Resumen
Precision farming has been suggested for the strategic management of agricultural crops on a smaller scale than total farm area, based on the use of information technology and agronomic know-how. In order to be successful in this activity, it is necessary to be able to manage the variation of plant and soil variables; this requires that data be collected and analyzed for decision-making process. The modeling of interactions between soil, weather and crop growth would be a natural choice for the integration and search for management strategies. However, further study of modeling for precision farming is needed, with regard to soil-structure and integrated-plant properties. An experiment was conducted in Angatuba, São Paulo, Brazil (23o33?S; 48o18?W; 670m), in which a maize crop was observed during two growing seasons (1999-2000) in two different areas where yield-related properties of soil and plant variables were monitored. Specific-site analyses were carried out in one area, using a simulation model based on a single sampling of soil properties. In the other area, the implementation potential of this new technology - with modeling and simulation as supporting tools - was discussed, based on the spatial structure of plant and soil variables, obtained in intensive sampling. Results from field observations suggest that modeling of biophysical processes is a fundamental tool for the implementation of precision farming, based on a realistic work scale and using sampling strategies from multiple sources, which should include remote sensing imagery. In the present case, grid soil sampling on a field scale was shown to be inconsistent with precision-farming needs, in terms of the extension of the application of modeling based on point processes, to the space of an area