Characterizing fast herbicide transport in a small agricultural catchment with conceptual models

JOURNAL OF HYDROLOGY

卷:586

Article

Ammann, Lorenz^1,2  Doppler, Tobias³  Stamm, Christian¹  Reichert, Peter^1,2  Fenicia, Fabrizio¹ 

[1] Eawag Swiss Fed Inst Aquat Sci & Technol, Dubendorf, Switzerland

[2] Swiss Fed Inst Technol, Dept Environm Syst Sci, Zurich, Switzerland

[3] Swiss Water Assoc, VSA, Glattbrugg, Switzerland

关键词: Pesticide transport;    Experimentalist knowledge;    Controlled application;    Conceptual model;    High-frequency concentration data;    Bayesian inference;   

DOI  :  10.1016/j.jhydrol.2020.124812

来源: Elsevier

【 摘 要 】

Herbicide pollution in headwater streams due to agricultural practices is a major environmental concern and is characterized by episodic peak concentrations from fast transport paths. We rely on previous experimental studies in a small (1.2 km(2)) agricultural catchment in the Swiss Plateau, to model dynamic diffuse herbicide pollution with emphasis on fast transport paths in a conceptual modelling framework at the catchment scale. We show how experimentalists' understanding of the fate of herbicides (perceptual model) can be translated into conceptual models considering sorption, degradation, and fast transport of water and chemicals facilitated by impervious surfaces, file drains and artificial shortcuts. Different types of experimental data (streamflow, high-frequency concentration measurements, and soil-water distribution coefficients) are used in a joint Bayesian inference of model parameters. We assess the ability of different spatial configurations of hydrological response units in explaining observed heterogeneity in transport behaviour of two corn herbicides. Thereby, we find that (1) relatively simple conceptual models can provide a realistic description of herbicide fate in small agricultural catchments, (2) accounting for spray drift onto hard surfaces is necessary to avoid a severe model bias during the first rainfall event after application, and (3) including catchment-specific experimentalist knowledge about important elements like artificial shortcuts and file drains leads to a reduction in uncertainty of 30% compared to the more conventional conception of the proximity to the stream as the dominant risk factor.

【 授权许可】

Free