Repository logo

Seeing the river through the trees: using cottonwood dendrochronology to reconstruct river dynamics in the Upper Missouri River Basin




Schook, Derek Michael, author
Rathburn, Sara, advisor
Friedman, Jonathan, committee member
Wohl, Ellen, committee member
Covino, Tim, committee member
Denning, Scott, committee member

Journal Title

Journal ISSN

Volume Title


Understanding the past is critical to preparing for the future, especially regarding rivers where extreme events and gradual changes underlie modern forms and processes. Both biological and human communities rely on the abundant resources provided by rivers and floodplains, particularly in dry regions of the western U.S. where water limits growth. To expand temporal perspectives on river processes, I reconstructed flow, channel migration, and riparian forest growth patterns in the Upper Missouri River Basin. Flow reconstructions typically use tree rings from montane conifers. However, I used riparian plains cottonwoods (Populus deltoides ssp. monilifera) directly connected to the alluvial water table to reconstruct flow on the Yellowstone (n = 389 tree cores), Powder (n = 408), and Little Missouri Rivers (n = 643). A two-curve Regional Curve Standardization approach was used to remove age-related growth trends from tree rings at each site. The flow reconstructions explained 57-58% of the variance in historical discharge and extended back to 1742, 1729, and 1643, respectively. Low-frequency flow patterns revealed wet conditions from 1870 to 1980, a period that includes the majority of the historical record. Two 19th century droughts (1816-1823 and 1861-1865) and one pluvial (1826-1829) were more severe than any recorded, revealing that risks are underestimated when using the instrumental period alone. These are the first flow reconstructions for the Lower Yellowstone and Powder Rivers, and they are the farthest downstream among Rocky Mountain rivers east of the Continental Divide. Cottonwood-based flow reconstructions were possible because the trees used river-connected groundwater, and tree ring width strongly correlated with March-June flow magnitude at the Yellowstone River (r = 0.69). Beyond the site-level growth patterns typically used to reconstruct flow, I found that biological and spatial characteristics affected how individual trees responded to flow and climate. Older trees contained stronger signals of non-growing season flow, precipitation, and temperature, which challenges the common dendrochronological assumption of stable tree ring-climate relationships through time. Although trees both near and far from the channel were better correlated to spring flow than precipitation, more distant trees had a stronger relative connection to precipitation, suggesting that greater distance decreases the ability of river water to fulfill transpirative demands. Like annual growth, cottonwood establishment is related to river flows, and tree age indicated fluvial processes including channel migration. I quantified nearly two centuries of channel migration on the Powder River by integrating measured channel cross-sections (1975-2014), air photos (1939-2013), and transects of aged cottonwoods (1830-2014). The combined data revealed that channel migration rates were lower (0.81 m/yr) in the recent and intensively studied cross-section period compared to the longer air photo (1.52 m/yr) and cottonwood (1.62 m/yr) periods. On the Powder River, extreme floods such as those in 1923 and 1978 increase subsequent channel migration rates and initiate decades of channel morphological adjustments. Across the study rivers, data indicate that fundamental fluvial processes have responded to climatic and watershed pressures. By identifying and quantifying past events, diverse research approaches improve understanding of the river, floodplain, and riparian forest processes that are essential to the persistence of these valuable ecosystems.


Rights Access



Associated Publications