A drinking water treatment plant in Viimsi, Estonia was monitored over three years for iron, manganese, radium-226, radium-228, and their daughter nuclides in order to determine the efficiency of the treatment process, get an insight of the removal mechanisms and interactions between radium, iron, and manganese, and assess the overall longevity and performance of the technology and possible build-up of NORM from the treatment process. During the study, samples were collected from raw water, first and second stage filtrate, consumer water, backwash water, and filter materials. The results show a consistent removal efficiency for iron and manganese, as well as an average of over 85% removal for radium with a slight decline with time. Backwash process has been optimized for maximum radium removal from the filters, while keeping the radium concentrations in the backwash water below exemption levels. However, accumulation of radium and thorium occurs in the filter material, exceeding exemption levels in the top layer of the filter columns in less than a year. By the end of the observation period, activity concentrations in the top layer of the columns were above 30 000 Bq/kg for Ra-226 and Ra-228, and around 15 000 Bq/kg for Th-228. Radionuclides are not homogenously distributed in the filter columns. In order to estimate the average activity concentrations in the filter media, the height distribution of radionuclides has to be accounted for. Two years and two months after commissioning of the treatment plant average activity concentrations of Ra isotopes in the filter columns were in a range of 10 000 Bq/kg while Th-228 activity concentration was roughly 3500 Bq/kg.