Quintana‐Seguí et al. (2019) The Utility of Land-Surface Model Simulations to Provide Drought Information in a Water Management Context Using Global and Local Forcing Datasets

Identification

• Journal: Water Resources Management
• Year: 2019
• Authors: Pere Quintana Seguí, Anaïs Barella-Ortiz, Sabela Regueiro-Sanfiz, Gonzalo Miguez‐Macho
• DOI: 10.1007/s11269-018-2160-9

Research Groups

• Observatori de l’Ebre (Universitat Ramon Llull - CSIC), Roquetes, Spain
• Universidade de Santiago de Compostela, Santiago de Compostela, Spain

Short Summary

This study evaluates the utility and uncertainty of Land-Surface Models (LSMs) driven by different resolution atmospheric forcing datasets (local vs. global) for simulating drought propagation (precipitation to soil moisture and streamflow) in mainland Spain (1980–2014), concluding that model structure uncertainty is the dominant factor affecting drought indices (SSMI and SSI) and propagation scales.

Objective

• To evaluate the usefulness of Land-Surface Models (LSMs) as tools to provide drought information (precipitation, soil moisture, and streamflow) to water managers, including the temporal and spatial structures of drought and its propagation through the system.
• To evaluate uncertainties due to forcing datasets (comparing global low-resolution products with local high-resolution products) and model structure.
• To detect areas of improvement in land-surface modeling to better meet the needs of water managers in Spain.

Study Configuration

• Spatial Scale: Mainland Spain (complex orography, Atlantic and Mediterranean climate influences). Simulations were performed at 5 km and 2.5 km resolution, aggregated to 5 km for comparison.
• Temporal Scale: 1980–2014 (35 years) for general simulations; 1982–2005 for streamflow-based calculations due to data availability and model spin-up. Monthly time series were used to calculate standardized indices.

Methodology and Data

• Models used:
  • SASER (SAfran-Surfex-Eaudysee-Rapid): Hydrometeorological platform based on SURFEX (Météo France's surface modeling platform).
    • ISBA-3L (3L): Three-layer soil description (force restore approach).
    • ISBA-DIF (DIF): Multilayer soil description (more complex).
    • RAPID: River routing model.
  • LEAFHYDRO (LHD): LEAF model modified to incorporate groundwater processes, including a dynamic water table and lateral groundwater flow. Uses kinematic wave method for river routing.
  • SIMPA (SMP): Monthly, conceptual, distributed hydrological model used by MAPAMA (used as reference for naturalized flow).
• Data sources:
  • Forcing Datasets (Atmospheric Analyses):
    • SAFRAN (SFR): Spanish meteorological analysis system (5 km resolution, reference dataset).
    • SAFRAN Low Resolution (SLR): SAFRAN data aggregated to 30 km resolution.
    • eartH2Observe (E2O): Global 0.25 degree forcing dataset derived from ERA-Interim.
    • MSWEP (MSW): E2O with its precipitation substituted by the MSWEP precipitation dataset (merges gauge, satellite, and reanalysis data).
  • Validation Data: Daily streamflow data from the MAPAMA database (naturalized flow estimates provided by SIMPA).
  • Drought Quantification: Standardized indices calculated using a nonparametric methodology: Standardized Precipitation Index (SPI), Standardized Soil Moisture Index (SSMI), and Standardized Streamflow Index (SSI).
  • Comparison Metrics: Root Mean Square Difference (RMSD) and Pearson Correlation coefficient ($r$).

Main Results

• Model Structure vs. Forcing Uncertainty: Model structure uncertainty (i.e., the choice between DIF, 3L, and LHD) is the most important source of uncertainty in simulating both SSMI and SSI, and in determining the temporal scale of drought propagation ($n_x$).
• Soil Moisture Drought (SSMI):
  • Differences between simulated SSMI are large, especially in the deep soil zone (RMSD up to 0.96, correlation as low as 0.51).
  • The forcing dataset has a larger impact on the overall uncertainty (spatial and temporal) of SSMI results than the model structure when comparing all simulations.
  • Drought Propagation to Soil Moisture ($nx$): The temporal scale ($nx$) at which precipitation anomalies (SPI) correlate best with SSMI is highly dependent on the model structure. DIF and 3L show overly homogeneous $n_x$ patterns (mostly 2 months), while LHD, which includes groundwater, shows a richer spatial pattern (3 to 12 months).
• Hydrological Drought (SSI):
  • SFR-DIF generally showed the best skill scores compared to the reference naturalized flow (SMP-SMP), with RMSD values ranging from 0.6 to 0.8. However, an RMSD of 1.0 (one standard deviation) was observed in some areas, indicating high uncertainty.
  • LHD exhibited larger RMSDs and lower correlations, likely due to a high baseflow bias impacting its ability to reproduce drought status.
  • Drought Propagation to Streamflow: DIF and 3L models showed $nx$ values that were too low and too homogeneous compared to the SMP reference, indicating a generally fast response of rivers to precipitation deficits. LHD, due to groundwater buffering, presented higher $nx$ values, suggesting slower dynamics.
• Forcing Dataset Performance: The MSWEP (MSW) precipitation product, which includes satellite observations, represents a large improvement compared with the E2O product (which does not include satellite data).

Contributions

• First comparison of the SURFEX (SASER) and LEAFHYDRO land-surface models in Spain.
• First study about drought in Spain using high-resolution land-surface models.
• Quantified that model formulation (structure, especially groundwater representation) is the dominant uncertainty source for drought propagation scales ($n_x$) and SSI skill, while forcing datasets contribute significantly to the spatial structure uncertainty of SSMI.
• Demonstrated that current state-of-the-art LSMs in Spain require significant improvement, particularly in groundwater processes and their interaction with soil and rivers, before they can reliably inform water managers.

Funding

• EU-FP7 eartH2Observe project (Grant agreement no. 603608).
• Spanish Ministry of Science, Innovation and Universities and the European Regional Development Fund (Grants CGL2013-47261-R and CGL2017-85687-R).
• Contribution to the HyMeX program (Hydrological Cycle in the Mediterranean Experiment).

Citation

@article{QuintanaSeguí2019Utility,
  author = {Quintana‐Seguí, Pere and Barella-Ortiz, Anaïs and Regueiro-Sanfiz, Sabela and Miguez‐Macho, Gonzalo},
  title = {The Utility of Land-Surface Model Simulations to Provide Drought Information in a Water Management Context Using Global and Local Forcing Datasets},
  journal = {Water Resources Management},
  year = {2019},
  doi = {10.1007/s11269-018-2160-9},
  url = {https://doi.org/10.1007/s11269-018-2160-9}
}

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