| Citation |
Moore, R.D., D.M. Allen, and K. Stahl (2007) Climate Change and Low Flows: Influences of Groundwater and Glaciers, University of British Columbia and Simon Fraser University. Final Report for Climate Change Action Fund Project A875, 17 August 2007. Prepared for: Climate Change Action Fund, Natural Resources Canada. |
| Abstract/Description or Keywords |
Streamflow patterns in BC can be broadly classified as rain-dominated, hybrid (rain and snow), snow-dominated, and glacierized, each having distinct low flow seasons. In this study, we chose to focus on the summer/early autumn low flow period, as that has particular economic and ecological significance. Variability in stream discharge was measured over a low flow season in Bertrand Creek, Pepin Brook and Fishtrap Creek, which drain the Abbotsford-Sumas aquifer in southwest British Columbia and northwest Washington State. Discharge was measured across a series of profiles, both repeating across a central line (at-a-station measurements) and along consecutive sections from upstream to downstream (successive downstream measurements) to assess the repeatability of the low flow measurements. As the flow in these small streams decreases, the coefficient of variation of the mean discharge tends to increase when the flow is very low (i.e., <0.1 m3/s), although the trends for Fishtrap are weak and less reliable likely due to the higher flows at that site. Overall, variations in measurements are of sufficient magnitude that subtle changes in streamflow due to, for example, climate change may be difficult to detect. Unglacierized catchments generally exhibited declining trends in September streamflow through most of British Columbia, but not in August. Glacierized catchments, on the other hand, show the opposite pattern, with dominantly negative trends in August but not September. The decreases in September flows are broadly consistent with a decline in September precipitation over the period studied. For all four regimes, the most important control on August streamflow is August precipitation. August streamflow is also positively but more weakly related to lagged variables including July precipitation and the previous winter's precipitation. There is clear evidence for the effect of multi-year storage change only for the glacierized catchments. Glacier retreat over the last few decades has apparently reduced glacier area sufficiently so as to reduce meltwater generation and thus streamflow. Snowmelt-dominated catchments showed no tendency to trends or serial correlation in the model residuals, though the runs test suggested a substantial number of stations had non-random residuals. Both raindominated and hybrid catchments tended to have positive residuals in August, suggesting either an increasing trend in groundwater storage or a spurious cause such as model mis-specification. The majority of groundwater well observation records in BC start in the late 1970s or 1980s, providing a total of 20 to 30 years of record. A select few of these wells monitor aquifers in pristine areas that reflect natural variability; the others have been influenced by human activity making them less representative. Climate change detection requires long term records because natural climate variability phenomena, such as the Pacific Decadal Oscillation (PDO), last over multiple decades. Long term fluctuations of precipitation, air temperature, and stream flow can be reflected in groundwater level fluctuations, but the hydrogeology of a given aquifer plays a unique role in the interaction of climate, surface water, and groundwater. By comparing the relationships between groundwater, climate, and surface water within and between groups of well records from the two major hydro-climatic zones in BC, a system for detecting the influence of climate change and variability on groundwater in the absence of long term records is defined, and suitable correlation coefficients are applied to evaluate the strength of these interactions. Different aquifer types are assessed with respect to vulnerability to climate change influences. Overall, summer groundwater levels appear to have lowered across the province, despite an increase in winter precipitation and recharge during the same time period. Due to the limited availability of long well records near gauged streams, the attribution of whether and how these changes have affected low flows proved difficult. The available groundwater observation wells vary substantially in terms of aquifer properties and the hydraulic connection to rivers and for many, we suspect the natural records may be altered by changes in the abstraction patterns. The few examples where streamflow and groundwater observations are available in the same basin and are nearby furthermore provide some of the exceptions to the trends found. In Abbotsford, even though a field study and a modelling study showed that summer low flow is fed by groundwater, low flows increased while over the same time period groundwater levels decreased. The same was found for Grand Forks, where previous studies have shown that streamflow is fed by groundwater in the summer. However, the recharge mechanisms are different in both examples. A methodology was developed to simulate the transient effects of glacier retreat on streamflow patterns by coupling a semi-distributed hydrological model (HBV-EC) with a glacier mass balance and glacier scaling model. This study investigated the sensitivity of streamflow to changes in glacier cover for the Bridge River basin in British Columbia, which is an important water source for one of BC Hydro's hydro-electric facilities. The model is driven into the future assuming three different types of climate scenarios: a continuation of the current climate and a moderate warming scenario realized with a weather generator, as well as a set of downscaled GCM scenarios with greenhouse gas forcing. Considering specifically the effect of changing glacier area in the basin, the modelled glacier mass balance is used to re-scale the glacier every decade using a universal volume-area-scaling relation. The model application shows strong reductions in glacier area and summer streamflow even under the assumption of a continuation of the present climate. The results of this study suggest that climate warming and associated glacier retreat will have significant implications for water resources and aquatic ecology. Glacier-fed rivers are likely to experience a shift from a glacial regime with high flows in mid and late summer, with an associated moderating effect on stream temperature, to a regime that responds to the summer dry period with streamflow recession, low flows and increased temperatures. |
