Abstract
Abstract. The new PAGES2k global compilation of temperature-sensitive proxies offers an unprecedented opportunity to study regional to global trends associated with orbitally driven changes in solar irradiance over the past 2 millennia. Here, we analyze pre-industrial long-term trends from 1 to 1800 CE across the PAGES2k dataset and find that, in contrast to the gradual cooling apparent in ice core, marine, and lake sediment data, tree rings do not exhibit the same decline. To understand why tree-ring proxies lack any evidence of a significant pre-industrial cooling, we divide those data by location (high Northern Hemisphere latitudes vs. midlatitudes), seasonal response (annual vs. summer), detrending method, and temperature sensitivity (high vs. low). We conclude that the ability of tree-ring proxies to detect pre-industrial, millennial-long cooling is not affected by latitude, seasonal sensitivity, or detrending method. Caution is advised when using multi-proxy approaches to reconstruct long-term temperature changes over the entire Common Era.
Highlights
Apart from documentary archives (Pfister et al, 1999), our estimate of climate variability prior to the systematic collection of instrumental measurements in the mid-nineteenth century relies on climate-sensitive proxy data (Christiansen and Ljungqvist, 2017; Frank et al, 2010; Jones et al, 2009; Smerdon and Pollack, 2016)
Pre-industrial cooling remains significantly stronger in glacier ice and marine and lake sediment records compared to tree-ring records at different latitudinal bands
Organizing the chronologies with respect to latitude confirms that glacier ice and marine and lake sediment records from the midlatitudes (30–60◦ N) show an enhanced cooling compared to their high-latitude counterparts (60–90◦ N), whereas the Northern Hemisphere (NH) tree-ring composites lack any significant cooling (Fig. S6)
Summary
Apart from documentary archives (Pfister et al, 1999), our estimate of climate variability prior to the systematic collection of instrumental measurements in the mid-nineteenth century relies on climate-sensitive proxy data (Christiansen and Ljungqvist, 2017; Frank et al, 2010; Jones et al, 2009; Smerdon and Pollack, 2016). We analyze the PAGES2k collection of temperature-sensitive proxy records to understand why the mean tree-ring record lacks a pre-industrial millennial-scale cooling trend that is otherwise preserved in ice core and lake and marine sediment data. Given that the PAGES2k database contains no information regarding the detrending method used to produce the tree-ring chronologies in its collection, we assume all were produced using different detrending methods and that those methods are applied to differently structured tree-ring datasets (i.e., the temporal distributions of short and long tree-ring measurements series, indicative of young and old trees, over the past 2000 years are not the same) If this is the case, such disparities will affect the database chronologies’ long-term variability, causing the tree-ring mean to lack millennial-scale trends (Briffa et al, 2013; Büntgen et al, 2017; Linderholm et al, 2014).
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
