Julie Edwards
Ph.D. Student
Global Change GIDP
Research Location
Birmensdorf, Switzerland
On Site Research Dates
May 15 to June 12, 2023
Title of Research Project:
High-resolution temperature reconstruction from the North American treeline using quantitative wood anatomy
Temperatures in the Arctic have one of the fastest warming rates on the planet, and they are
projected to increase at more than twice the rate of the global mean temperature change over the
next century. The Arctic also experiences some of the largest magnitude summer temperature
anomalies following large explosive volcanic eruptions and is influenced by large-scale modes of
climate variability. Therefore, it is a critical region for observing and understanding both externally
forced and internal climate system variability and concomitant changes in forests, carbon,
cryosphere feedback, and extreme events. However, the short length of instrumental and satellite
observations provide only limited opportunities to observe multidecadal or centennial variability,
extreme events, and climate anomalies associated with rare large volcanic eruptions.
Paleoclimate reconstructions fill these observational gaps by using networks of proxy observations
that can be used to estimate past temperatures. Tree-ring data, in particular, provide annually
resolved and absolutely dated information about the growing season conditions over multiple
centuries or millennia. However, there is a paucity of long tree-ring chronologies from the high
latitudes of North America. Therefore, paleotemperature reconstructions over North America
remain severely limited by this striking lack of regional and highly temperature-sensitive tree-ring
data spanning the last millennium.
For my Ph.D. dissertation, I propose to develop novel high resolution
millennial-scale tree-ring data from the North American treeline by applying recent
advances in an emerging subfield of dendrochronology (the study of tree rings) called quantitative
wood anatomy. These techniques not only provide us with a cellular-scale, and therefore, a highly
precise anatomical determination of wood density, but also provide a wealth of additional intraannual
information that can be used to refine the seasonal climate signal and precisely detect the
timing of volcanic cooling anomalies. We can leverage the precise seasonal climate signals of these
novel data in combination with climate model simulations to estimate not only past temperatures,
but also atmospheric circulation features associated with temperature variability and extremes.
Quantitative wood anatomy uses image analysis of high-resolution tree-ring thin-section digital
photomicrographs to measure the characteristics of individual cells in the rings. This method
produces an extremely large amount of data (millions of cells) from various measurements and
positions of each cell in each ring across each tree sample. Data from quantitative wood anatomy
can provide more information on past climate than what can be gained from traditional tree-ring
width alone. My preliminary data analysis shows that anatomically measured wood density is
highly correlated with temperature in a high-latitude forest near the Firth River in Alaska, a result
that cannot be seen by simply looking at tree-ring width.
The creation of quantitative wood anatomy data requires specialized equipment that is available
only at a small number of institutions. One such institution is the Dendrosciences research unit at
the Swiss Federal Institute for Forest, Snow, and Landscape Research (WSL) in Birmensdorf,
where the group leader Dr. Georg von Arx pioneered quantitative wood anatomy techniques. The
WSL Dendrosciences group uses a Zeiss Axio Scan.Z1 slide scanner to create high-quality digital
images of wood thin-sections. The Z-stack imaging capability of the Zeiss Axio Scan creates
images with nearly perfect focus, which is essential because minor inaccuracies in the image focus
can lead to a bias in the final wood anatomy measurements. The WSL Dendrosciences group
calibrated their slide scanner specifically to optimize tree-ring thin-section image creation. The thinsection
images are then analyzed using the image analysis software ROXAS, which was created by
Dr. von Arx, to perform the specific task of automatically measuring wood cell characteristics with
minimal manual adjustment. ROXAS software is continuously updated with community efforts in
quality control testing.
During my visit to the WSL Dendrosciences group, I will learn more about operating their state-ofthe-
art slide scanner and the ROXAS program. Dr. von Arx and I planned my visit to the WSL
Dendrosciences group from May 15 to June 12, where I will complete three objectives.
I am currently located at the Laboratory of Tree-Ring Research (LTRR) at the University of
Arizona, which is another world-renowned center for dendrochronology (the study of tree rings),
yet we currently lack the specialized equipment required to generate quantitative wood anatomy
data independently. To address this, my PhD advisor Dr. Kevin Anchukaitis is acquiring a Zeiss
Axio Scan.Z1 slide scanner for the LTRR next year. As part of this equipment acquisition, I will
calibrate the slide scanner for tree-ring thin-section digital photomicrographs and will be
responsible for training other users. At the WSL, I will learn the best practices for imaging wood
thin sections and learn the solutions to the roadblocks and pitfalls encountered by the
Dendrosciences group. I will then return to the LTRR with this knowledge to train users of the new
scanner. I will learn how to operate the slide scanner to capture images using my own thin-section
samples to create data for my Ph.D. research. This will have the dual benefit of generating my own
data while being trained in slide-scanner operation.
I will also bring wood samples to the WSL Dendrosciences group to learn about their sample
preparation techniques. I am currently collaborating with a master’s student in our lab group to
research the wood anatomy of Guatemalan Hartweg's pine and the endangered Guatemalan fir.
While it has been previously impossible to reconstruct past temperatures using tree-ring widths in
this region, we are investigating whether there is a temperature signal that can be obtained through
quantitative wood anatomy techniques. However, because of the age and composition of these tree
samples, obtaining high-quality thin-section slides has been challenging. During my visit to WSL, I
will learn if the Dendrosciences group has any recommendations for getting high-quality thinsections
from these samples. I will also present on the progress that the master’s student has made
in her research on Guatemalan tree-rings and climate.
Finally, I will learn about recent updates to the ROXAS software and will participate in software
quality assurance while running my tree-ring thin-section digital photomicrographs through the
program and generating my own data. The ROXAS software is currently dependent on an expensive
commercial image analysis software; however, the WSL Dendrosciences group is currently
programming a free open-source version. By the time of my visit in May, there will be a working
model of the software that I can assist with in the quality control and quality assurance processes.
As a result of my visit, I will complete several products and deliverables. The research I generate
during my visit will be used in my Ph.D. dissertation, which will become two additional journal
publications. I will also generate data on the wood anatomy of Guatemalan conifers for the master’s
student’s thesis, which will also be published in a peer-reviewed journal. I will add the laboratory
techniques and methods I learn at the WSL into the training documents I have already created for
the LTRR. This compiled report will be done in collaboration with the WSL Dendrosciences group
and LTRR research faculty and will be circulated to other research groups that would like to
conduct quantitative wood anatomy research.