Department of Earth and Ocean Sciences University of British Columbia 6339 Stores Road Vancouver BC V6T 1Z4 Canada

This case study briefly turns the spotlight upon a topic with a broader significance that may prove to belie its perhaps modest appearance: differences in the winter groundwater responses of glacierized and glacier-free watersheds to climatic warming. Glacial runoff could be altered significantly by projected anthropogenic climate warming, and observed long-term trends in surface water resources under historical warming conditions can be dramatically different between glacierized and nival catchments. Virtually no attention has been devoted, however, to potential differential trends in groundwater availability. Such differential trends, if present, may be particularly important from an ecological perspective. Groundwater resources control quality and quantity of cold-regions winter aquatic habitat, largely by sustaining the (typically very low) winter baseflows of freshet-driven rivers, with powerful and direct impacts upon fish populations and biogeography (see Reynolds, 1997; Cunjak et al., 1998). Temporal trends in winter groundwater supplies, and interbasin variability in such trends arising from the presence or absence of watershed glacial cover, would thus strongly imply parallel fluvial hydrological and ecological effects.

Consistent multidecadal piezometric measurements from aquifers undisturbed by human utilization are very rare, but usably long streamflow records from sufficiently pristine watersheds are relatively common. As winter baseflow is a direct reflection of hydrogeological conditions, such data offer a surrogate measure of groundwater resource availability and also 'speak' directly to riverine impacts. Time series of annual minimum daily discharge were therefore constructed from historical streamgauge records (Environment Canada, 1999) for five glacierized and four nival basins in southwest Yukon and northwesternmost British Columbia (Fig. 29.1). These watersheds possess a number of properties advantageous to an empirical analysis of the effects of watershed glacierization upon water resource responses to historical climatic warming, and have been used previously for the purpose. Fleming & Clarke (2003) provide a detailed description of the study area and data; two key points are that the region is experiencing a climatic warming trend, and that systematic differences in the hydrological responses of glacierized and nival catchments are indeed directly attributable to presence or absence of glacial cover. Note that surface and subsurface hydrological catchment areas appear to correspond reasonably well within the study area (see Owen, 1967). Each of the nine time series was normalized by its own maximum value to facilitate interbasin comparisons, and winter baseflow trends, b, were determined for each

Alaska

White River Big Creek

1975-99 1975-99

/ glacial nival

Kluane River 1953-95 glacial

Yukon

Dezadeash River 1953-99

nival Takhini River cS® WCIintock River 0 glacial 1966^4

O nival

Wheaton River, 1966-99, nival

Dezadeash River 1953-99

nival Takhini River cS® WCIintock River 0 glacial 1966^4

O nival

Wheaton River, 1966-99, nival

Wann River, 1964-93, glacial

British Columbia

Figure 29.1 Gauge locations and corresponding hydrologi-cal record durations.

catchment using a rank-based approach resistant to outliers (Theil, 1950)

where xi, i = 1, N is the annual time series of length N. Results are shown in Table 29.1. The median trend value for glacierized catchments is somewhat higher than that of the nival basins, but the difference is not statistically significant (P > 0.05, 2-tailed Mann-Whitney test; see also Fleming & Clarke, 2003). The ranges are also similar. However, the overall trend pattern appears to reveal a stark contrast: glacierized catchments yield uniformly positive trends, whereas the various nival watersheds exhibit a mixture of both progressively increasing and decreasing winter baseflows.

The likelihood that this pattern arises from chance trends in random time series is assessed using a non-parametric Monte Carlo randomization technique. In a given realization, random baseflow records are generated by scrambling the temporal order of the observations in each of the nine original time series. Trends are then estimated using Equation (1). The trend pattern thus generated is taken to match the observed pattern if the following condition is satisfied

ß > 0 all l = 1, Ng} and ß(ßm > 0 any m = 1, Nn ) and

where Ng and Nn are the number of glacierized and nival basins, respectively. A total of 105 such realizations are synthesized. The probability, P, that the observed trend pattern occurs by chance is the proportion of the ensemble that satisfies Equation (2). By this technique, P = 0.02; the observed pattern is statistically significant.

Although streamflow data are utilized, the baseflow trend pattern is reflective of changes in hydrogeological conditions and must be interpreted in that light. I posit that the observed pattern arises from systematic basin-to-basin variability in the net balance of trends in aquifer recharge and aquifer properties. Fleming & Clarke (2003) demonstrated that freshet magnitudes here are increasing in glacierized watersheds but decreasing in nival watersheds. This essentially results from the relative importance of trends in temperature and precipitation and their effects upon evapotranspiration and glacial melt production, and implies increased (decreased) summertime aquifer recharge in glacial (nival) watersheds. However, aquifer properties are also believed to be changing due to the permafrost impacts of historical climatic warming. These watersheds lie in areas of sporadic (valley bottoms) to continuous (alpine zones) permafrost, which exerts powerful control over hydrogeological properties. Permafrost may be degrading regionally in various fashions under warming conditions, which has been hypothesized to increase the connectivity and storage capacity of aquifers, resulting in greater winter baseflow (Woo, 1990; Michel & van Everdingen, 1994). Warming-induced trends in aquifer recharge and aquifer storage capability are in the same direction for glacierized catchments, resulting in uniformly positive baseflow trends. In contrast, such hydrogeo-logical trends lie in opposite directions for nival catchments, so that the net direction of groundwater resource impacts is sensitive to interbasin variability in basic geological characteristics, leading to a mixture of positive and negative baseflow trends in different nival watersheds. Implications for basin-to-basin consistency in the direction of climatically induced changes in aquifer-stream coupling and winter aquatic habitat availability are clear.

The above interpretation, although plausible, is by necessity conjectural. Two important steps toward a deeper understanding of glacial influences upon groundwater responses to climatic shifts are to replicate the analyses presented here and by Fleming & Clarke (2003) in other regions, and to obtain year-round piezo-metric measurements in undisturbed fluvial aquifers to better constrain aquifer-stream interactions in glacial and permafrost environments, enabling more effective and thorough use of available long-term baseflow datasets.

Table 29.1 Estimated baseflow trends (10 3yr ')

Glacial watersheds

Nival watersheds

Wann Alsek

Kluane

White

Takhini

Big Creek M'Clintock

Dezadeash

Wheaton

1.2 5.2

9.7

6.2

4.5

-0.2 0

4.6

-3.3

0 0

Post a comment