No abstract

Abstract. We describe new reconstructions of northern extratropical summer temperatures for nine subcontinental-scale regions and a composite series representing quasi "Northern Hemisphere" temperature change over the last 600 years. These series are based on tree ring density data that have been processed using a novel statistical technique (age band decomposition) designed to preserve greater long-timescale variability than in previous analyses. We provide time-dependent and timescale-dependent uncertainty estimates for all of the reconstructions. The new regional estimates are generally cooler in almost all precalibration periods, compared to estimates obtained using earlier processing methods, particularly during the 17th century. One exception is the reconstruction for northern Siberia, where 15th century summers are now estimated to be warmer than those observed in the 20th century. In producing a new Northern Hemisphere series we demonstrate the sensitivity of the results to the methodology used once the number of regions with data, and the reliability of each regional series, begins to decrease. We compare our new hemisphere series to other published large-regional temperature histories, most of which lie within the 1 • confidence band of our estimates over most of the last 600 years. The 20th century is clearly shown by all of the palaeoseries composites to be the warmest during this period.

A detailed description is presented of the statistical patterns of climate forcing of tree growth (annual maximum latewood density and ring-width time series), across a network of 387 specially selected conifer sites that circle the extra-tropical Northern Hemisphere. The in‘ uence of summer temperature dominates growth. A mean April–September response is optimum for describing the major forcing signal over the whole densito-metric network, though a shorter June–July season is more relevant in central and eastern Siberia. The ring- width chronologies also have a shorter optimum (June–August) seasonal signal, but this is much weaker than the density signal. The association between tree-ring density and precipitation variability (as measured by partial correlations to account for the correlation between temperature and precipitation) is considerably weaker than with temperature. The ring-width response to precipitation is dominated by ‘noise’ and local site in‘ uences, though a negative response to winter precipitation in northern Siberia is consistent with the suggestion of an in‘ uence of delayed snowmelt. Average correlations with winter temperatures are small for all regions and correlations with annual temperatures are positive only because of the strong link with summer temperatures. Reconstructions of summer temperature based on composite regional density chronologies for nine areas are presented. Five regions (northwestern North America, NWNA; eastern and central Canada, ECCA; northern Europe, NEUR; northern Siberia, NSIB; and eastern Siberia, ESIB) constitute an arbitrary ‘northern’ division of the network, while the four other regions (western North America, WNA; southern Europe, SEUR; central Asia, CAS; and the Tibetan Plateau, TIBP) make up the ‘southern’ part. We also present two larger composite regional reconstructions comprising the data from the five higher-latitude (HILAT) and four lower-latitude (LOLAT) areas respectively; and a single series made up of data from all regions (ALL), which is highly correlated with Northern Hemisphere mean summer temperature. We calculate time-dependent uncertainty ranges for each of these reconstructions, though they are not intended to represent long timescales of tempera ture variability (>100 years) because the technique used to assemble the site chronologies precludes this. Finally, we examine in more detail the reduced sensitivity in the tree-growth data to decadal-timescale summer-temperature trends during the last 50 years, identified in earlier published work.

In recent years there has been a notable increase in the number of research projects engaged in building supra-long (multi-millennial) tree-ring chronologies. Together with a growing awareness of the potential for anthropogenic climate change, this work is shifting the focus of dendroclimatology. Instead of a more traditional interpretation of tree-ring data in terms of annual-to-decadal timescale climate variability the emphasis is increasingly placed on century timescale changes. We review a number of problems with the interpretation of low-frequency climate change in tree-ring derived data. Perhaps the most significant is the high-pass filtering effect of "standardization" techniques commonly used in chronology construction to remove age-related sample bias in the original tree growth measurement data. These techniques effectively remove low-frequency variability and with it the evidence of long-term climate change. Other forcings may also be 'corrupting' the climate signal in the recent period (that used for calibrating the climate signal). Differences in the origin of the samples or changes in site ecology may also impart an inhomogeneity in the response of tree growth through time, hence violating the fundamental assumption of uniformitarianism that underpins proxy climate research.

In the current debate on the magnitude of modern-day climate change, there is a growing appreciation of the importance of long, high-resolution proxies of past climate1-3. Such records provide an indication of natural (pre-anthropogenic) climate variability, either singly at specific geographical locations or in combination on continental and perhaps even hemispheric scales4. There are, however, relatively few records that are well dated, of high resolution and of verifiable fidelity in terms of climate response, and conspicuously few that extend over a thousand years or more5. Here we report a tree-ring-based reconstruction of mean summer temperatures over the northern Urals since AD 914. This record shows that the mean temperature of the twentieth century (1901-90) is higher than during any similar period since AD 914