No clear trends could be observed and the correlation
varies from year to year. Except for 1996 when the R2 was
0.82, there were other years when the R2 was as low as 0.28
(1997) or even 0 in 2000. One possible reason is that the
NDVI becomes saturated at high LAI, so that above some
threshold, further increases in LAI do not alter the NDVI. In
some years, the LAI may increase beyond the threshold, in
other years this may not be the case. Thus, pooling NDVI
data over the years during this stage and without prior
examination of this effect would diminish the strength of the
LAI–NDVI relationship