Objectives

In the crop modelling community, the needs are increasing for automation in the processes of input and output management, control over simulations, parallelization, and model evaluation. In addition, parameter estimation and evaluation of crop models is still a major issue (Seidel et al. 2018 ; Wallach et al, 2021).
The SticsRPacks project, initiated in 2018, aims to develop a suite of R packages for managing crop models. The project is coordinated by S. Buis (INRAE EMMAH). The project team consists of five members from two institutes: S. Buis (INRAE EMMAH), P. Lecharpentier (INRAE AgroClim), M. Giner (CIRAD AIDA), C. Midingoyi (CIRAD AIDA), and R. Vezy (CIRAD AMAP).
 

CroptimizR and CroplotR: generic packages for  parameter estimation and evaluation of crop models

Two packages, CroptimizR and CroPlotR, are designed to offer dedicated software solutions to the crop modelling community to easily implement, compare and automate parameter estimation and evaluation processes of crop models. These packages propose a generic interface for linking crop models with a set of up-to-date parameter estimation algorithms and for automatically generating ad-hoc plots and statistics for their evaluation. They are the reference software used within the AgMIP Calibration project for multi-model calibration exercises (see e.g. Wallach et al., 2023). To date, they have been integrated as standard tools in STICS, APSIM NextGen and SiriusQuality2 crop models.

SticsOnR and SticsRFiles: specific packages for the STICS crop model

Two other packages developed within the SticsRPacks project, SticsOnR and SticsRFiles, are dedicated to the STICS crop model. They enable its control from within R, including reading/writing input/output files, retrieving information on parameters and variables and running simulations parallelized over CPUs. They are teached in the STICS training sessions and widely used within the STICS community.

Availability

All these packages are open source, released under the LGPL license, and freely available on GitHub. They follow a strict development workflow with shared concepts, enforced code style, and automatic tests. An online tutorial for the SticsRPacks packages is available here.

Partnership