A Python package for automating PEST (Parameter ESTimation) file creation for DSSAT crop model calibration using time series data.

dpest is a Python package designed to automate the creation of PEST (Parameter Estimation and Uncertainty Analysis) control files for calibrating DSSAT (Decision Support System for Agrotechnology Transfer) crop models. Currently, dpest is capable of calibrating DSSAT wheat models only. It generates template files for cultivar and ecotype parameters, instruction files for OVERVIEW.OUT and PlantGro.OUT, and the main PEST control file. A utility module is also included to extend PlantGro.OUT files for complete time series compatibility.

For detailed usage instructions and module references, see the full documentation on: dpest.readthedocs.io/en/latest.

• Automated generation of PEST control files (.PST).
• Compatibility with DSSAT wheat models.
• Creation of PEST template files (.TPL) for cultivar and ecotype parameters.
• Generation of PEST instruction files (.INS) for key DSSAT output files.
• Utility module to extend PlantGro.OUT files for complete time series.

dpest can be installed via pip from PyPI.

pip install dpest

If you use dpest in your research, please cite our article in the Journal of Open Source Software:

APA:

Vargas-Rojas, L., & Wang, D. R. (2025). dpest: Streamlining Creation of PEST input files for DSSAT Crop Model Calibration. Journal of Open Source Software, 10(111), 8188. https://doi.org/10.21105/joss.08188

BibTeX:

@article{Vargas-Rojas_dpest_Streamlining_Creation_2025, author = {Vargas-Rojas, Luis and Wang, Diane R.}, title = {dpest: Streamlining Creation of PEST input files for DSSAT Crop Model Calibration}, journal = {Journal of Open Source Software}, year = {2025}, volume = {10}, number = {111}, pages = {8188}, doi = {10.21105/joss.08188}, url = {https://joss.theoj.org/papers/10.21105/joss.08188}, month = jul }

For additional citation formats, see the CITATION.cff file in this repository or use the "Cite this repository" button in the GitHub About section.

1. DSSAT Software: DSSAT (version 4.8) must be installed on your system.
2. DSSAT Experiment Data: You'll need a DSSAT experiment to work with. The example below uses the SWSW7501WH N RESPONSE wheat experiment.
3. PEST Software: You’ll need to have PEST (version 18) installed.

To quickly explore how dpest works in practice, you can open and run the example Jupyter notebook included in the repository:

examples/wheat/ceres/usage_example.ipynb

This notebook walks through:

• Loading DSSAT outputs (OVERVIEW.OUT, PlantGro.OUT)
• Generating template (.TPL) and instruction (.INS) files
• Creating the main PEST control file (.PST)
• Validating the setup using PEST check commands

To run the notebook:

  pip install dpest[notebook]
  jupyter notebook examples/wheat/ceres/usage_example.ipynb

Note: The [notebook] extra installs ipykernel and notebook dependencies required to open the notebook.

For a detailed, step-by-step example, please refer to the official documentation:
Complete Example on Read the Docs

The following steps provide a brief overview of how to use dpest.

1. Locate DSSAT Genotype Files: The cultivar (.CUL) and ecotype (.ECO) files are typically included with the DSSAT installation. These are usually found in the C:\DSSAT48\Genotype\ directory.

2. Run a DSSAT Simulation: Execute a DSSAT simulation for your chosen wheat experiment using the CERES model. This will generate the necessary output files (OVERVIEW.OUT and PlantGro.OUT), typically found in the C:\DSSAT48\Wheat\ directory.

   2.1. Launch DSSAT.
   2.2. Click "Selector".
   2.3. Expand "Crops" and select "Wheat".
   2.4. In the "Data" panel select the "SWSW7501.WHX" experiment.
   2.5. Click "Run" button in the toolbar.
   2.6. In the "Simulation" popup window, choose "CERES" as the crop model.
   2.7. Click "Run Model" and wait for the simulation to finish.
   

3. Import the dpest Package:

   • To import the entire dpest package:

   import dpest
   

   • To import specific modules:

   from dpest.wheat.ceres import cul, eco
   from dpest.wheat import overview, plantgro
   from dpest import pst
   from dpest.wheat.utils import uplantgro
   
   # Now you can use the functions directly:
   cul(...)
   eco(...)
   overview(...)
   plantgro(...)
   pst(...)
   uplantgro(...)
   

4. Use the Modules

   • cul(): This module creates PEST template files (.TPL) for CERES-Wheat cultivar parameters. Use it to generate template files for cultivar calibration.
   • eco(): This module creates PEST template files (.TPL) for CERES-Wheat ecotype parameters. Use it to generate template files for ecotype calibration.
   • overview(): This module creates PEST instruction files (.INS) for reading observed (measured) values of key end-of-season crop performance metrics and key phenological observations from the OVERVIEW.OUT file. The instruction file tells PEST how to extract model-generated observations from the OVERVIEW.OUT file, compare them with the observations from the DSSAT A file, and adjust model parameters accordingly.
   • plantgro(): This module creates PEST instruction files (.INS) for reading observed (measured) values of plant growth dynamics from the PlantGro.OUT file. The instruction file tells PEST how to extract time-series data from the PlantGro.OUT file, compare that data with the time-series data provided in the DSSAT T file, and adjust model parameters accordingly.
   • pst(): enerates the main PEST control file (.PST) to guide the entire calibration process. It integrates the PEST template files (.TPL) and PEST instruction files (.INS), defines calibration parameters, observation groups, weights, PEST control variables and model run command.
   • uplantgro(): modifies the DSSAT output file (PlantGro.OUT) to prevent PEST errors when simulated crop maturity occurs before the final measured observation. This ensures PEST can compare all available time-series data, even when the model predicts maturity earlier than observed in the field.

