Annual and seasonal variability of sensible and latent heat fluxes above a coastal Douglas-fir forest, British Columbia, Canada

Two years of continuous eddy covariance measurements were used to characterize the seasonal and annual variability of the latent and sensible heat fluxes above a 50-year-old, 33 m tall coastal Douglas-fir forest on the east coast of Vancouver Island, Canada. The total annual evaporation was found to be very conservative for this temperate coniferous rainforest despite variability in weather between the years (432 mm in 1998 and 435 turn in 1999). Winter evaporation was a significant component of the annual total, on average 27% of the mean 434 turn per year. Seasonal variations in the magnitude and direction of the sensible heat flux above the canopy were linked to changes in the surface conductance to water vapour transfer. The wet canopy tended to act as a sink for sensible heat, especially throughout the winter months, resulting in an average daily 24 h Bowen ratio of - 1.7. This contrasted dramatically with summer daytime turbulent exchange, which was usually dominated by upward sensible heat flux during the summer months (April-September; inclusive). The total 24 h Bowen ratio for the summer was 1.1, with daily 24 h values reaching a maximum of 3.1 for dry-canopy conditions. The average rate of transpiration was 1.7 mm per day reaching a maximum of 3.7 mm day, while canopy conductance ranged from 1 to 30 mm s(-1). Although the total winter and summer evaporation was similar between years, differences in the timing of the maximum evaporation rates and flux partitioning patterns resulted in considerable variability within the seasons. The cooler, cloudier and wetter weather of 1999 maintained relatively low evaporation rates and low Bowen ratios through the entire summer, while clear skies and hot and dry conditions in 1998 resulted in greater evaporation until mid-July. After this time, drought and a decrease in canopy conductance reduced evaporation and led to a daily daytime Bowen ratio as high as 3.8. (C) 2002 Elsevier Science B.V. All rights reserved.