Raster Dataset: Frost-affected corn areas in Western Paraná, Brazil, for the 2020/2021 crop season (GEEadas methodology)

File description

This multi-band GeoTIFF (raster) file contains spatial data regarding the 2020/2021 Second-Season Corn crop. It provides a dual-layer analysis: first identifying where corn was planted, and second, classifying the specific impact of frost events or crop status in those areas. It provides the underlying data for the paper 'GEEadas: GEE-based automatic detection of adverse-frost stress', published in Remote Sensing Applications: Society and Environment. Reference for the publication: Chaves, M., Adami, M., Oldoni, L., Rodigheri, G., Prudente, V., Santana, C., Garcia, A., Covre, R. & Sanches, I. (2025). GEEadas: GEE-based Automatic Detection of Adverse-frost Stress. Remote Sensing Applications: Society and Environment, 101799. https://doi.org/10.1016/j.rsase.2025.101799. Link: https://www.sciencedirect.com/science/article/pii/S2352938525003520. The method to generate this file was described in the paper. The processing chain was published in the GitHub of the Agricultural Remote Sensing Laboratory (AgriRS Lab) of the National Institute for Space Research (INPE) https://github.com/agrirslabinpe/GEEadas.

Band breakdown

This raster is composed of two distinct bands, each serving a specific analytical purpose:

Band 1: Crop Mask (Corn Presence)

This band acts as a binary mask to identify the agricultural land use, specifically isolating corn fields from other land covers.

• Value 1 (No corn): Areas identified as other land uses (forest, water, urban, other crops) or where corn was not detected.

• Value 2 (Corn): Confirmed areas of corn cultivation.

  • Usage: Use this band to calculate the total planted area or to mask other datasets to only analyze corn fields.

Band 2: Frost Impact & Crop Status

This band provides a detailed classification of the condition of the corn fields identified in Band 1, specifically focusing on damage caused by the frost events of 2021.

• Value 1 (Harvested/Senescent): Corn fields that were either already harvested or had reached natural senescence (drying) before the frost events occurred. These areas effectively "escaped" the frost damage.

• Value 2 (Frost 25/05): Corn crops that suffered damage specifically from the frost event on May 25th.

• Value 3 (Frost 30/06): Corn crops that survived the first event but were damaged by the subsequent frost on June 30th.

• Value 4 (Not affected): Corn fields that remained healthy/green through these dates and did not show significant spectral signatures of frost damage.

How to use in GIS (QGIS/ArcGIS)

Since the information is split across two bands, you will need to adjust your visualization settings:

1. To view the Corn Map:

   • Open the Layer Properties.

   • Set the Band to Band 1.

   • Style with distinct colors for values 1 and 2.

2. To view the Frost Damage Map:

   • Open the Layer Properties.

   • Set the Band to Band 2.

   • Use a Paletted/Unique Values renderer.

   • Assign specific colors to the classes (e.g., Grey for Harvested, Red for Frost 25/05, Orange for Frost 30/06, and Green for Not affected).

Analytical Note: You can assume that the classification in Band 2 is spatially coincident with the "Corn" pixels in Band 1. Pixels where Band 1 is "No Corn" likely have a specific fill value or are ignored in Band 2.

Google Earth Engine project: https://ee-victorohden.projects.earthengine.app/view/geeadas