Constraining topsoil pesticide degradation in a conceptual distributed catchment model with compound-specific isotope analysis (CSIA)

Abstract

Predicting pesticide dissipation at the catchment scale using hydrological models is challenging due to limited field data distinguishing degradative from non-degradative processes. This limitation hampers the calibration of key parameters, such as biodegradation and volatilisation half-lives (DT50) and the carbon-water partition coefficient (KOC), often leading to equifinality and reducing confidence in predictions of pesticide persistence in topsoil and transport from agricultural fields to catchment outlets. This study examines the use of pesticide compound-specific isotope analysis (CSIA) data to improve model predictions of pesticide persistence in agricultural topsoil and off-site transport at the catchment scale. The study was conducted in a 47 ha crop catchment using the pre-emergence herbicide S-metolachlor. A new conceptual distributed hydrological model, PiBEACH (Pesticide isotope BEACH (Bridge Event And Continuous Hydrological)), was developed to simulate daily pesticide dissipation in soils and its transport to surface waters. The model integrates changes in the carbon isotopic signatures (δ13C) of S-metolachlor during degradation to constrain key parameters and reduce equifinality. Model and parameter uncertainties were estimated using the generalised likelihood uncertainty estimation (GLUE) method. Incorporating δ13C data and S-metolachlor concentrations from topsoil samples reduced the uncertainty in the estimated degradation half-life, DT50, by more than half, yielding a value of 18 ± 4 d. This approach also significantly decreased uncertainty in six key metrics of pesticide persistence and transport. Between the day of application (day 0) and day 115, the modelled mass balance components, ranked by relative contribution, were as follows: degradation accounted for the majority at 82 % ± 21 %, followed by the remaining bioavailable mass in the topsoil at 12 % ± 8 %. Leaching contributed 4 % ± 17 %, while export to the river outlet accounted for 2 % ± 6 %. The irreversibly sorbed mass represented 1.1 % ± 2.0 %, and volatilisation was minimal (<1 %). The results highlighted the fact that moderate targeted sampling efforts can identify degradation hotspots and hot moments in agricultural soil when stable-isotope fractionation is integrated into the model. Overall, integrating CSIA data into the PiBEACH model significantly enhances the reliability of pesticide degradation predictions at the catchment scale. In addition, PiBEACH, which accounts for spatial and seasonal variations in topsoil pesticide concentrations, enables coupling with distributed event-based hydrological models such as OpenLISEM-pesticide (OLP) to capture intra-event pesticide transport dynamics more accurately.