Grandjouan et al. (2026) An original approach combining biogeochemical signatures and a mixing model to discriminate spatial runoff-generating sources in a peri-urban catchment

Identification

• Journal: Hydrology and earth system sciences
• Year: 2026
• Date: 2026-02-03
• Authors: Olivier Grandjouan, Flora Branger, Matthieu Masson, Benoît Cournoyer, Nicolas Robinet, Pauline Dusseux, Angélique Dominguez Lage, Fabien Thollet
• DOI: 10.5194/hess-30-591-2026

Research Groups

• INRAE, UR Riverly, Villeurbanne, France
• Univ Lyon, UMR Ecologie Microbienne (LEM), Université Claude Bernard Lyon 1, VetAgro Sup, France
• UMR CNRS 5194 Pacte, Université Grenoble Alpes, Cermosem, Mirabel, France
• Institut d’Urbanisation et de Géographie Alpine, Université Grenoble-Alpes, CNRS, PACTE, Grenoble, France
• INSA Lyon, DEEP, UR7429, Villeurbanne, France (present address of the first author, also associated with PhD funding)

Short Summary

This study developed an original approach combining biogeochemical signatures and a Bayesian mixing model to spatially decompose streamflow into eight runoff-generating sources in a peri-urban catchment, revealing highly variable source contributions influenced by hydro-meteorological conditions and land use.

Objective

• To identify runoff-generating sources linked to both vertical and spatial characteristics (e.g., geology, land use) in a small peri-urban catchment and estimate their contributions to streamwater under contrasted hydro-meteorological conditions using a biogeochemical mixing model.

Study Configuration

• Spatial Scale: Ratier peri-urban sub-catchment (19.8 km²), Mercier stream (tributary, 7.8 km²), located west of Lyon, France. Perceptual model improved at the hillslope scale.
• Temporal Scale:
  • Source sampling campaigns: 8 campaigns between February 2022 and March 2023.
  • Streamwater sampling during hydrological events: 6 events sampled between March 2019 and March 2023.
  • Streamwater sampling during dry weather: Monthly monitoring from March 2017 to December 2019.
  • Hydro-meteorological data records: Rainfall and air temperature since 1997 (Pollionnay), 2005 (Col de la Croix du Ban), 2009 (Col de la Luère); continuous hydrological monitoring since 1997 (Ratier) and 2010 (Mercier).

Methodology and Data

• Models used:
  • Bayesian mixing model (MixSIAR package in R) for hydrograph separation.
  • Hierarchical Clustering Analysis (HCA) for sample classification.
  • Kruskal-Wallis test and Dunn post hoc test for parameter discrimination.
  • Linear Discriminant Analysis (LDA) coupled to a Wilks lambda approach for tracer selection.
  • Perceptual hydrological model (improved).
• Data sources:
  • Biogeochemical data: 35 parameters including major ions (Ca²⁺, K⁺, Mg²⁺, Na⁺, Cl⁻, SO₄²⁻), silica (SiO₂), 15 trace metal elements (Al, As, B, Ba, Cd, Co, Cr, Cu, Li, Mo, Ni, Pb, Rb, Sr, Ti, U, V, Zn), Dissolved Organic Carbon (DOC), UV-Visible indicators (S1, S2), High Pressure Size Exclusion Chromatography (HPSEC) indicators (Mn-254, A0-254, A1-254, A2-254, A3-254), and 2 host-specific microbial DNA targets (HF183, rum-2-bac).
  • Source samples: 38 water samples from 8 identified runoff-generating sources (Forest-1, Forest-2, Grassland, Agriculture, Colluvium Aquifer, Urban and Road Surface Runoff, Sewer System, Quick Surface Runoff).
  • Streamwater samples: 10 to 12 samples per hydrological event (6 events) and 24 samples for dry weather conditions, collected at Mercier and Ratier outlets.
  • Hydro-meteorological data: Rainfall and air temperature from pluviometric stations (Pollionnay, Col de la Croix du Ban, Col de la Luère); discharge from gauging stations (Mercier, Ratier).
  • Geospatial data: Geological, field capacity, land use, agricultural activities, and sewer system maps of the catchment.
  • Rainwater composition data: Used to infer Quick Surface Runoff (SUR) signature (Pollionnay and Ecully pluviometric stations).

