HR: 1340h
AN: B23C-1491    [Abstracts]
TI: Estimating Fractional Cover of Photosynthetic Vegetation, Non-Photosynthetic Vegetation and Soil in Savannas Using the EO-1 Hyperion and MODIS Sensors
AU: GUERSCHMAN, J P
EM: juan.guerschman@csiro.au
AF: CSIRO Land and Water, GPO Box 1666, Canberra, ACT 2601, Australia
AU: * HILL, M J
EM: hillmj@aero.und.edu
AF: University of North Dakota, Department of Earth Systems Science and Policy, University of North Dakota, Grand Forks, ND 58202, United States
AU: BARRETT, D J
EM: damian.barrett@csiro.au
AF: CSIRO Land and Water, GPO Box 1666, Canberra, ACT 2601, Australia
AU: RENZULLO, L
EM: luigi.renzullo@csiro.au
AF: CSIRO Land and Water, GPO Box 1666, Canberra, ACT 2601, Australia
AU: MARKS, A
EM: alan.marks@csiro.au
AF: CSIRO Land and Water, GPO Box 1666, Canberra, ACT 2601, Australia
AU: BOTHA, E
EM: elizabeth.botha@csiro.au
AF: Ensis Forest Biosecurity and Protection, PO Box E4008, Kingston, ACT 2604, Australia
AB: Monitoring the fractional cover of photosynthetic vegetation (PV), non-photosynthetic vegetation (NPV) and bare soil (BS) in savannas is important for carbon and water modeling, grazing management, fire risk assessment and erosion control. We developed a simple method for resolving their fractional cover with hyperspectral imagery, by combining the Normalized Difference Vegetation Index (NDVI) which measures vegetation greenness and the Cellulose Absorption Index (CAI), which quantifies the intensity of the cellulose-lignin feature at 2.0-2.2 μm, and then applying a linear unmixing. We applied this method to three EO-1 Hyperion scenes acquired during the 2005 growing season in a site in northern Australia. Data from field measurements and from fire scar maps provided a means for qualitatively validating the results obtained. We then explored the potential of the MODIS-TERRA sensor for resolving vegetation fractional cover. We generated synthetic MODIS data from the Hyperion images and also used actual MODIS reflectance from the MOD09 product, concurrent with the Hyperion images. We found that the MODIS-TERRA sensor, despite not being able to quantify the cellulose feature directly, can be used for mapping fractional cover. This is due to the fact that vegetation, regardless of its photosynthetic status, has a lower reflectance at 2.1 μm (MODIS band 7) than at 1.6 μm (MODIS band 6), compared to soils, which have a relatively flat spectra at those wavelengths. We propose using the ratio of band 7 to band 6 together with the NDVI for resolving the proportions of PV, NPV and BS. We tested the method in 10 independent savanna sites across Australia where grass curing is continuously monitored and found very good agreement both in space and in time between observed and modeled fractional cover. Finally, we developed a prototype of an operational product based on the MOD43 product (Nadir BRDF-Adjusted Reflectance 16-Day composites) and discuss its strengths and limitations.
DE: 0414 Biogeochemical cycles, processes, and modeling (0412, 0793, 1615, 4805, 4912)
DE: 0428 Carbon cycling (4806)
DE: 0430 Computational methods and data processing
DE: 0439 Ecosystems, structure and dynamics (4815)
DE: 0480 Remote sensing
SC: Biogeosciences [B]
MN: 2007 Fall Meeting