3.2.2 Monthly Scale
Figure 11 displays monthly averaged turbulent heat fluxes over the whole study period. Latent heat fluxes always remained positive despite some condensation events in June (Sect. 3.2.1). LE peaked in September, when the heat storage reached its maximum (Fig. 13). It remained above 70 W m−2 from August to October.LE began to rise in July when stratification in the reservoir set in. It then sharply increased in August all the way to December, during which 84% of the annual evaporation occurred. This is consistent with the findings from Blanken et al. (2011) for Lake Superior, where 89% of the annual evaporation occurred when there was unstable atmospheric stratification. It is also consistent with the Rouse et al. (2003) study at Great Slave Lake, where 85% to 90% of evaporation occurred during a period when the atmosphere was unstable. However, in our study,LE decreased gradually due to the declining vapour pressure deficit that followed the decline in air temperature. In winter and spring, LE stayed below 20 W m−2 and reached its minimum in June, during the ice-free period (Fig. 6).
Sensible heat fluxes were negative from February to July, with values less than −10 W m−2 and reaching as low as −20 W m−2 when the reservoir was colder than the air. On average, values remained positive from August to January. In June, the reservoir was ice-free while the water column was under vernal transition, with upper layers that were not warm enough to stratify (i.e., below 3.98°C, the temperature of maximum water density). The water surface temperature stayed far below the air temperature, preventing H from becoming positive and enabling LE to become negative. Despite the high net radiation term around the summer solstice, the heat content of the water body only slowly rose because of the high specific heat and density of water. In summer and fall, the surface layers were warmer than the air above the water. Hincreased steadily until December, reaching 80 W m−2, and then decreased abruptly in January as the water surface froze. From May to August, net radiation (Rn ) was high, contributing mostly to the storage of heat in the reservoir (ΔHS ), while turbulent heat fluxes were relatively low. Nearly all the energy brought in byRn was used to increase the heat storage term. During the fall and early winter, Rn declined rapidly while LE and H increased, fuelled by the energy stored in the reservoir, firmly establishing the heat release period of the reservoir. We observed a three-month delay between the maximum summer net radiation that occurred in June and the maximum latent heat flux. The delay was six months between the peak net radiation and the maximum sensible heat flux. We also observed different delays between the maximum surface water temperature and the maximum LE andH , which were delayed by one and four months, respectively.