Variability in Maize Yield Due to Differences in Soil Texture and the Effect of Water and Nitrogen.

The WISE3.1 database provides a homogenized set of primary profiles soil data relevant for a wide range of environmental studies including assessments of crop production. Soil texture is frequently measured to estimate characteristics related to water holding parameters and plant available water. The objective of this study was to determine the relationship between maize yield and soil textural classes and the effects of water and nitrogen. Soil profiles for 9613 locations around the world from the ISRIC-WISE3.1 database were converted to the DSSAT format and used for simulating maize yield with the CSM-CERES-Maize model. Crop management from an experiment that was conducted at the University of Florida, Gainesville, Florida was used as input. Weather data were generated using the WGEN weather generator for a period of 30 years. There were four hypothetical treatments, including rainfed low nitrogen, rainfed high nitrogen, automatic irrigation low nitrogen and automatic irrigation high nitrogen. Simulated yield was classified into 11 different textural classes. Analysis of variance and multiple range tests were used to compare the treatments.  Crop management, weather, and the soil profile characteristics contributed to the explanation of the variability in yield. The means for the textural classes as well as the nitrogen applications differed significantly. The application of nitrogen increased significantly the yield compared to the treatment with irrigation and low nitrogen. Nevertheless, the variability within each texture class was high as an indication of the variability due to other soil characteristics. There was no strong interaction between the soil textural classes and the four treatments since the pattern across the texture classes was similar but at a different scale. In conclusion the yield variability depended on the soil texture, weather, and management of the crop; however, more detailed studies are needed to explain the yield variability inside of each textural class.