AbstTactWater vapour and CO2 fluxes were measured using the eddy correlation method above and below the overstorey of a 21-m tall aspen stand in the boreal forest of central Saskatchewan as part of the Boreal Ecosystem-Atmosphere Study (BOREAS). Measurements were made at the 39.5-m and 4-m heights using 3-dimensional sonic anemometers (Kaijo-Denki and Solent, respectively) and closed-path gas analysers (LI-COR 6262) with 6-m and 4.7-m long heated sampling tubing, respectively. Continuous measurements were made from early October to mid-November 1993 and from early February to lateSeptember 1994. Soil CO2 flux (respiration) was measured using a LI-COR 6000-09 soil chamber and soil evaporation was measured using lysimetry.The leaf area index of the aspen and hazelnut understorey reached 1.8 and 3.3, respectively. The maximum daily evapotranspiration (£) rate was 5-6 mm d^^ Following leaf-out the hazelnut and soil accounted for 22% of the forest £. The estimated total £ was 403 mm for 1994. About 88% of the precipitation in 1994 was lost as evapotranspiration.During the growing season, the magnitude of half-hourly eddy fluxes of CO2 from the atmosphere into the forest reached 1.2 mg CO2 m'^ s^* (33 |imol C m~^ s"^) during the daytime. Downward eddy fluxes at the 4-m height were observed when the hazelnut was growing rapidly in June and July. Under well-ventilated night-time conditions, the eddy fluxes of CO2 above the aspen and hazelnut, corrected for canopy storage, increased exponentially with soil temperature at the 2-cm depth. Estimates of daytime respiration rates using these relationships agreed well with soil chamber measurements. During the 1994 growing season, the cumulative net ecosystem exchange (N££) was -3.5 t C ha"^ y"^ (a net gain by the system). For 1994, cumulative NEE, ecosystem respiration (K) and gross ecosystem photosynthesis (GEP = R-NEE) were estimated to be -1.3, 8.9 and 10.2 t C ha"^ y~^, respectively. Gross photosynthesis of the hazelnut was 32% of GEP.

Both species showed a decrease in canopy conductance as the saturation deficit increased and both showed an increase in canopy conductance as the photosynthetic active radiation increased. There was a linear relationship between forest leaf area index and forest canopy conductance. The timing, duration, and maximum leaf area of this deciduous boreal forest was found to be an important control on transpiration at both levels of the canopy. The full-leaf hazelnut daytime mean Priestley and Taylor [1972] a coefficient of 1.22 indicated transpiration was largely energy controlled and the quantity of energy received at the hazelnut surface was a function of aspen leaf area. The full-leaf aspen daytime mean c• of 0.91 indicated some stomatal control on transpiration, with a directly proportional relationship b6tWeen forest leaf area and forest canopy conductance, varying c• during much of the season through a range very sensitive to regional scale transpiration and surface-convective boundary layer feedbacks. IntroductionThe boreal forest represents one of the world's largest yet least understood ecosystems. Of the estimated 48.5 million km 2 total land area of the world's forests, 12.0 (25%) is covered by boreal forest, second only to the 17.0 (35%) covered by tropical rain forest. Boreal forest net primary productivity (ex---1 pressed as dry matter) accounts for an estimated 9.6 Gt yr (13%) of the world's forests 73.9, exceeding both temperate deciduous 8.4 (11%) and temperate coniferous 6.5 (9%) forests [Salisbury and Ross, 1978].•University of British Columbia, Vancouver, Canada. 2Atmospheric Environment Service, Downsview, Ontario, Canada.•Yale University, New Haven, Connecticut. Continuing with the work presented by Black et al. [1996], the objectives of this paper are (1) to describe the diurnal and seasonal patterns of the aspen overstory and hazelnut understory energy balance, (2) to describe the diurnal and seasonal patterns of canopy water vapor conductances for both the aspen and the hazelnut, and (3) to relate the canopy conductances to the ambient meteorological conditions at both the canopy and the regional levels. Site DescriptionThe study site (

Abstract.A boreal deciduous forest in Saskatchewan, Canada, sequestered 144i65, 80i60, 116-t-35 and 290-t-50 g C m -2 y-• in 1994, 1996, 1997 and 1998, respectively. The increased carbon sequestration was the result of a warmer spring and earlier leaf emergence, which significantly increased ecosystem photosynthesis, but had little effect on respiration. The high carbon sequestration in 1998 was coincident with one of the strongest E1 Nifio events of this century, and is considered a significant and unexpected benefit.

To evaluate the carbon budget of a boreal deciduous forest, we measured CO2 fluxes using the eddy covariance technique above an old aspen (OA) forest in Prince Albert National Park, Saskatchewan, Canada, in 1994 and 1996 as part of the Boreal Ecosystem‐Atmosphere Study (BOREAS). We found that the OA forest is a strong carbon sink sequestering 200 ± 30 and 130 ± 30 g C m–2 y–1 in 1994 and 1996, respectively. These measurements were 16–45% lower than an inventory result that the mean carbon increment was about 240 g C m–2 y–1 between 1919 and 1994, mainly due to the advanced age of the stand at the time of eddy covariance measurements. Assuming these rates to be representative of Canadian boreal deciduous forests (area ≈ 3 × 105 km2), it is likely they can sequester 40–60 Tg C y–1, which is 2–3% of the missing global carbon sink. The difference in carbon sequestration by the OA forest between 1994 and 1996 was mainly caused by the difference in leaf emergence date. The monthly mean air temperature during March–May 1994, was 4.8 °C higher than in 1996, resulting in leaf emergence being 18–24 days earlier in 1994 than 1996. The warm spring and early leaf emergence in 1994 enabled the aspen forest to exploit the long days and high solar irradiance of mid‐to‐late spring. In contrast, the 1996 OA growing season included only 32 days before the summer solstice. The earlier leaf emergence in 1994 resulted 16% more absorbed photosynthetically active radiation and a 90 g C m–2 y–1 increase in photosynthesis than 1996. The concomitant increase in respiration in the warmer year (1994) was only 20 g C m–2 y–1. These results show that an important control on carbon sequestration by boreal deciduous forests is spring temperature, via the influence of air temperature on the timing of leaf emergence.

Abstmct. Observations were made of turbulence in an extensive deciduous forest on level terrain using a vertical array of seven three-dimensional sonic anemometer/thermometers within and above the canopy. Data were collected through the period of leaf fall and over a range of thermal stabilities. A bulk canopy drag coefficient was nearly independent of the density of the forest but decreased greatly with the onset of nocturnal stability. The depth of penetration of momentum' into the forest increased with leaf fall but, again, was greatly curtailed by stable conditions. Turbulent velocities decreased with increasing depth in the forest but relative turbulence intensities increased to midcanopy levels. Leaf density influenced turbulence levels but not as strongly as did thermal stability. Thermal effects were adequately described by the single parameter h/L, where h is the canopy height and L is the Monin-Obukhov' length. The longitudinal and vertical velocity correlation coefficient was larger in magnitude than expected in the upper layers of the forest but decreased to a small value in the lowest layers where the Reynolds stress was small. The ratio u,,,/u*, where u* is the local friction velocity, reflected changes in the uw correlation, becoming smaller than usual in the upper canopy layers. It is believed that these effects result from the intermittent, spatially coherent structures that are responsible for a large fraction of the momentum flux to the forest.