Water footprint and impact of hydroelectric reservoirs in the northern boreal forest of Quebec on the regional climate

In Québec, 96% of the electricity generated comes from the province’s abundant water resources which, in the form of thousands of rivers and lakes, cover close to 12% of its territory. Large bodies of water such as lakes and reservoirs have properties that are very different from the surrounding land. They store heat during the warm season, and reduce local surface roughness and diurnal temperature variations. They are also an important source of water that is transferred to the atmosphere through evaporation. Through these properties lakes and reservoirs have an impact on the regional climate. Yet, to date few studies have performed a detailed analysis of this impact in the boreal forest environment.

• Analyze the hydroclimatological changes associated with the impoundment of the La Grande Complex reservoirs in northern Québec with regard to water balance and energy balance variables at different spatial and temporal scales.

• Produce two climate simulations over a domain covering Quebec at a spatial resolution of 12 km, the first without the hydroelectric reservoirs—corresponding to land cover before flooding—and the second with the reservoirs, using the fifth generation of the Canadian Regional Climate Model (CRCM5) coupled with the FLake and CLASS V3.6 models (Canadian LAnd Surface Scheme);

• Validate the simulations with data from weather stations in the region and with Hydro-Québec data from the Eastmain-1 project;

• Evaluate the impacts of the reservoirs on air temperature, water and energy balance components as well as precipitation recycling.

The first stage of the project assessed the ability of CRCM5 to simulate historical climate in the La Grande River region in northern Québec. On comparison with the region’s 38 weather stations (Figure 1), the model shows a slight cold bias throughout the year which is more pronounced in spring. Simulated precipitation largely exceeds observations (by more than 48 %); under-measurement of snowfall at the weather stations is partly responsible for this overestimation but does not explain it entirely.

At the watershed scale, the model reproduces seasonal energy budget terms quite well, although an overestimation of upwelling shortwave radiation and sensible and latent heat fluxes is seen in April and May, due to a tendency to overestimate snow cover duration as reported in previous studies. However, these differences have little effect on the outcome of the analysis as they are present in both configurations (with and without reservoirs).

Figure 1. Outline of the Grande River watershed and its hydroelectric plants (black diamonds). The weather stations and Eastmain-1 eddy flux towers used for the study are represented by red circles and green triangles respectively.

In the second stage, a comparison of the two CRCM5 simulations provided an evaluation of the impacts the reservoirs. It turns out that the impacts of the impoundment of the reservoirs on temperature, the energy balance and evapotranspiration are limited to the grid cells that include the reservoirs. As shown in Figure 2, in summer, the presence of the reservoirs has a slight cooling effect (less than 0.5°C), as the decrease in maximum daily temperature (between –0.6°C and –1.3°C) is slightly greater that the increase in minimum temperature (between +0.5°C and +0.8°C). In winter, the reservoirs cause an increase in air temperature (between +0.4°C and +1.0°C) due to an increase in minimum daily temperature (between +1.2°C and +1.8°C).

The presence of the reservoirs also increases net shortwave radiation at the surface throughout the year, especially in the spring (11%) and summer (3.8%). During these seasons, sensible and latent heat fluxes are reduced or virtually unchanged. These transfers of surface heat increase in fall: sensible heat flux increases by 60% and latent heat flux increases by an average of 25% over the reservoirs. In winter, sensible heat flux and latent heat flux over the reservoirs increase by 210% and 36% respectively.

Figure 2. Modeled seasonal changes in daily mean temperature due to the presence of reservoirs, derived from the two CRCM5 simulations (with and without reservoirs). Model seasonal internal variability is represented by the grey bars.

An analysis of water cycle components at the watershed scale shows that the reservoirs have little impact on precipitation or seasonal runoff. Any slight changes can be attributed to the internal variability of CRCM5. Net evaporation at the La Grande hydroelectric complex is estimated at 1.7 m3 /GJ. In fact, while evaporation increases significantly at the reservoir level (an average of 47 mm/year), this increase is only 2 % (5.9 mm/year) at the watershed scale, given its large surface area.

Finally, the presence of the reservoirs does not influence the rate of evaporation recycling. In both CRCM5 simulations (with and without reservoirs), about 13 % of annual evaporation returns to the watershed as precipitation. Thus, of the 5.9 mm of additional evaporation due to the presence of a reservoir, only 3.6 mm will precipitate outside of the watershed in a year.

In summary, at the watershed scale, reservoir impoundment for hydroelectric power production does not significantly increase evapotranspiration in comparison to the pre-impoundment landscape.

Hydro-Québec Research Institute (IREQ)