Nuñez-Ibarra et al. (2026) From grid to ground: how well do gridded products represent soil moisture dynamics in natural ecosystems during precipitation events?

Identification

• Journal: Hydrology and earth system sciences
• Year: 2026
• Date: 2026-04-02
• Authors: Daniel Nuñez-Ibarra, Mauricio Zambrano-Bigiarini, Mauricio Galleguillos
• DOI: 10.5194/hess-30-1813-2026

Research Groups

• Center for Climate and Resilience Research, Universidad de Chile, Santiago, Chile
• Department of Civil Engineering, Universidad de la Frontera, Temuco, Chile
• Facultad de Ingeniería y Ciencias, Universidad Adolfo Ibáñez, Santiago, Chile
• Data Observatory Foundation, ANID Technology Center, Santiago, Chile
• Master's Program in Territorial Management of Natural Resources (GESTREN), Universidad de Chile, Santiago, Chile
• Departamento de Ecosistemas y Sistemas Naturales, Pontificia Universidad Católica de Chile, Santiago, Chile
• Departamento de Ingeniería Hidráulica y Medio Ambiente, Pontificia Universidad Católica de Chile, Santiago, Chile

Short Summary

This study evaluates four gridded soil moisture (SM) products against in situ observations from ten natural ecosystems in central and southern Chile to assess their ability to represent SM dynamics, especially during precipitation events. It finds that ERA5 and ERA5-Land generally outperform other products, particularly in humid regions, while highlighting challenges in arid areas and the diagnostic value of event-based SM signatures.

Objective

• To evaluate how well 3-hourly gridded products represent surface soil moisture (SSM) and root zone soil moisture (RZSM) dynamics in natural ecosystems of the Southern Hemisphere during precipitation events.
• To assess the reliability of gridded datasets in representing soil moisture (SM) signatures (rising time and amplitude) across diverse hydroclimatic conditions.

Study Configuration

• Spatial Scale: Ten natural monitoring sites in central and southern Chile (between 32.01° S and 39.64° S), covering five semi-arid and five humid hydroclimatic conditions. Gridded products have native spatial resolutions ranging from 9 km to 31 km.
• Temporal Scale: Two years (1 January 2022 to 31 December 2023) at a 3-hourly temporal resolution.

Methodology and Data

• Models used: ERA5, ERA5-Land, SMAP-L4 (SPL4SMAU), GLDAS-Noah.
• Data sources:
  • In situ observations: Kimün-Ko SM monitoring network in Chile, using TEROS 10 and TEROS 12 capacitance sensors at depths from 10 cm to 200 cm, and hourly rain gauge data.
  • Gridded products: ERA5 (ECMWF reanalysis), ERA5-Land (ECMWF land reanalysis), SMAP-L4 (NASA Soil Moisture Active Passive Level-4, assimilates L-band brightness temperature), GLDAS-Noah (NASA Global Land Data Assimilation System, Noah Land Surface Model).
  • Ancillary data: Global-AI-PET-v3 (aridity index), SRTMv4.1 (elevation), CLDynamicLandCover.V2 (land cover), CLSoilMaps (soil properties), Harmonized World Soil Database (HWSD) V1.2, GlobCover 2009.

Main Results

• ERA5 and ERA5-Land consistently outperformed SMAP-L4 and GLDAS-Noah across most metrics and regions, with ERA5-Land showing particular strength in humid areas.
• SMAP-L4 achieved the best surface soil moisture (SSM) performance in selected northern arid locations based on the modified Kling-Gupta efficiency (KGE').
• GLDAS-Noah generally performed the worst, with the exception of moderate correlation values in southern root zone soil moisture (RZSM).
• During the first precipitation event of the year, all gridded products systematically overestimated both rising times and amplitudes in the arid north, indicating difficulties in capturing SM responses under dry antecedent conditions.
• Gridded products showed better agreement with in situ measurements during intense precipitation events, especially in humid regions, with reduced amplitude and timing errors.
• All gridded products performed better for RZSM than for SSM, exhibiting higher KGE' values and lower errors across most sites, likely due to the more buffered dynamics of deeper soil layers.
• The deseasonalised Spearman rank correlation proved effective in identifying inconsistencies in temporal dynamics, particularly in arid regions with strong seasonal cycles.
• Event-based SM signatures (amplitude and rising time) revealed systematic discrepancies not captured by traditional statistical metrics, demonstrating their complementary diagnostic value.

Contributions

• Provides the first comprehensive evaluation of 3-hourly gridded soil moisture datasets for native ecosystems in South America, a largely under-monitored and under-evaluated region.
• Offers an exhaustive assessment of both surface and root-zone soil moisture responses to precipitation events, revealing key spatial patterns and product-specific strengths and limitations across contrasting hydroclimatic conditions.
• Highlights the diagnostic value of event-based soil moisture signatures (amplitude and rising time) as a complement to traditional statistical metrics for a process-oriented evaluation of SM dynamics.
• Recommends the routine use of deseasonalised Spearman rank correlation in future validation efforts to improve robustness and comparability, especially in regions with strong seasonal cycles.

Funding

This research was supported by the Chilean Agencia Nacional de Investigación y Desarrollo (ANID) through the following projects: ANID-Fondecyt Regular 1212071 and 1210932, ANID-PCI NSFC 190018, ANID-FONDAP 1523A0002, ANID DO21000, and Fondecyt 1250718.

Citation

@article{NuñezIbarra2026From,
  author = {Nuñez-Ibarra, Daniel and Zambrano-Bigiarini, Mauricio and Galleguillos, Mauricio},
  title = {From grid to ground: how well do gridded products represent soil moisture dynamics in natural ecosystems during precipitation events?},
  journal = {Hydrology and earth system sciences},
  year = {2026},
  doi = {10.5194/hess-30-1813-2026},
  url = {https://doi.org/10.5194/hess-30-1813-2026}
}