Evaluation of Spatial Productivity Patterns in an Annual Grassland during an AVIRIS Overflight

Evaluation of Spatial Productivity Patterns in an Annual Grassland During an AVIRIS Overflight*

John A. Gamon¹, Christopher B. Field² and Susan L. Ustin³ ¹California State University, Los Angeles 5151 State University Drive Los Angeles, CA 90032 ²Carnegie Institution of Washington Department of Plant Biology 290 Panama Street Stanford, CA 90035 ³Department of Land, Air and Water Resources University of California, Davis Davis, CA 95616

1. Background

In May, 1991, coincident with an AVIRIS overflight, we completed a ground-based study covering 9 hectares of an annual grassland. There were two goals to this ground study:

1) Obtain ecologically and physiologically meaningful data for relating AVIRIS images to canopy structure, biochemistry and physiology.

2) Evaluate the suitability of the 20-m AVIRIS pixel size for depicting detailed spatial patterns of productivity.

2. Materials And Methods

A 9-hectare annual grassland at Stanford University's Jasper Ridge Biological Preserve was sampled every 20 m for biomass, leaf area index (LAI), intercepted photosynthetically active radiation (PAR), chlorophyll, nitrogen, surface temperature and spectral reflectance. These measurements coincided with the May 14 and 15, 1991, AVIRIS overflights, providing a basis for biological interpretations of the AVIRIS images. Measurements of photosynthetic fluxes by eddy correlation were also completed during the overflight.

Three 20x20 m pixels were chosen for detailed ground sampling (every 5-m). The goal was to evaluate the 20-m AVIRIS pixel size by examining the semivariance in productivity estimates as a function of between-sample distance.

3. Results

Images derived from ground-based measurements at the 20-m scale depict the influence of topography, soil and vegetation type on productivity in the grassland (Fig. 1). Preliminary geostatistical tests of the ground data also indicate that the 20-m between-sample distance (comparable to the AVIRIS pixel size) is suitable for studying spatial productivity patterns in this relatively uniform grassland. Different results might have been obtained if the study had been conducted on adjacent chaparral and woodland, composed of complex mosaics of individual canopies 1-10 meters in diameter.

Good correlations were obtained between ground-based reflectance indices (SR and NDVI) and several ecologically or physiologically significant indicators of productivity, including biomass, LAI, canopy chlorophyll and nitrogen content (Fig. 2). Ground data will he compared with endmember and residual images derived from mixture models to estimate productivity patterns from AVIRIS images (Ustin et al., this volume).

Results from this study will be compared with AVIRIS images obtained at other dates and at other sites and will be compared to CO₂ flux measurements that were made concurrently with the overflight. These studies should assist in development and validation of ecological and physiological models for AVIRIS image interpretation.

4. Acknowledgments

This work was partly supported by the A. W. Mellon Foundation, and by the enthusiastic efforts of Nona Chiariello, Geeske Joel, Barbara Mortimer, Robert Rousseau and Riccardo Valentini.

1998, Center for Spatial Technologies and Remote Sensing (CSTARS)
University of California, Davis