234 
reflectance of the two species, the canopy cover data were analysed. The seasonal variations in percentage 
canopy' cover were calculated by' averaging the cover estimates derived from the four hemispherical 
photographs taken on each date at each tower. At the ash site canopy cover shows smooth variations, which are 
a function of the discontinuous structure of the tree canopy', and the "compound" nature of the ash leaves which 
develop gradually. At the beech site canopy cover variations are more abrupt, and correspond to the presence or 
absence of "simple" beech leaves on the continuous tree canopy, which appear in a synchronous bud-burst, and 
are lost in a simultaneous abscission. 
Figure 3. Seasonal variations in mean visible and near-infrared reflectance, and canopy cover. 
Figure 3 shows that at the ash site, seasonal fluctuations in visible reflectance are broadly the inverse of canopy 
cover, while for near-infrared this relationship is directly proportional, however, some important seasonal 
discrepancies exist. The downwards-looking canopy photographs show that between May and July, the 
herbaceous and shrub layers dev elop to produce a complete ground coverage of photosynthetically active 
foliage, which has a limited number of leaf layers and substantial intra-canopy shadow, and therefore does not 
cause a substantial increase in near-infrared reflectance, but is still capable of absorbing most visible 
irradiation, and producing low reflectance values. The dense herbaceous layer is not accounted for in the 
estimation of percentage canopy cover, as the camera was placed at 1.3m above ground level. Thus, percentage 
canopy cover rises slowly along with near-infrared reflectance, while visible reflectance decreases very rapidly. 
After a maximum in August, near-infrared reflectance begins to decrease, while canopy' cover remains high. 
This is caused by the die-back of the herbaceous and shrub layers, but this reduction in the number of leaf 
layers was not recorded as a decrease in percentage canopy cover. In addition, during October and November, 
as the tree leaves begin to senesce, their cell walls begin to disintegrate, thereby reducing the amount of cell- 
wall \ air interface which is largely responsible for the internal scattering of near-infrared radiation After 
November, visible reflectance follows the inverse of canopy cover, as the tree leaves fall to the woodland floor 
and exhibit a slow increase in visible reflectance as the remaining photosynthetic pigments are destroyed 
during decay. Similarly, near-infrared reflectance decreases quite rapidly during leaf abscission as the number 
of layers of leaves decrease, then further as the internal structure of the leaves is further broken down during 
decay. Between March and April, canopy cover begins to increase as the hazel understorey develops, producing