Climate change scenarios were downloaded from the Pacific Climate Impacts Consortium (PCIC). Three future time series were assessed:
2020s (2010-2039), 2050s (2040-2069), and 2080s (2070-2099). The observed daily climate time series for 1961-90 was used as the base-line period. IPCC (2007) recommended using more than one GCM in impact assessments to demonstrate how a range of air tempera-ture and precipitation changes may affect a given region. In order to provide sensitivity analyses of predicted climate change impacts on streamflow, five scenarios were selected using the method de-veloped by Barrow & Yu (2005). These scenarios were chosen from the 41 climate scenarios provided by PCIC containing both air tem-perature and precipitation predictions. Following Barrow & Yu (2005), the climate scenario selection was based on mean air temperature change and percent precipitation change for the 2050 spring (March – April - May) time period. Figure 3 presents the selection method presenting data for the CRW. The selection of GCM scenarios is based on the creation of four quadrants separated by the median air temperature change (here: +1.75°C) and the median precipita-tion changes (here: +12.0%). Five scenarios from GCM output were selected based on their projection of the range of possible future climates: for the BCW hotter-drier, hotter-wetter, median, warmer-drier, warmer-wetter, and for the CRW hotter-wetter, warmer-wetter, median, hotter-wettest, and warmer-wettest (Figure 3). Rather than using just one climate scenario, these five scenarios were needed to provide a representation of the range of predicted possible future climates in the respective study area, thus enabling the simulation of future streamflow scenarios without prejudice.
Due to the large size of the GCM climate grid cells of 400 by 400 km, the GCM output needed to be downscaled to a regional climate to evaluate climate change impacts at a watershed scale. To con-Figure 2. Key elements of the ACRU agro-hydrological modeling
system
struct regional climate change scenarios based on GCM output, a widely used procedure, the delta method, was applied (Hay et al., 2000). The delta method of downscaling uses projected monthly changes in air temperature and precipitation based on results from each selected scenario (Figure 3) to perturb the observed 1961 to 1990 climate record.
Results
Whilst a wide range of variables and behaviours were analysed (Forbes et al., 2010; Nemeth et al, 2010), only a few key results are reported here. For the BCW, the mean annual simulation results are summarized in Table 1. The water balance components (Precipiation
= Actual Evapotranspiration + Runoff ) do not sum up to 100%, as there are storage changes in the watershed for groundwater, soil moisture, and snow pack. Annual precipitation volume increased in the majority of scenarios, except those scenarios which projected a decrease in annual precipitation (HD in the 2020s and 2080s scenari-os, WD in the 2020 scenario). In all scenariscenari-os, a greater volume of the rainfall was simulated, while concurrently snowfall was simulated to be reduced The proportion of snowfall to total precipitation de-creases from the historical (1961-1990) 48.9% to an average of 41.7%
for 2010-2039, an average of 39.8% for 2040-2069, and an average of 38.1% for 2070-2099. In all scenarios, potential evapotranspiration was simulated to increase above the baseline simulation, which can be attributed to the increase in mean annual air temperatures across the scenarios. In all scenarios, the changes in AET are related to the changes in precipitation (Table 4). This is an interesting result, as it indicates that in this semi-arid, water limited region future changes in AET depend more on precipitation changes than on air temperature changes. In this environment, actual evapotranspiration is limited by available soil moisture rather than atmospheric demand. The simulated changes in mean annual streamflow (Q), relative to the baseline, showed an increase for most simulations. This is due to an increase in precipitation with a concurrent lesser increase in actual evapotranspiration.
Figure 3. Selection of five GCM scenarios
Period Rain Snow Total P PET AET Q
Baseline 1961-90 218 242 460 959 431 25
HD
2010-39
228 207 435 1111 411 20
HW 245 221 466 1106 435 28
MD 250 212 462 1089 433 25
WD 245 203 448 1089 422 22
WW 282 212 494 1085 461 29
HD
2040-69
252 207 459 1156 430 26
HW 242 218 460 1180 423 34
MD 266 202 468 1141 440 25
WD 269 198 467 1216 436 28
WW 281 209 490 1106 460 27
HD
2070-99
240 203 443 1208 415 26
HW 267 220 487 1245 444 40
MD 280 195 475 1179 445 28
WD 310 187 497 1302 463 32
WW 303 231 534 1116 493 38
Table 1. Mean annual water balance components for the Beaver Creek watershed (all values in mm).
