Release of copper from a naturally aged copper roof on a shopping centre building in a suburban site of Stockholm has been measured during different rain events after its interaction with the internal drainage system and storm drains made of cast iron and concrete. Concentrations of copper removed by means of urban storm water from a nearby parking space have been determined for comparison. Predictions and measurements of the chemical speciation of released copper are discussed compared to the total concentration, and to threshold values for freshwater and drinking water. The results clearly illustrate that the major part of the released copper from the roof is readily retained already during transport through the internal drainage system of the building, a pathway that also changes the chemical speciation of released copper and its bioavailable fraction. Most copper, not retained by cast iron and concrete surfaces, was strongly complexed to organic matter. The median concentration of free cupric ions and weak copper complexes was less than, or within the range of reported no effect concentrations, NOECs, of copper in surface waters. The parking space contributed with significantly higher and time-dependent concentrations of total copper compared to measured concentrations of copper from the roof after the interaction with the drainage system. Most copper in the surface runoff water was strongly complexed with organic matter, hence reducing the bioavailable fraction significantly to concentrations within the NOEC range. Dilution with other sources of urban storm water will reduce the released concentration of copper even further. The results illustrate that already the internal drainage system and the storm drains made of cast iron and concrete act as efficient sinks for released copper which means that any installation of additional infiltration devices is redundant.