Geostatistical simulation of the spatial variability of an irrigated maize farm plot

Abstract

This paper proposes a methodology to spatialize the STICS crop model at the soil-unit scale in an agricultural field. The soil's apparent resistivity is used as auxiliary information in a geostatistical model of the soil depth. Stochastic simulations of some STICS input parameters showing spatial variability, as the soil depth, are used in order to spatialize this model. The methodology was tested on an irrigated maize field in Petite Beauce (France). Apparent resistivities at three different depths were determined automatically with MUlti-depth continuous electrical profiling (MUCEP) on a 1.4-ha area. Soil characteristics and irrigation were also measured on a grid to access to the field spatial variability of STICS inputs. Comparison of simulated yield and water contents at the end of harvesting with the field measurements is carried out for two different irrigation conditions, thanks to an irrigation restriction imposed on one subarea. The results show that the methodology is promising, but improvements are needed for practical uses.