Telehealth has been proposed to reduce rural prenatal care barriers, but its equity effects depend on whether predicted accessibility translates into observed utilization. We examine whether predicted and revealed in-person and telehealth prenatal care accessibility diverge across the urban-rural continuum of South Carolina and whether these measures align with ZIP Code Tabulation Area (ZCTA)-level birth-related need, measured as average annual live births by residence. Using Medicaid claims data for 2020–2021, predicted accessibility was estimated using floating catchment area methods, with the telehealth specification incorporating broadband availability and subscription constraints; revealed accessibility was operationalized as observed prenatal visit volume within 30-min catchments across 424 ZCTAs. Urban-rural differences were assessed with Mann-Whitney U tests, and associations with birth volume were examined using spatial lag and spatial error models. Predicted in-person accessibility did not differ significantly between urban and rural ZCTAs (p = .67), whereas revealed in-person accessibility was significantly lower in rural areas (p < .001), indicating that predicted opportunity did not translate uniformly into observed utilization. Both predicted and revealed telehealth accessibility also showed significant urban-rural disparities (p < .001), with 36.76% of rural ZCTAs recording zero predicted telehealth accessibility compared with 7.53% of urban ZCTAs. These findings show that telehealth expanded potential access but operated within existing in-person care networks and did not equalize access. Improving rural prenatal care access will therefore require stabilizing in-person obstetric capacity, strengthening digital infrastructure, and supporting provider adoption of telehealth.
Data centres are the physical backbone of the digital economy and one of the world's fastest-growing infrastructure classes. Their siting places new demands on land, electricity, water and skilled labour. Place-based policies such as the UK's AI Growth Zones and China's ‘Eastern Data Western Computing’ programme seek to attract, steer or restrain this development, yet the evidence base on what drives location choice is thin. This paper models the spatial determinants of 5821 data centres worldwide using an inhomogeneous Poisson point process combined with density-based clustering (DBSCAN). Data centres are highly concentrated with 76.2% inside 258 clusters at a median nearest-neighbour distance of 0.86 km. Each standard-deviation increase in distance from the nearest airport, city or observed cluster reduces expected density by 46.9%, 43.5% and 23.5%, while geopolitical stability raises it by 16.2%. Submarine cable proximity is weak, and renewable electricity share does not predict location once other factors are controlled. Climate runs counter to expectations. Warmer places with greater cooling costs host more facilities. The results extend the agglomeration thesis of economic geography to the compute layer of the digital economy. For governments that wish to host data centres, cluster-affirming strategies aligned with urban hierarchies, airport access and stable institutions run with the grain of revealed preferences, whereas dispersion strategies premised on cool climates or renewable-rich peripheries face structural headwinds. Findings apply chiefly to multi-tenant colocation facilities, and hyperscale campuses are under-represented.
Understanding temporal patterns is essential for exploring dynamic processes and predicting trends in criminological research. However, a major challenge lies in identifying the coexistence of multiple temporal patterns within criminal processes, which are often influenced by temporal scale and the Modifiable Temporal Duration Problem (MTDP). Existing studies primarily focus on a single temporal pattern, often failing to identify and compare multiple temporal patterns in crime activity. This study proposes a novel data model that converts time into a categorical variable to mitigate the influence of temporal scale. Building on this data model, we develop an innovative method, Time-Constrained Geographically Weighted Co-location Quotient (TC-GWCLQ), to identify and compare both periodic and temporal association patterns in crime incidents. Using crime incident data from Greater Manchester collected between 2014 and 2018, we empirically demonstrate the method's capability to uncover multiple coexisting temporal patterns in theft-from-the-person crime. The results reveal that, in June 2016, periodic association exerted a stronger influence than temporal association. Spatial clustering of these patterns further reveals pronounced spatial heterogeneity across urban areas, indicating that the temporal effects on crime vary across different locations and temporal scales. The proposed TC-GWCLQ framework not only advances the analysis of temporal crime patterns but also provides a scalable exploratory tool for large geospatial datasets, supporting deeper insights into spatiotemporal behaviour in criminology and related disciplines.
Small Island Developing States (SIDS) are increasingly vulnerable to compound marine–terrestrial heatwaves (CMTHWs),a newly recognized form of compound climate extreme that simultaneously affects marine and terrestrial environments. This study presents a comprehensive assessment of CMTHWs across SIDS using a multi-decadal global gridded dataset (1982–2023). Combining event-based heatwave detection with an Event Coincidence Analysis (ECA) framework, we quantify the spatiotemporal characteristics at 25 km pixel resolution, including the frequency, duration, intensity, and population exposure of CMTHWs. Three major findings emerge: (1) Regional-based CMTHW characteristics across SIDS exhibited pronounced temporal variability during 1982–2023, with mean summer frequency, duration, and intensity reaching maxima during 1998–2010 (15.58 events in 1998, 17.08 days in 2006, and 0.69 °C in 2004) before declining to 4.73 events, 8.0 days, and 0.40 °C, respectively, by 2023; (2) Northern Hemisphere SIDS, particularly those in the Caribbean, exhibited relatively high CMTHW probabilities, with Saint Lucia recording the highest probability (51.15%), indicating persistent land–ocean thermal coincidence; (3) Coastal populations in SIDS, including Grenada, Kiribati, and Niue, experienced increasing exposure to CMTHWs, particularly in regions associated with lower humidity anomaly, such as the southern tropical Pacific. These findings improve understanding of the spatial and temporal characteristics of CMTHWs in SIDS and provide a basis for future assessments of compound climate risks in vulnerable coastal regions.
Existing telehealth accessibility models increasingly account for broadband availability, affordability, and demographic variations, yet largely overlook regulatory constraints that determine whether patients may legally access care across administrative boundaries. To address this gap, we proposed a Regulatory Two-Step Virtual Catchment Area (R-2SVCA) model, which encodes licensure laws and regulations into spatial accessibility measures. Using U.S. telepsychology and the Psychology Interjurisdictional Compact (PSYPACT) as a case study, we demonstrated the framework and compared the R-2SVCA model against a traditional borderless model (as the baseline). The results show that regulatory constraints redistributed accessibility along state lines: populations dependent on cross-border providers experienced significant losses in accessibility due to diminished supply, whereas those served by in-state providers gained accessibility as outbound service was constrained. Furthermore, PSYPACT participation inverted the conventional rural-urban accessibility gradient—benefiting rural residents in member states while disadvantaging their counterparts in non-member states. By explicitly accounting for licensure barriers, this framework provides a robust tool for evaluating telehealth policy and strategically targeting interventions in high-need border regions.
