Noise Emission Approximation through Open Geospatial Data

Traffic noise is one of the two biggest environmental health burdens in Europe. Excessive and chronic noise exposure leads to serious illnesses and impairs people’s general quality of life. Therefore it is necessary to establish a comprehensive quantification procedure that not only helps to monitor the situation but also serves as a basis for action planning. Developed approaches of the European Noise Directive 2002/49/EG turned out to be insufficient in this context, as they only cover a limited area of specific localities and exclude a majority of the population. The objective of this thesis is to contribute to an ongoing research project funded by the German Federal Environment Foundation (DBU) in cooperation with the German Aerospace Center (DLR), called "Mapping Noise Propagation From Space", which aims at developing a cost-effective modeling process for comprehensive noise mapping. However, decent road noise emission values that can be integrated into the model are still missing, at which point this thesis ties in. The approach is to develop a Land Use Regression (LUR) model to predict missing road noise emissions at the example location of the German city of Koblenz (Germany, Rhineland-Palatinate), by applying a multiple linear regression. Aggregated road traffic noise immission data consisting of the day–evening–night noise level indicator Lden is the dependent variable, and information derived from publicly available data is used as predictors. The main sources are the database OpenStreetMaps (OSM) and various freely available Open Government Data (OGD). As a result of an iterative pre-selection process and multiple imputation for missing values in the OSM overall data, a model was created consisting of seven different predictor variables. With an R2 of 0.74 and a standard error of 6.99, the result finally leads to a road noise emission approximation of 26 percent in total