cc: Keith Briffa date: Thu, 23 Jun 2005 14:49:36 +0100 from: "Betts, Richard" subject: RE: sudden response to vegetation change? to: Tim Osborn , "Tett, Simon" Tim, Yes, this probably is due to land use, and is probably an artifact of the particular diagnostic chosen for the run rather than being a result of climate change. That particular soil moisture diagnostic represents soil moisture in the root zone. When deforestation is imposed, the root zone shrinks so that diagnostic shows less soil moisture. The soil moisture content per unot root depth may not have actually changed though. There is an alterntive diagnostic (soil moisture content in layers) which is independent of the plant roots, but unfortunately I think that wasn't included in the run. Cheers, Richard Dr Richard Betts Manager, Ecosystems and Climate Impacts Hadley Centre for Climate Prediction and Research Met Office, Fitzroy Road, Exeter, EX5 2SN, UK Tel: +44 (0)1392 886877 Fax: +44 (0)1392 885681 richard.betts@metoffice.gov.uk http://www.metoffice.gov.uk -----Original Message----- From: Tim Osborn [mailto:t.osborn@uea.ac.uk] Sent: Thursday, June 23, 2005 2:35 PM To: Betts, Richard; Tett, Simon Cc: Keith Briffa Subject: sudden response to vegetation change? Dear Simon and Richard, I have a question regarding some very sudden changes in soil moisture in the ALL250 HadCM3 run. Please see the attached PDF plots. They all come from 9 grid boxes in and around Fargo in North Dakota. The exact lat-lon coordinates for the 9 boxes are given in the header of the ASCII data file which contains the soil moisture data only. If you look at the soil moisture plots, you'll see very sudden reductions in some boxes at different times. Why? If "real", I can only guess that this is due to vegetation change imposed during ALL250. These are annual means with a 30-yr filter also shown. If you look at the data file you'll see that the second time series (from 97.5 W, 50 N) has the change between 1910 and 1911 affect all months. And it is huge relative to the pre-1910 mean and relative to interannual variability. The first time series in the file doesn't show such a change. Should such changes be so sudden, and with such contrast between neighbouring grid boxes? I guess it might if the Ramankutty & Foley land use data have such discontinuities in. For interest I've also included time series of new surface SW radiation and air temperature from the same 9 boxes. They show trends rather than sudden changes, perhaps more related to tropospheric sulphate aerosol cooling than to vegetation change? Anyway, there's apparently no anthropogenic warming here in HadCM3. Any comments on the apparent discontinuities? I'm wondering whether I should use these grid box time series? I could average the 9 together to get a North Dakota time series, which would smooth the change out, but given the underlying behaviour, I'm not sure I'd be confident in this either. Cheers Tim