Degree Name

Doctor of Philosophy


School of Health and Society


Epidemiological evidence is growing on the beneficial health effects of residential green spaces. Simultaneously, in the last decade, life course theories among epidemiologists have focussed on exploring potential influences of green spaces on pregnancy outcomes. However, research to assess the impacts of green spaces on perinatal health is still limited. Research on green space and pregnancy outcomes must be translated into evidence that can be adapted and used in the development of planning guidelines and monitoring of their implementation thereafter. This depends, to some extent, upon identifying critical thresholds of green spaces which is likely to promote perinatal outcomes. However, very little is known about critical thresholds of green spaces and role of green space types on birth outcomes. Furthermore, the role of air pollution in this complex association remains unclear. The aim of this thesis, therefore, is to investigate association between residential green space and birthweight by addressing above mentioned gaps in the literature. This thesis analyse association between multiple green space metrics and birthweight, the role of air pollution in this association, and the modification by some of the individual and neighbourhood characteristics. It is our belief that findings of this thesis have policy relevance and will inform planning guidelines for healthier pregnancy outcomes. Birth data was obtained from Perinatal Data Collection. It covered all live births occurred between January 2016 and December 2017 in Sydney Metropolitan Area, New South Wales, Australia. Multiple green space metrics were used in this research. (1) Green space quantity data which was used to answer the second and third empirical question was derived from “parkland” mesh block in the ABS land-use classification for 2011. Parkland mesh block contains areas classified as nature reserves such as and excludes agricultural lands. (2) Data on green space types which was used to answer the third empirical question was obtained from Pitney Bowes Ltd and contained percentages of total green space, tree canopy, open grass excluding tree canopy and shrub. Air pollution data was provided by The Common wealth Scientific and Industrial Research Organisation (CSIRO) and included daily averages of NO2, PM2.5, SO2 and O3. Three empirical studies were conducted. In the first study, mixed effect linear and logistic models accounting for potential covariates were developed to examine association between green space quantity and birthweight, low birthweight, gestational age and preterm. Also, the potential modification effect by area-level socio-economic status and mother’s country on this association was examined by fitting interaction terms and stratified models. In the second study, combined effect of green space and air pollution on birthweight were examined by fitting interaction terms and stratified models. In the final study, effect of green spaces types on birthweight was explored. Furthermore, interaction between green space types and air pollution was investigated by fitting interaction terms and stratified models. The result of the first study demonstrated statistically significant association between green space quantity and birthweight, with stronger association was observed for those lived in areas with more than 40% green space. Interaction between green space quantity and socio-economic status and mother’s country of birth was identified, indicating stronger association between green space quantity and birthweight among overseas born mothers and those lived in affluent areas. Results of the second study revealed statistically significant association between higher levels of NO2, PM2.5 and SO2 and lower birthweight. We observed Interactions between green space quantity and NO2 and PM2.5 (p < 0.05), indicating that increasing green space may help to mitigate harms of moderate and low levels of NO2 and PM2.5 on birthweight. Greener areas were associated with birthweight increase for the areas with low to moderate level of air pollution. In highly polluted areas there was no association between green space and birthweight (p > 0.05). Final study demonstrated that increase in every green space type was associated with increase in birthweight, however, after adjusting model for air pollution the associations attenuated yet remained significant for tree canopy and other vegetation type. We identified interaction effect between green space metrics and air pollutants were identified (p < 0.05), suggesting that association between NO2, PM2.5 and birthweight was weaker in the areas with more tree canopy.

These findings indicate that promoting green areas in the urban area can help birthweight increase. More specifically, our findings suggest that living in areas with at least 40% green space cover would potentially help improving birthweight. However, in the complex relationship of environment and health, simply increasing green space is likely to be insufficient to mitigate air pollution and elicit potential health benefits. Therefore, several multi-scale factors from vegetation type to urban form which is likely to influence the role of green spaces in shaping healthy environments should be carefully considered. Finally, replication of this findings in contrasting settings (e.g., low-middle income countries) are warranted to draw clear conclusion.

FoR codes (2020)

440605 Health geography, 330413 Urban planning and health, 420203 Environmental epidemiology

This thesis is unavailable until Tuesday, January 31, 2023



Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.