Dissemination of antibiotic resistance in receiving environments under a changing climate: A modeling exercise

Antibiotic resistance in rivers has become a global problem, particularly due to the discharge of wastewater treatment plant (WWTP) effluents into these systems. These effluents contain residual antibiotics, antibiotic-resistance genes (ARGs), and antibiotic-resistant bacteria (ARB). While watershed-scale models are commonly used to address other water quality issues, they have not typically been used to address antibiotic resistance. In this study, we present a new model called SWAT-ARB (SWAT- Antibiotic-Resistant Bacteria) that can simulate antibiotic resistance in E. coli at the watershed scale. SWAT-ARB is an adaptation of the widely-used SWAT (Soil and Water Assessment Tool) model, which is a physically-based, watershed-scale hydrological model. We used SWAT-ARB to study the receiving environments of WWTPs in the Adyar River basin in India, Crab Creek in the United States, and the Upper Viskan basin in Sweden. We analyzed the simulations of resistant fractions (the ratio of resistant E. coli concentration to total E. coli concentration) in the streamflow at different flow levels. We also examined the long-term trends of resistant fractions to understand how rising temperatures may impact resistance. We found that in the Adyar and Crab Creek basins, the resistant fractions were largely influenced by temperature rather than flow and wash-off processes, while in the Upper Viskan basin, the resistant fractions were affected by both temperature and flow conditions. In a simulation where we only increased temperatures by 2 °C in the bacteria sub-routine, we found that the Adyar basin showed a decrease in resistant fractions of up to 17 % in dry conditions, while Crab Creek showed increases of 17.5-24.1 % and Upper Viskan showed increases of 4.6-33.5 % across flow classes. Under future climate scenarios (SSP 2-4.5 and SSP 5-8.5), Adyar's resistant fractions decreased by up to 55.5 % as temperatures approached the bacterial growth inhibition threshold, while Crab Creek's resistant fractions increased by up to 175 % as temperatures remained within the optimal 10-20 °C growth range. Our results suggest that the SWAT-ARB model could be further improved by incorporating temperature-dependent parameters into the resistance simulation component.