FAIRagro UC6 Crop Simulation Modelling Workflows

Work in progress — this repository is under active development. Workflows, tooling, and documentation are subject to change.

Introduction

The repository contains computational workflows for crop simulation modelling created using the FAIRagro Scientific Workflow Infrastructure (SciWIn) tool.

Workflows are expressed in CWL v1.2 and executed with s4n, which runs each tool step inside a Docker container.

Implemented Workflow

The full pipeline takes field observations (UAV-derived NDVI, crop calendar, and field metadata) and produces ready-to-run DSSAT crop simulation files alongside a results plot:

Requirements

Dependency	Purpose
s4n	CWL workflow runner
Docker	Container runtime used by s4n for each tool step
raster2sensor	Process and manage temporal vegetation indices from UAV images using the OGC SensorThings API standard
phenocover	Phenology analysis and ground-cover estimation
csmTools	ETL functions for DSSAT data integration

Quick Start

1. Copy config-example.yml to config.yml and fill in your FROST credentials:

   cp config-example.yml config.yml

2. Run the full pipeline:

   ./run-demo.sh

   The script calls s4n execute local ./workflows/demo/demo.cwl config.yml and moves outputs to outputs/.

Note: config.yml is gitignored — secret credentials are never committed.

Repository Structure

fairagroUC6-workflows/
│
├── config-example.yml          # Template for workflow configuration (copy → config.yml)
├── run-demo.sh                 # Entry-point script: executes demo.cwl via s4n
├── workflow.toml               # Workflow metadata (name, version)
│
└── workflows/
    ├── raster2sensorTools/
    │   └── fetch-ndvi.cwl              # Download NDVI time series from FROST server
    │
    ├── phenocoverTools/
    │   └── phenology-analyzer.cwl      # Wheat phenology analysis with weather integration
    │
    ├── csmTools/                       # Individual ETL steps (all backed by csmtools-cli image)
    │   ├── get-field-data.cwl          # Extract field metadata from ICASA template
    │   ├── identify-production-season.cwl
    │   ├── get-sensor-data.cwl         # Download observations from FROST server
    │   ├── convert-sensor-data-icasa.cwl
    │   ├── get-weather-data.cwl        # Download weather from NASA POWER
    │   ├── convert-nasa-data-icasa.cwl
    │   ├── get-soil-profile.cwl        # Extract soil data from SoilGrids
    │   ├── lookup-gs-dates.cwl         # Identify growth stage dates from phenology
    │   ├── convert-gs-dates-icasa.cwl
    │   ├── assemble-icasa-dataset.cwl  # Merge all data into a single ICASA dataset
    │   ├── convert-icasa-dssat.cwl     # Convert ICASA → DSSAT file format
    │   ├── normalize-soil-profile.cwl
    │   ├── calculate-initial-layers.cwl
    │   ├── build-simulation-files.cwl  # Generate DSSAT .filex, .sol, .wth files
    │   ├── run-simulations.cwl         # Execute DSSAT crop simulations
    │   └── plot-results.cwl            # Plot simulated vs. observed growth
    │
    ├── demo/
        └── demo.cwl            # Top-level workflow wiring all steps above