02188naa a2200181 a 450000100080000000500110000800800410001902400460006010000220010624501270012826000090025552015990026465300160186365300230187970000170190270000170191977300700193619388472021-11-11       2012    bl uuuu     u00u1 u    #d7 ahttps://doi.org/10.1029/2012JG0019822DOI1 aVASQUES, G. de M.  aInfluence of the spatial extent and resolution of input data on soil carbon models in Florida, USA.h[electronic resource]  c2012  aUnderstanding the causes of spatial variation of soil carbon (C) has important implications for regional and global C dynamics studies. Soil C predictive models can identify sources of C variation, but may be influenced by scale parameters, including the spatial extent and resolution of input data. Our objective was to investigate the influence of these scale parameters on soil C spatial predictive models in Florida, USA. We used data from three nested spatial extents (Florida, 150,000 km2; Santa Fe River watershed, 3,585 km2; and University of Florida Beef Cattle Station, 5.58 km2) to derive stepwise linear models of soil C as a function of 24 environmental properties. Models were derived within the three extents and for seven resolutions (30?1920 m) of input environmental data in Florida and in the watershed, then cross-evaluated among extents and resolutions, respectively. The quality of soil C models increased with an increase in the spatial extent (R2 from 0.10 in the cattle station to 0.61 in Florida) and with a decrease in the resolution of input data (R2 from 0.33 at 1920-m resolution to 0.61 at 30-m resolution in Florida). Soil and hydrologic variables were the most important across the seven resolutions both in Florida and in the watershed. The spatial extent and resolution of environmental covariates modulate soil C variation and soil-landscape correlations influencing soil C predictive models. Our results provide scale boundaries to observe environmental data and assess soil C spatial patterns, supporting C sequestration, budgeting and monitoring programs.  aSoil carbon  aSoil carbon models1 aGRUNWALD, S.1 aMYERS, D. B.  tJournal of Geophysical Research Geosciencesgv. 117, n. G4, 2012.