central China are possible to bring an acceptable weather to the totality area between Chongqing and Wuhan.

The totality area over India are possible to be under clouds produced by a 500hPa low pressure over northern Gulf of Bengal. Areas over east of Ryukyu Islands are possible to share a cloudless weather granted by the western Pacific subtropical high.

The remark of good weather between Chongqing and Wuhan is exceptional. The remark of unsatisfied weather over east of Wuhan is also consistent with the actual situation. Unfortunately, although "right guess" did occasionary appeared in the following bulletins, the guesses remained unstable until July 17 (bulletin No. 11), i.e. before the employed models depicted a similar situation.

On July 18, 4 days prior to the event, the author published his "final prediction" of the weather on July 22 (bulletin No. 12) before flying to Shangri-La for the eclipse. Combining with the revision (on July 21, bulletin No. 13) and the actual weather conditions listed above, we may sumarize as below:

• Forecast for India goes "mainly cloudy over western and eastern totality area, may be mostly clear around Patna". The forecast generally matches the actual situation despite being somewhat too opmistic, as the satellite data shows the weather was better over east while western India was covered by heavy clouds.
• Forecast for western Sichuan goes "mostly cloudy over west, including big Shangri-La (Deqin, Derong, Daocheng), may be suitable for eclipse for some lucky area. More clouds over eastern part (Gongga Shan/Minya Konka, Ya'an, Emei Shan)" The forecast generally matches the actual situation but had been somewhat too pessimistic. Satellite data and observing reports showed the weather to be partly to mostly clear over the western part. It was indeed cloudy over most area of eastern part, although observers to the west of Gongga Shan/Minya Konka were blessed by a clear region.
• Forecast for Chengdu-Chongqing Plain goes "mainly cloudy with a few showers. There is a weak chance that the weather to be better over east", the chance was upgraded to moderate in the revision bulletin. Generally matches despite its conservativeness.
• Forecast for Hubei Province goes "weak chance for western area (like Enshi) to be better; mainly overcast over the entire province, but clouds may be slightly thinner (may permits a partial eclipse observation) over the south since it locates further from the shearline". The chance was upgraded to moderate in the revision bulletin. Actual weather condition highly matched the trend of this forecast, however it was somewhat too pessimistic even after the revision. The area over west and south was excellent, even some areas in Wuhan did witness the totality.
• Forecast for South Anhui goes "...the southern area of South Anhui might be slightly more promising since it will keep a distance from the shearline...", this matched the case. But according to the satellite data, the area well away from Yangtze Valley also enjoyed a cloudless sky, at least in the morning.
• Forecast for pan-Yangtze River Delta goes "generally overcast with rain". Although a number of cities (such as Shanghai) around this area were indeed overcast with rain, lucky observers around Hangzhou did witness the totality thru a large cloudless region. The author failed to note the possibility of the appearance of such a significant cloudless region.
• Forecast for Ryukyu Islands goes "mainly cloudy, with a few or some sunshines". It seems there were not much observers at this region, but satellite images depicted this forecast was about right.

CONCLUSION

The author notices that in this compaign, the forecast for "possible opmistic" situation (such as the situation for Chongqing and western Hubei Province) were somewhat too conservated. What is more, comparision between actual weather and forecast by automatic numeric model shows the accuracy of the later was not as satisfying as human forecast (such as Fig. 14). Although now being applied into using by 7Timer! and providing a good assistment on assessing the weather situation during this compaign, the 35km NCEP/GFS still needs to incorporate closely with human experiences for a most reliable forecast on great event. The 12km MM5/GFS was even more unsatisfying than 35km NCEP/GFS  and further adjustment is required on the model. After all, from the discussion above we can depict that, the compaign was generally successful, accuracy of the forecast was agreeable for most areas.

ACKNOWLEDGEMENTS

• The author is happy to learn that a number of observers had made the right decisions according to his forecast. Total solar eclipses are rare, but not as rare as meteor storms. No matter whether you had witness this eclipse, sooner or later you will have an excellent eclipse somewhere else.
• Original Fengyun 2-C and Fengyun 2-D images are kindly provided by the National Satellite Meteorological Center (NSMC). The original Fengyun 2-D images are kept by Cao Zhiqiang of NSMC.
• The author would like to thank colleague Zhou Linjiong for his valuable discussions and helps of this work. The author would also like to thank Johnson Lau, Dickson Fu, Zhang Linna, Robert Yen, Chen Sisi, Shi Liaoshan and Cheng-Su Diandian for their valuable discussions, contributions and helps.
• The author would like to thank Dr. Cui Chenzhou for providing web service for 7Timer!.
• The author would like to thank all of you for the supports of his work.