Chen et al. (2025) Baseflow index dynamics and its non-monotonic drivers from a spatiotemporal heterogeneity perspective

Identification

• Journal: Journal of Hydrology
• Year: 2025
• Date: 2025-11-29
• Authors: Hsin-Yu Chen, Hsin-Fu Yeh
• DOI: 10.1016/j.jhydrol.2025.134692

Research Groups

• Department of Resources Engineering, National Cheng Kung University, Tainan, Taiwan

Short Summary

This study analyzed the spatiotemporal dynamics and non-monotonic drivers of the Baseflow Index (BFI) in 60 Taiwanese catchments from 1960 to 2022, revealing significant non-stationarity and identifying the simple daily intensity index (SDII) as the most dominant driver.

Objective

• To analyze the spatiotemporal dynamics and non-monotonic driving mechanisms of catchment attributes on the Baseflow Index (BFI) in Taiwanese catchments, addressing the often-overlooked spatial and temporal heterogeneity of these mechanisms.

Study Configuration

• Spatial Scale: 60 catchments across Taiwan.
• Temporal Scale: 62 years, from 1960 to 2022.

Methodology and Data

• Models used: Geographically and Temporally Weighted Regression (GTWR) as the primary model; compared against Geographically Weighted Regression (GWR), Pooled Regression, and Random Forest.
• Data sources: Observation data from 60 Taiwanese catchments used to calculate Baseflow Index (BFI); static physiographic factors; time-varying rainfall characteristics; and land cover change data.

Main Results

• The Baseflow Index (BFI) exhibits significant spatial and temporal non-stationarity, showing both increasing and decreasing trends over time and statistically significant spatial clustering.
• The Geographically and Temporally Weighted Regression (GTWR) model demonstrated superior performance compared to GWR, pooled regression, and even outperformed the random forest method in analyzing BFI dynamics.
• Most catchment attributes showed non-monotonic effects on BFI, with their coefficients varying between positive and negative depending on the specific time and location.
• In contrast, the simple daily intensity index (SDII) consistently showed a negative influence on BFI across all locations and time periods, indicating a monotonically negative effect.
• Both GTWR and Random Forest analyses consistently identified the simple daily intensity index (SDII) as the most dominant driver of BFI.

Contributions

• Provides deeper insights into the drivers of BFI by explicitly accounting for spatiotemporal heterogeneity using the GTWR model.
• Highlights the necessity of considering spatiotemporal non-stationarity and non-monotonic effects of drivers in hydrological signature analysis.
• Identifies the simple daily intensity index (SDII) as a consistently dominant and monotonically negative driver of BFI.
• Offers valuable findings to support sustainable water resources management and land-use planning under changing hydroclimatic conditions.

Funding

• Not specified in the provided text.

Citation

@article{Chen2025Baseflow,
  author = {Chen, Hsin-Yu and Yeh, Hsin-Fu},
  title = {Baseflow index dynamics and its non-monotonic drivers from a spatiotemporal heterogeneity perspective},
  journal = {Journal of Hydrology},
  year = {2025},
  doi = {10.1016/j.jhydrol.2025.134692},
  url = {https://doi.org/10.1016/j.jhydrol.2025.134692}
}