A predictive runoff rate model for copper has been refined and used to generate copper runoff maps for Europe. The new model is based on laboratory and field runoff data and expresses the runoff rate R (g m⁻² yr⁻¹) through two contributions, both with a physical meaning:Input parameters are the SO₂ concentration (μg m⁻³), pH, amount of rain (mm yr⁻¹), and surface angle of inclination (θ). The first contribution originates from dry periods between rain events (the first-flush effect) and the second from the rain events. The dry term has been refined in comparison to the original model by assuming a mass balance between measured corrosion mass loss, calculated copper retention in the patina and predicted copper runoff. The refined model predicts 76% of all reported runoff rates, worldwide, within 35% from their measured value. This includes sites with low SO₂ concentration, where the original model erroneously predicted higher runoff rates than corrosion rates. Based on environmental data from the EMEP programme for the years 1980–2000, the new model has been used to derive runoff rate maps for Europe with 50 × 50 km grid resolution. The runoff mapping shows a substantial reduction in runoff rate over the investigated time period, and with copper runoff rates now generally less than 2 g m⁻² yr⁻¹.