Effects of Water Flow Variation in Large Rivers Exacerbated by Drought

Model shows frequent fluctuation in river flows, caused by dam operations, lead to greater changes in water temperature and biogeochemical reaction rates in river sediments.

The Science

Biogeochemical activity in the hyporheic zone (HZ), sediments where the flowing waters of a river mix with shallow groundwater, supports many of the biological processes that occur within a watershed. Through the creation of a cross sectional (2-D) model of the Columbia River Hanford Reach’s HZ, PNNL researchers, led by Xuehang Song and Xingyuan Chen, found that low flow conditions contribute to warmer waters in the HZ. This, in turn, increases the rate of biogeochemical activity in the sediments. Long-term analysis shows this effect is exacerbated during times of drought.


The Impact

Thermal and biogeochemical dynamics in the HZ are important to fluvial ecology, such as thermal refugia for fish spawning and growth, benthic food production, and nitrate removal. These results can enable natural resource managers to more accurately assess the ecological consequences of long-term frequent water flow variation in riverine systems. In turn, this information will inform dam operations in the context of river and watershed management planning.


Studies of thermal changes in HZs have largely focused on short-term analysis of steady state flow conditions in smaller streams. This study is among the first to model and conduct field analyses in a large river system with high frequency in flow variation. Large fluctuations in water flow levels are a common phenomenon in most river systems with hydroelectric dam operations. To assess the long-term impact of these fluctuations, PNNL researchers created a cross sectional (2-D) thermal-hydro-biogeochemical model of the Columbia River Hanford Reach’s HZ with data supported by field monitoring.

Researchers assessed multiple years’ worth of flow level fluctuation data seeking the most powerful variations, signals unique to dam operations. Inland ground water monitoring data was also used to track the hydraulic gradients driving flow in and out of the HZ. By comparing natural variations against dam-induced differences in flow level, the researchers tracked, over time, the change in temperature, carbon consumption, and other biogeochemical-relevant variables.

Through numerical simulation the model shows a long-term persistent cold-water zone in the riverbed after winter, verified by observational data from a multi-depth thermistor array. Frequent stage fluctuations when the mean flow level is low-particularly under drought conditions during summer and early fall-enhanced heat exchange between the river and the HZ, reaching a maximum temperature difference between 5° to 10°C. All biogeochemical reactions in the HZ were enhanced by increasing nutrient supply and creating more oxygenated conditions. Total carbon consumption, a primary indicator of biogeochemical activities in the HZ, increased by almost 20%. In addition, the model demonstrated that the variable properties of riverbed sediment, such as permeability, influence water residence times and nutrient supplies by controlling flow paths. These variables also determine the spatial distribution of biogeochemical reaction hot spots in the HZ.

Already working towards further improvements to this model, PNNL researchers are expanding the scope of their work from one 2-D cross-sectional analysis to a 3-D analysis of the entire Columbia River Hanford Reach.

Principal Investigator(s)

Xingyuan Chen
Pacific Northwest National Laboratory


Funding for this research came from the DOE Office of Science Office of Biological and Environmental Research’s  Subsurface Biogeochemical Research program for the PNNL Subsurface Biogeochemical Research SFA.


Song, X., X. Chen, J. Stegen, G. Hammond, H-S Seob, H. Dai, E. Graham, and J. Zachara. “Drought Conditions Maximize the Impact of High-Frequency Flow Variations on Thermal Regimes and Biogeochemical Function in the Hyporheic Zone.” Water Resources Research 54(10), 7361-7382 (2018). DOI: 10.1029/2018WR022586