Search

• A framework for considering the heterogeneity within land parcels is proposed. • Using hash functions to handle the heterogeneity of driving factors within parcels. • Using GAT to build different neighborhood and obtain interaction relationships. • Different hash functions and neighborhood have a significant impact on the results. Vector cellular automata (VCA) models, which excel at representing spatiotemporal dynamics of irregularly shaped land parcels, have been widely employed in land use change simulations. However, current research faces the following issues: (1) most VCA models neglect the spatial heterogeneity of driving factors within each land parcel when evaluating the environmental driving effects; (2) when calculating the neighborhood effects, simple statistics of land use types in neighboring parcels are often used, overlooking the influence of driving factors within neighboring parcels; (3) the ability to explore the interactions between land parcels is often limited. To address the aforementioned issues, this study proposes a HashGAT-VCA model for investigating urban land use changes. The model utilizes a Hash function to encode the non-uniform distribution of each driving factor within each irregularly shaped land parcel into a fixed-length vector, and constructs a graph structure between land parcels based on their spatial topological relationships. By employing a Graph Attention Network (GAT), the model explores the mechanisms of environmental driving effects and inter-parcel interactions to calculate the probability of land use change for each parcel. The proposed HashGAT-VCA model was applied to simulate urban land use changes in Shenzhen, China, from 2009 to 2012. Compared to other VCA models, the HashGAT-VCA demonstrated higher simulation accuracy. The results indicated that HashGAT-VCA can effectively capture the impacts of the heterogeneously distributed driving factors and the interactions between land parcels on land use changes. Additionally, this study simulated land use patterns for the years 2025 and 2030 under ecological control strategies, providing decision support for urban planning. [ABSTRACT FROM AUTHOR]

Vector-based cellular automata (VCA) models have been applied in land use change simulations at fine scales. However, the neighborhood effects of the driving factors are rarely considered in the exploration of the transition suitability of cells, leading to lower simulation accuracy. This study proposes a convolutional neural network (CNN)-VCA model that adopts the CNN to extract the high-level features of the driving factors within a neighborhood of an irregularly shaped cell and discover the relationships between multiple land use changes and driving factors at the neighborhood level. The proposed model was applied to simulate urban land use changes in Shenzhen, China. Compared with several VCA models using other machine learning methods, the proposed CNN-VCA model obtained the highest simulation accuracy (figure-of-merit = 0.361). The results indicated that the CNN-VCA model can effectively uncover the neighborhood effects of multiple driving factors on the developmental potential of land parcels and obtain more details on the morphological characteristics of land parcels. Moreover, the land use patterns of 2020 and 2025 under an ecological control strategy were simulated to provide decision support for urban planning. [ABSTRACT FROM AUTHOR]