HR: 0830h
AN: B41C-0900    [PDF]
TI: Environmental Controls on Multi-Scale Soil Nutrient Variability in the Tropics: the Importance of Land-Cover Change
AU: * Holmes, K W
EM: karen@geog.ucsb.edu
AF: Department of Geography, 3611 Ellison Hall University of California, Santa Barbara, CA 93106 United States
AU: Kyriakidis, P C
EM: phaedon@geog.ucsb.edu
AF: Department of Geography, 3611 Ellison Hall University of California, Santa Barbara, CA 93106 United States
AU: Chadwick, O A
EM: oac@geog.ucsb.edu
AF: Department of Geography, 3611 Ellison Hall University of California, Santa Barbara, CA 93106 United States
AU: Matricardi, E
EM: matricard@msu.edu
AF: Department of Geography, Michigan State University, East Lansing, MI 48823 United States
AU: Soares, J V
EM: vianei@ltid.inpe.br
AF: Instituto de Pesquisas Espaciais (INPE), Caixa Postal 515 Av. dos Astronautas 1758, Sao Jose dos Campos, SP 12227-010 Brazil
AU: Roberts, D A
EM: dar@geog.ucsb.edu
AF: Department of Geography, 3611 Ellison Hall University of California, Santa Barbara, CA 93106 United States
AB: The natural controls on soil variability and the spatial scales at which correlation exists among soil and environmental variables are critical information for evaluating the effects of deforestation. We detect different spatial scales of variability in soil nutrient levels over a large region (hundreds of thousands of km2) in the Amazon, analyze correlations among soil properties at these different scales, and evaluate scale-specific relationships among soil properties and the factors potentially driving soil development. Statistical relationships among physical drivers of soil formation, namely geology, precipitation, terrain attributes, classified soil types, and land cover derived from remote sensing, were included to determine which factors are related to soil biogeochemistry at each spatial scale. Surface and subsurface soil profile data from a 3000 sample database collected in Rond“nia, Brazil, were used to investigate patterns in pH, phosphorus, nitrogen, organic carbon, effective cation exchange capacity, calcium, magnesium, potassium, aluminum, sand, and clay in this environment grading from closed canopy tropical forest to savanna. We focus on pH in this presentation for simplicity, because pH is the single most important soil characteristic for determining the chemical environment of higher plants and soil microbial activity. We determined four spatial scales which characterize integrated patterns of soil chemistry: less than 3 km; 3 to 10 km; 10 to 68 km; and from 68 to 550 km (extent of study area). Although the finest observable scale was fixed by the field sampling density, the coarser scales were determined from relationships in the data through coregionalization modeling, rather than being imposed by the researcher. Processes which affect soils over short distances, such as land cover and terrain attributes, were good predictors of fine scale spatial components of nutrients; processes which affect soils over very large distances, such as precipitation and geology, were better predictors at coarse spatial scales. However, this result may be affected by the resolution of the available predictor maps. Land-cover change exerted a strong influence on soil chemistry at fine spatial scales, and had progressively less of an effect at coarser scales. It is important to note that land cover, and interactions among land cover and the other predictors, continued to be a significant predictor of soil chemistry at every spatial scale up to hundreds of thousands of kilometers.
DE: 1615 Biogeochemical processes (4805)
DE: 1625 Geomorphology and weathering (1824, 1886)
DE: 1803 Anthropogenic effects
DE: 1869 Stochastic processes
DE: 9360 South America
SC: Biogeosciences [B]
MN: 2003 Fall Meeting