Temperate deciduous forests exhibit dramatic seasonal changes in surface exchange properties following on the seasonal changes in leaf area index. Nearly continuous measurements of turbulent and radiative fluxes above and below the canopy of a red oak forest in central Massachusetts have been ongoing since the summer of 1991. Several seasonal trends are obvious. Global solar albedo and photosynthetically active radiation (PAR) albedo both are good indicators of the spring leaf emergence and autumnal defoliation of the canopy. The solar albedo decreases throughout the summer, a change attributed to decreasing near-infrared reflectance since the PAR reflectance remains the same. Biweekly satellite composite images in visible and near-infrared wavelengths confirm these trends. The thermal emissions from the canopy relative to the net radiation follow a separate trend with a maximum in the midsummer and minima in spring and fall. The thermal response number computed from the change in radiation temperature relative to the net radiation is directly related to the Bowen ratio or energy partition. The subcanopy space follows a different pattern dictated by the presence of the canopy; there the midday sensible heat flux is a maximum in spring and fall when the canopy is leafless, while subcanopy CO2 flux is maximum in midsummer. Subcanopy evapotranspiration did not have a distinct seaasonal peak in spring, summer, or fall. The temperature dependence of the respiration rate estimated from the eddy correlation subcanopy CO2 flux is comparable to that found using nocturnal flux measurements.
The surface energy balance follows a seasonal pattern in which the ratio of turbulent sensible heat flux to the net radiation (QH/Q*) is a maximum in the spring and fall (0.50.6), while the latent heat flux (QE) peaks in midsummer (QH/Q* = 0.5). This pattern gives rise to a parabolic growing season shape to the Bowen ratio with a minimum in early August. Growing season changes in the canopy resistance (Rc), related to the trends in the Bowen ratio, are more likely to be predicted using the thermal channels of remote sensing instruments than the shorter-wavelength bands.