Ecosystem structure along bioclimatic gradients in Hawaii from imaging spectroscopy

摘要

The Hawaiian Islands contain more than two-thirds of the life zones delineated by Holdridge, L.R., 1947. Determination of world plant formations from simple climate data. Science, 105, 367-368, and is thus an ideal testing ground for remote sensing studies of ecosystem function and structure. We tested the generality of imaging spectroscopy with "tied" red-edge and shortwave-infrared (RESWIR2) spectral mixture modeling for automated analysis of the lateral distribution of plant tissues and bare substrate across diverse bioclimatic gradients in Hawai'i. Unique quantities of the fractional cover of photosynthetic and non-photosynthetic vegetation (PV, NPV) and bare substrate identified fundamental differences in ecosystem Structure across life zones. There was a similar to 20-fold increase in fractional PV cover with a 10-fold increase in inean annual precipitation (<= 250-2000 mm year(-1)). This rate of increase diminished from 2000 to 3000 mm year(-1) of rainfall, suggesting that photosynthetic canopy cover may be limited by water saturation at 3000 mm year(-1). The amount of exposed surface senescent material (NPV) remained nearly constant at similar to 50% in ecosystems with a mean annual precipitation < 1500 mm year(-1). Thereafter, NPV steadily declined to a minimum of similar to 20% at 3000 mm year(-1) of rainfall. Bare substrate fractions were highest (similar to 50%) at precipitation levels < 750 mm year(-1), then declined to < 20% in the 750-1000 mm year(-1) zones. The combination of low bare substrate and high NPV cover in the 750-1000 mm year-1 rainfall zones identified these areas as high fire risk. Remotely sensed fractional cover of PV+NPV was poorly correlated with canopy leaf area index (LAI), showing the uniqueness of the lateral structural measurement afforded by automated RE-SWIR2 spectroscopy approaches. The results indicate the accuracy, precision and applicability of imaging spectroscopy for ecological research across a wide range of bioclimatic conditions. (c) 2005 Elsevier Inc. All rights reserved.