Multispectral imaging of vegetation at Biosphere 2

Abstract

Remote sensing plays a key role in the study of Earth’s biological processes on a global scale and can also be used in experimental settings. Here we report initial results of a pilot project to explore the potential of remote sensing to measure and monitor vegetation within the thornscrub mesocosm of the Biosphere 2 experimental facility in Oracle, AZ, USA. Goals of the project were to investigate methods for mapping the spatial distribution of diverse species within the facility and to monitor vegetation changes over time. We show that the total cross-sectional area of vegetation within a scene can be estimated from digital visible and near-infrared images obtained from simple and inexpensive camera systems. Plant species can be distinguished on the basis of morphology and appearance in false-color images, allowing estimation of the fractional cross-sectional area occupied by each plant type. Changes in cross-sectional areas can be used to monitor temporal variations such as those caused by growth or senescence, possibly providing an alternative to destructive testing techniques. Practical considerations complicate the derivation of biologically useful information from remotely sensed data inside Biosphere 2. Future research will focus on the relationship between the cross-sectional areas derived from image analysis and important parameters such as Leaf-Area-Index (LAI) and growth rate.