Abstract. Most studies considering water quality pollution in the upper reaches of the uMngeni Catchment have relied on the 10 physical grab sampling of water and the subsequent laboratory analysis of chemical determinants. However, this provides limited spatial and temporal information. Thus, the objectives of this study are to assess the capability of the Hydrological Predictions for the Environment (HYPE) model in simulating streamflow, dissolved inorganic nitrogen (DIN) and total phosphorus (TP), in the fast developing uMngeni Catchment in KwaZulu-Natal province, South Africa. The model was set up and calibrated, following a stepwise approach and then validated. Results indicated that the simulation of discharge is most 15 sensitive to the parameters related to evapotranspiration and the water-holding capacity of the soil, while DIN and TP are affected by plant uptake and initial pools of nutrients. DIN is also affected by denitrification. Runoff was captured well during the calibration (1989)(1990)(1991)(1992)(1993)(1994)(1995) and validation periods , with a Nash-Sutcliffe efficiency (NSE) greater than 0.0 in eight of the nine stations and a Pearson's correlation coefficient (r) of > 0.5 at all the sub-catchments. High streamflow events were represented well, low streamflows were over-simulated. The accumulative streamflows were over-predicted in the 20 downstream sub-catchments, with an absolute percentage of bias (PBIAS) of > 25 %. The transport and dynamics of DIN and TP vary differently and they are driven by hydrological and biochemical processes. The concentration of TP follows the pattern of the streamflow, whereas DIN shows an inconsistent variation. The values of DIN decrease from upstream to downstream, while the TP values increase from the headwaters to the outlet of the catchment. Agricultural activities were found to be the largest source of DIN, while the TP is mainly ascribed to the point sources of pollution. 25