date: Mon Jan 17 10:10:42 2005 from: Phil Jones subject: Fwd: Re: Fwd: Rainfall trends to: "Klein Tank, Albert" Albert, Maybe you have this attachment - maybe not. I thought I would send on as it relates to how we should or should not refer to Michaels et al (2004) which came out in IJC in last issue of 2004. Cheers Phil Date: Sun, 16 Jan 2005 14:54:32 -0500 From: "Thomas R Karl" User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.4) Gecko/20030624 Netscape/7.1 (ax) X-Accept-Language: en-us, en To: Thomas R Karl CC: Phil Jones , Pasha Groisman Subject: Re: Fwd: Rainfall trends X-UEA-MailScanner-Information: Please contact the ISP for more information X-UEA-MailScanner: Found to be clean X-UEA-MailScanner-SpamScore: ss Sorry Phil --- pdf version attached Thomas R Karl wrote: Phil, Thanks for the article and note. I refer you to our latest work (lead author Pasha) which is attached. Here you will see we have expanded our network to include all the data we have been digitizing over the past few years. We have far more stations than we had in 1998 and were able to search for trends in the higher quantiles with longer return periods. Our results clearly show stronger trends compared to the mean the higher in the distribution you go (see Tables 1 through 4). In addition, we find (I calculated this from Table 3 in our paper) that when you compare 1910-1970 with 1971-99 you find relative to the mean annual total precip during the two periods the 99.9 percentile value accounted for 12% more precip in the later period (relative to the mean of that period) compared with the 99.9 percentile and the mean of the earlier period. For the 99th percentile it was 9% higher in the later period and for the 95 percentile it was 5% more. Also you can see the trends at the higher percentiles increasing much fast thatn for the mean. I note Pat's Fig. 4a the sign of the change is consistent with our results (not sure about magnitude because he uses inches (WHO REVIEWED THIS AND LET HIM GET AWAY WITH INCHES!!)), but when he divides by the total annual precip the results are quite different. I am not sure how much his small network of 129 stations contributes to unstable results. Second, if you look at our Table 1 we show the national trend in the mean to be about 6% per Century (consistent with IPCC and the US National Assessment). I suspect his method of averaging is misleading him, and a few stations with high 10-day precip compared to the mean within the regions are dominating and causing unstable results. We found that in looking at the extremes one really has to be careful how one does area averaging. Phil, if you simply look at the trends for the 95, 99 and 99.9 percentiles all our increasing at a much faster rate from 17 up to 33 percent per century. The intepretation is straightforward, the heaviest precip events are increasing faster than the mean (again consistent with Katz's theoretical work and Pasha's earlier empirically based model we featured in IPCC 2001). . So... I am not sure we need to respond to Pat's work as I think our latest is much more comprehensive. What do you think? P.S. I think we are close to sending our Vertical Temp Report out next week. I will send you, Kevin, and Susan copies. It was far more work than I ever imagined and we are just in our first review stage! As you can see thinks are pretty busy ---- seem to be working too much as I am sure you are!---- at least on weekends I am up here on the Mountain which makes up for a lot. Regards, Tom Prof. Phil Jones Climatic Research Unit Telephone +44 (0) 1603 592090 School of Environmental Sciences Fax +44 (0) 1603 507784 University of East Anglia Norwich Email p.jones@uea.ac.uk NR4 7TJ UK ----------------------------------------------------------------------------