A new approach to crop model calibration: Phenotyping plus post‐processing

Abstract

Crop models contain a number of genotype-dependent parameters, which need to be estimated for each genotype. This is a major difficulty in crop modeling. We propose a hybrid method for adapting a crop model to new genotypes. The genotype-dependent parameters of the model could be obtained by phenotyping (or gene-based modeling). Then field data for example from variety trials could be used to provide a simple empirical correction to the model, of the form a +b times an environmental variable. This approach combines the advantages of phenotyping, namely that the genotype-specific parameters have a clear meaning and are comparable between genotypes, and the advantages of fitting the model to field data, namely that the corrected model is adapted to a specific target population. It has the advantage of being very simple to apply, and furthermore gives useful information as to which environmental variables are not fully accounted for in the initial model. In this study, this empirical correction is applied to the SUNFLO crop model for sunflower, using field data from a multi-environment trial network. The empirical correction reduced mean squared error (MSE) on the average by 54% for prediction of yield and by 26% for prediction of oil content, compared to the initial model. Most of the improvement came from eliminating bias, with some further improvement from the environmental term in the regression.

Forecasting Sunflower Grain Yield by Assimilating Leaf Area Index into a Crop Model

“…This can explain the over-estimation of yield forecast and constitutes a line of thought to improve the SUNFLO model in the context of on-farm experiments. However, the model accuracy on fields without non-modeled yield limiting factors was in the range of previous independent experimental validations (RMSE 4-7 q•ha −1 ) as reported in [18,[50][51][52].…”

Tackling G × E × M interactions to close on-farm yield-gaps: creating novel pathways for crop improvement by predicting contributions of genetics and management to crop productivity

“…Building on the early demonstrations of successful applications of prediction methodology for maize G × E × M interactions in the US corn-belt, there are nascent opportunities emerging to consider broader applications for other crops and production systems. These developments are stimulating advances in the integrated approaches to crop modelling, phenotyping, machine learning and high-performance computing to harness “Big Data” from combinations of designed and on-farm empirical studies to enable prediction-based agriculture (Holzworth et al 2014 ; Brown et al 2014 ; Ramirez-Villegas et al 2020 ; Casadebaig et al 2020 ; Bogard et al 2020 ; Sinclair et al 2020 ; Ersoz et al 2020 ; Washburn et al 2020 ; Stöckle and Kemanian 2020 ; Cooper et al 2021 ).…”

Interdisciplinary strategies to enable data-driven plant breeding in a changing climate

“…7,[70][71][72][73] Because CGMs depend on physiological parameters that may vary genetically, experimental estimation of these parameters limits the application of CGMs at the scale of breeding programs. Recent work has demonstrated some success with in silico estimation of these parameters from experimental data, 74,75 but challenges remain. 76 Alternatively, yield predictions can be improved using models that are agnostic to the actual values of the CGM parameters.…”

“…The difficulty of measuring the physiological parameters required by CGMs means that these models generally incorporate only a small sample of available genetic variation at best (see, for example, Padilla and Otegui 85 ). Promising results for expansion of this genetic sample have been obtained using combinations of HTP, statistical models, ML, and CGMs, 72,74,75 although challenges remain due to the large number of combinations of CGM parameters that can lead to the same output. 76 To a certain extent, these all represent partial statistical fixes.…”