The concept of light-use efficiency (LUE) is the underlying basis for estimating carbon exchange in many ecosystem models, especially those models that utilize remote sensing to constrain estimates of canopy photosynthesis. An understanding of the factors that control the efficiency with which forest canopies harvest available light to fix carbon via photosynthesis is therefore necessary for the development of useful production efficiency models. We present an analysis of observations of daily LUE for 2004 in a northern hardwood stand at the Bartlett Experimental Forest CO2 flux tower, White Mountains, New Hampshire (USA). We used eddy covariance measurements to estimate gross carbon exchange (GCE), and radiometric instruments mounted above and below the canopy to estimate the fraction of incident photosynthetically active radiation absorbed by the canopy (fAPAR). Both GCE and fAPAR show strong seasonal and day-to-day variability that contribute to temporal variation in LUE. During the middle of the growing season, when fAPAR is relatively constant, day-to-day variation in LUE is largely explained (r2 = 0.85) by changes in the ratio of diffuse to total downwelling radiation, but is not strongly correlated with any other measured meteorological variable.
We also calculated top-of-canopy NDVI based on measurements of reflected radiation at 400–700 and 305–2800 nm. Seasonal variation in this broadband NDVI paralleled that of the 500 m MODIS pixel containing the flux tower. The relationship between broadband NDVI and fAPAR is approximately linear during green-up, but non-linear during autumn senescence. This seasonal hysteresis has implications for the use of remote sensing indices (such as NDVI or EVI) in satellite estimation of fAPAR for production efficiency modeling.