There is strong shift in seasonality. With a few exceptions, the future streamflow regime is simulated to change towards much increased streamflow in winter, a smaller increase in spring, a decline in sum-mer, and a potentially severe decline in fall.
For the CRW, results are summarized in Figure 4. Peak streamflows are simulated to increase and occur earlier in all five scenarios and all three time periods. Also, starting approximately in Week 40 (early October), and lasting until approximately Week 12 (late March), baseflow is simulated to increase significantly. Of importance is the period between approximately Week 27 (mid July) and Week 39 (late September), when future streamflow is reduced, ranging from be-tween about -25% in the 2020s period to about -33% in the 2080s period. As expected, simulations further in the future result consist-ently in a wider spread of predicted streamflow behaviour, resulting in increased uncertainties. While PET is simulated to increase from an average of all five scenarios of 11.5% in the 2020s to 27% in the 2080s, AET is projected to increase from 2.5% to 8.5% respectively.
Soil moisture is simulated to very slightly increase (from an aver-age 0.5% in the 2020s to an averaver-age of 2.6% in the 2080s). Ground-water recharge is projected to increase from an average of 1.6% in the 2020s to 8.0% in the 2080s, resulting in higher baseflow and increased low flows (Figure 4). Snow water equivalent is simulated to strongly decrease, from an average of -47% in the 2020s to -66%
in the 2080s. This results in a shift from a hydrological regime that is snow melt dominated to one that is dominated by stormflows.
While mean annual water yields are simulated to increase by 1.1%
in the 2020s, ranging from an increase of 3.7% to a decrease of 2%, they are projected to increase by an average of 11.5% in the 2080s, ranging from 7.5% to 17.8%.
Figure 4. Simulated future weekly mean streamflows and percent changes
Discussion
In the Beaver Creek watershed, water yield was simulated to be quite stable, despite future increases in air temperature. This is explained by the limitation of actual evapotranspiration due to restrictive soil moisture availability. Both watersheds exhibited an increase in pre-cipitation during the winter and spring, and either a decline (BCW) or lesser increase (CRW) in precipitation during the summer and fall.
This resulted, in the BCW, in a decrease in soil moisture and subse-quent decrease in groundwater recharge, with sunsesubse-quently lower baseflow during summer and fall. Whilst groundwater recharge was simulated to be increased in the CRW, it happens also much earlier, with the consequence of being exhausted in the summer and early fall, resulting in a decline in summer and fall streamflows. The BCW, with a runoff coefficient of just 5%, is vulnerable to climate change. In contrast, the CRW has a runoff coefficient of 58% and is predicted to receive increased precipitation, making it more resistant to harmful climate changes. The CRW streamflows are also influenced by the declining glacier melt contributions, as the peak in glacier melt was likely surpassed about a decade ago, and the steadily declining glaci-ated area and glacier volume will result in a declining glacier melt contribution, until all glaicers have completely melted.
A major problem in simulating climate change is the shortcoming of all GCMs in their current ability to predict changes in terms of frequency (number of days with precipitation) and intensity (amout of rainfall or snowfall on a given day). As most hydrological processes are non-linear, a change in either intensity or frequency (or both) would change elements of the hydrological cycle. For example, a shift towards more frequent, but less intensive, precipitation events would result in increased evapotranspirational losses, drier soils, and less runoff; and vice versa.
Another deficiency of the GCMs applied here is their inability to sim-ulate the impacts of the Pacific Decadal Oscillation (PDO), which has an approximate 60-year cycle and a major influence on the weather in southern Alberta, as do other cycles such as the Arctic Oscillation and the El Nino Southern Oscillation (ENSO). All of these long-term cycles are associated with distinct long- to medium term weather patterns (drier or wetter, warmer or colder) within the study water-sheds, which can temporarily (for a year or up to several decades) override the long-term climate change trends. Failure to incorporate these multi-decadal cycles increases the inherent uncertainties as-sociated with global climate models.
Only when future climate models successfully integrate these long-term cycles and include future changes of precipitation intensities and frequencies, we can expect to simulate more reliable future pre-dictions of climate change impacts on watershed behaviour.
