date: Thu Jun 23 15:43:29 2005 from: Tim Osborn subject: Re: NH Borehole Timeseries to: Jason Smerdon , Henry Pollack Dear Jason and Henry, (1) Data---------------------------------------- I've at last found time to extract time series from the HadCM3 and ECHO-G (==ECHAM4/HOPE) model simulations of the last 500 and 1000 years, respectively. To avoid focussing on single grid boxes, I've actually extracted the closest box to the lat/lon coordinates you provided, plus the 8 surrounding boxes. For ECHO-G there is a single simulation ("erik") from 1000-1990 forced by solar, volcanic and greenhouse gas variations. For HadCM3 there are two runs, "nat" 1500-1999 forced by solar and volcanic, plus "all" 1750-1999 forced by solar, volcanic and anthropogenic forcings. Thus to recreate the best simulation of the real climate since 1500, you need to take years 1500-1749 from "nat" and tack on 1750-1999 from "all". I haven't done this because I thought you might first just look at the recent period of "erik" and "all" and see how they compare with Fargo observations (but see below about what you might do with the data). The variables I've provided for each run are: airtm = 1.5m air temperature precp = precipitation snowd = depth of lying snow swrad = net surface short-wave radiation plus for HadCM3 soilt = soil temperatures at 4 different depths all in one file. The depths are not give in the files: levels 1 to 4 have thicknesses 0.1, 0.25, 0.65 and 2.0 m respectively. Thus the middles of levels 1 to 4 are at 0.050, 0.225, 0.675 and 2.000 m below the surface. while for ECHO-G tsoi* = soil temperatures at * levels, where * = 1 to 5. The depths of the middles of these five levels are given in each file. I do have a few other variables, including (e.g.) soil moisture. But I didn't want to overload you and also soil moisture seems to behave strangely in HadCM3 - I'm checking it out with the Hadley Centre. (2) Plots--------------------------------------- To give you an idea of what the data look like, I've also included plots of the mean seasonal cycles and of annual time series from each variable (in separate files) and for each of the 9 grid boxes (in separate plots in each file). (3) Climate behaviour--------------------------- You'll see that ECHO-G begins relatively warm, which is partly an artefact of the initial conditions being too warm - so don't spend a long time worrying about that. Also, ECHO-G has large warming from LIA to present - but no tropospheric sulphate aerosol cooling was included, so that explains why it is greater warming that in other models. I recently submitted a paper describing these issues which I can send if you want. It makes comparing long-term trends with palaeodata rather complex. You'll see that HadCM3 "all" actually cools at this location. I think that this is because of the relatively strong tropospheric sulphate aerosol cooling included in this model (see large falls in swrad). Not also that some big vegetation changes may have been imposed in the HadCM3 "all" simulation - makes a big difference to soil moisture, and possibly contributes to the cooling simulated near Fargo in this model. In other areas warming does occur, and in fact warming dominates when looking at the whole hemisphere or globe. (4) What to do with the data?-------------------- We didn't really discuss what you would do with the data in detail. It might be interesting to look at the land-air coupling simulated near Fargo by these models, in comparison with observations. I've provided snow and precip in case you wanted to look at how these affect the coupling in the models. In HadCM3 there are large changes in vegetation imposed at various times in some of the 9 boxes - mainly between 1870 and 1910. Maybe a response in land-air coupling might be apparent in HadCM3 by comparing 1750-1850 with 1920-present? Anyway, you're free to investigate the data in whatever way you wish and hopefully there might be results interesting enough for a paper? This single site data isn't really appropriate for looking at long-term trends, given the atypical behaviour of this region in HadCM3 and the lack of sulphate in ECHO-G. I'm still working on the borehole simulations and will send you some results to look at in a few weeks. Sorry for the long email - I tried to be comprehensive, but if you need further explanation then please ask. I've just realised that all these time series data and plots come to about 20 MB -- too big to email. I've put them on our anonymous ftp site: ftp.cru.uea.ac.uk cd /people/timosborn/borehole4pollacksmerdon You'll want all the files in this directory. Cheers Tim