su rui ke

Yes, it will be close - I'm 90% sure the final figure is going to be 16.6 +/- 0.1C now. It could come right down to the last day 'deciding' whether the record falls or not. I honestly can't see why some people are still talking about it possibly falling by a whole degree in these last 6 days. That would require a CET of 12C from now until Saturday and none of those days looks like having a CET anywhere near as low as that!

But the number of days available to produce such a drop is fast running out! From yesterday's 18Z I get a CET of 15.5 for 23rd-27th and from today's 00Z 16.0 for the same period. That produces CETs of 16.6 and 16.7, respectively by the end of the 27th. This weekend, at least, looks warm enough for the CET to stay the same or maybe even rise very slightly. After that there is a cooling trend but still with maxima in the mid- to high-teens by 28th/29th and the nights don't look particularly cold. I still think the final CET will be closer to the record than it is to 16!

Yes, at a total of 26.55 degrees, isn't it? The next is 2002 with a cumulative Oct-Dec figure of 25.06. As far as I can see, those are the only two years exceeding a total of 25 degrees for the 3 months.

Making a fairly conservative estimate of 16.0C for the September CET, we would (I think) need to collect another 25.5 degree-months in October-December to reach the all-time record. I haven't had time to do the calculations yet, but I don't think too many years have managed that, if any. Of course, if September is slightly warmer than 16.0C it would make achieving the record slightly easier.

I think you're edited comments are nearer the mark, West. A 5-day GFS-based calculation shows it staying more or less steady until at least the 26th. That would mean a daily average of 10C was needed in the last four days to bring it down to 16C.

I would think the record would be more likely than not if it's still 17.1 by the end of the 27th. If it turns out the way you suggest, the last three days would have to average below 12.1 for the record not to be broken - not ridiculously cold but even at this stage you would have to say there is a very good chance of the record going.

I notice that in the whole of the 1960s only two months (July 1967 and July 1969) had a CET higher than than the record September CET (16.6). So, this month really has been as warm - or warmer - than a typical 60s summer in CETland.

It looks even more likely now if the GFS's predictions of a 'heatwave' for southern parts from Thursday onwards come true!

Well, after a breezy but not unpleasant day here things have deteriorated quite a bit in the last hour or so with heavy squally showers.

Using the GFS as a guide, the CET could still be at or above 16.6C by the 20th - which would still be above the 1971-2000 July average. It would, of course, be expected to fall towards the end of the month but around 16 doesn't look impossible. (This is not a forecast, by the way, just a comment on the trend!)

You're probably going to see some effects even there in Shanghai though. There's a strong breeze here and the wind is forecast to continue strengthening today - maybe to gale force by evening. How is it where you are?

Sorry I wasn't able to give you any more news on this. It was just my luck to be leaving Ningbo as news about this typhoon was starting to come through. You probably all know more about it than me by now as I have been on the road for the last few days. Parmenides is also right about the heat - I was in Shanghai on Sunday, which I think was the hottest day of the year - 38.6C. One thing I did hear was that about 1,000,000 people were evacuated ahead of this storm. This was not some remote rural area that was hit. Not many people abroad may have heard of Wenzhou but it is one of the most developed and densely populated parts of the eastern coast.

Still not feeling much effect here in Ningbo. It's turned into quite a pleasant evening actually because - as usually happens with a typhoon not too far away - it's slightly cooler than normal for the time of year. I just heard a report that the total number of people evacuated in Zhejiang has reached 980,000. Also the forecast track continues to be northwest turning more westerly later, taking Saomai over northern Jiangxi and Hunan.

OK, I see what you mean. Yes, that stretch of coast between Fuzhou and Wenzhou seems to get hit several times a year. I would think it must have one of the highest rates of typhoon/hurricane landfalls anywhere in the world. I've been here for nearly six years (teaching) and will be here for at least another year - though I suspect it will be longer. I can't say that it's the weather that keeps me here though - it's either too hot, too cold or it's raining! (Well, it can be nice in autumn and spring, I suppose.) Good luck with your next trip to China, although I hear those train tickets to Lhasa are very difficult to get. You might need to use some guanxi (connections) to get any.

