Integrating random forests and propagation models for high-resolution noise mapping

Reductions in community noise levels in vancouver, Canada, during pandemic lockdown and association with land cover type

BACKGROUND

Urban transportation noise is a major public concern because of its adverse effects on health. The determinants of urban noise exposure have not been widely explored but the "natural experiment" presented by the COVID-19 lockdowns presented a unique opportunity. This study examined the relationship between environmental characteristics and urban noise pollution during the COVID-19 related lockdown in Metro Vancouver, Canada, from March 21st to May 18th, 2020.

METHODS

We used noise exposure data from the Vancouver International Airport (YVR) noise management program, comparing the noise levels during "Phase One" of the COVID-19 lockdown in 2020 to the corresponding time period in 2019 from 21 Noise Monitoring Terminals (NMTs) located throughout Metro Vancouver. We modelled the relationship between the change in noise level and the physical NMT environments, including land cover, and total length of roads at four different time periods (24Hr, daytime, evening and nighttime) and within three different buffer zones (100 m, 250 m, and 500 m).

RESULTS

Of 59,472 hourly measurements of community noise, the 24-h noise level was reduced by an average of 2.20 dBA between 2019 and 2020. Higher proportions of greenspace, barren areas, and soil-cover around NMTs resulted in stronger noise reductions and higher density of building, pavement, and water weakened the amount of noise reduction. Proximity of high-volume traffic roads (highways) were associated with weaker noise reduction.

CONCLUSION

The COVID-19 related lockdown was associated with reduced noise in Metro Vancouver, and the relative reduction depended on the types of the environment surrounding the NMT. Future research on the effects of urban environmental characteristics on geographic inequality in noise levels and health consequences of the COVID-19 related lockdown is merited.

Tree characteristics and environmental noise in complex urban settings - A case study from Montreal, Canada

Field studies have shown that dense tree canopies and regular tree arrangements reduce noise from a point source. In urban areas, noise sources are multiple and tree arrangements are rarely dense. There is a lack of data on the association between the urban tree canopy characteristics and noise in complex urban settings. Our aim was to investigate the spatial variation of urban tree canopy characteristics, indices of vegetation abundance, and environmental noise levels. Using Light Detection and Ranging point cloud data for 2015, we extracted the characteristics of 1,272,069 public and private trees across the island of Montreal, Canada. We distinguished needle-leaf from broadleaf trees, and calculated the percentage of broadleaf trees, the total area of the crown footprint, the mean crown centroid height, and the mean volume of crowns of trees that were located within 100m, 250m, 500m, and 1000m buffers around 87 in situ noise measurement sites. A random forest model incorporating tree characteristics, the normalized difference vegetation index (NDVI) values, and the distances to major urban noise sources (highways, railways and roads) was employed to estimate variation in noise among measurement locations. We found decreasing trends in noise levels with increases in total area of the crown footprint and mean crown centroid height. The percentages of increased mean squared error of the regression models indicated that in 500m buffers the total area of the crown footprint (29.2%) and the mean crown centroid height (12.6%) had a stronger influence than NDVI (3.2%) in modeling noise levels; similar patterns of influence were observed using other buffers. Our findings suggest that municipal initiatives designed to reduce urban noise should account for tree features, and not just the number of trees or the overall amount of vegetation.