date: Thu, 10 Apr 2008 06:52:00 -0600 from: "Joel Smith" subject: RE: longterm river flow (2) to: "Phil Jones" , "Tom Wigley" HI Phil Thanks and that's very interesting. We did something similar in Boulder, Colorado. A few years ago the city looked at its vulnerability to a 300 year reconstruction by Connie Woodhouse (then of NOAA, now at the University of Arizona). We combined a new 400+ year reconstruction with GCM output. We derived proxy temperature and precipitation in the reconstruction by matching reconstructed flow years with "nearest neighbors" in the part of the reconstructed flow record that overlaps with observations. We then applied the monthly changes in precipitation and temperature from a wide range of GCM output. In this case the combination is very interesting. Boulder has low vulnerability to the reconstructed flows (with regard to drought). It also has low vulnerability to the imposition of climate change on the historic observed climate record. But, the combination of GCM output and the reconstruction can cause more frequent violations of drought criteria yours, Joel Joel B. Smith Stratus Consulting Inc. P.O. Box 4059 Boulder, CO 80306-4059 USA Tel: 1-303-381-8218 Fax: 1-303-381-8200 jsmith@stratusconsulting.com www.stratusconsulting.com ______________________________________________________________________________________ From: Phil Jones [mailto:p.jones@uea.ac.uk] Sent: Thu 4/10/2008 2:38 AM To: Tom Wigley Cc: Joel Smith Subject: Re: longterm river flow (2) Tom, The EA work in 2006/7 resulted in 3 EA reports and 3 summaries. I can't find the first referred to in Jones et al 2006 as Cole and Marsh (2005). This one that I don't have is mainly placing recent droughts in a longer context with EWP and other long rainfall and groundwater level series. In this report we took the riverflow reconstructions from the 2006 paper in IJC (for the River Eden nr Carlisle and the River Ouse to Denver Sluice). We used these reconstructions to approximate inflows to reservoirs on the Ouse catchment and in the Lake District (Eden). The relevant Water Authorities then used their resource models with daily inflows to see how their systems responded to flows over the last 200 years. To get daily flows, we had some modern records, so we took monthly daily sequences with roughly the same mean flows as reconstructed at the two gauging stations. As there was between 20-50 years of daily flows, there was some repetition of sequences to cover the 200 years. We then rerun the whole process with several futures from 3 RCM simulations (chosen to be from 3 different driving GCMs (HadCM3, ECHAM4 and Arpege). These changes to rainfall and T were applied to the whole sequence for 200 years, so combining a 'future' with the long historic record which encompassed natural variability. The 'future' precip changes were applied directly to the historic rainfall sequences, For future actual evapotranspiration (required by the statistical rainfall/runoff model) I developed a simple water balance model, based in rainfall and temperature. Modified temperature produced monthly sequences of Thornethwaite PET, which then produced modified AET from the simple water balance model. The Wade et al reports then look at implications for the two Water Authorities. The historic droughts in the 19th century (with modern abstractions) were sometimes worse (particularly on the Ouse) than recent droughts. These didn't get much worse in the future as winter rainfall went up in the RCMs, and the catchment has a long memory. Future droughts got worse in the Lake District as summers got drier and here the river memory was shorter. Should all be described in the reports. We didn't ever write this up - except for the these EA reports which are on the EA web site. No idea how these have been applied in the two Authorities or by the EA. Cheers Phil At 03:46 10/04/2008, Tom Wigley wrote: >Phil, > >Can you send me any reports or papers on the latest >long term riverflow reconstructions you've done. > >Has any of this been used in the context of future change? >In other words, if one just added future projections >to present (say the last 50 years), then the results >would be different from the case if one added the future >to a wider range of "present" based on observed variability >over a number of centuries. > >More specifically, if the change in flow were a reduction >of X units, and if there were a time a few hundred years >ago when the natural flow was Y less than today, then >a combination of an anthropogenic reduction of X and a >natural reduction of Y would be doubly bad. > >So -- big question -- has the UK looked at the combined >effect of X *and* Y? > >Thanks for your help, >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 ----------------------------------------------------------------------------