Strategic Noise Mapping with GIS for the Universitat Jaume I Smart Campus: best methodology practices

Dissertation submitted in partial fulfillment of the requirements for the Degree of Master of Science in Geospatial Technologies.Noise is a type of pollution often overlooked in conversations about pollution, which usually center on air, water and waste management. However, it has not been missed by decision makers in the European Union (EU). There are laws to keep noise levels down, and schools are a target specifically mentioned in the European Environmental Noise Directive (END). Strategic noise mapping can identify problem areas and help evaluate situations. This thesis project explores and compares various approaches in an attempt to offer useful information to the noise mapping field based on the results of the analysis. The measurements used commonly in studies are taken by professionals using professional equipment. Either teams physically enter the environment to manually take measurements or they collect data wirelessly from fixed sensors. Both of these methods are expensive due to the manpower or equipment. In addition, these methods are limited in the number of measurements in space and time that they can represent. One option is to use citizens with smart phones to record noise measurements. Involving the public to gather information is commonly called crowdsourcing, Volunteered Geographic Information (VGI) or Public Participatory GIS (PPGIS). Three applications for Android smart phones were tested and compared to a certified, calibrated professional sound level meter. Also, mapping noise by taking sample noise measurements without also mapping noise sources may not provide the full picture. The second objective of this thesis was to apply sound attenuation and combination rules in ArcGIS to create a noise source map and compare the results to the common spatial interpolation methods. The comparisons of smart phone measurements with the professional sound level measurements revealed that they are not comparable quality. Each ANOVA and t-Test revealed statistically significant differences. This is mostly attributed to the phone’s hardware, which varies between mobile device models and versions. The geostatistical interpolation tools delivered noise maps which had similar accuracy rates for predicting measurement points according to the cross validation methods used. The best (most accurate) prediction model was indeed the kriging method. The author successfully applied sound attenuation equations to create a multiple noise source propagation and combination interpolation toolset in ArcGIS. This can be used for an infinite number of noise sources. The fit of the actual measurement points in the noise source attenuation noise map was very similar although slightly higher than that of to the geostatistical methods