ESTIMATION OF FOREST LEAF-AREA INDEX USING REMOTE-SENSING AND GIS DATA FOR MODELING NET PRIMARY PRODUCTION

Ecosystem models can be used to estimate potential net primary product ion (pNPP) using GIS data, and remote sensing input of actual forest l eaf area to such models can provide estimates of current actual net pr imary, production (aNPP). Comparisons of pNPP and aNPP for a given sit e or regional landscape can be used to identify forest stands for diff erent management treatments, and may provide new information on wildli fe habitat, forest diversity and growth characteristics. Leaf area est imates may be obtained from satellite imagery through correlation with physiologically-based vegetation indices such as the Normalized Diffe rence Vegetation Index (NDVI). However, in areas with high Leaf Area I ndex (LAI), vegetation indices usually saturate at leaf areas greater than about 4. In predominantly deciduous(hardwood) and mixedwood stand s remote sensing estimates may be influenced by understory and other f actors. We examined digital Landsat TM imagery and GIS data in the Fun dy Model Forest of southeastern New Brunswick to determine relations t o forest leaf area index within different stand structures or covertyp es. The image data were stratified using GIS covertype information pri or to development of LAI predictive equations using spectral reflectan ce, and the prediction of LAI from Landsat TM imagery was improved wit h reference to estimates of stem density which are standard forest inv entory information contained in GIS databases. Actual stand LAI was co mpared to assumed maximum LAI values for several species and sites usi ng an ecosystem process model (BIOME-BGC) which relies on climate, soi ls and topographic information also obtained from the GIS. Subsequent comparison of pNPP and aNPP revealed that even disturbed sites in this environment can reach close to maximum site potential. Specific sites with suboptimal species composition were identified. A future refinem ent of this approach is to classify the imagery independently of the G IS, which assumes a homogeneous covertype for each polygon in the syst em, and thus improve still further the aNPP estimates through higher c overtype and LAI estimation accuracy.