Assessing water quality along rivers is vital for watershed management. The Bayesian Maximum Entropy method of modern geostatistics provides a powerful framework to model the space/time variability of water quality and perform a statistical assessment of all river miles. This talk presents joint work with Prahlad Jat that describes how river distances and flow can be incorporated in the covariance models used to describe the variability of surface water quality. Two case studies are then used to demonstrate the use of these covariance models for the assessment of surface water chloride in Maryland and fecal coliform in North Carolina.