Grillakis et al. (2021) Regionalizing Root‐Zone Soil Moisture Estimates From ESA CCI Soil Water Index Using Machine Learning and Information on Soil, Vegetation, and Climate

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

This study addresses the limitation of shallow sensing depth (2–5 cm) in the ESA CCI soil moisture dataset by developing a methodology to estimate root-zone soil moisture (RZSM). By calibrating the Soil Water Index (SWI) using in situ observations and leveraging machine learning techniques with global physical descriptors, the researchers successfully derived RZSM for the period 2001–2018, demonstrating good agreement with established reanalysis products like ERA5 Land, particularly over mid-latitudes.

Identification

• Authors: Manolis Grillakis, Aristeidis Koutroulis, Dimitrios D. Alexakis, Christos Polykretis, Ioannis Ν. Daliakopoulos
• Group/Lab: Not specified in the abstract
• Citation: Not specified in the abstract
• DOI: Not specified in the abstract

Objective

• To overcome the limitation of shallow sensing depth (2–5 cm) in the ESA CCI soil moisture dataset by estimating and calibrating the Soil Water Index (SWI) against in situ observations, and subsequently leveraging machine learning to regionalize this calibration globally to derive root-zone soil moisture (RZSM).

Study Configuration

• Spatial Scale: Global scale.
• Temporal Scale: RZSM assessment period: 2001–2018. Underlying ESA CCI data span nearly four decades.

Methodology and Data

• Models used: Soil Water Index (SWI) model; Machine learning techniques (used for regionalizing the calibration based on physical descriptors).
• Data sources: European Space Agency Climate Change Initiative (ESA CCI) satellite-observed soil moisture (uppermost 2–5 cm); International Soil Moisture Network (ISMN) in situ observations (for calibration); Physical soil, climate, and vegetation descriptors; Comparison datasets: European Centre for Medium-Range Weather Forecasts (ECMWF) ERA5 Land and Famine Early Warning Systems Network Land Data Assimilation System (FLDAS) reanalyses soil moisture data sets.

Main Results

• A robust methodology was developed to estimate and calibrate the Soil Water Index (SWI) using ISMN data, which was then regionalized globally via machine learning techniques incorporating physical descriptors.
• The resulting root-zone soil moisture (RZSM) product for the period 2001–2018 showed good overall agreement when compared against the ERA5 Land and FLDAS reanalysis datasets.
• The strongest agreement between the derived RZSM and the established reanalysis products was observed over mid-latitudes.

Contributions

• Provides a novel, calibrated methodology to extend the utility of the long-term ESA CCI surface soil moisture data to estimate deeper root-zone soil moisture (RZSM).
• Generates a new, large-scale RZSM dataset (2001–2018) that is homogenized and suitable for supporting large-scale hydrological and climate studies.
• Demonstrates the effective use of machine learning to regionalize complex hydrological calibrations globally using physical soil, climate, and vegetation characteristics.

Funding

• European Space Agency (ESA) Climate Change Initiative (CCI) (Implied, as the initiative provides the core data and context).

Citation

@article{abstract2026Regionalizing,
  author = {abstract, Not specified in the},
  title = {Regionalizing Root‐Zone Soil Moisture Estimates From ESA CCI Soil Water Index Using Machine Learning and Information on Soil, Vegetation, and Climate},
  journal = {Unknown Journal},
  year = {2026},
  doi = {Not specified in the abstract},
  url = {https://doi.org/Not specified in the abstract}
}

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