References
Ů Bathurst, J.C., Ewen, J., Parkin, G., O’Connell, P.E., Cooper, J.D., 2004. Validation of catchment models for predicting land-use and climate change impacts. 3. Blind validation for internal and outlet responses. Journal of Hydrology, 287 (1-4), 74-94.
Ů Barrow, E., Yu, G., 2005. Climate Scenarios for Alberta. A Report Prepared for the Prairie Adaptation Research Collaborative (PARC) in co-operation with Alberta Environment. Alberta Envi-ronment, Regina, Saskatchewan.
Ů Forbes, K.A., Kienzle, S.W., Coburn, C.A., Byrne, J.M. and Rasmus-sen, J., 2010. Modelling the impacts of selected GCM derived climate scenarios on the future hydrology of a hybrid watershed in the Oldman River watershed, Alberta, Canada.
Ů Gan, T.Y., 1998. Hydroclimatic trends and possible climatic warm-ing in the Canadian Prairies. Water Resources Research, 34(11), 3009–3015.
Ů Hay, L.E., Wilby, R.L., Leavesley, G.H., 2000. A Comparison of delta change and downscaled GCM scenarios for three mountain-ous basins in the United States. Journal of the American Water Resources Association, 36(2), 387-397.
Ů IPCC (2007) Climate Change, 2007: The Physical Science Basis.
Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change.
Solomon S, Qin D, Manning M, Chen Z, Marquis M, Averyt KB, Tignor M, Miller HL (eds.). Cambridge University Press, Cam-bridge, United Kingdom and New York, NY, USA.
Ů Kienzle, S.W., 2008. A new temperature based method to sepa-rate rain and snow. Hydrological Processes, 22: 5067-5085.
Ů Kienzle, S.W., 2010. Effects of area under-estimations of sloped mountain terrain on simulated hydrological behaviour. Hydro-logical Processes.
Ů Kiker, G.A., Clark, D.J., Martinez, C.J., Schulze, R.E., 2006. A Java-based, object-oriented modeling system for southern African hydrology. Transactions of the ASABE, 49(5): 1419-1433.
Ů Loukas, A., Quick, M.C,. 1996. Effect of Climate Change on Hydro-logic Regime of Two Climatically Different Watersheds. Journal of Hydrologic Engineering, 1 (2), 77-87.
Ů Loukas, A., Vasiliades, L., Dalezios, N.R., 2002. Potential climate change impacts on flood producing mechanisms in southern British Columbia, Canada using the CGCMA1 simulation results.
Journal of Hydrology, 259, 163-188.
Ů Nurmohamed, R., Naipal, S., De Smedt, F., 2007. Modeling hy-drological response of the Upper Suiname river basin to climate change. Journal of Spatial Hydrology, 7(1), 1-22.
Ů Schulze, R.E., 1995. Hydrology and Agrohydrology: a text to accompany the ACRU 3.00 agrohydrological modelling system.
Report TT69/95. Water Research Commission, Pretoria, RSA.
Ů Schulze, R.E., Perks, L.A., 2003. The Potential Threat of Significant Changes to Hydrological Responses in Southern Africa as a Result of Climate Change: A Threshold Analysis on When These Could Occur, and Where the Vulnerable Areas are. In: Schulze, R.E. (Ed.) Modelling as a Tool in Integrated Water Resources Man-agement: Conceptual Issues and Case Study Applications. Water Research Commission, Pretoria, RSA, WRC Report 749/1/02.
Chapter 14, 250-258.
Ů Smithers, J., Schulze, R.E., 1995. ACRU Agrohydrological Model-ling System User Manual. WRC Report TT 70/95, Water Research Commission, Pretoria.
Ů Toth, B., Pietroniro, A., Conly, F.M., Kouwen, N., 2006. Modelling climate change impacts in the Peace and Athabasca catch-ment and delta: I-hydrological model application. Hydrological Processes, 20, 4197-4214.
Ů Wheaton, E., 2001. Limited Report. Changing Climates: Exploring Possible Future Climates of the Canadian Prairie Provinces, SRC Publication No. 11341-3E01, Government of Canada.
Ů Whitfield, P.H., Fraser, D., Cohen, S., 2003. Climate Change Im-pacts on Water in Georgia Basin / Puget Sound – Special Issue.
Canadian Water Resources Journal, 28(4), 523-529.