259-3 Modeling Soil Carbon In a Hydrologically Complex Region: Lessons Learned In Florida.



Tuesday, October 18, 2011: 8:40 AM
Henry Gonzalez Convention Center, Room 209, Concourse Level

Gustavo M. Vasques1, Sabine Grunwald2, D. Brent Myers3, Nicholas B. Comerford4, James O. Sickman5 and Willie G. Harris2, (1)National Center of Soil Research, Embrapa - Brazilian Agricultural Research Corporation, Rio de Janeiro, Brazil
(2)Soil and Water Science, University of Florida, Gainesville, FL
(3)Agricultural Research Service, United States Department of Agriculture, Columbia, MO
(4)North Florida Research and Education Center, University of Florida, Quincy, FL
(5)Environmental Sciences, University of California, Riverside, Riverside, CA
Florida offers an extraordinary opportunity to investigate the influence of climate, hydrology, parent material, vegetation and land use on soil formation due to its contrasting landscape dynamics. This region is dominated by sandy soils and greatly influenced by water dynamics and human activities. In search for the main sources of spatial variation of soil carbon (C) in Florida, we present our experience with spatial and spectral modeling of soil C. Using legacy and newly collected data at three scales covering, respectively, a cattle research farm (5.58 km2), a representative mixed-use watershed (3585 km2), and the state as a whole (150,000 km2), we estimated C based on laboratory reflectance properties of the soil, and on soil-landscape relationships between C and soil-forming factors. Laboratory soil reflectance explained up to 86% of the soil C variance (validation), whereas landscape properties (e.g. land use, soil drainage, slope, and geology) allowed upscaling of soil C by explaining its field to regional variation. We further unveiled soil C pattern-process links, observing strong correlation between soil C and hydrological (e.g. wetlands, and available water capacity) patterns. Finally, by testing different modeling strategies, we assessed (i) the detrimental effect of decreasing the spatial extent, and increasing the pixel resolution, respectively, on soil C-landscape relationships, and (ii) the somewhat unpredictable behavior of soil C models when derived in one geographic region and applied in other regions. We explain some of the complexity of soil C-reflectance and soil C-landscape correlations, providing knowledge to support the development of soil C modeling, management, and decision-making strategies.
See more from this Division: S05 Pedology
See more from this Session: Symposium--Spatial Predictions In Soils, Crops and Agro/Forest/Urban/Wetland Ecosystems: I