3. Annual change of NDVI 
Annual variation of NDVI diagram (Fig. 6) has three 
characteristic parts: steep rise during leafing and 
blooming, high values in vegetation period and slow 
decline during yellowing and falling of leaves. Local 
minima noticed during vegetation period are not 
simultaneous even in neighbouring woods (llovski dol, 
Opeke i Srnolov). This could be a mark of developed 
cumulus cloudiness typical for summer afternoons. 
Leafing is the most distinct phenological phase in the 
annual variation of NDVI. In this phase NDVI rises from 
low values, about 0.1, steeply to 0.4, with average 
gradient +15 mNVDI/day (MNDVI, one thousandth of 
Table 2 NDVI values on the phenological phases starting day in observed oak-woods. 
  
NDVI). Even local maximum values have no such steep 
rises. During the second half of vegetation period we 
noticed slow descent of NDVI values. Yellowing of 
leaves begins around 0.4 and has average decline of 
—4 mNDVi/day. These values are not so distinct and 
could be misinterpreted with some of local minima. 
Table 2 shows NDVI values for all observed woods on 
the day of phenological phases beginning. Exact values 
of NDVI at the phase appearance day are computed 
using bi-cubic spline interpolation. According to annual 
change of NDVI, reliable determination of leafing and 
yellowing is possible, but blooming and appearance of 
acorn have not distinct mark at the variation of NDVI. 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
   
   
    
  
  
; Appearance of Yellowing of 
First leafing First blooming acom leaves 
Spacva 0.269 0.460 0.568 0.344 
Cesma 0.353 0.384 0.408 0.428 
llovski dol 0.326 0.428 0.469 0.476 
Srnolov 0.290 0.329 0.427 0.312 
Opeke 0.291 0.384 0.453 0.397 
Average annual change of NDVI for all five oakwoods 
NDVI Y 
0.600 | 
0.500 
Average NDVI 
0.400 | 
0.300 | 
0.200 
0.100 Standard deviation of NDVI 
333 259 SEEEESSSSS SOS mq = es 3 
SSS SSESSSSSSS b oen = s 5 2 
s040245 Q00450520545 SEE 5 508 d 
  
  
  
Figure 6 Average annual change of NDVI computed from data of all five woods. 
Average annual change of NDVI is computed 
from all five oak-woods data. Acceptable low standard 
deviation shows that this is a good representative for 
observed oak-woods with the advantage of smoothing 
local extreme values. Regularities in the annual change 
of NDVI show that they could be used for leafing and 
yellowing begginings approximation for oak-woods far 
Intemational Archives of Photogrammetry and Remote Sensing. Vol. XXXII, Part 7, Budapest, 1998 
from phenological stations. There is a practical usage 
potential of this knowledge because oak-woods are 
wide spread in the lowlands of Croatia. NDVI changes 
for longer periods (few years) could produce data about 
development or degradation of wood communities 
caused by natural phenomena or human activities (acid 
rain, for example). 
753