AbstractGas exchange across the air–water boundary of streams and rivers is a globally large biogeochemical flux. Gas exchange depends on the solubility of the gas of interest, the gas concentrations of the air and water, and the gas exchange velocity (k), usually normalized to a Schmidt number of 600, referred to as k600. Gas exchange velocity is of intense research interest because it is problematic to estimate, is highly spatially variable, and has high prediction error. Theory dictates that molecular diffusivity and turbulence drives variation in k600 in flowing waters. We measure k600 via several methods from direct measures, gas tracer experiments, to modeling of diel changes in dissolved gas concentrations. Many estimates of k600 show that surface turbulence explains variation in k600 leading to predictive models based upon geomorphic and hydraulic variables. These variables include stream channel slope and stream flow velocity, the product of which, is proportional to the energy dissipation rate in streams and rivers. These empirical models provide understanding of the controls on k600, yet high residual variation in k600 show that these simple models are insufficient for predicting individual locations. The most appropriate method to estimate gas exchange depends on the scientific question along with the characteristics of the study sites. We provide a decision tree for selecting the best method to estimate k600 for individual river reaches to scaling to river networks.This article is categorized under: Water and Life > Nature of Freshwater Ecosystems Science of Water > Water Quality Water and Life > Methods