from: Tim Osborn <t.osbornatXYZxyz.ac.uk>
subject: ad hoc detection progress
I've just typed up and expanded upon the progress report for the Hamburg meeting. In case you're reading e-mail, I've copied it as text below. Please let me know if I've got anything wrong!
Task 1: Paleo observations
Work item A: Reconciling proxies
1. Intercompare high-frequency vs. low-frequency records
A brief discussion of these issues is included in Jones et al. (2001, to appear in Science). A comment (Osborn & Briffa, 2001) on the paper by Harris & Chapman (2001, Geophys. Res. Lett.) has been submitted to Geophys. Res. Lett., pointing out that their comparison of borehole and surface temperature records is flawed because one of the records they use, by Overpeck et al. (1997, Science), had not been calibrated. Thus the borehole records still suggest much cooler temperatures in the 17th and 18th centuries than the surface reconstructions indicate.
2. Numerical & spectral differences between high & low frequency records
3. Physical reasons for the differences
Briffa et al. (2001, J. Geophys. Res.) address the possible loss of low-frequency signal in tree-ring networks by re-processing them using a method that removes age-related artefacts but retains century time scale information, though with the potential cost of increased noise early in the 600-year record. Enhanced century-scale information is captured using this method compared with traditional approaches, though verification of the low-frequencies is not possible.
Work item C: Paleo temperature data set
1. Re-assess each proxy's usefulness for temperature
Briffa et al. (2001, submitted to Holocene) has carried out a detailed study of the climate signals in networks of tree-ring density and tree-ring width across the northern hemisphere. Spatial/regional differences in their usefulness for reconstructing temperature are identified, but the major result of the comparison is the clear superiority of the density over the ring width for reconstructing local temperature or highly resolved temperature patterns. At larger spatial scales, the superiority of density over width is maintained in North America, but is weaker over northern Eurasia.
2. Develop long paleo proxy reconstructions
Updated comparison of various quasi-hemispheric reconstructions has been pulished in Briffa et al. (2001, J. Geophys. Res.), and a subset in Jones et al. (2001, to appear in Science), the latter with a comparison of estimated uncertainty ranges.
3. Estimate the errors in these proxy temperature series
Time-dependent (and timescale-dependent) uncertainty ranges have been estimated for the Jones et al. (1998) reconstruction of northern hemisphere temperature, though based solely on calibration statistics. This will be published in Jones et al. (2001, to appear in Science).
Task 2: Natural forcing and response
Work item A: Volcanic forcing record
2. Refine volcanic composites for each hemisphere
Task 4: Anthropogenic responses
Work item A: Detection of changes in extremes
1. Assess which changes in extreme events are predicted by models
Some work in progress (supported by a different project) has determined simulated changes in precipitation extremes over Europe, finding that HadCM3 suggests coherent patterns with zonal structure (plus some land-sea structure), that shows meridional shifts with season.
4. Compare simulated changes with observational counterparts
Jones et al. (2001, J. Geophys. Res.) have attempted to remove the influence of changing station and SST sample numbers on the variance of the gridded, monthly temperature data set. This adjusted data set can now be used to assess inter-monthly and inter-annual variance changes, with more robustness than the unadjusted data set. Yan et al. (2001, to appear in J. Geophys. Res.) and Yan et al. (2002, to appear in Clim. Change) have analysed long instrumental records from Europe and China to identify changes in "weather" variability. A recent study by Frich et al. has obtained values for 10 indices of climate variability/extremes from many station across many countries; these indices may not be optimum in signal-to-noise terms, but we highlight their study because these indices will probably provide the best observational coverage available. Finally, work in progress is comparing simulated and observed changes in precipitation intensity over the UK.
5. Determine feasibility of detection of changes in variability/extremes
Other long instrumental records for Europe are available.
Publications produced with partial support from this project
Hegerl GC, Jones PD and Barnett TP (2001) Effect of observational sampling error on the detection of anthropogenic climate change. J. Climate 14, 198-207.
Jones PD, Osborn TJ and Briffa KR (2001) The evolution of climate over the last millennium. Science (in press).
Osborn TJ and Briffa KR (2001) Comments on the paper by RN Harris and DS Chapman "Mid-latitude (30N-60N) climatic warming inferred by combining borehole temperatures with surface air temperatures". Submitted to Geophysical Research Letters.