BigFoot Project Background & Summary

MODIS (the Moderate Resolution Imaging Spectrometer) is the principal high temporal frequency mapping sensor on-board NASA's Earth Observation System (EOS), Terra. The MODIS instrument views the entire Earth's surface every one to two days, acquiring data in 36 spectral bands at spatial resolutions from 250 m to 1 km (Running, et al. 1994). The unprecedented volume of data that will be collected has led the MODIS instrument team to develop a number of derived data products, with the intent of reducing the burden of data processing on the user. A series of land product algorithms were selected by open competition, and the MODIS Land Discipline Group (MODLand) has been charged with the development of these algorithms, the generation of the associated data products, and their validation. Included among the MODLand products are surface reflectance, spectral vegetation indices, land cover, the absorbed fraction of photosynthetically active radiation (FPAR), leaf area index (LAI), net primary productivity (NPP), and land surface temperature. These, and other MODIS products, will play an important role in measuring and monitoring surface variables. Validation of these global data products is crucial to establish the accuracy of the data products for the scientific user community, and to provide feedback for improving the data processing algorithms.

The BigFoot project grew from a workshop held in 1996 which was attended by ecologists and scientists of related disciplines, primarily from the Long Term Ecological Research (LTER) Network. The purpose was to explore validation protocols and scaling issues that would lead to an improved understanding of several MODLand products. The BigFoot field sites are also EOS Land Validation Core Sites and are part of the FLUXNET program. The sites have active science programs concentrating on CO₂, water vapor, and energy exchange using flux tower measurements. The "footprint" over which gas flux data are collected varies, but is roughly 1 km or less. For the BigFoot analysis, this footprint will be extended to 25 km² to include multiple 1 km MODIS cells, hence the project name. BigFoot investigators will focus on validation of the MODLand land cover, LAI, FPAR, and NPP products. We will develop fine grain (25 m resolution) surfaces of land cover, LAI, FPAR, and NPP, aggregate these to 1 km resolution, then assess the similarities and differences between these surfaces and the MODLand products.

Our goals, in addition to providing MODLand product validation, are:

Explore the errors and information losses that accrue when extrapolating field data to coarse grain (1 km) surfaces
Determine if there is a fundamental grain size at each site, above which error rates accelerate when modeling land cover, LAI, FPAR, and NPP
Develop a better understanding of the climatic and ecological controls on net primary production and carbon allocation within and among biomes
Learn how flux tower measurements of net ecosystem exchange (NEE) and field measurements of NPP co-vary, and how to translate between them using ecological models.

We will be working at four field sites: a boreal forest, a tallgrass prairie, a mixed deciduous-conifer forest, and a mixed corn and soybean agricultural system.

The core BigFoot products will be: field-collected data sets of land cover, LAI, FPAR, NPP, and related variables; and 25 km² surfaces at 25 m spatial resolution of land cover, LAI, FPAR, and NPP for each site. These surfaces will be developed from field data, Landsat ETM+ imagery, image classification and related statistics, geostatistical analysis, and ecosystem process models. Errors in each data layer will be characterized using independent field data.

Explicit examination of scaling from field measurements, to fine grain (25 m) surfaces, to coarse grain (1 km) MODLand grids is a central theme of BigFoot. The fine grain surfaces will be aggregated to several resolutions, up to 1 km, to determine if there is a grain size above which information loss rapidly increases. Through these analyses, we hope to characterize error due to scaling differences versus error due to algorithm definitions. It is theoretically possible for no single MODIS grid cell to accurately estimate land cover, LAI, FPAR, or NPP, but for multiple cell estimates within and across sites to be accurate. A cross-site comparison of MODLand and BigFoot surfaces will permit us to assess MODLand data product accuracy and gain an understanding of the source of errors at both the site and cross-site level.