A bstract Continuous high frequency water quality monitoring is becoming a critical task to support water management. Despite the advancement s in sensor technologies, certain variables cannot be easily and/or economically monitored in - situ and in real time. In these cases, surrogate measures can be used to make estimations by means of data - driven models. In th is work, variables that are commonly measured in - situ are used as surrogates to estimate the concentration s of nutrients in a rural catchment and in an urban one, making use of machine learning models, specifically Random Forests . The results are compared with those of linear modelling using the same number of surrogates, obtaining a reduction in the Root Mean Squared Error (RMSE) of up to 60.1% . Th e profit from including up to seven surrogate sensors was computed, concluding that adding more than 4 and 5 sensors in each of the catchments respectively was not worthy in terms of error improvement. Keywords water monitoring, water quality, surrogate parameters, random forests, soft - sensors, machine learning 2 1. Introduction Waterb odies must maintain a good ecological and chemical status in order to protect human health, preserve water supply and safeguard natural ecosystems and biodiversity. The assessment of the ecological status of these waterbodies in a coherent and comprehensiv e way would benefit from improving water quality monitoring progra mmes (Voulvoulis et al., 2017) . To date, many substa nces like major nutrients (nitrogen (N) and phosphorus (P)) are mostly monitored by means of analytical discrete campaigns with low sampling frequenc y . Nutrient monitoring is of great importance to reduce the risk of eutrophication, a water quality problem that leads to numerous negative impacts like public health issues, fish mortality and unhealthy ecosystems, among others .