Seasonal patterns in leaf reflectance red-edge characteristics

Leaves from ten tree species, including three conifers from a woodlot in southern Ontario were sampled at weekly intervals for a period of 150 days spanning the phenological events in deciduous trees of leaf development and expansion (flushing), leaf maturity and leaf senescence. The highly diverse seasonal red-edge reflection patterns were studied collectively and individually both from the perspective of long-term trends and relatively short-term or oscillating trends. The phenological events characteristic of deciduous trees were most effectively described using the red-edge position (λpg) and the chlorophyll-well position (λog ) derived from the inverted Gaussian model. Moreover, these parameters appeared consistent with what is known about the seasonal turnover of leaf chlorophyll and with other parameters R550 or Rog which more specifically quantify leaf chlorophyll. By these terms of reference chlorophyll declines considerably earlier in the season than the onset of other physiological or structural changes normally associated with senescence. Both parameters λPR and λog and the real red-edge position (λpr) conjointly provide specific information about each species and differentiate clearly deciduous from coniferous growth forms. Individual species variation in these and other variables was more ‘short term’ than ‘long term’, providing support for the perception that the long-term trends as visualized by λpg and λog are species-specific and can probably be correlated with ecological parameters such as tolerance. Short-term or oscillating variations were observed in all species most specifically in the shoulder reflectance (Rgr), the red-edge parameter (λpr ) and, to a lesser extent, in the inverted Gaussian model parameters, the wavelength at minimum reflectance (λog) and the average red-edge position (λpg). The generally synchronous behaviour of Rsr between all species, and of reflectance and spectral parameters within specific species, suggest the oscillations were not attributable to sampling error. Instead they appeared to be associated with rainfall and temperature events. Furthermore, systematic uncoupling of short-term variations between the spectral and the reflectance parameters observed in other species suggest different casual factors for spectral and reflectance parameters of the red edge.