Skip to content

Latest commit

 

History

History
 
 

Crop_models

README.md

Crop simulation models and sensitivity analysis

Seminar, Introduction to Digital Agro.

License issues Rg

Presentation

Crop models - pdf

Google Colab

Open Seminar_Crop_Model.ipynb in Google Colab!

Open In Colab

Open Seminar_Sensitivity_analysis.ipynb in Google Colab!

Open In Colab

Dependencies

PCSE/WOFOST - Python Crop Simulator Environment

https://pcse.readthedocs.io/en/stable/

SAlib - Sensitivity Analysis Library in Python

https://salib.readthedocs.io/en/latest/

Installation

Clone this repository and create new conda env on your local machine

git clone https://github.com/EDSEL-skoltech/Intro_to_Digital_Agriculture

Create new env with pcse package for crop models WOFOST

cd Crop_models

conda env create -f py3_pcse.yml

conda activate py3_pcse

Meta

Mikhail Gasanov – [email protected]

License

Distributed under the MIT license. See LICENSE for more information.

Homework

In our lab work, we will use the crop simulation model WOFOST. Documentation for package and model: https://pcse.readthedocs.io/. Your task will be to conduct any crop growth simulations for several years for the Moscow region or any other. It will also be necessary to conduct a sensitivity analysis and optimize the irrigation strategy. For sensitivity analysis, we will use the SALib package. For optimization, we will use Nevergrad package.

All the necessary data is available in the GitHub repository and in Google Colabs.

Part 1 – Crop Yield Prediction (PCSE/WOFOST) - 4 points

Assess the yield of one of the crops for the Moscow region over several years (potatoes, sugar beets or others)

Generate plots for biomass growth, crop yield and seasonal weather for one year
Plot weather dynamics (T min, ET0 or others)

Agromanagement for Moscow Potato crop and NPK fertilization

"name": "Potato", 
"latitude": 54.85,
"longitude": 38.85, 
"crop_start": "2019-04-20", 
"crop_end": "2019-09-15", 
"crop_name": "potato", 
"npk_events": ["2019-06-22"], 
"npk": [[90,10,90]]

Available crops

from pcse.fileinput import YAMLCropDataProvider
cropd = YAMLCropDataProvider()
cropd.print_crops_varieties()

Part 2 – Sensitivity Analysis (SALib) - 4 points

  1. Perform sensitivity analysis of one of the model blocks (crop, soil, agromanagement *) with SALib. You can choose one of the methods that you consider necessary (Sobol, FAST, ...). Generate samples – In report provide the size of the resulting matrix and the sample size (N). Conduct parameter sensitivity analysis - In report provide S1 and ST indices.
    List of available crop parameters: df=pd.read_excel("./data/ScalarParametersOfWofost-Potential.xlsx") display(df)
  2. Generate plots (Hist, etc.) *3) Speed-up sensitivity analysis with multiprocessing *4) Estimate the required number of simulations to obtain reliable values of the sensitivity indices. Try to estimate the sample size at the confidence interval of the sensitivity indices.

Part 3 – Optimization (Nevergrad) - 2 points

The task is to compare several optimizers from the Nevergrad package with the Random Search method. To do this, you can use the visualizations provided in the notebook. It is also necessary to determine the required size of the budget.

Contributing

  1. Fork it (https://github.com/EDSEL-skoltech/Intro_to_Digital_Agriculturefork)
  2. Create your feature branch (git checkout -b feature/fooBar)
  3. Commit your changes (git commit -am 'Add some fooBar')
  4. Push to the branch (git push origin feature/fooBar)
  5. Create a new Pull Request