Dari et al. (2023) Regional data sets of high-resolution (1 and 6 km) irrigation estimates from space

Identification

• Journal: Earth system science data
• Year: 2023
• Authors: Jacopo Dari, Luca Brocca, Sara Modanesi, Christian Massari, Angelica Tarpanelli, Silvia Barbetta, Raphael Quast, Mariëtte Vreugdenhil, Vahid Freeman, Anaïs Barella-Ortiz, Pere Quintana Seguí, David Bretreger, Espen Volden
• DOI: 10.5194/essd-15-1555-2023

Research Groups

• Department of Civil and Environmental Engineering, University of Perugia, Perugia, Italy
• Research Institute for Geo-Hydrological Protection, National Research Council (CNR-IRPI), Perugia, Italy
• Department of Geodesy and Geoinformation, Research Unit Remote Sensing, Vienna University of Technology (TU Wien), Vienna, Austria
• Earth Intelligence, Spire Global, Luxembourg
• Observatori de l'Ebre (OE), Ramon Llull University – CSIC, Roquetes, Spain
• School of Engineering, The University of Newcastle, Callaghan, New South Wales, Australia
• European Space Agency (ESA), ESRIN, Frascati, Italy

Short Summary

This study presents the first regional-scale, high-resolution (1 km and 6 km) gridded data sets of irrigation water use derived from satellite soil moisture observations over three major river basins (Ebro, Po, and Murray–Darling), demonstrating satisfactory performance in quantifying irrigation amounts and timing, particularly in semiarid regions.

Objective

• To develop and present regional-scale, high-resolution (1 km and 6 km) data sets of irrigation water amounts estimated from satellite soil moisture observations over three major river basins characterized by varying irrigation practices and climatic conditions.

Study Configuration

• Spatial Scale: Regional scale, covering the Ebro river basin (Spain, $\approx 86\,000\,\text{km}^2$), the Po valley (Italy, $\approx 78\,000\,\text{km}^2$), and the Murray–Darling basin (Australia, $\approx 1\,000\,000\,\text{km}^2$). Output grids are 1 km (European sites) and 6 km (Australian site).
• Temporal Scale: Weekly temporal aggregation. The European data sets cover January 2016 to July 2020, and the Australian data set covers April 2017 to July 2020.

Methodology and Data

• Models used: SM-based (soil-moisture-based) inversion approach, which relies on inverting the soil water balance equation to estimate total water input (rainfall + irrigation). The approach incorporates drainage (power-law equation) and actual evapotranspiration (soil-moisture-limited potential rate).
• Data sources:
  • Soil Moisture (SM):
    • Ebro and Po basins (1 km): Sentinel-1 observations processed using the RT1 (first-order Radiative Transfer) model.
    • Murray–Darling basin (6 km): CYGNSS (Cyclone Global Navigation Satellite System) observations (Spire's GNSS-R product).
  • Potential Evapotranspiration (PET): GLEAM (Global Land Evaporation Amsterdam Model) v3.5b product ($0.25^\circ$ spatial resolution).
  • Rainfall: ERA5-Land reanalysis data set (9 km spatial resolution).
  • Land Cover: Corine Land Cover (CLC2018, 25 m) for Europe; ESA CCI land cover map (300 m) for Australia, used to mask agricultural areas.
  • Benchmark Data: In situ irrigation rates (volumes/depths) collected from agricultural districts in the Ebro basin (SAIH Ebro), Po valley (CER consortium), and Murray–Darling basin (IIOs).

Main Results

• Ebro Basin (1 km, semiarid climate): Satisfactory performance with median validation metrics (14-day aggregation): Root-Mean-Square Error (RMSE) of $12.4\,\text{mm}/14\,\text{d}$, Pearson correlation ($r$) of 0.66, and bias of $-4.62\,\text{mm}/14\,\text{d}$ (slight underestimation).
• Murray–Darling Basin (6 km, semiarid climate): Very good performance with median validation metrics (monthly aggregation): RMSE of $10.54\,\text{mm}/\text{month}$, $r$ of 0.77, and bias of $-3.07\,\text{mm}/\text{month}$.
• Po Valley (1 km, humid subtropical climate): Validation was limited due to small pilot district size and data availability, showing a tendency to overestimate benchmark rates (e.g., Formellino: RMSE $14.86\,\text{mm}/14\,\text{d}$, bias $8.65\,\text{mm}/14\,\text{d}$).
• Accuracy: For the Ebro basin, yearly cumulated estimates were within $\pm 30\%$ of the benchmark in 9 out of 15 cases. For the Murray–Darling basin, seasonal estimates were within $\pm 30\%$ in 5 out of 10 cases.
• Limitations: The products may suffer from false irrigation rates during non-irrigated seasons, particularly in the Po valley, attributed potentially to rainfall overestimates under humid conditions.

Contributions

• Generation and public release of the first regional-scale, high-resolution (1 km and 6 km) gridded irrigation data sets derived from satellite observations (Sentinel-1 and CYGNSS).
• Successful application and validation of the SM-based inversion approach across diverse climatic and irrigation environments (Mediterranean and Australian semiarid/subtropical).
• Demonstration of the feasibility of using high-resolution satellite soil moisture products for detailed water resource management in agriculture, meeting the spatial resolution requirements ($\le 1\,\text{km}$) for monitoring agricultural practices.

Funding

• European Space Agency (ESA) under the Irrigation+ project (contract no. 4000129870/20/I-NB).

Citation

@article{Dari2023Regional,
  author = {Dari, Jacopo and Brocca, Luca and Modanesi, Sara and Massari, Christian and Tarpanelli, Angelica and Barbetta, Silvia and Quast, Raphael and Vreugdenhil, Mariëtte and Freeman, Vahid and Barella-Ortiz, Anaïs and Quintana‐Seguí, Pere and Bretreger, David and Volden, Espen},
  title = {Regional data sets of high-resolution (1 and 6 km) irrigation estimates from space},
  journal = {Earth system science data},
  year = {2023},
  doi = {10.5194/essd-15-1555-2023},
  url = {https://doi.org/10.5194/essd-15-1555-2023}
}

Generated by BiblioAssistant using gemini-flash-latest (Google API)