abstract
- Direct measurements of evaporation from Great Slave Lake, Northwest Territories, Canada, over a six-year period are compared in terms of the influence of time scales on the magnitude and control of the process. Based on measurements using the eddy covariance method, both the latent and sensible heat fluxes were consistently small following ice break up, but increased dramatically as ice formation commenced. Half of the total evaporative water loss occurred over only 20% of the observation periods. Large evaporation events occurred over periods typically lasting nearly three days, and were controlled by the product of the horizontal wind speed and the difference between vapor pressure at the water surface and atmosphere. Both a stepwise linear regression and principal component analysis were used to determine the sensitivity of the relationship between evaporation and its controls to various sampling intervals. Varying the sampling interval from one to six hours in 1-hr steps had no significant effect on the coefficients used to predict the 24-hr evaporation totals. The dichotomy between lake and land surfaces in terms of temporal changes in heat source and sink, and the longer duration for processes occurring over the lake than land are discussed.