The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B7. Beijing 2008 
is more efficient than the traditional method of using original 
spectra reflectance to explore the change of reflection curve. 
The relationship between REP and chlorophyll content will be 
studied further in the following section. 
Wavelength (nm) 
Figure 1. The reflectance spectrum and derivative reflectance 
spectrum of all species mean spectrum reflectation 
3.2 Relationship Between REP and Chlorophyll Content 
In this study, in order to explore the capability of using REP to 
estimate the chlorophyll content of leaves, the correlation 
coefficient related with red edge inflection point, REP and 
chlorophyll, and carotenoid of each species was analyzed as 
shown in Table 1. 
Table 1. Regression equations relating the parameters of red 
edge of leaves to their pigment contents for Daphniphyllum 
glaucescen, Michelia compressa, Illicium dunnianum and 
Machilus kusanoi 
Trees 
Regression equations 
R 2 
Daphniphyllum 
glaucescen 
Chi = 98.584X.red -66761.259 
Car = 42.565X.red -28546.904 
0.508** 
0.696** 
Michelia compressa 
Chi = 215.619Xred -149796.414 
Car = 74.906Xred -51566.680 
0.666** 
0.617** 
Illicium dunnianum 
Chi = 147.675Xred -102017.431 
Car = 66.055Xred -45729.653 
0.503** 
0.669** 
Machilus kusanoi 
Chl = 113.183X.red -77126.258 
Car = 42.886Xred -29021.724 
0.774** 
0.770** 
All species 
Chi = 133.989Xred -91865.439 
Car = 49.718Xred -33753.059 
0.577** 
0.541** 
**,P<0.01 
From Table 1, there is a good correlation between the red edge 
parameter and the total contents of chlorophyll and carotenoid, 
and have reached at 0.01 significant level; while on the aspect 
of the correlation between REP and the total chlorophyll 
content, Michelia formosanaand Illicium dunnianum reached a 
great level. 
the location of red edge comparatively near longer wavelengths 
range is an index that shows healthy growing. 
Based on these preceding results, the samples of the 4 species 
were comprehensively studied for the relationship between REP 
and chlorophyll content and carotenoid, and are illustrated in 
Fig. 2. From Fig. 2, correlation between some indices was 
obvious not only in the analysis of a single species, but also in 
the mixed samples of 4 species. From the above results a 
conclusion may be drawn that whether an analysis is conducted 
on a single species or a group of tree species, chlorophyll and 
carotenoid content can be calculated with their linear 
relationship with REP; REP can faithfully reflect the thickness 
of pigment in plants’ leaves. These results agree with the 
studies of most reports, although many researches were about 
crops (Dalezios et al., 2001; Pattey et al., 2001; Thenkabail et 
al., 2000). 
Figure 2. Relationship between leaf red edge inflection point, 
chlorophyll content and carotenoid content of mixed species. 
Pigments extracted with acetone from Daphniphyllum 
glaucescen, Michelia compressa, Illicium dunnianum and 
Machilus kusanoi 
The REP shift to shorter wavelengths due to pigment 
degradation, as expected and previously reported (Horler et al., 
1983; Vogelmann et al., 1993). Despite these described may 
change in the canopy scale (due to the canopy structural 
change), it can be seen that the REP indicator of pigment 
content could be detected from the experiment of REP value in 
leaf scale. And there were many suitable simulate models for 
broad leaf canopies to link leaf optical properties with canopy 
structural characteristics through radiative transfer modeling 
(e.g., PROSPECT model, SAIL model, LIBERTY model, 
BOREAS model). The next section describes the results 
obtained that the relationship between several vegetation 
indices calculated from the leaf spectral reflectation and leaf 
pigment content measurements in different terrains. 
From the above results we understand that the average REP 
always moved towards longer wavelength with the increase of 
chlorophyll content, which is called “red shift”; the REPs of all 
species were in the range of 698 nm-723.2 nm; from the view 
angle of plant’s physiological condition, when the leaves are 
senescing or when they are under adverse environmental 
conditions, their chlorophyll contents may fall gradually, and 
the location of the red edge tent to move towards shorter 
wavelength, which is called “blue shift”. This result shows that 
3.3 Vegetation Index and Chlorophyll Content 
This study analyzed the relationship between pigment content 
and NDVI, mNDVI, SR, and mSR; the results of a single tree 
species are in Fig. 3. From the results we understand that the 
correlation between 2 modified indices such as mNDVI and 
mSRand chlorophyll content were much better than the 
ordinary vegetation indices i.e. NDVI and SR. After the factor 
of different leaf structures is taken into calculation according to 
the calculation principle for modified vegetation indices, the 
233