MEASUREMENT AND REMOTE-SENSING OF LAI IN ROCKY-MOUNTAIN MONTANE ECOSYSTEMS

We estimated leaf area index (LAI) for Glacier National Park, Montana, U.S.A., by using various methods to measure LAI in the field and corr elating these values with Landsat Thematic Mapper data. Ground-based L AI were estimated using sapwood-LAI allometric equations and optical i nstruments, including the LAI-2000 and a Decagon ceptometer. Optical-b ased LAI estimates contain nonrandom self-shading therefore, allometri c and optical LAIs were compared to calculate coefficients to correct optical LAI data within similar vegetation types and canopy structural conditions. Least-squares regression models were constructed from poo led the ground-based allometric and corrected optical LAI values and f rom Landsat Thematic Mapper vegetation indices. Average LAI and satell ite indices for defined slope, aspect and elevation classes were used in the model calculation, as point estimates were generally poor. The normalized difference vegetation index and a mid-infrared corrected si mple ratio had the ''best fit'' with field LAI values. We applied thes e two models to the Thematic Mapper indices and tested LAI estimation with independent field LAI data. In addition, we tested the effect of spatial resolution on satellite-estimated LAI values by averaging the Thematic Mapper data into 250 x 250 m grid cells (pixels). Our results showed that the normalized difference vegetation index provided the b est estimate of LAI and decreased in accuracy with coarser pixels. The corrected simple ratio index overestimated LAI largely because of dif ficulty deriving the appropriate reflectance scale of mid-infrared cor rection to apply to this index at the larger landscape scale investiga ted here. However, mid-infrared correction of the Thematic Mapper indi ces was a good indicator of understory canopy cover.