Andréassian et al. (2025) Time shift between precipitation and evaporation has more impact on annual streamflow variability than the elasticity of potential evaporation

Identification

• Journal: Hydrology and earth system sciences
• Year: 2025
• Date: 2025-10-21
• Authors: Vazken Andréassian, Guilherme Mendoza Guimarães, Alban de Lavenne, Julien Lerat
• DOI: 10.5194/hess-29-5477-2025

Research Groups

• Université Paris-Saclay, INRAE, HYCAR Research Unit, Antony, France
• CSIRO, Canberra, Australia

Short Summary

Using 4122 catchments from four continents, this study investigates how annual streamflow variability depends on climate variables (rainfall and potential evaporation) and on the synchronicity between precipitation and potential evaporation. The analysis reveals that, across diverse climates, the time shift between precipitation and evaporation is the second most important factor, after precipitation, in explaining annual streamflow anomalies, significantly improving prediction models.

Objective

• To quantify catchment response to climatic variations, specifically annual streamflow anomalies, by investigating the dependency on rainfall, potential evaporation, and the synchronicity between precipitation and potential evaporation.
• To improve the prediction of streamflow elasticity by introducing anomalies in the synchronicity between precipitation and potential evaporation as a predictor, alongside variability in rainfall and potential evapotranspiration.

Study Configuration

• Spatial Scale: 4122 catchments from nine countries (Australia, Brazil, Denmark, France, Germany, Great Britain, Sweden, Switzerland, United States) across four continents (Europe, Australia, North America, South America).
• Temporal Scale: Hydrological years (1 October to 30 September in the Northern Hemisphere; 1 April to 31 March in the Southern Hemisphere). Average length of catchment time series is 39 years, totaling 162,005 station-years.

Methodology and Data

• Models used: Linear regression models (Ordinary Least Squares - OLS) were used to compute streamflow elasticities. Two main equations were solved on a catchment-by-catchment basis:
  • ΔQ_n = e_Q/P ΔP_n + e_Q/E0 ΔE0_n (classical approach)
  • ΔQ_n = e_Q/P ΔP_n + e_Q/E0 ΔE0_n + e_Q/Λ ΔΛ_n (including synchronicity)
• Data sources:
  • Catchment data from various large-sample datasets (e.g., CAMELS-AUS, CABra, CAMELS-DK, CAMELS-FR, CAMELS-DE, CAMELS-GB, CAMELS-CH, CAMELS-US, CAMELS-SE).
  • Catchments were selected based on having more than 20 complete hydrological years, minimal interannual memory, and minimal regulation (dam storage less than 10 millimeters equivalent volume).
  • Precipitation data: Best quality product recommended by original dataset authors, avoiding exclusively satellite-based estimates.
  • Potential evaporation (E0): Recomputed daily for all catchments using the Oudin et al. (2005) formula, requiring only extraterrestrial radiation and air temperature.
  • Synchronicity index (Λ): A modified version of the seasonality index by de Lavenne and Andréassian (2018), calculated monthly and aggregated annually, expressed in millimeters per year (mm yr⁻¹).

Main Results

• Annual streamflow anomalies (ΔQ) are strongly positively correlated with annual precipitation anomalies (ΔP) across all countries and climate classes.
• Annual streamflow anomalies are clearly negatively correlated with the synchronicity index anomaly (ΔΛ) for all countries, implying that years with lower synchronicity (precipitation and potential evaporation more out of phase) result in greater streamflow.
• The synchronicity index anomaly (ΔΛ) is empirically verified as the second most important factor explaining annual streamflow anomalies, after precipitation, but before potential evaporation.
• Introducing ΔΛ as an additional predictor in the linear regression models improves the prediction of annual streamflow variability, with an average additional explained variance (adjusted R²) of 6% globally, ranging from 3% to 10% depending on the country or climate group.
• The synchronicity anomaly (ΔΛ) provides a statistically significant contribution to the regression for 64% of the catchments, significantly more than potential evaporation (23%).
• The inclusion of ΔΛ does not degrade the physical realism of the elasticity coefficients for precipitation (eQ/P) and potential evaporation (eQ/E0), and slightly increases the proportion of physically realistic eQ/P (from 93% to 94%) and eQ/E0 (from 6% to 11%).
• Humid catchments (Humidity index > 2) generally show less sensitivity to P-E0 seasonality.

Contributions

• Empirically demonstrates, using a large and diverse dataset of 4122 catchments across four continents, that the synchronicity between precipitation and potential evaporation is a crucial, previously underemphasized, factor in explaining annual streamflow variability.
• Quantitatively shows that the impact of the time shift between precipitation and evaporation on annual streamflow variability is more significant than the elasticity of potential evaporation.
• Introduces and validates a novel approach to improve streamflow elasticity prediction by incorporating a synchronicity index (ΔΛ) as an additional, highly effective predictor in linear regression models.
• Provides a robust, data-driven framework for understanding catchment response to climatic variations, with implications for water resource management under changing climate conditions.

Funding

• Agence Nationale de la Recherche (projects CIPRHES ANR-20-CE04-0009 and DRHYM ANR-22-CE56-0007).

Citation

@article{Andréassian2025Time,
  author = {Andréassian, Vazken and Guimarães, Guilherme Mendoza and Lavenne, Alban de and Lerat, Julien},
  title = {Time shift between precipitation and evaporation has more impact on annual streamflow variability than the elasticity of potential evaporation},
  journal = {Hydrology and earth system sciences},
  year = {2025},
  doi = {10.5194/hess-29-5477-2025},
  url = {https://doi.org/10.5194/hess-29-5477-2025}
}