date: Tue Nov 30 16:04:12 2004 from: Keith Briffa subject: Re: Model stuff to: Martin Todd Martin have done minor edits to fix typos etc At 15:30 30/11/2004, you wrote: Here's the near final version of section 3 (3) Comparison of proxy records to GCM model outputs 3.1 Description of GCM model experiments We use output from two state-of-the-art experiments using two (UK and German) coupled ocean-atmosphere GCMs in which best estimates of (i) natural forcings only (ii) natural plus anthropogenic radiative forcings are applied over the past 500 (UK) to 1000 (German) years. By using results from two GCMs we will be able to quantify some of the uncertainty in past climate simulations associated with model climate sensitivity. The GCMs are HadCM3 (Gordon et al, 2000, Pope et al, 2000) and ECHO-G (Zorita et al., 2003). Output from HadCM3, will be made available by the Hadley Centre. The first experiment (NAT500) simulating the period 1500 to 2000 is driven with natural forcings alone (changes in solar irradiance, volcanic aerosol and orbital parameters (mainly the date of perihelion). In this experiment land cover and greenhouse gases are set to pre-industrial conditons. The second experiment (ALL250) over the period 1750-2000 includes all natural forcings plus additional anthropogenic forcings. These are the changes in well-mixed greenhouse gases , sulphate aerosols (both their direct effect on scattering solar radiation and their indirect effect on cloud albedo), changes in vegetation (mainly the replacement of forest with pasture) and changes in tropospheric and stratospheric ozone. Simon Tett at the Hadley Centre will provide data on near surface temperature and other diagnostics. The ECHO-G experiments are very similar except that there is no land use change , geographical structure to volcanic aerosol forcing and no specification of tropospheric aerosols. The 'natural forcings only' experiennt runs for 1000 years to the present. A very preliminary analysis suggests that anthropogenic forcings have had a substantial effect on Arctic Siberian climate during the 20th century. For the illustrative study region the ALL250 simulation indicates a warming of about 0.750C since 1950 compared to the NAT500 simulation, Fig 2 NATURAL500 (solid) and ALL250 (dotted) simulations for 1500-2000 AD (Tett ref) for 60-115E and 65N to 75N 3.2 Palaeo-data - model comparisons 3.2.1 GCM evaluation Prior to proxy-model comparisons we will determine the extent to which the HadCM3 and ECHO-G GCMs accurately reproduce (a) the basic state of the regional climate (b) the timescales and structure of regional climate variability and the associated dynamics. We will undertake an equivalent analysis to that with observational data described in Section 1.4 using data from the GCM control and experimental runs and compare with results from the analysis of observations. 3.2.2. Paleo data-model comparison We will test the veracity of the simulated model climate variability against proxy data by quantifying the coherence in regional climate variability estimates from HadCM3 and ECHO-G and multi-proxies at decadal-centennial timescales. We will compare the regional climate history in the GCMs and proxies in terms of (a) the magnitude of variance in (b) the relative variance at different timescales (c) the phase relationships and coherence of variability. This will employ a variety of timeseries approaches (including cross spectral analysis and wavelet analysis),correlation and regression. We will utilise techniques for proxy-model comparison developed within the EU-funded SO&P projects (Simulations, Observations & Paloclimatic data: climate variability over the last 500 years) which account for the error in the GCM and proxy estimate. This will determine the statistical uniqueness of recent trends relative to past periods. We will also determine whether the spatial pattern of lake response bwteen the sites is consistent with that recorded in the spatially more complete tree ring data and the GCM output. Using the timing of climate response to the natural and anthropogeic forcing history as simulated by the GCMs we will be able to attribute causes to variability observed in the proxy records. From our analysis we will determine whether the observed lake ecosystem response is more consistent with climate variability/change simulated by GCMs under natural or natural plus anthropogenic forcing. **************************** Martin Todd University Lecturer Department of Geography UCL (University College London) 26 Bedford Way London WC1 8HR email m.todd@geog.ucl.ac.uk ******************************** -- Professor Keith Briffa, Climatic Research Unit University of East Anglia Norwich, NR4 7TJ, U.K. Phone: +44-1603-593909 Fax: +44-1603-507784 [1]http://www.cru.uea.ac.uk/cru/people/briffa/