AbstractHerbicides, such as MCPA and clopyralid, may be transported to surface waters via runoff, which can have unintended environmental consequences. A split herbicide application strategy, wherein applications are spread across a season, may improve herbicide effectiveness, although impacts of this strategy on runoff mitigation have not been investigated. Therefore, this study aimed to (1) quantify the impact of split‐dose applications of MCPA and clopyralid on herbicide losses in surface runoff and (2) assess the impact of split‐dose applications of MCPA on the quantity and classification of MCPA‐degrading soil bacteria. Intact grassed soil sods were placed in 1 m‐long × 0.25 m‐wide × 0.1 m‐deep laboratory flumes, onto which either MCPA or clopyralid were applied in one full‐dose (13.5 kg MCPA ha−1; 2 kg clopyralid ha−1) or two split‐doses (each 6.75 kg MCPA ha−1; 1 kg clopyralid ha−1) 42 days apart. On days 2, 7 and 21 following herbicide applications, flumes were subjected to controlled rainfall simulations at an intensity of 11 mm h−1, and the herbicides in the runoff were quantified. MCPA and clopyralid concentrations in the runoff were highest immediately after the initial application. Both herbicides were below the limit of detection (0.1 μg l−1 for MCPA and 0.45 μg l−1 for clopyralid) by 44 days. No herbicides were detected in the runoff following the second split‐dose application. For MCPA, this was attributed to an adaptation in the microbial community with the emergence of bacteria possessing the tfdA class III gene in the soil. These results support split‐dose herbicide application as a strategy for agricultural management.