|Abstract or Summary
- Streamflow patterns are a result of the interaction of many factors, including climate, vegetation, geology, and topography. Analyses indicate that streamflow patterns have changed around the United States over the past century, raising questions about the possible role of climate variability as a driver of water yield. This thesis examines streamflow trends in two separate analyses: a study of streamflow trends and the physical characteristics of large basins in the Willamette River Basin, and a study of streamflow trends in small headwater basins around the United States.
In the Willamette River basin, as in much of the western US where summer precipitation is low, declining streamflow will exacerbate the task of water
management to meet varying demands for water. Although regional climate warming over the past half-century in the Pacific Northwest is a plausible cause of declining streamflow, climate is only one of several factors that determine the hydrologic regime of the Willamette River basin. This study quantified trends in streamflow from 1950 to 2010 in ten sub-basins of the Willamette, including seven 60-600 km² predominantly forested sub-basins above dams in the Willamette National Forest, and three gages downstream of dams in urban areas of the Willamette valley (Albany, Salem, and Portland). Trends at the annual, seasonal, and daily time scales were estimated using linear regression with appropriate transformation of variables and the Mann-Kendall non-parametric test.
Trends in streamflow and runoff ratios in the Willamette Basin were compared to geology, topography, and changes in forest cover above dams. Data on forest cover by age class were obtained from Willamette National Forest. Up to 29% of area in sub-basins above dams was clearcut since the 1930s. From 1 to 51% of basin area above dams is on young (<7 ma) volcanic rocks of the High Cascades, which have longer water residence times than the older (7-25 ma) volcanic rocks of the western Cascades, which make up the remainder of basin area above dams. From 71 to 93% of basin area above dams is above 800 m elevation, where seasonal snowpacks are controlled by winter temperature and precipitation.
Over the study period (1950-2010), precipitation in the Willamette River basin did not change significantly at the seasonal or annual time scale. Declining
streamflows in spring and summer were found in low elevation basins above dams. These trends may be attributed to climate-warming-induced increases in the rain:snow ratio and earlier snowmelt, or changes in water use by vegetation in response to climate warming, or both. More rapid declines in summer and fall discharge and runoff ratios occurred in basins which experienced relatively rapid conversion of old to young forest in the past half-century and have relatively large proportions of young forest today. Declining flows can be attributed to increasing evapotranspiration in young forests and shallow flow paths in Western Cascades-dominated basins that intersect the rooting zone. High-elevation basins with greater percentages of High Cascades geology are less susceptible to declining discharge, even if they have substantial areas in young forest.
In contrast to the trends in streamflow above dams, trends over time in streamflow downstream of dams in the Willamette basin have been largely influenced by management of reservoirs and flow. Flow downstream of the dams has become more homogenized throughout the year, with higher flows in the fall and declines in the later winter/early spring. However, discharge has declined at the annual time scale, and in the spring and summer. These trends may be a result of flow management, decreasing spring streamflow upstream of the dams, or a recent trend of relatively dry years.
Although historical water yields may not be representative of future water yields, long term historical climate and discharge records from instrumented sites
provide the basis for understanding the effects of climate variability on water yield from headwater systems. This study quantified trends in climate and streamflow data from eight predominantly forested headwater basins in the US to understand the complex relationships between climate change, basin geography, vegetation, history, and hydrology. Data were obtained for up to 70-year periods from the Long Term Ecological Research (LTER) network and US Forest Service system of Experimental Forests and Ranges (EFR). Trends in precipitation, temperature, discharge, and runoff ratios were calculated for annual, seasonal, and daily data using linear regression of appropriately transformed data and Mann-Kendall non-parametric tests. Linear regression with appropriate transformation of variables produced similar results to the Mann-Kendall analysis.
The first day of spring (defined as the last day of freezing temperature) has moved earlier by 0.31 to 1.98 days/year at all sites, except the alpine site. Most sites experienced significant increases in spring minimum air temperature, but only a few sites experienced an associated change in spring runoff ratios. Spring runoff ratios decreased in sites with snowpacks and near-zero mean spring air temperature (AND). Fall baseflow increased at the alpine, non-forested site (NWT), apparently as a result of increased permafrost melt. Earlier snowmelt and decreases in spring discharge and runoff ratios are expected to lead to increases in winter streamflow and reductions in summer streamflow. However, despite increases in summer minimum air temperature at most sites, summer runoff ratios did not change at most sites, and some sites (HBR, FER) experienced increases in summer and fall runoff ratios. The expected reduction
in summer runoff ratios in response to warmer temperature at these sites may have been mitigated by reductions in leaf area due to disturbance or forest succession after logging in the early 1900s. These findings indicate that both climate change and hydrologic response to climate change varies across the US. The lack of expected streamflow responses apparently is the result of ecosystem processes that produce counteracting trends in vegetation water use in response to climate trends; these processes deserve further study.