Spatial estimation of veterinary antibiotics from husbandry to soil to river with resistance risk assessment in urban agglomerations

Abstract

The rapid development of husbandry has increased veterinary antibiotic pollution risks, making emission reduction essential for sustainable growth. This study presents a novel method to quantitively explore antibiotic pathways from livestock to soils and then to rivers at grid scale in the Shandong Peninsula urban agglomeration. Validation demonstrated the method can relatively precisely and accurately estimate the spatial distributions and transports of veterinary antibiotics compared to previous methods and observations. Random Forest model yielded robust livestock distribution estimates (R² > 0.945 for internal cross-validation, R² > 0.652 for external validation), with the concentration of bovine, swine, and poultry in the west of the study area. In addition, the spatial distributions of nighttime light index and cropland cover were significantly correlated to the livestock distribution. The spatial distribution of antibiotics in soils followed the livestock distribution pattern, mainly driven by poultry and bovine framing. The estimation of antibiotic loads to rivers, based on sand transport and surface runoff, demonstrated moderate to high accuracy with R² = 0.377 for sulfadiazine (SDZ) and R² = 0.602 for ciprofloxacin (CFX). Antibiotic loads to soils primarily determined its loads to rivers which accounted for only 1.24%–2.68% of total antibiotic loads. Fluoroquinolone levels were one to two orders of magnitude higher than sulfonamide levels. Resistance risks were in the moderate to high for CFX and enrofloxacin (EFX) in a few rivers, while SDZ and sulfamonomethoxine (SMM) had low risks for all rivers. This refined method provides valuable insights for managing livestock-related pollution and offers a scientific basis for regional pollution control, with potential applications in the regions facing similar challenges, such as data scarcity of the soil physicochemical properties.