Hyperspectral remote sensing tools for quantifying plant litter and invasive species in arid ecosystems

By: , and 
Edited by: Prasad S. ThenkabailAlfredo Huete, and John G. Lyon



Green vegetation can be distinguished using visible and infrared multi-band and hyperspectral remote sensing methods. The problem has been in identifying and distinguishing the non-photosynthetically active radiation (PAR) landscape components, such as litter and soils, and from green vegetation. Additionally, distinguishing different species of green vegetation is challenging using the relatively few bands available on most satellite sensors. This chapter focuses on hyperspectral remote sensing characteristics that aim to distinguish between green vegetation, soil, and litter (or senescent vegetation). Quantifying litter by remote sensing methods is important in constructing carbon budgets of natural and agricultural ecosystems. Distinguishing between plant types is important in tracking the spread of invasive species. Green leaves of different species usually have similar spectra, making it difficult to distinguish between species. However, in this chapter we show that phenological differences between species can be used to detect some invasive species by their distinct patterns of greening and dormancy over an annual cycle based on hyperspectral data. Both applications require methods to quantify the non-green cellulosic fractions of plant tissues by remote sensing even in the presence of soil and green plant cover. We explore these methods and offer three case studies. The first concerns distinguishing surface litter from soil using the Cellulose Absorption Index (CAI), as applied to no-till farming practices where plant litter is left on the soil after harvest. The second involves using different band combinations to distinguish invasive saltcedar from agricultural and native riparian plants on the Lower Colorado River. The third illustrates the use of the CAI and NDVI in time-series analyses to distinguish between invasive buffelgrass and native plants in a desert environment in Arizona. Together the results show how hyperspectral imagery can be applied to solve problems that are not amendable to solution by the simple band combinations normally used in remote sensing.

Additional publication details

Publication type Book chapter
Publication Subtype Book Chapter
Title Hyperspectral remote sensing tools for quantifying plant litter and invasive species in arid ecosystems
DOI 10.1201/b11222-23
Year Published 2012
Language English
Publisher CRC Press
Publisher location Boca Raton, FL
Contributing office(s) Southwest Biological Science Center
Description 34 p.
Larger Work Type Book
Larger Work Title Hyperspectral remote sensing of vegetation
First page 361
Last page 394
Online Only (Y/N) N
Additional Online Files (Y/N) N