Nutrient pollution represents one of the costliest and most challenging threats to water quality in the United States and is a major problem elsewhere in the world. Water quality trading is drawing significant interest from policy makers as a mechanism to manage nutrient loads cost-effectively. Overlooked in nutrient markets developed to date and in standard guidance for market design are the fundamental dynamics of nutrient pollution delivery processes. This paper develops rules to efficiently manage pollution trading between sources that differ in terms of how fast their pollution arrives at degraded waters. We develop a system of optimal time-varying discharge permits and lag trade ratios, as well as a simpler, second-best system of time-invariant permits and trade ratios that achieve optimal pollution loads in the long run steady state. Our second-best system accounts for lags in a way that is both pragmatic and grounded in economic theory. Results from a numerical simulation indicate that the total costs of the second-best system are within 0.2% of the first-best optimum, suggesting that the efficiency loss associated with the second-best design may therefore be of little practical concern.