Wednesday, August 6, 2008 - 1:30 PM

COS 76-1: Airborne analysis of landscape-level canopy structure and biogeochemistry along a substrate age gradient across the Hawaiian Islands

Peter M. Vitousek, Stanford University, Gregory P. Asner, Carnegie Institution, Oliver A. Chadwick, University of California, and Sara C. Hotchkiss, University of Wisconsin.

Background/Question/Methods

We determined canopy heights and nitrogen concentrations in 2 by 2 km landscapes centered on four long-term research sites arrayed along a substrate age gradient across the Hawaiian Islands, asking: How well do the sites represent the landscapes in which they are embedded?  How does canopy structure and chemistry on the landscape scale vary, what controls that variation, and how does within-landscape variation compare with differences observed along the substrate age gradient?  We used airborne wave-form LiDAR to determine canopy heights, and the Airborne Visible and Infrared Imaging Spectrometer (AVIRIS) to measure canopy nitrogen; both were flown as part of the Carnegie Airborne Observatory.  Canopy properties were analyzed for a circle 50 m in radius focused on each long-term site, for the entire 2 x 2 km landscape regardless of land cover, and (after stratification) for native-dominated vegetation on constructional geomorphic surfaces within each landscape. 

Results/Conclusions

At all three scales, differences among the landscapes were consistent with key features of earlier ground observations. For example, airborne observations showed that native forests on the two intermediate-aged landscapes have significantly higher canopy N concentrations (1.26% and 1.23% N) than the youngest (0.97 % N) or oldest landscapes (0.93% N).   Also, the two younger landscapes have taller canopies – averaging 12.2 m and 15.4 m, versus 4.8 m and 4.0 m in the two older landscapes.   However, synoptic analyses also demonstrated substantial variability in canopy characteristics across these landscapes, only some of which is associated with recognized causes such as erosion/topography, biological invasion, pastures, or forest plantations.  Canopy heights become progressively more heterogeneous as well as shorter in native-dominated communities on the older landscapes; the coefficient of variation for canopy heights on constructional geomorphic surfaces increases progressively from 23% to 41% to 69% to 78%, from the youngest to the oldest landscape.  Disturbance-succession dynamics and an increase in the area dominated by the mat-forming fern Dicranopteris linearis contribute to these patterns, in addition to the progressive changes in soil fertility and hydrology that have been the focus of past site-based research on this gradient.