Urban-STA4CLC: Urban Theory-Informed Spatio-Temporal Attention Model for Predicting Post-Disaster Commercial Land Use Change
Ziyi Guo, Yan Wang · 2025
Read Urban-STA4CLC: Urban Theory-Informed Spatio-Temporal Attention Model for Predicting Post-Disaster Commercial Land Use Change by Ziyi Guo, Yan Wang (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-08976.txt.
Introduction
Ziyi Guo and Yan Wang build a model to predict how disasters reshape commercial land use in a 2025 paper. Hurricanes and wildfires disrupt economic activity more and more, and shops and services can be hit hard because they depend on customer visits. Past models capture poorly how people and land change affect each other.
Resilient urban planning theory captures those links better. The authors design a spatio-temporal attention model based on that theory and tune it for post-disaster settings. Planners could use such models to guide aid to rebuild toward the commercial areas most likely to recover or fail.
Catalog record
| CREATOR | Ziyi Guo, Yan Wang |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 47 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-08976
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Urban-STA4CLC: Urban Theory-Informed Spatio-Temporal Attention Model for Predicting Post-Disaster Commercial Land Use Change Ziyi Guo, Yan Wang Abstract . Natural disasters such as hurricanes and wildfires increasingly introduce unusual disturbance on economic activities, which are especially likely to reshape commercial land use pattern given their sensitive to customer visitation. However, current modeling approaches are limited in capturing such complex interplay between human activities and commercial land use change under and following disturbances. Such interactions have been more effectively captured in current resilient urban planning theories. This study designs and calibrates a U rban Theory-Informed S patio- T emporal A ttention Model for Predicting Post-Disaster C ommercial L and Use C hange (Urban-STA4CLC) to predict both the yearly decline and expansion of commercial land use at census block level under cumulative impact of disasters on human activities over two years. Guided by urban theories, Urban-STA4CLC integrates both spatial and temporal attention mechanisms with three theory-informed modules. Resilience theory guides a disaster-aware temporal attention module that captures visitation dynamics. Spatial economic theory informs a multi-relational spatial attention module for inter-block representation. Diffusion theory contributes a regularization term that constrains land use transitions. The model performs significantly better than non-theoretical baselines in predicting commercial land use change under the scenario of recurrent hurricanes, with around 19% improvement in F1 score (0.8763). The effectiveness of the theory-guided modules was further validated through ablation studies. The research demonstrates that embedding urban theory into commercial land use modeling models may substantially enhance the capacity to capture its gains and losses. These advances in commercial land use modeling contribute to land use research that accounts for cumulative impacts of recurrent disasters and shifts in economic activity patterns. 1. Introduction Natural disasters present significant challenges to urban systems by causing damage to physical built environment and affecting human activities ( Guo et al. 2025 ) . The short-term consequences of recurrent or compound shocks can be accumulated and lead to long-term transformations in urban form ( Safabakhshpachehkenari et al. 2025 ; Nguyen et al. 2024 ) . This phenomenon is particularly pronounced in commercial land use change in coastal areas. Natural disasters introduce additional challenges, including disrupted physical infrastructure, raised maintenance and insurance cost, and declined customer activities ( Guo et al. 2025 ; Guo and Wang 2025 ) . These factors can undermine commercial viability and market value, which potentially leads to temporary or permanent decline in commercial land use if not effectively mitigated. For instance, following Hurricane Ian’s landfall on Florida, U.S. in 2022, coastal commercial districts struggled with prolonged recovery periods, with some areas facing vacancy even after two years ( Florida 2025 ) . The decline of commercial areas is often regarded as both a warning sign and a driver of broader cascading deterioration in local economy and vitality ( Yin et al. 2011 ) . With the expectation of recurrent hurricanes in the coastal communities, accurate modeling the impact of natural disasters to land use change is has become essential to meet several critical