Pérez-Vega et al. (2025) Monitoring Irrigated Agriculture Using Remote Sensing and Census Data: A Case Study from Guanajuato, Mexico

Identification

• Journal: SN Computer Science
• Year: 2025
• Date: 2025-12-26
• Authors: Azucena Pérez-Vega, Jean-Francois Mas
• DOI: 10.1007/s42979-025-04591-0

Research Groups

• Departamento de Geomática e Hidraúlica, Universidad de Guanajuato, Mexico.
• Área de análisis espacial y modelación, Centro de Investigaciones en Geografía Ambiental, Universidad Nacional Autónoma de México (UNAM), Mexico.

Short Summary

This study evaluates the consistency between official agricultural census data and multi-sensor remote sensing data (MODIS, Landsat, and Sentinel-2) for monitoring irrigated crop dynamics in Guanajuato, Mexico. The research demonstrates that integrating coarse-resolution temporal data with high-spatial-resolution imagery provides a robust framework for managing water resources in semi-arid regions.

Objective

• To assess the effectiveness and consistency of integrating official census data with various remote sensing platforms to monitor irrigated agricultural patterns and water consumption across different crop cycles.

Study Configuration

• Spatial Scale: Municipality of Pénjamo, Guanajuato, Mexico (Central Mexico), focusing on irrigation areas exceeding 10,000 hectares.
• Temporal Scale: Long-term analysis spanning 20 years (2003–2023), covering two distinct agricultural cycles: autumn/winter (dry season) and spring/summer (rainy season).

Methodology and Data

• Models and Indices:
  • Normalized Difference Vegetation Index (NDVI) for vegetation health.
  • Evapotranspiration (ET) products for water consumption analysis.
  • A custom Irrigated Agriculture Index ($I_{irr}$) based on the temporal difference between maximum and minimum NDVI values.
  • Linear regression models for spectral unmixing of coarse-resolution pixels.
• Data sources:
  • Satellite: MODIS (MOD13A1 and MOD16A2 at 500 m resolution), Landsat 8 (30 m resolution), and Sentinel-2 (10–60 m resolution).
  • Census: Agri-Food and Fisheries Information Service (SIAP) statistical data.
  • Climate/Cartography: National Water Commission (CONAGUA) and National Institute of Statistics and Geography (INEGI).
  • Software: R (packages: terra, tmap, MODISstp, rsat) and Climex.

Main Results

• Identified a strong correlation ($R^2 = 0.82$) between MODIS-derived NDVI values and the proportion of irrigated crop areas, validating the use of coarse-resolution data for large-scale monitoring.
• Quantified significant fluctuations in irrigated areas: 6,000 to 29,000 hectares during the autumn/winter cycle and 10,000 to 24,000 hectares during the spring/summer cycle.
• Sentinel-2 data revealed complex agricultural patterns and short-cycle crop rotations that were undetectable in coarse MODIS imagery.
• Observed a shift in crop dominance: since 2015, sorghum areas have decreased and been largely replaced by corn in the spring/summer cycle.
• Detected discrepancies between official census reports and satellite-derived estimates, particularly in peak years like 2008.

Contributions

• Demonstrates a "multisensory approach" that combines the high temporal frequency of MODIS with the high spatial resolution of Sentinel-2 and Landsat to overcome individual sensor limitations.
• Provides a scalable methodology for "unmixing" coarse satellite data to estimate sown areas in regions where census data may be incomplete or infrequent.
• Offers a practical framework for water governance in regions facing critical groundwater depletion (Guanajuato's availability is only 845 $m^3$/inhabitant/year compared to the national average of 3,705 $m^3$/inhabitant/year).

Funding

• UNAM-PAPIIT project (Reference code: IN112823).
• Dirección de Apoyo a la Investigación y el Posgrado, Universidad de Guanajuato.

Citation

@article{PérezVega2025Monitoring,
  author = {Pérez-Vega, Azucena and Mas, Jean-Francois},
  title = {Monitoring Irrigated Agriculture Using Remote Sensing and Census Data: A Case Study from Guanajuato, Mexico},
  journal = {SN Computer Science},
  year = {2025},
  doi = {10.1007/s42979-025-04591-0},
  url = {https://doi.org/10.1007/s42979-025-04591-0}
}

Generated by BiblioAssistant using gemini-3-flash-preview (Google API)