Skip to content

Preprocessing Input Data

Before running CropSuiteLite, raw climate and soil variables must be transformed into climatological daily raster products.

Climate dataset transformation

The original NEX-GDDP-CMIP6 data is provided at 0.25° resolution. This module reprocesses it to 0.05° to match CropSuiteLite's spatial requirements. While this example uses a 20-year climatological period, this duration is configurable.

This step generates daily climatological raster datasets for temperature and precipitation.

To perform the transformation, define the parameters in yaml_configurations/crop_suite_datasets.yaml. Key settings include:

  • Reference Mask: Defines the target resolution and coordinate grid (reference_mask_layer_path).
  • Clip Extent: The geographic bounding box.
  • Periods: The time windows for climatological aggregation.
  • SSPs: The emission scenarios to process.

Example configuration:

YAML
GENERAL_INFO:
  process_climate: True
  clip_extent: [-26,-37, 56, 40] # AFRICA
  reference_mask_layer_path: 'data/africa_mask_005.tif' 

CLIMATE:
  input_path: 'climate/nex-gddp-cmip6_AFRICA/'
  output_path: 'climate/nex-gddp-cmip6_AFRICA_cs_005/'
  periods: 
    - [2021,2040]
    - [2041,2060]
    - [2061,2080]
    - [2081, 2100]
  ssps: ['ssp585']

Once the configuration file is ready, run the download script:

Bash
python datasets/create_spatial_datasets.py -config ./yaml_configurations/crop_suite_datasets.yaml

After processing the climate data (e.g., for the access-esm1-5 model under ssp585), the output directory structure should look like this:

Text Only
CropSuiteLite/
│
├── climate/
│    ├── nex-gddp-cmip6_AFRICA_cs/
│    └── nex-gddp-cmip6_AFRICA_cs_005/
│         ├── access-esm1-5_ssp585_2021_2040/
│         │    ├── Prec_avg.tif
│         │    └── Temp_avg.tif
│         ├── access-esm1-5_ssp585_2041_2060/
│         │    ├── Prec_avg.tif
│         │    └── Temp_avg.tif
│         ├── access-esm1-5_ssp585_2061_2080/
│         │    ├── Prec_avg.tif
│         │    └── Temp_avg.tif
│         └── access-esm1-5_ssp585_2081_2100/
│              ├── Prec_avg.tif
│              └── Temp_avg.tif

Other datasets

In addition to climate variables, CropSuiteLite requires soil data (from SoilGrids) and terrain information (SRTM elevation). To mask and resample these datasets, update the same configuration file (crop_suite_datasets.yaml) by enabling soil processing and specifying the directories.

YAML
GENERAL_INFO:

  process_soil: True
  clip_extent: [-26,-37, 56, 40] # AFRICA
  reference_mask_layer_path: 'data/africa_mask_005.tif' 

SOIL:
  input_path: '../soilgrids/'
  output_path: '../soilgrids_005/'

Run the processing script:

Bash
python datasets/create_spatial_datasets.py -config ./yaml_configurations/crop_suite_datasets.yaml