5. Create Template Files and Instruction Files: Use the dpest modules to generate PEST template files (.TPL) for cultivar and ecotype parameters, and PEST instruction files (.INS) for the OVERVIEW.OUT and PlantGro.OUT files.

import dpest

# 1. Create CULTIVAR parameters TPL file
cultivar_parameters, cultivar_tpl_path = dpest.wheat.ceres.cul(
    P = 'P1D, P5', # How the user should enter the parameters
    G = 'G1, G2, G3', 
    PHINT = 'PHINT',
    cultivar = 'MANITOU',
    cul_file_path = 'C:/DSSAT48/Genotype/WHCER048.CUL'
)

# 2. Create OVERVIEW observations INS file
overview_observations,  overview_ins_path = dpest.wheat.overview(
    treatment = '164.0 KG N/HA IRRIG', #Treatment Name
    overview_file_path = 'C:/DSSAT48/Wheat/OVERVIEW.OUT' #Path to the OVERVIEW.OUT file
)
# 3. Create PlantGro observations INS file
plantgro_observations, plantgro_ins_path = dpest.wheat.plantgro(
    treatment = '164.0 KG N/HA IRRIG', #Treatment Name
    variables = ['LAID', 'CWAD', 'T#AD'] #Variables to calibrate
    plantgro_file_path = 'C:/DSSAT48/Wheat/PlantGro.OUT', #Path to the PlantGro.OUT file
)

6. Create the PEST Control File: Use the dpest.pst() module to generate the main PEST control file (.PST).

# 4. Create the PST file
dpest.pst(
    cultivar_parameters = cultivar_parameters,
    dataframe_observations = [overview_observations, plantgro_observations],
    model_comand_line = r'py "C:\pest18\run_dssat.py"', #Command line to run the model
    input_output_file_pairs = [
        (cultivar_tpl_path, 'C://DSSAT48/Genotype/WHCER048.CUL'), #Template file and the file to be modified
        (overview_ins_path , 'C://DSSAT48/Wheat/OVERVIEW.OUT'), #Instruction file and the file to be modified
        (plantgro_ins_path , 'C://DSSAT48/Wheat/PlantGro.OUT') #Instruction file and the file to be modified
    ]
)

7. Run PEST: Calibrate the model using PEST.

In addition to automating the creation of PEST control and template files for DSSAT CERES-Wheat model calibration, dpest includes a growing set of utility functions under dpest.utils for directly modifying existing .pst (PEST control) files.

Unlike libraries such as pyEMU, which parse .pst files into Python objects and rewrite the entire file when saving any change, dpest.utils performs lightweight, line-by-line edits to preserve the original structure and untouched content.

These utilities:

• Edit existing .pst files in place, without reconstructing or redefining unrelated fields.
• Are model-agnostic: they work with any .pst file, not just DSSAT-related ones.
• Only modify model-independent control parameters (e.g., optimization settings), leaving parameter and data blocks untouched.

This makes them ideal for quickly tuning optimization settings or cleaning up .pst files generated by dpest or other tools.

# Load the dpest.utils
from dpest.utils import noptmax, rmv_splitcols

# Path to the .pst file
pst_file_path = './ENTRY1/PEST_CONTROL.pst'

# Increase the number of optimization iterations (NOPTMAX) in a .pst file
noptmax(pst_file_path, new_value=50)

# Remove SPLITTHRESH/SPLITRELDIFF/SPLITACTION columns from a .pst file parameter groups section
rmv_splitcols(“PEST_CONTROL.pst”)

Full List of Utility Functions:
For the complete reference of available utility functions and their descriptions, see the dpest ReadTheDocs Utils page.

A detailed and up-to-date test coverage report for the dpest codebase is available on Codecov:

View detailed coverage report on Codecov

These instructions will guide you through setting up your environment, installing dependencies, and running tests for the dpest package.

• Python 3.10 or higher installed
• Git installed

git clone https://github.com/DS4Ag/dpest.git
cd dpest

On Windows:

py -3.10 -m venv .venv
.\.venv\Scripts\activate

On macOS/Linux:

python3.10 -m venv .venv
source .venv/bin/activate

PDM is a modern Python package and dependency manager recommended for development and testing.

Install PDM using pipx (recommended):

python -m pip install --user pipx
pipx ensurepath
pipx install pdm

Or install PDM using pip:

pip install --user pdm

pdm install

pdm install --dev

pdm run pytest tests/ -v

We welcome contributions from the community!

• To report issues or request features, please use GitHub Issues.
• To contribute code, fork the repository, create a branch, and submit a pull request.
• For questions or support, open an issue or participate in GitHub Discussions if enabled.
• All participants are expected to follow the Contributor Covenant Code of Conduct
• For detailed instructions on how to contribute, please see our contribution guidelines.

dpest is released under the GNU General Public License v3.0 only.