Using satellite imagery to assess plant species richness: The role of multispectral systems
- 16 July 2007
- journal article
- Published by Wiley in Applied Vegetation Science
- Vol. 10 (3) , 325-331
Abstract
Question: The use of variations in the spectral responses of remotely sensed images was recently proposed as an indicator of plant species richness (Spectral Variation Hypothesis, SVH). In this paper we addressed the issue of the potential use of multispectral sensors by testing the hypothesis that only some of the bands recorded in a remotely sensed image contain information related to the variation in species richness.Location: Montepulciano Lake, central Italy.Methods: We assessed how data compression techniques, such as Principal Component Analysis (PCA), influence the relationship between spectral heterogeneity and species richness and evaluated which spectral interval is the most adequate for predicting species richness by means of linear regression analysis.Results: The original multispectral data set and the first two non‐standardized principal components can both be used as predictors of plant species richness (R2∼ 0.48; p < 0.001), confirming that PCA is an effective tool for compressing multi‐spectral data without loss of information. Using single spectral bands, the near infrared band explained 41% of variance in species richness (p < 0.01), while the visible wavelengths had much lower prediction powers.Conclusions: The potential of satellite data for estimating species richness is likely to be due to the near infrared bands, rather than to the visible bands, which share highly redundant information. Since optimal band selection for image processing is a crucial task and it will assume increasing importance with the growing availability of hyperspectral data, in this paper we suggest a ‘near infrared way’for assessing species richness directly from remotely sensed data.Keywords
This publication has 37 references indexed in Scilit:
- Mapping the species richness and composition of tropical forests from remotely sensed data with neural networksEcological Modelling, 2006
- PREDICTING WOODY-PLANT SPECIES RICHNESS IN TROPICAL DRY FORESTS: A CASE STUDY FROM SOUTH FLORIDA, USAEcological Applications, 2005
- Remote sensing of upland vegetation: the potential of high spatial resolution satellite sensorsGlobal Ecology and Biogeography, 2004
- From space to species: ecological applications for remote sensingTrends in Ecology & Evolution, 2003
- Species richness along multiple gradients: testing a general multivariate model in oak savannasOikos, 2003
- Radiometric normalization of multitemporal high-resolution satellite images with quality control for land cover change detectionRemote Sensing of Environment, 2002
- Quantitative tools for perfecting species listsEnvironmetrics, 2002
- Interannual variability of NDVI and species richness in KenyaInternational Journal of Remote Sensing, 2002
- Satellite imagery as a tool for monitoring species diversity: an assessmentJournal of Applied Ecology, 1999
- The influence of principal component analysis on the spatial structure of a multispectral datasetInternational Journal of Remote Sensing, 1999