Main Results

• Eight runoff-generating sources were identified: Forest-1 (FOR-1), Forest-2 (FOR-2), Grassland (GRA), Agriculture (AGR), Colluvium Aquifer (AQU), Urban and Road Surface Runoff (URB), Sewer System (SEW), and Quick Surface Runoff (SUR).
• A selection of 15 biogeochemical tracers (7 major parameters, 6 dissolved metals, 2 Dissolved Organic Matter (DOM) characteristics) was sufficient to discriminate these sources.
• Dry weather conditions: The AQU source was predominant in the Ratier catchment (up to 85 % of total runoff, >500 m³ d⁻¹), while AGR (up to 40 %) and GRA (up to 50 %) dominated the Mercier catchment. SEW contributions were significant in both catchments (10-50 % in Mercier, <10 % in Ratier), with estimated volumes ranging from 30 m³ d⁻¹ (low flow) to 1000 m³ d⁻¹ (high flow).
• Hydrological events:
  • Small winter events: FOR-1/FOR-2 were major contributors in Mercier (up to 31 %), while URB (up to 38 %) and SUR (up to 35 %) were significant at Ratier. SEW contributions remained stable (12-26 % in Mercier, 4-13 % in Ratier).
  • Summer storm events: GRA, URB, and SUR were predominant (>40 %), with SUR contributions varying widely (20-65 % in Ratier).
  • Major events: AGR was the predominant source (33-66 %), with significant SUR (up to 47 %) and URB (up to 35 %) contributions. SEW volumes reached 900-2000 m³ during major events.
• Runoff generation and source activation are strongly dependent on land use, season, event type, rainfall spatial variability, geological/pedological characteristics, and urban water infrastructures.
• Spatially limited sources like urban areas and sewer systems can dominate streamflow composition and contaminant fluxes despite their limited spatial extent.
• An improved perceptual hydrological model for the Ratier and Mercier catchments was proposed, integrating both natural and anthropogenic controls on runoff generation.

Contributions

• Developed an original approach for spatial decomposition of streamflow in a peri-urban catchment using a Bayesian mixing model with classical and innovative biogeochemical tracers.
• Successfully identified and discriminated eight runoff-generating sources based on geological, pedological, land use characteristics, and anthropogenic infrastructure.
• Quantified the highly variable contributions of these sources to streamwater under contrasting hydro-meteorological conditions, highlighting the dynamic nature of runoff generation.
• Improved a previously designed perceptual hydrological model at the hillslope scale, providing a better understanding of peri-urban catchment hydrological behavior.
• Demonstrated the potential of tracer-based mixing models to validate spatially distributed hydrological models and anticipate the influence of land use, urbanization, and climate change on runoff generation processes.
• Emphasized the critical role of hydrological connectivity, water management infrastructure, water storage capacity, and water transit times in controlling source activation and streamflow composition, beyond simple area-weighted parametrizations.

Funding

• PhD funding by EUR 45 H2O Lyon.
• CHYPSTER research project, funded by the French National Research Agency (ANR-21-CE34-0013-01).
• IDESOC project, granted by the ZABR – Rhone Basin LTSER within the Water Agency RMC – RB LTSER funding agreement.

Citation

@article{Grandjouan2026original,
  author = {Grandjouan, Olivier and Branger, Flora and Masson, Matthieu and Cournoyer, Benoît and Robinet, Nicolas and Dusseux, Pauline and Lage, Angélique Dominguez and Thollet, Fabien},
  title = {An original approach combining biogeochemical signatures and a mixing model to discriminate spatial runoff-generating sources in a peri-urban catchment},
  journal = {Hydrology and earth system sciences},
  year = {2026},
  doi = {10.5194/hess-30-591-2026},
  url = {https://doi.org/10.5194/hess-30-591-2026}
}