Nie et al. (2025) Identifying Hydrological Analogous Year from Magnitude and Temporal Patterns

Identification

• Journal: Water Resources Management
• Year: 2025
• Date: 2025-12-26
• Authors: Shiqi Nie, Ganggang Zuo, N. Wang, Jiancang Xie
• DOI: 10.1007/s11269-025-04344-x

Research Groups

State Key Laboratory of Water Engineering Ecology and Environment in Arid Area, Xi’an University of Technology, Xi’an, Shaanxi Province, China.

Short Summary

This study proposes a comprehensive Similarity Index (SI) that integrates six existing methodologies to identify hydrologically similar years by balancing magnitude consistency and temporal pattern alignment. The SI family demonstrates superior cross-station stability, while a hybrid Cosine Similarity + Euclidean Distance (CS + ED) method achieves the highest overall similarity score, both significantly improving runoff forecasting accuracy.

Objective

• To develop and evaluate a comprehensive Similarity Index (SI) that effectively balances magnitude consistency and temporal pattern alignment for identifying hydrologically similar years, overcoming limitations of traditional single-metric approaches.

Study Configuration

• Spatial Scale: Yellow River Basin, China, covering a drainage area of 752,400 km². Data from 15 hydrological stations distributed across the upper, middle, and lower reaches of the basin.
• Temporal Scale: Monthly-scale hydrological data from January 1961 to December 2018. Target years for analogous year identification were 2016–2018.

Methodology and Data

• Models used:
  • Six base similarity/distance metrics: Euclidean Distance (ED), Dynamic Time Warping (DTW), Fréchet Distance (FD), Hausdorff Distance (HD), Longest Common Subsequence Similarity (LCSS), and Cosine Similarity (CS).
  • Proposed comprehensive Similarity Index (SI), integrating the six base metrics.
  • Weighted SI variants: SI(EWM) using the Entropy Weight Method (EWM) and SI(PCA) using Principal Component Analysis (PCA).
  • Hybrid scheme: CS + ED.
  • Runoff prediction model: Multiple linear regression.
• Data sources:
  • Monthly-scale datasets of natural runoff, precipitation, potential evaporation (PET), and surface 7 cm soil moisture.
  • Data sourced from the National Tibetan Plateau Data Center (https://www.tpdc.ac.cn/home), obtained by interpolating China’s gridded data based on latitude and longitude.

Main Results

• The definition of hydrological similarity is highly method-dependent, with significant divergence in identified analogous years across different methods for some stations.
• The CS + ED hybrid method achieved the highest mean Overall Score (0.8347) across all stations, demonstrating the most effective balance between magnitude and temporal pattern similarity.
• The SI family (SI, SI(EWM), SI(PCA)) exhibited superior cross-station stability, with SI(PCA) showing the lowest coefficient of variation (4.24%) in overall scores.
• SI(PCA) achieved the highest mean Pattern Score (0.8342), while CS + ED and standalone ED achieved the highest mean Magnitude Scores (0.8365 and 0.8356, respectively).
• In a practical runoff forecasting experiment, the proposed SI method demonstrated a 13.47% improvement in Root Mean Square Error (RMSE) compared to the worst-performing method (LCSS). SI(PAC) and SI(EWM) achieved 13.14% and 11.20% RMSE improvements, respectively, significantly outperforming traditional methods.

Contributions

• Introduction of a novel integrated Similarity Index (SI) framework that systematically combines six complementary metrics to effectively balance magnitude and temporal pattern considerations in hydrological similarity assessment.
• Development of a dedicated two-dimensional evaluation system employing robust statistical indicators (Kendall’s τ, Theil-Sen slope for pattern similarity; Normalized Root Mean Square Error, Nash-Sutcliffe Efficiency, Q-Q plot correlation for magnitude similarity) to independently assess pattern alignment and magnitude fidelity.
• Comprehensive benchmarking of ten identification schemes across diverse hydrological stations, providing unique insights into methodological performance and stability.
• Demonstration of the practical significance of the proposed framework through substantial improvements in monthly runoff forecasting accuracy, highlighting its strong potential for operational hydrological applications.

Funding

• National Natural Science Foundation of China (Grant No. 52309034)
• China Postdoctoral Science Foundation (Grant No. 2022M722561)
• Water Conservancy Science and Technology Project of Shaanxi Province (Program No. 2025slkj-10)

Citation

@article{Nie2025Identifying,
  author = {Nie, Shiqi and Zuo, Ganggang and Wang, N. and Xie, Jiancang},
  title = {Identifying Hydrological Analogous Year from Magnitude and Temporal Patterns},
  journal = {Water Resources Management},
  year = {2025},
  doi = {10.1007/s11269-025-04344-x},
  url = {https://doi.org/10.1007/s11269-025-04344-x}
}