A biogeochemical approach to build a perceptual hydrological model for a small peri-urban catchment

<p>The origin and transport of water in peri-urban catchments is complex to model as they are affected by multiple anthropogenic modifications of water pathways (surface imperviousness, sewer overflow releases&#8230;), especially in a context of fast growing urbanization. The hydrological dynamics are also&#160; impacted by natural and agricultural land use patterns. Perceptual models aim at reproducing our understanding of a catchment behaviour and can be useful to illustrate the impact of such spatial contrast and human-induced modifications on a catchment hydrological dynamics. Conservative geochemical and microbiological tracers can be linked to the hydrological processes and water pathways to enhance this understanding and to build-up the hydrological perceptual model of a catchment.</p><p>From 2017 to 2019, a monthly monitoring of geochemical and microbiological tracers was conducted at the Ratier catchment (19 km&#178;) near Lyon (France). Surface waters were collected and analysed for major chemical parameters (cations, anions, dissolved organic carbon and conductivity), dissolved metals, stable isotopes (²H et ¹⁸O), and microbial parameters (total bacterial counts, microbial source tracking DNA datasets, species &#8211; specific DNA trackings). Using these datasets, a step-by-step statistical approach was undertaken, and used to build-up the perceptual hydrological model. The main steps were: (1) group correlated biochemical parameters to reduce redundancy in the dataset, (2) compute the main indicators illustrating the hydro-climatologic dynamics during the sampling campaigns (e.g. antecedent index precipitation, average daily flow) based on the hypothesis of a two-component catchment (groundwater and subsurface flow), and (3) perform a principal component analysis to link the biogeochemical dataset to the computed hydro-climatologic indicators and the runoff processes.</p><p>Results revealed a differentiation of the datasets in two groups matching groundwaters and subsurface waters. Groundwaters showed two geochemical profiles linked to the two main geological formations of the catchment. Subsurface waters showed more variable biogeochemical patterns highly influenced by land use and soil properties. This step-by-step statistical approach led to a better understanding of the dynamics of the water pathways and these insights were then used to build-up the hydrological perceptual model of the catchment. As a next step, such a model should help in the evaluation and improvement of a distributed hydrological model.</p>