Effects of sampling effort on the estimation of spatial gradients in a tropical reservoir impacted by an oil refinery

We investigated the effects of increase in sampling effort (30–1043 sampling points) on the accuracy of assessment of the spatial patterns of surface-water quality in a eutrophic tropical reservoir. The investigation was carried out during the dry season, when previous investigations showed that the spatial heterogeneity is more stable. A multi-parameter Yellow Springs Instruments probe coupled to a TechGeo D-GPS was used. This system is equipped to measure and store in a continuous recording mode, several physical and chemical parameters linked to geographical coordinates obtained with a precision of less than 1 m. We used different geostatistical approaches to determine the optimal number of sampling points required to reflect the real spatial patterns of water quality in the system. This approach was tested in a small tropical reservoir (Ibirité) that receives effluents from an oil refinery (the state-owned REGAP oil refinery, PETROBRAS) located near the city of Belo Horizonte. The study showed not only that the spatial patterns of water quality are significantly affected by sampling effort but also it was demonstrated that the establishment of an adequate sampling program is a critical point for the precise identification of source points of pollution. The results of this investigation enabled us to demonstrate the potential uses and limits of this method for rapid assessment of the water quality of lakes and reservoirs that receive external inputs of water contaminants or nutrients.