This paper summarizes the results of eddy covariance measurements made over the last four years above a boreal aspen (Populus tremuloidesMichx.) forest in Prince Albert National Park, Saskatchewan. This site, which was one of the ten BOREAS tower flux sites, is one of the Canadian Boreal Ecosystem Research and Monitoring Sites (BERMS) and part of the Ameriflux Network. In 1994, the stand was 70 years old and the trees were 21 m tall. The understory is mainly hazelnut (Corylus cornuta Marsh) about 2 m high. Half-hour fluxes were measured from October 10 to November 15, 1993, February 2 to September 20, 1994 and April 20, 1996 to present. Eddy covariance instrumentation consists of a 3-axis sonic anemometer (Kaijo TR-61B (1993 and 1994) and Gill Instruments R2/3 (1996 to present)) and a temperature-controlled infrared gas analyzer (LI-COR Inc. 6262) with a heated sampling tube 3-4 m long. Measurements are made at the 39-m height on a 37-m tall scaffold tower. Fluxes are calculated at the site at the end of each half hour. All raw data are saved. Since 1996, the eddy covariance system has operated automatically, which includes once daily calibration and transfer of fluxes via cell phone/modem to our laboratory.
So far in this study, soil moisture deficiency has not significantly affected carbon sequestration or water use. Both, however, were strongly affected by leaf emergence date. This was demonstrated by the results in 1996, in which mean air temperature during March-May was 4.8 degrees C less than in 1994, resulting in leaf emergence being about 3 weeks later than in 1994. This was the main reason that net ecosystem productivity in 1996 was 30% less than in 1994.
A rectangular hyperbolic function best described the relationship between gross ecosystem photosynthesis and incident photosynthetic photon flux. Photosynthetic capacity was highest in 1994 probably because of above-average growing conditions. Annual photosynthetic energy conversion efficiency was 2.3-2.5%. Ecosystem respiration, estimated from nighttime CO2 flux measurements, was closely related to soil temperature at the 2-cm depth and the relationship was very similar for the three years. There is, however, some uncertainty in this relationship because eddy fluxes on calm nights were lower than expected respiration rates. This is either the result of temporary storage of CO2 in soil and snow or CO2 removal by mass flow associated with cold air drainage. Consequently annual carbon sequestration was estimated by either (i)assuming nighttime flux measurements were correct or (ii) replacing low nighttime fluxes with estimates using the relationship between fluxes measured during high nighttime windspeeds and soil temperature. Values of annual carbon sequestration estimated using the first approach were 200, 140 and 160 g C (carbon) per square meter for 1994, 1996 and 1997, respectively. The corresponding values using the second approach were 100, 20 and 80 g C per square meter. This emphasizes the importance of resolving the uncertainty in fluxes during low windspeed conditions at night. Despite the uncertainty, these figures strongly suggest that this forest is a moderate or, at least, a weak carbon sink.