Antimicrobial resistance (AMR) is a growing global health threat driven by antibiotics misuse in medicine, agriculture, and aquaculture. The One Health framework highlights the need to understand AMR dissemination across human, animal, and environmental interfaces. Lakes receiving treated and untreated wastewater may be potential AMR hotspots, yet the spatial and temporal dynamics of AMR contamination in lakes remain poorly understood. This study investigates AMR patterns in Vidy Bay, Lake Geneva - a potable water source and recreational site influenced by the Lausanne sewage treatment plant (STEP) effluent. Because in lakes effluent disperses three-dimensionally, modulated by hydrodynamics and thermal stratification, sampling was guided by a hydrodynamic model using Lagrangian particle tracking and temperature profiles. Water samples were collected at 2, 25, and 30.5 m depths, at various distances from the outfall (m to km scale), and across seasons. Samples were processed and analyzed using qPCR and 16S rRNA amplicon sequencing. The AMR indicator (intI1) and clinically relevant antibiotic resistance genes (ARGs) were at least 100-fold higher in STEP effluent compared to lake samples. ARG abundances in the lake were elevated near the wastewater discharge at depths predicted by lake stratification and transported in the model-predicted direction. Wastewater-associated phyla showed similar patterns. Stratification and seasonal effects shaped ARG distributions and bacterial communities. While the model helped predict plume movement, it failed to capture details of spatial dispersion, highlighting the complex lake hydrodynamics and importance of integrating modeling with empirical data to assess environmental AMR risks.