Hydrology and hillslope processes explain spatial variation in tree‐ring responses to the 1983 earthquake at Borah Peak, Idaho, USA

Dendrogeomorphic approaches have been used to identify tree‐ring responses to earthquakes, and have the potential to contribute new paleoseismic information. However, this potential is limited by the relatively low intensity of damage from earthquakes. In this paper, we present a comparison of tree‐ring responses in multiple species to the Mw 6.9 1983 Borah Peak, Idaho earthquake across three adjacent but distinct sites with different hydrologic settings, and different types and intensities of co‐seismic mass movement. We analyzed samples from 166 trees for a broad range of growth responses, including suppressions, releases, reaction wood, and traumatic resin ducts. The proportions of trees recording growth disturbances at the sites with clear hydrologic changes and/or co‐seismic mass movement were twice as high (39–42%) as the site where these processes were less evident (19%), and probability density functions of percentage change in growth for all crossdated trees echoed these results. Growth suppressions and reaction wood were common across sites, whereas traumatic resin ducts were only associated with rockfalls and landslides at one site. Growth releases were also primarily restricted to one site, and in contrast to previous studies, most of these probably reflected direct improvements in growing conditions via documented increased flow in springs and streams, rather than the indirect influence of reduced competition via the death of neighboring trees. These results highlight the importance of site selection in dendroseismology, and particularly suggest that focusing on sites with potential co‐seismic hydrologic changes and mass movement, sampling trees with and without external evidence of earthquake impacts, and understanding the connection between responses and site conditions may facilitate greater contributions of dendrogeomorphic approaches to paleoseismology. © 2018 John Wiley & Sons, Ltd. Tree‐ring responses to the earthquake were more than twice as common in sites with hydrologic changes or co‐seismic mass movement. Growth releases were primarily associated with increases in flow from streams and springs rather than the death of neighboring trees. Traumatic resin ducts in Douglas‐fir and Engelmann spruce were associated with co‐seismic rockfalls and landslides. Applications of dendrogeomorphology to paleosesimology should include sites with potential hydrologic changes or mass movement, and trees without external evidence of earthquake impact