A high-throughput X-ray micro-CT pipeline for dendrochronological applications

Tree-ring data provides essential insights into historical climate conditions and current ecosystem dynamics. Wood density measurements complement traditional tree-ring width series by extracting additional climatic information embedded within tree rings, with maximum latewood density serving as the gold standard for summer temperature reconstructions. However, the labor-intensive wood sample preparation required by conventional techniques is one of the main factors limiting the more widespread use of tree-ring densitometry. X-ray micro-computed tomography (XμCT) offers a novel, non-destructive, 3D densitometry technique that enables the simultaneous study of tree-ring width and wood density at high resolution and with minimal sample preparation. Despite its potential, the lack of a streamlined and time-efficient XμCT workflow has hindered its broader application in the past. Here we present a highly optimized XμCT pipeline aimed at large-scale tree-ring densitometry studies, capable of processing large amounts of increment cores with a low demand for human labor time (2.4 - 6.4 minutes per 35 cm increment core for 16 cores and more). Key parts include large and efficient sample holders that fit 4 to 48 samples per scan, optimized and standardized scan settings (helical XμCT at 15, 30, and 60 μm voxel size), and three custom software packages that facilitate scan processing, ring indication, and cross-dating. Overall, this optimized XμCT toolchain represents a significant step forward in high-resolution tree-ring densitometry, enabling large-scale studies with a wide spatial extent, high replication, and/or long temporal range. As an example application of this toolchain for a palaeoclimatological task, we demonstrate it on samples of Bristlecone pine (Pinus longaeva) trees, that are known for their exceptional longevity. We extracted maximum latewood density from 51 cores from the California White Mountains to build a chronology that correlates significantly (r = 0.66, p < 0.01) with warm-season (March-September) temperature over a large spatial extent. Previously it was impossible to achieve this because of gnarly and twisted growth of these ancient trees, that makes traditional X-ray densitometry methods inapplicable for maximum latewood density extraction. This led to the first X-ray CT-based temperature reconstruction (1625–2005 CE) [1]. Good reconstruction skill (RE = 0.51, CE = 0.32), which was not possible to obtain based on tree-ring width , shows that extending maximum latewood density record across the full length of the Bristlecone samples archive could yield a robust warm-season temperature proxy for the American Southwest over several millennia.