cc: dave lister date: Wed Apr 30 17:31:26 2008 from: Phil Jones subject: Re: Rain Days - Russia to: Ian Harris , Tim Osborn Harry, Tim, What this looks like is a different threshold for the rainday data for the period before 1984 from that after 1989. As all these stations start in 1936, the data come from daily precip data, so someone (could have been Mark) has calculated the rainday counts directly. I suspect this was for 0.1mm. From CLIMAT/MCDW the threshold is likely 1.0mm so this is why the day counts have reduced. As this has likely happened at all the 250+ stations across the fUSSR that are in the daily dataset we used to have, I'd suggest doing (1), but need to extend it a little. Don't do just Russia, but all fUSSR. You can get this by doing all WMO IDs beginning in 2 or 3. The fUSSR has all countries have all WMO Ids from 20-39. The issue will occur in other fUSSR countries, but is less important further south, as there it rains less, but when it does it generally more than 1.0mm, so the value of the threshold doesn't matter that much. Cheers Phil At 15:52 30/04/2008, Ian Harris wrote: On 30 Apr 2008, at 15:41, Tim Osborn wrote: At 14:27 30/04/2008, Ian Harris wrote: 2388400 6160 9000 63 BOR RUSSIA (ASIA) 1936 2007 -999 -999 2926300 5845 9215 78 JENISEJSK RUSSIA (ASIA) 1936 2007 -999 -999 2928200 5842 9740 134 BOGUCANY RUSSIA (ASIA) 1936 2007 -999 -999 Their datasets are attached. Not attached? Durrrrrrr. Attached. Looks like the data added 1990 and after is significantly lower than the previous data (on which the normals would be based). This is the geneaology of the current rd0 database: wet.0311061611.dtb + rdy.0709111032.dtb (MCDW composite) + rdy.0710151817.dtb (CLIMAT composite with metadata added) V V wet.0710161148.dtb However, as I've now understood that this problem existed in 2.1, it's back to square one. we have two courses of action, I think. Incidentally, Tim - I've checked and I must have imagined a big MCDW archive - both CLIMAT and MCDW bulletins were only acquired from Jan 2003. presumably on the basis that Tim M. had already incorporated them for the period prior to 2003? Mitchell & Jones (2005), Table 1, indicates that he did indeed do that for 1990-2002 for MCDW and for 1994-2002 for CLIMAT. Their Figure 1, though rather blurred, suggests that it is MCWD not CLIMAT that dominates for wetdays. well the series look to diverge from 1990 rather than 1994, so that figures. 1. We could delete all rd0 data after 1990 and rely on synthetics. This may introduce noticeable incongruities in certain areas, especially as we're changing methodology just after the normals period. 2. We could just derive rd0 from the precip data. After all, it's a pretty good relationship. How about I do both, and we compare? It won't take too long.. I would suggest doing (1) first; then do (2) if incongruities are evident in (1). However, the drop in Russia occurs in 1990. By eye, I thought it was 1991, but by inspecting the data, it is clear that they have dropped by summer 1990 and perhaps part-way through spring 1990. Therefore can you drop all rd0 from the start of 1990 onwards, rather than after 1990 (which could mean 1991 onwards). I started like that, then realised we'd lose stations that needed that extra year to qualify! However, see the attachment - the new data kicks in in May 1990 so you are right. I'll back track again and re-do with a Dec 1989 cutoff. Cheers Harry Phil, what do you think? Cheers Tim Dr Timothy J Osborn, Academic Fellow Climatic Research Unit School of Environmental Sciences University of East Anglia Norwich NR4 7TJ, UK e-mail: t.osborn@uea.ac.uk phone: +44 1603 592089 fax: +44 1603 507784 web: [1]http://www.cru.uea.ac.uk/~timo/ sunclock: [2]http://www.cru.uea.ac.uk/~timo/sunclock.htm 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 ----------------------------------------------------------------------------