Integration of vegetation indices and thermal measurements for ecosystem modeling.

Abstract

This study represents a continuation of the historical trend in the analysis of the relationships between climate and vegetation. The focus of the work is on an examination of the association of functioning plants and climatic processes and on the interaction of structured vegetation communities with their environment. Satellite observations and remote sensing technology provide new opportunities for studying the geo-biosphere at high temporal frequencies and for large geographic areas. Integration of spectral vegetation indices and thermal measurements from space is the prime methodology used in this study. The integration of thermal information with vegetation indices has the potential to result in an important contribution to ecological remote sensing. This study has investigated the triangular distribution pattern which has been repetitively observed in spectral domains defined by thermal measurements and vegetation indices. Results show that the triangular shape of an AVHRR data cluster can be consistently observed on different continents and at different times of the year. The triangular data pattern has also been observed using a seasonally averaged data set, representing the mean thermal and biomass conditions of a growing season. The triangular shape of the data spread in feature space clearly suggests an ecosystem triangle model for simulation of the global ecology. A closer analysis of 15 test sites representing the major biomes in the U.S. Southwest study area has shown that each biome has a stable territory in the two dimensional space of temperature and biomass. If the territory of each test site in the triangle can be viewed as a vegetation cell, the ecosystem as a whole may be modeled by using this cell structure. The ecosystem triangle model and the concept of the cell structure have been applied in a vegetation classification exercise. The significant improvement achieved in the vegetation classification supports the conclusion that the ecosystem triangle model is a reflection of surface biomes, and may be used as a tool to study the structure, organization, and function of the biosphere.