Crop Establishment SIMulator: A Qualitative Aggregative Model to Predict the Role of Phytobiomes on Field Crop Establishment

Abstract

The definition of phytobiomes can be transposed to any agroecosystem and applies to any phase of crop cycles. Here, we study the crop establishment phase using a generic modeling framework to assess the potential role of phytobiomes on field crop establishment. We first developed a generic model called Crop Establishment SIMulator (CESIM) that takes into account cropping practices, seed and seedling characteristics, seedbed components (physical chemical and biological), and weather, as well as their interactions. All these variables were integrated in a qualitative aggregative hierarchical network to predict the quality of field crop establishment. CESIM has 38 basic (input variables) and 20 aggregated (19 state variables and 1 output variable) attributes for a total of 58 attributes. The prediction quality of the model was evaluated for a dataset of 231 field observations across four states of Australia and experimental results obtained in the last 40 years. Accuracy of predictions of the final attribute (i.e., crop establishment) was 91% and explained 29% of variability of the dataset, as described by the quadratic weighted Cohen’s κ. CESIM represents a unique and original generic model capable of taking into account a large number of variables and their interactions to predict the quality of field crop establishment. This model is flexible, transparent, and user friendly and, therefore, is suitable both for academic and nonacademic users. CESIM can be used across a wide range of situations to perform not only the ex-ante assessment of potential establishment quality of a given crop but also ex-post assessment.