<p>Eutrophication, driven by urbanization and industrialization, poses a serious threat to water bodies. Although numerical models are valuable tools for investigating water quality dynamics, their effectiveness is limited by infrequent water sampling. This study presents a framework combining Generalized Additive Models (GAMs) and the Environmental Fluid Dynamics Code (EFDC) to assess the seasonal contributions of total nitrogen (TN) and phosphorus (TP) loads into the Des Hurons River (DHR), the main tributary of Lake St. Charles, Quebec City’s primary source of drinking water. GAMs were utilized to reconstruct missing water quality data from both point sources (the Stoneham-and-Tewkesbury wastewater treatment plant (WWTP)) and non-point sources (sub-watersheds with multiple individual septic tank systems). These estimates were then used to develop a 1D EFDC model, simulating hydrodynamics and water quality dynamics along the DHR between 2011 and 2017. GAMs effectively captured seasonal patterns of ammonia nitrogen (NH₄) and dissolved oxygen (DO), though their performance declined for TP and nitrite/nitrate (NO₂/NO₃) due to limited observations. Meanwhile, the EFDC model successfully reproduced temperature and DO dynamics, with reduced accuracy for NH₄, TP, and NO₂/NO₃. Winter was identified as the most critical period, when TN and TP concentrations at the DHR outlet exceeded water quality standards. These winter exceedances were likely associated with higher seasonal nutrient concentrations in the WWTP effluent, suggesting a stronger seasonal influence of this point source during winter. Additionally, annual load quantifications indicated that point sources contributed about ≈15%, while the majority originated from non-point sources influenced by meteorological conditions. This study demonstrates the usefulness and feasibility of the combined GAMs-EFDC framework in a data-limited region and highlights the need for better winter monitoring.</p>