Spatial and Temporal Evolution of the Characteristics of Spatially Aggregated Elements in an Urban Area: A Case Study of Wuhan, China

Zhihao Sun, Dezhi Kang, Hongzan Jiao, Ya Yang, Wei Xue, Hao Wu, Lingbo Liu, Yuwei Su, Zhenghong PengISPRS Int. J. Geo Inf 2023 · DOI

TL;DR

This paper introduces a novel framework using the Block2Vec model to analyze the spatial and temporal evolution of urban element agglomeration in Wuhan, China. By integrating spatial interactions with element density, the study reveals a dynamic urban development pattern and provides insights for optimizing urban planning.

What Problem Does This Paper Address?

Urbanization is a complex process characterized by the agglomeration and dispersion of various urban elements. Understanding these spatial patterns and their evolution is crucial for effective urban planning and management. While previous studies have focused on individual urban elements (e.g., land use, services) or used density-based methods, they often overlook the critical aspect of spatial interaction between elements. This limits a comprehensive understanding of urban structure and development. This paper aims to bridge this gap by proposing a method that captures both the volume and spatial interaction of urban elements to better understand their agglomeration characteristics and temporal evolution.

Methods

The study employs a multi-stage methodology:

  1. Spatial Feature Extraction with Block2Vec: The core innovation is the use of the Block2Vec model, adapted from natural language processing. Block2Vec constructs "semantic sequences" of Points of Interest (POIs) within Traffic Analysis Zones (TAZs). It considers not only the POIs within a TAZ but also their spatial proximity to neighboring TAZs, thus capturing spatial interactions. This is achieved using an encoder-decoder architecture based on Long Short-Term Memory (LSTM) networks.
  2. K-Means Clustering: The extracted high-dimensional spatial semantic features for each TAZ are then clustered using the K-Means algorithm with cosine distance. This groups TAZs with similar agglomeration characteristics into distinct categories (high, medium, low).
  3. Spatial-Temporal Analysis: The study analyzes the evolution of these clusters over five years (2017-2021) by examining changes in the number of TAZs in each category. Spatial autocorrelation statistics (Moran's I and Local Moran's I) are used to assess spatial patterns and changes in agglomeration.

Datasets

  • Point of Interest (POI) Data: Collected continuously from 2017 to 2021 through the Amap Development Platform for the urban development area of Wuhan. This data includes type classification and geographic coordinates. Over five years, an average of 537,511 POIs were acquired annually.
  • Planned Road Network Data: Acquired from the Wuhan Municipal Bureau of Natural Resources and Planning to delineate Traffic Analysis Zones (TAZs), which serve as the minimum research unit.
  • 1-meter Resolution Remote Sensing Images: Obtained from Google Earth map data for different years to visually compare land cover changes.

Key Insights and Ideas

  • Beyond Density: The Power of Spatial Interaction: The study highlights that simply looking at the density of urban elements is insufficient. Incorporating spatial interaction, as done by Block2Vec, provides a richer understanding of how elements cluster and influence each other.
  • Dynamic Urban Structure: Wuhan's urban element agglomeration doesn't follow a simple concentric circle. Instead, it exhibits a decentralized pattern with clusters forming at specific nodes, aligning with China's "multi-centered, networked, and clustered development" strategy.
  • Peripheral Growth vs. Central Decline: While the urban development zone expanded significantly, the study observed increased agglomeration in peripheral areas, suggesting outward growth. Conversely, some central areas showed a decrease in service facilities, indicating potential shifts in urban function.
  • Block2Vec's Efficacy: The Block2Vec model proved effective in capturing comprehensive geographic information, allowing for objective and nuanced descriptions of urban element agglomeration.

Key Results

  • Circular Agglomeration: Urban elements exhibit a distinct circular structure in their spatial agglomeration, with variations across different element types.
  • Urban Expansion: The Wuhan urban development zone experienced significant expansion between 2017 and 2021.
  • Peripheral Agglomeration: Increased agglomeration was primarily observed in the surrounding areas of the city, indicating outward development.
  • Central Area Changes: Some central areas showed weaker element agglomeration and a reduction in various service facilities.
  • Validation: The analysis of typical areas, supported by remote sensing imagery, validated the identified agglomeration patterns and their changes over time.

Limitations and Future Work

  • Time Period: The study covers a relatively short five-year period (2017-2021). Longer-term analysis would provide a more comprehensive understanding of urban expansion and renewal trends.
  • Data Sources: While POI data is valuable, incorporating additional data sources like remote sensing and street view imagery could enhance precision and provide a more holistic view.
  • Future Directions: The authors suggest applying this framework to other cities with varying development levels to examine their urban element agglomeration and evolutionary characteristics. Further research could also focus on longer time horizons and integrate more diverse data sources.

Summary updated

BibTeX
@article{sun2023spatial,
  author = {Zhihao Sun and Dezhi Kang and Hongzan Jiao and Ya Yang and Wei Xue and Hao Wu and Lingbo Liu and Yuwei Su and Zhenghong Peng},
  title = {{Spatial and Temporal Evolution of the Characteristics of Spatially Aggregated Elements in an Urban Area: A Case Study of Wuhan, China}},
  journal = {ISPRS Int. J. Geo Inf.},
  year = {2023},
  doi = {10.3390/ijgi12110448},
}
Figure