Hi Peter, Nothing much happening here - still breezy with occasional light showers. The wind is maybe slightly stronger than it was this morning. We keep getting promised heavy rain later but there's no sign of that yet. As landfall now looks likely to be in this province - maybe to the south of Wenzhou - there has been a lot of coverage on local TV today but, so far, I can't really add any more to what you already know. I don't think Shanghai is in much danger by the way. Anyway, I'm heading up there tomorrow!

Ningbo weather bureau reports winds were 60 m/s at 1200 (0400 GMT) and says it is the strongest typhoon for 40 years. Looking bad for southern Zhejiang/northern Fujian. It's breezy and showery here - definitely 'outer' typhoon-type weather. At the moment I'm 341 km from the centre!

Unusual for us here to be warned of possible torrential rain and near gale force winds with a likely landfall again in Fujian, but that's the case for tomorrow. Is Saomai really such a large system? Still dry and clear at the moment.

Going off topic a bit, but, as you ask, I was just given some today as a present. I didn't know they were 'famous' though! 'Fenghua' is correct, by the way. It's a rural district south of Ningbo and is where Jiang Jie Shi (Chiang Kai Shek) was born.

Yes, since I wrote my last message I've changed my mind a bit! I agree, weighting by share if land mass is better but the accuracy will depend on the density of stations and location. Also, if the density of stations were the same over the whole country, weighting by area or number of stations would give the same answer. The point I was trying to make is that we don't know that there isn't the sort of bias you are warning about in the individual averages for England, Scotland, etc. I mean, we assume that it isn't biased towards one corner of the country but it could be! So, yes, it is interesting - and important - to try and find out what these 'averages' actually mean.

But isn't they way Mr Data suggests how the individual 'national' averages are calculated anyway - adding up and dividing by the number of stations? I mean, they aren't weighted for the areas they nominally cover, are they? Using the first method is then just an extension of this. Surely the accuracy of the average depends more on the density of reporting stations used in the first place rather than which of those methods you use to calculate the average?

Ah, but this is about trying to integrate a variable over time so both methods give an approximation to the mean. They are assuming that the temperature changes in steps - the only difference is that the traditional method assumes that the temperature stays at its minumum value for 12 hours, jumps up to the maximum for the day, stays there for 12 hours and then goes to the next minimum ... and so on. The computerised integration method still assumes that the temperature changes in steps but each step lasts for a shorter period of time. To get the true mathematical average of a quantity that varies with time - like temperature - you would have to take an infinite number of readings, add them up and divide by the total length of time the observations were being made over (1 month or whatever)! Anyway, as is often the case, as long as everyone uses the same method there isn't a problem.

When I quoted that I meant to imply that I was sure what you actually said would have been qualified somehow. Thanks for the confirmation! I get 19.69C now doing the same thing! Are you starting from 19.97C? What about today's values?

Thanks, WIB. Yes, it would probably save a lot of argument if both records were broken! I agree, it is just possible given that the forecast temperatures for the weekend seem to be rising slightly and - very importantly, as you said - Saturday night into Sunday looks like it might be warm enough to hold the CET almost steady right up till the end of Sunday. It will then all depend on how much Monday dents the average. By the way, I saw Philp Eden being quoted in the Glasgow Herald predicting the CET would end up at 19.9 and, therefore, break the record - so I guess he must have meant the Hadley figure!

Shouldn't we really say that the Hadley series did not exist before 1974 - which, strictly speaking, means that you cannot compare the Hadley series with the pre-1974 data? If both series do turn out this month to give a CET of around 19.7C, then the main argument is going to be over the 'warmest on record' tag given to this month. According to the Manley series, 2006 will simply be in second place behind 1983 (19.9C). Using the Hadley series, to be consistent, all we can then really say is that this was the warmest July since 1974 with 1983 in second spot on 19.5C. However, the next warmest July was far behind any of these months - 18.8 in 1783. So, it seems that had the Hadley series existed at that time it would have had this and all other July's in third place or below, the discrepancy between the two series never being as much as 0.8C in any month. (?). So, by this reasoning it would be reasonable to say this is the warmest July on record if you follow the Hadley CET and it all comes down to which series you regard as a better measure of average temperatures in England. I realise I have slightly contradicted my opening statement but I hope this makes some sort of sense!

My impression is that it has been a much wetter than normal July here but I don't have any statistics to hand - maybe I will be able to come up with some later. There is nearly always flooding somewhere in China in the summer - again it's hard to say how this year compares with others. What was that about peaches?