Palumbo et al. (2025) Precipitation, moderated by spring temperature and vegetation, drives runoff efficiency in the Upper Colorado River Basin, USA

Identification

• Journal: Communications Earth & Environment
• Year: 2025
• Date: 2025-12-29
• Authors: David Palumbo, Subhrendu Gangopadhyay, Upmanu Lall
• DOI: 10.1038/s43247-025-03136-w

Research Groups

• Department of Earth and Environmental Engineering, Columbia University, New York, NY, USA
• Bureau of Reclamation, Washington, DC, USA
• Bureau of Reclamation, Denver, CO, USA
• Water Institute, Julie Ann Wrigley Global Futures Laboratory, Arizona State University, Tempe, AZ, USA

Short Summary

This study uses causal inference with historical data to identify drivers of surface runoff efficiency in the Upper Colorado River Basin. It finds that runoff efficiency is primarily driven by precipitation and snow accumulation, moderated by spring temperature and vegetation phenology, with summer temperature not emerging as a statistically significant direct driver.

Objective

• To understand how surface runoff efficiency (RE), defined as the amount of surface flow generated per unit precipitation, changes with precipitation and seasonal temperature conditions, including through vegetation and evapotranspiration dynamics, in the Upper Colorado River Basin.

Study Configuration

• Spatial Scale: Upper Colorado River Basin (UCRB), USA, which drains approximately 647,500 km². The study focuses on 16 sub-basins within the UCRB, ranging from approximately 778 km² to 51,800 km² (average 11,655 km²), with average elevations from 1,981 m to 3,353 m.
• Temporal Scale:
  • Full period: Water years 1906 through 2020 (115 years).
  • NDVI – SWE period: Water years 1983 through 2020 (38 years).
  • Seasonal analysis for temperature and Normalized Difference Vegetation Index (NDVI) (March-June, March-May, July-September, and fall).

Methodology and Data

• Models used:
  • Bayesian Network (BN) based causality analysis.
  • Principal Component Analysis (PCA).
  • Multiple Linear Regression (MLR) modeling.
  • Pearson's correlation analysis.
  • Wavelet analysis.
• Data sources:
  • Naturalized streamflow: Bureau of Reclamation (water years 1906–2020).
  • Precipitation and Surface Air Temperature: NOAA Monthly U.S. Climate Gridded (NClimGrid), 5 km × 5 km resolution (water years 1896–present).
  • Normalized Difference Vegetation Index (NDVI): Global Inventory Modeling and Mapping Studies - 3rd Generation V1.2 (GIMMS-3G+), approximately 8 km × 8 km resolution (water years 1983–2021).
  • Sea Surface Temperature (SST), Wind, Precipitation and Precipitable Water, Radiation, and Atmospheric Pressure: NCEP—NCAR Reanalysis 1 Data, NOAA Extended Reconstructed SST V5.
  • Snow Water Equivalent (SWE): National Snow and Ice Data Center (NSIDC) gridded, 4 km × 4 km resolution (water years 1982–present).

Main Results

• Surface runoff efficiency (REt) in the UCRB is most strongly influenced by total water year precipitation (Pt).
• In the full period (1906–2020), spring average temperature (TMAMJt) modulates REt, with prior water year precipitation (Pt-1) and prior water year summer temperature (TJASt-1) having weaker influences.
• In the NDVI–SWE period (1983–2020), peak snow water equivalent (SWEMAXt) and spring average NDVI (NDVIMAMt) emerge as primary and secondary determinants of REt, respectively. Pt's influence on REt is mediated by SWEMAXt, which in turn influences spring temperature and subsequently spring vegetation activity.
• Runoff efficiency increases with higher precipitation and snow accumulation, accompanied by cooler spring temperatures and delayed vegetation phenology (attenuated biomass). Conversely, it decreases with lower precipitation and snow, or warmer springs, when vegetation activity is accelerated or amplified.
• Summer temperature (TJASt) does not emerge as a statistically significant direct driver of REt when accounting for winter-spring precipitation, temperature, and vegetation dynamics.
• Years with extreme winter-spring precipitation are associated with distinct large-scale atmospheric circulation patterns and sea surface temperature anomalies, indicating the influence of larger-scale climate drivers.

Contributions

• Provides a data-driven causal inference framework using Bayesian Networks to identify and quantify the direct and mediated drivers of runoff efficiency in the UCRB, offering an alternative to physics-based simulation models.
• Clarifies the hierarchical and multivariate dependence structure of runoff efficiency, establishing precipitation as the dominant driver and highlighting the modulating role of spring temperature and vegetation dynamics through snow water equivalent and spring NDVI.
• Demonstrates that summer temperature, often cited as a key driver of increased evapotranspiration and reduced streamflow, does not hold statistical significance as a direct driver of runoff efficiency when conditioned on winter-spring hydroclimatic and vegetation variables.
• Links basin-scale runoff efficiency dynamics to large-scale ocean-atmosphere circulation patterns and sea surface temperature anomalies during extreme wet and dry years.

Funding

• Bureau of Reclamation

Citation

@article{Palumbo2025Precipitation,
  author = {Palumbo, David and Gangopadhyay, Subhrendu and Lall, Upmanu},
  title = {Precipitation, moderated by spring temperature and vegetation, drives runoff efficiency in the Upper Colorado River Basin, USA},
  journal = {Communications Earth & Environment},
  year = {2025},
  doi = {10.1038/s43247-025-03136-w},
  url = {https://doi.org/10.1038/s43247-025-03136-w}
}