planning needs. First, disaster preparation planning demands for early detection of systematic vulnerabilities in commercial areas that lie within and beyond traditional hazard maps, which eventually inform efficient resource allocation ( Guo et al. 2025 ) . Second, scenario-based modeling is needed to anticipate continued and unforeseen disaster impacts, informing long-term adaptive land use and economic development planning ( Hao et al. 2024 ) . Third, it is necessary to evaluate the effectiveness of proposed planning strategies aiming at proactive and resilient urban development, such as infrastructure development or zoning adjustments, prior to implementation under disasters impact ( Linardos et al. 2022 ) . Despite the abundant modelling effort that intend to capture the land use change, the integration of recurrent disaster’s impacts on land use change with a focus on commercial land use has been rare. Changes in commercial land use are unique for two main reasons. First, they exhibit particular sensitivity to fluctuations in customer demand, especially in the context of disasters ( Guo et al. 2025 ; Colaço and De Abreu E Silva 2024 ) . Second, commercial land uses tend to cluster together, forming functional patches ( Ballantyne et al. 2022 ) . Their changes are often influenced by spatial spillover effects, including both negative externalities from nearby competition and positive agglomeration economies ( Hsu et al. 2025 ; Liu et al. 2024 ) . Thus, a specific-designed modelling approach is needed to capture the pathway of disaster impact on commercial land use through economic activity disturbances, built environment disruption among others. Current land use change modeling, which predominantly relies on conventional approaches such as cellular automata (CA) and system dynamic model (SDM), remains limited in its ability to capture changes in built-up land uses or atypical transition pathways ( Liu et al. 2021 ; Dede et al. 2021 ) . These models also tend to focus on the slow-changing built environment and are insensitive to temporal changes in visitation within a year, such as weekly or monthly fluctuations, which can also influence land use change. Moreover, while some large-scale models incorporate disasters, they often treat them merely as initial conditions or boundary constraints, rather than fully integrating their dynamic role in shaping land use change ( Mardalena 2024 ; Wang and Yan 2023 ) . Figure 1. Architecture of Multi-Relational Spatio-Temporal Attentive Network The growing application of deep learning techniques promise to capture the complex relationship between socio-economic factors, built environments, natural disasters, human activity change and land use patterns across space and time ( Colaço and De Abreu E Silva 2024 ; Hao and Wang 2023 ; Hou et al. 2021 ; Khalid et al. 2024 ) . However, most of the extant efforts are detached from existing planning theories. Given the massive, fluctuating, and heterogeneous nature of urban data, general-purpose deep learning models often face difficulties in extracting meaningful patterns tailored to specific urban challenges that are frequently better understood through theoretical frameworks. In particular, general-purpose models often struggle to capture the complex interplay among disasters, human activity, and land use change due to several key theoretical challenges. The theory of resilience has successfully quantified the impact of disasters using various formulations, such as the exponential decay function ( Chambers et al. 2019 ; Wang et al. 2020 ; Wang et al. 2018 ) . General temporal mechanisms might conflate disaster-induced behavioral shifts with regular fluctuations in visitation. Moreover, spatial economic theories highlight competition for customer visitation among same-sector businesses ( Doan and Lim 2016 ) . General deep spatial models typically define neighborhoods using physical adjacency alone, overlooking functional linkages between business districts that competition effects ( Kipf and Welling 2017 ; Veličković et al. 2018 ) . Finally, although land use change frequently exhibits spatial diffusion patterns, existing models rarely differentiate between expansion and decline, nor do they enforce spatial coherence through diffusion-based constraints. These limitations highlight the need for a domain theory-informed designs that explicitly integrates economic interdependence, disaster impact accumulation, and directional propagation into a cohesive predictive model ( Hoffer et al. 2022 ; Kang et al. 2023 ) . To address these challenges, we propose the Urban Theory-Informed Spatio-Temporal Attention Model for Predicting Post-Disaster Commercial Land Use