SEMI EMPIRICAL-INDICES TO ASSESS 
CAROTENOIDS/CHLOROPHYLL a RATIO FROM LEAF SPECTRAL 
REFLECTANCE 
J. PENUELAS 1 , F. BARET 2 , I. FILELLA 1 
1 I.R.T.A. 
Carretera de Gabrils s/n 
Cabrils 08348, Barcelona (Spain). 
2 I.N.R.A. Bioclimatologie 
B.P. 91, 
84 143 Montfavet Cedex (France). 
ABSTRACT: 
The ratio between carotenoids and chlorophyll a concentrations (C c /C a ) is indicative of the physiology and 
phenology of plants. Reflectance R of a wide range of leaves from several species and conditions were measured 
with high spectral resolution spectroradiometers for 400 nm<A<800 nm. The performances of three pigment 
indices (i) simple ratio pigment index: SRPfcR^lR * 2 , (ii) normalized difference pigment index: NDPI =(R* J - 
R* 2 )I(R^- 1 +R*' 2 ), and (iii) the structure insensitive pigment index: SIPI=(R 800 -R* J )I(R 800 +R* 2 ) were tested to 
estimate the C c lC a pigment ratio. For each pigment index, every set of wavebands [Xj, X 2 ] was systematically 
tested. Good sets for SRPI and NDPI were observed in the blue-red domain [400 nm</^,<530 nm, 600 
nm<A 2 <700 nm] where both chlorophyll and carotenoid absorb concurrently in the blue, while only chlorophyll 
absorbs in the red. There were also good sets in the red edge [600 nm<^,<700 nm, 700 nm<X 2 <800 nm] where 
there is no absorption by carotenoids. This surprising result was explained by the correlation between 
chlorophyll a and carotenoid concentrations. However, the semi-empirical estimation of the C c IC a ratio is 
provided by SIPI in the blue-red domains [400 nm<X,<530 nm, 600 nm<X 2 <700 nm]: 
C c /C= 4.44-6.77 exp(-0.48 [R^PR 445 )!^ 800 ^ 680 )) 
This index minimizes the confounding effects of leaf surface and mesophyll structure. These reflectance pigment 
indices provide new insight in the use of remote sensing for the assessment of physiology and phenology of 
vegetation. 
KEY WORDS: Carotenoids and chlorophyll a, Reflectance indices, Vegetation assessment. 
1.INTRODUCTION 
The determination of photosynthetic pigments is one of the most frequently performed analysis in plant ecology 
and plant physiology. Pigment concentrations are routinely used as indicators of physiological status, detritus 
content and grazing process (Margalef,, 1974). The absolute and relative concentrations of these photosynthetic 
pigments is one of the drivers of leaf photosynthetic potential. Thus, changes in the pigment concentration 
would relate strongly to the physiological status of the plant and therefore its productivity. When plants are 
subjected to stresses, the change in pigment concentration may be small and follows variations of canopy 
structure. However, some severe stresses may result eventually to bleaching (Young and Britton,, 1990). 
Photosynthetic pigments are intimately associated to the biochemical reactions of photosynthesis. They also 
govern light transfer in the leaf and thus drive leaf optical properties such as reflectance or transmittance. Leaf 
optical properties are obviously very attractive for remote sensing of plant physiology and photosynthetic 
capacity (Field el al ., 1993). 
Chlorophyll estimates are widely used in remote sensing of marine ecosystems but very few 
applications were developed for terrestrial ecosystems. Penuelas et al. (1993a and b) show some potential use for 
environmental stress detection. Chlorophyll concentration is generally derived using leaf reflectance in the 
maximum absorption in the red at 675nm through empirical relationships (Benedict and Swidler, 1961; 
Wallihan, 1973; Hardwick and Baker, 1973; Macnicol et al ., 1976). Conversely, Thomas and Gausmann (1977) 
proposed to use reflectance at 550 nm because low reflectance at 675 nm increases the scatter of the relationship. 
Jacquemoud and Baret (1990) explained this apparent contradiction in terms of leaf reflectance sensitivity to 
341