Manikanta et al. (2026) Understanding data sufficiency and temporal informativeness for hydrological model calibration in data scarce regions

Identification

• Journal: Journal of Hydrology
• Year: 2026
• Date: 2026-09-12
• Authors: Velpuri Manikanta, Surya Kiran Guniganti, Daneti Arun Sourya, Rathinasamy Maheswaran
• DOI: 10.1016/j.jhydrol.2026.136424

Research Groups

• Department of Civil Engineering, Indian Institute of Technology, Hyderabad, India

Short Summary

The study develops a predictive framework using Classification and Regression Tree (CART) and Symbolic Regression to estimate the minimum amount and optimal timing of calibration data required for hydrological models based on physical catchment characteristics.

Objective

• To determine if the amount and timing of calibration data required for reliable hydrological model performance can be predicted a priori from a catchment's physical and meteorological attributes.

Study Configuration

• Spatial Scale: 671 catchments in the United States (477 rainfall-dominated and 194 snow-dominated).
• Temporal Scale: Calibration periods ranging from 1 month to 1 year.

Methodology and Data

• Models used: Classification and Regression Tree (CART) analysis and Symbolic Regression.
• Data sources: Observed streamflow records, dynamic meteorological inputs, and static catchment attributes.

Main Results

• Dominant Controls: Mean annual precipitation, snow fraction, aridity, and soil depth were identified as the primary drivers of performance loss when using sparse calibration data.
• Data Sufficiency by Regime:
  • Rainfall-dominated catchments: 23% (110/477) achieved benchmark performance with only one month of calibration data.
  • Snow-dominated catchments: 44% (85/194) required a full 12 months of data to adequately represent snowmelt and storage dynamics.
• Sensitivity Factors: Higher aridity and greater variability in evapotranspiration increased the model's sensitivity to sparse calibration.
• Metric Influence: The choice of objective function significantly impacted calibration performance due to different metrics emphasizing different segments of the hydrograph.

Contributions

• Provides a process-oriented, data-driven framework that allows hydrologists to anticipate calibration data requirements a priori, which is particularly valuable for Prediction in Ungauged Basins (PUB) and data-scarce regions.

Funding

• Not specified in the provided text.

Citation

@article{Manikanta2026Understanding,
  author = {Manikanta, Velpuri and Guniganti, Surya Kiran and Sourya, Daneti Arun and Maheswaran, Rathinasamy},
  title = {Understanding data sufficiency and temporal informativeness for hydrological model calibration in data scarce regions},
  journal = {Journal of Hydrology},
  year = {2026},
  doi = {10.1016/j.jhydrol.2026.136424},
  url = {https://doi.org/10.1016/j.jhydrol.2026.136424}
}