At least one million healthy life years are lost every year from traffic-related noise in the western part of Europe according to the World Health Organization. Other diseases have been linked with environmental noise, such as sleep disturbance, heart illnesses or tinnitus. The Environmental Noise Directive 2002/49/EC (END) and the CNOSSOS-EU framework are the main instruments provided by the European Union (EU) to identify and combat the noise pollution in the European countries. The END asks the EU member states to publish noise maps and the consequent action plans every five years. In order to automatize the process of noise monitoring, several projects are aimed at implementing a noise-mapping system by means of a wireless acoustic sensor network. However, to analyze the impact of traffic noise on public health, noise events non-related to traffic should be removed from the noise map computation since they may bias significantly the result. This work evaluates the impact of individual anomalous noise events on the equivalent sound level computation in relation to their Signal-to-Noise Ratio (SNR) and duration. The analysis is conducted within the framework of the LIFE DYNAMAP project, which is aimed at representing the acoustic impact of road infrastructures through dynamic noise mapping. The experiments consider four hours of real-life acoustic data recorded in the urban area of Milan (Italy), defining as critical those anomalous noise events exceeding the A-weighted equivalent sound level computation by more than 2 dB in a 5-min integration time. The results evidence, on the one hand, the presence of critical anomalous noise events, and on the other hand, prove that both SNR and duration are relevant parameters to characterize these events, which should be removed to avoid biasing subsequent health studies or action plans.