A GLOBAL TRIANGLE MODEL 
FOR CHARACTERIZING VEGETATION 
IN RELATION TO ITS ECOLOGICAL ENVIRONMENT 
Li Jiang 
(Arizona Remote Sensing Center, University of Arizona, Tucson, AZ) 
Abstract 
This study is to propose a global triangle model based on the characteristics 
of data structure observed in the thermal and NDVI spectral spaces of NOAA- 
AVHRR data. The development of the global triangle model is an attempt to study 
vegetation in relation to its ecological environment, and to understand the Earth 
as an integrated system. — 
A scatter plot was produced with data from channel 2 and 4 of the NOAA- 
AVHRR scene of January 3, 1989, (Figure 1). Figure 1 is a "global view" of the 
West Africa in the near and thermal infrared (NIR and TIR) space. The view 
observed looks like a huge triangle containing a variety of plants, soils, and land 
surface covers. Similar triangle shape has been observed in five AVHRR data-sets 
of West Africa acquired on January 3 and 23, February 11 and 22, and march 23, 
1989. This same phenomena was observed when scenes were cut in half but still 
contained the various African landscapes. The triangle was also observed using 
AVHRR bands 2 and 5 in a similar fashion as observed using bands 2 and 4, since 
bands 4 and 5 are both thermal channels performing in adjoining spectral ranges 
of 10.5 - 11.5 pm, and 11.5 - 12.5 pm respectively. Four small sites(ocean, bare soil, 
dry grassland, and dense vegetation) selected based on visual interpretation of the 
image and NDVI value were used to make separate scattergraphs as illustrated 
in Figure 1 and 2. Soils occur at the bottom left corner of the triangle where 
minimum vegetation and high temperature are usual the case. The vegetation 
points lay at the top of the triangle where the pixels are characterized by high 
vegetation and relative low temperature. All water pixels crowd around the 
bottom right comer of the triangle where temperature and vegetation are both 
minimal. If the near-infrared band is replaced with a NDVI band, the 
scattergraph still looks like a triangle, but changes in shape and orientation 
become apparent. It can be seen from Figure 2, the global triangle is rotated 
clockwise in the NDVI and band 4 spectral plane. The high vegetation corner is 
tilted towards lower temperature regions, which is often true because of the 
cooling effect of green vegetation. The desert corner is lifted a little above the water 
comer since some vegetation, more or less, is always found in desert areas. The 
basic relationship and distribution of the triangular structure stay the same, and 
the thermal and NDVI model looks more reasonable than the triangle observed in 
the TIR and NIR space. A possible explanation for this improvement is that the 
NDVI is a better vegetation indicator than the near-infrared band alone. 
The DN values of deep ocean from the thermal bands were found to be quite 
constant both spatially and temporally, and may be used as a reference for 
comparison across time and space for vegetation and terrain analysis. DN values 
of desert areas vary greatly from day to day and from day to night, (Figure 4 and