Lombardi et al. (2025) Testing Machine Learning and Traditional Models for Tree-Ring-Based scPDSI Streamflow Reconstruction: A 1500-Year Record of the French Broad River, Tennessee, USA

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Water
• Year: 2025
• Date: 2025-11-18
• Authors: Ray Lombardi, Abel Andrés Ramírez Molina, Glenn Tootle
• DOI: 10.3390/w17223288

Research Groups

Not explicitly stated in the provided text.

Short Summary

This study reconstructed 1500 years of streamflow for the French Broad River using dendrochronological tools, identifying a significant hydrologic regime change in 1271 CE and an emerging trend of higher average flow with severe single-year droughts.

Objective

• To evaluate new dendrochronological tools and examine long-term streamflow trends for the French Broad River in eastern Tennessee.

Study Configuration

• Spatial Scale: French Broad River, eastern Tennessee, Appalachian region.
• Temporal Scale: 1500 years (500–1970 CE), with an instrumental record spanning less than a century.

Methodology and Data

• Models used: Linear regression, Random Forest, Deep Learning.
• Data sources: Tree-ring-derived self-calibrating Palmer Drought Severity Index (scPDSI), instrumental streamflow record (for calibration and validation).

Main Results

• Linear regression models provided skillful and stable streamflow reconstruction across calibration and validation periods.
• Random Forest and Deep Learning models achieved higher skill but showed some skill loss during validation, indicating overfitting.
• All models more reliably captured drought years than flood years, reflecting scPDSI's sensitivity to soil moisture and limitations for high-flow extremes.
• Trend analyses identified a significant change point in 1271 CE, separating a drought-dominated early period (500–1272 CE) from a wetter, less variable regime (1273–1970 CE).
• An emerging trend shows higher average flow interrupted by severe single-year droughts, consistent with regional evidence and projected changes in Appalachia.

Contributions

• Provides a millennial perspective on hydrologic extremes for the French Broad River, extending the instrumental record by 1500 years.
• Offers guidance on the application and limitations of paleohydrology tools (linear regression, machine learning with scPDSI) for water resource planning.
• Identifies a significant long-term hydrologic regime shift and an emerging trend relevant to climate change impacts in the Appalachian region.

Funding

Not explicitly stated in the provided text.

Citation

@article{Lombardi2025Testing,
  author = {Lombardi, Ray and Molina, Abel Andrés Ramírez and Tootle, Glenn},
  title = {Testing Machine Learning and Traditional Models for Tree-Ring-Based scPDSI Streamflow Reconstruction: A 1500-Year Record of the French Broad River, Tennessee, USA},
  journal = {Water},
  year = {2025},
  doi = {10.3390/w17223288},
  url = {https://doi.org/10.3390/w17223288}
}