, Volume 10, Issue 4, pp 536–549

Hyperspectral Remote Sensing of Canopy Biodiversity in Hawaiian Lowland Rainforests


  • Kimberly M. Carlson
    • Department of Global EcologyCarnegie Institution of Washington
    • Department of Global EcologyCarnegie Institution of Washington
  • R. Flint Hughes
    • Institute for Pacific Islands ForestryUSDA Forest Service
  • Rebecca Ostertag
    • Department of BiologyUniversity of Hawai’i
  • Roberta E. Martin
    • Department of Global EcologyCarnegie Institution of Washington

DOI: 10.1007/s10021-007-9041-z

Cite this article as:
Carlson, K.M., Asner, G.P., Hughes, R.F. et al. Ecosystems (2007) 10: 536. doi:10.1007/s10021-007-9041-z


Mapping biological diversity is a high priority for conservation research, management and policy development, but few studies have provided diversity data at high spatial resolution from remote sensing. We used airborne imaging spectroscopy to map woody vascular plant species richness in lowland tropical forest ecosystems in Hawai’i. Hyperspectral signatures spanning the 400–2,500 nm wavelength range acquired by the NASA Airborne Visible and Infrared Imaging Spectrometer (AVIRIS) were analyzed at 17 forest sites with species richness values ranging from 1 to 17 species per 0.1–0.3 ha. Spatial variation (range) in the shape of the AVIRIS spectra (derivative reflectance) in wavelength regions associated with upper-canopy pigments, water, and nitrogen content were well correlated with species richness across field sites. An analysis of leaf chlorophyll, water, and nitrogen content within and across species suggested that increasing spectral diversity was linked to increasing species richness by way of increasing biochemical diversity. A linear regression analysis showed that species richness was predicted by a combination of four biochemically-distinct wavelength observations centered at 530, 720, 1,201, and 1,523 nm (r2 = 0.85, p < 0.01). This relationship was used to map species richness at approximately 0.1 ha resolution in lowland forest reserves throughout the study region. Future remote sensing studies of biodiversity will benefit from explicitly connecting chemical and physical properties of the organisms to remotely sensed data.


AVIRISbiological diversityHawai’iimaging spectroscopyleaf pigments

Supplementary material

Copyright information

© Springer Science+Business Media, LLC 2007