Spatial accessibility in the 3D space, and the uncertain geographic morphological problem

Abstract

People traverse space to reach opportunities and resources for daily needs since neither the supply nor demands are uniformly distributed. Spatial accessibility to/for resources has thus become an important topic in social equity. However, errors exist when the reality is conceptualised and formalised in the virtual environment in which the presentation and methods are predominantly horizontal. Spatial accessibility investigation commonly ignores the topography and the difference in the response of travel costs to uphill and downhill directions. Such slope-unaware and horizontal approaches thus misestimate accessibility situations, especially in morphologically complex areas and when walking is modelled, to which the slope is highly sensitive. Current discourses mainly focus on the deficiency of the conventional usage of 2D measures and the disparity in travel costs between initial and return trips. This thesis aims to contribute to the ontology of imperfection when spatial accessibility is operationalised and provide an illustrative example of operationalising spatial accessibility in a volumetric way. Therefore, this study thus includes two empirical studies. The first study modelled the building-level walking accessibility to healthy food outlets and compared the results by employing 2D and 3D computations, least-effort and least-time principles, isotropy and anisotropy, and sequences (outbound first vs. inbound first). Firstly, the 2D approach can overestimate and underestimate travel costs due to origin-network mismatch, with overestimation being more likely in denser areas. Secondly, weighted by population, the 3D least-effort outbound travel time is 2.9 minutes longer for 10% of buildings than inbound travel, and another 10% are 2.5 minutes shorter. Thirdly, on average, inbound (outbound) travel time exceeds return time by 47.9 seconds (33.4 seconds) for 10% of buildings, with another 10% experiencing shorter return times by 48.5 seconds (73.5 seconds). These reflect the deficiency of using one-way outbound or outbound plus return trips to conclude the access situation. Lastly, we propose the Uncertain Geographic Morphological Problem to highlight the spatially dependent errors arising from the domains of data, method, assumption of human-environment interaction, and their junctions. The second study employed the volumetric approach to appraise the fine-scale spatial accessibility to healthy food in an ageing community in Hong Kong. Guided by the principles of least effort, our findings suggest that 95% of older adults can access healthy food within 913.3m due to a compact and transit-oriented built environment. However, nearly half (47%) of older adults encounter difficult pedestrian paths even with the least-effort method. From this, an Aggregated Accessibility Index (AAI) is devised to identify communities requiring improvement in promoting active living for an older person. Site visits were also conducted to validate the AAI and present real-world situations to better articulate the mobility challenges imposed on older adults. This study underscores the instrumental role of advanced spatial data computation in shaping age-friendly communities that prioritise and enhance spatial accessibility to healthy food, advocating for a nuanced approach to urban planning that addresses the diverse needs of ageing populations.