Abstract. The first direct measurements of evaporation from a large high-latitude lake, Great Slave Lake, Northwest Territories, Canada, were made using eddy covariance between July 24 and September 10, 1997, and June 22 and September 26, 1998. The main body of the lake was ice-free between June 20 and December 13, 1997, and June 1, 1998, and January 8, 1999, with the extended ice-free season in 1997-1998 coinciding with 4øC above normal air temperatures and an abnormally strong E1 Nifio. Measurements extending roughly 5.0 to 8.5 km across the lake were made from a small rock outcrop located near the main body of the lake. The lake was thermally stratified between midJuly and September, with the thermocline extending down to approximately 15 m. High winds were effective in mixing warm surface waters downward and, when accompanied by cold fronts, resulted in large, episodic evaporation events typically lasting 45 hours. The daily total evaporation was best described as a function of the product of the horizontal wind speed and vapor pressure difference between the water surface and atmosphere. Seasonally, the latent heat flux was initially negative (directed toward the surface) followed by a steady increase to positive values (directed away from the surface) shortly after ice breakup. The latent heat flux then remained positive for the remainder of the ice-free period, decreasing midsummer and then steadily increasing until freeze-up. The sensible heat flux was small and often negative most of the spring and summer yet switched to positive and began to increase in the early fall. Extrapolation of evaporation measurements for the entire ice-free periods gave totals of 386 and 485 mm in 1997 and 1998-1999, respectively.

Atmospheric gradient techniques were used to measure the net ecosystem exchange of CO2 for a subarctic sedge fen near Churchill, Manitoba, during the summer of 1994. This was the second driest and wannest summer since 1943. The mean daily temperature was 2°C above average, and the rainfall was 55% below normal. More than half of the rain fell after the main growth period. The fen landscape comprises hummocks and hollows. Equilibrium retention storage occurs at an average standing water depth of 80 mm above the hollow bottoms (water table reference is O). During the summer of 1994 the average water table position at −117 mm resided well below the zero equilibrium retention depth. Periodically this decreased to −265 mm, well below a 30‐year average depth of −70 mm. During the full summer period, measurements indicate that the fen was a source of CO2. Only during a relatively short period of most active photosynthesis in midseason was there a small net CO2 uptake. A deep and warm soil aerobic layer promoted a large respiration flux, and this exceeded the photosynthetic CO2 uptake of the stressed sedge community. Diurnally, changes in surface temperature and incident solar radiation can explain most of the changes in the net CO2 exchange. It is hypothesized that in 1994 photosynthesis was significantly decreased and the respiration loss enhanced by the hot, dry conditions. If this hypothesis is correct, by analogy, climate warming would need to be accompanied by a substantial rainfall increase to maintain a condition of net CO2 gain to this peatland.

No abstract

No abstract

This study details seasonal characteristics in the annual surface energy balance of upland and lowland tundra during the 1998-99 water year (Y2). It contrasts the results with the 1997-98 water year (Y1) and relates the findings to the climatic normals for the lower Mackenzie River basin region. Both years were much warmer than the long-term average, with Y1 being both warmer and wetter than Y2. Six seasons are defined as early winter, midwinter, late winter, spring, summer, and fall. The most rapid changes in the surface energy balance occur in spring, fall, and late winter. Of these, spring is the most dynamic, and there is distinct asymmetry between rates of change in spring and those in fall. Rates of change of potential insolation (extraterrestrial solar radiation) in late winter, spring, and fall are within 10% of one another, being highest in late winter and smallest in spring. Rates of change in air temperature and ground temperature are twice as large in spring as in fall and late winter, when they are about the same. Rates of change in components of the energy balance in spring are twice and 4 times as large as in fall and late winter, respectively. The timing of snowpack ripening and snowmelt is the major agent determining the magnitude of asymmetry between fall and spring. This timing is a result of interaction between the solar cycle, air temperature, and snowpack longevity. Based on evidence from this study, potential surface responses to a 1ЊC increase in air temperature are small to moderate in most seasons, but are large in spring when increases range from 7% to 10% of average surface energy fluxes.