Trends in Streamflow Above and Below Dams Across the Columbia River Basin from 1950 to 2012: Assessing Sub-basin Sensitivity

Climate change, combined with population growth, is expected to exacerbate
water scarcity globally. In the Columbia River basin (CRB), streamflow is managed for
multiple objectives with a network of dams and reservoirs distributed throughout the
basin that may mitigate climate change effects on water scarcity. This study quantified
trends in daily streamflow from 1950-2012 in 28 pairs of gages above and below dams in
the CRB, including the U.S. and Canada. Each gage pair consisted of an above-dam gage
that is primarily forested and has little apparent human disturbance and minimal flow
regulation or diversions, combined with a below-dam gage immediately downstream of a
major dam and reservoir. Monthly streamflow and precipitation for a total of 59 sites
was correlated with monthly indices of the Pacific Decadal Oscillation (PDO), the
Northern Pacific Index (NPI), the Pacific North America pattern (PNA), and the 700-
millibar wind speeds (u700) using Pearson's r. Long-term trends and patterns in daily
streamflow and climate data were estimated using linear regression, the Mann-Kendall
test, and wavelet analysis. Streamflow was generally weakly correlated with NPI, PNA,
and PDO, while positively correlated with upper elevation wind speed at low elevations
during wet months. Fewer than half of above-dam gages in the CRB have experienced
! !
increasing trends in springtime daily streamflow (mid-March to the beginning of May)
over the period 1950 to 2012, whereas a majority has experienced decreasing summer
and early autumn trends (mid- May to mid-October). These trends in above-dam basins
are consistent with trends expected from climate change, but they also may be affected by
legacies of past forest harvest or wildfire. Below dams, reservoir management appears to
have overwritten the signal of earlier snowmelt, except in the Boise sub-basin, but longterm
declines in late summer flows were evident at half of below-dam sites. Declining
summer flows below dams were attributable to a variety of factors, including changes in
reservoir management as well as reservoir management that propagates signals from
above dam catchments, such as climate change or forest-harvest legacy effects on
streamflow. There have been very few significant changes in annual flow volume
throughout the basin, and these trends represent a small percentage of annual flow
volume; thus, observed trends appear to be shifts in timing rather than shifts in the annual
water balance. The magnitude of 60-yr declines in late-summer flow in above-dam
catchments represents only a few percent of annual yield, and much less than the storage
capacity of the dams, but these long-term changes may be quite important for water yield
during the late-summer low-flow period. These findings are relevant for strategies to
adaptively manage water resources in light of the ongoing review activities around
possible renegotiation of the U.S.-Canada Columbia River Treaty.