Ciencia del suelo
versión On-line ISSN 1850-2067
GREGORET, María Celeste; DARDANELLI, Julio; BONGIOVANNI, Rodolfo y DIAZ-ZORITA, Martín. Site-specific corn nitrogen and soil water response model in haplustolls. Cienc. suelo [online]. 2006, vol.24, n.2, pp. 147-159. ISSN 1850-2067.
Appropriate strategies for nitrogen fertilization in corn production is specially important in semi-arid regions, where soil water and nutrients are limiting factors. Precision agriculture has the potential to adjust fertilizer rates according to soil types and/or other defined conditions within management zones. The objectives were: i) to determine the site specific response of corn to nitrogen and available water at planting, and ii) to determine the economic and agronomic optimal levels of nitrogen and water. Three management zones with different production potential were identified: rainfed high and low potential and high potential with irrigation. The data are from on-farm trials for the 2004-2005 crop season, in Manfredi (Córdoba) with a uniform N rate along strips (0, 32, 64, 96, 129 and 161 kg N ha-1 in the dry land and 0, 64, 137, 212, 279 and 351 kg N ha-1 in the irrigated area) and a randomized complete block design to estimate site-specific crop response functions. Spatial autocorrelation is taken into account in regression estimation of N response functions by landscape position, in the form a spatial autoregressive error structure. Results suggest that N response differs significantly by landscape position and by soil water availability. Yields ranged from 4.914 to 11,346 kg ha-1 in the zone of low productivity, between 4,956 and 12,204 kg ha-1 in the one of high productivity and between 7,830 and 14,387 kg ha-1 in irrigated plots. The agronomical (DOA) and economically (DOE) optimal N rates also differed among management zones, being DOA= 285, 184 and 162 kg ha-1 for the zones from greater to smaller productivity, meanwhile DOE values were 171, 111 and 90 kg ha-1, respectively. Initial soil nitrogen and available water inclusion allowed us to build a single N response model for the rainfed management zones. Besides, irrigated experiment maize performance was well predicted by the same model type but excluding available water (non-limiting factor), as input. The model performed well using either available N at 60 or 200 cm soil depth, and available water al 200 cm depth. This model allowed us to recommend fertilization rates using variable rate application, maximizing net return to N, and minimizing environment contamination risks, by over dose input (i.e. N fertilization) application. Our conclusions are preliminary because they come from a single year experiment, which did not include the whole range of expectated initial conditions. More scenarios are needed in order to validate the proposed model.
Palabras llave : Precision agriculture; Water availability; Management zones; Variable rate; Site-specific management; Argentina.