The distance between noise sources at a large industrial area and a local community can be in the order of several kilometers. At such distances it may not be clear which sources are the main contributors to possible noise complaints. A long-term monitoring project is described that measures the sound sources in the industrial area and the sound in the nearby residential area. This paper focuses on the time varying sound propagation that is needed to determine the industrial source strengths and the relevance of the sources for the nearby community. Data from a meteorological model is combined with measurements from four geographically distributed meteorological masts via data assimilation. In this way the wind and temperature, as a function of height and time, between all possible source and receiver locations can be determined. Next, the corresponding sound propagation for all transfer paths is obtained near real time as these have been calculated beforehand. It will be shown that this monitoring project captures the time varying industrial noise as perceived in the residential area, whereas a standard noise model uses a constant sound propagation based on an average meteorology. This approach makes a comparison with registered complaints over time meaningful.