Abstract
As spatial and socioeconomic processes are the two key aspects of urban development, revealing the relationship between these two key aspects is critical. Previous studies attempted to explain their correlation at the city or region level using built-up area metrics and nighttime light (NTL) data. However, more comprehensive studies on urban interior spatial characteristics and their relationship to NTL intensity are lacking in a three-dimension space. Using Luojia 1-01 nighttime light data, LiDAR digital surface model data, and other auxiliary data, this study applies an extreme gradient boosting regression model and Sharpley Additive exPlanations method to model and interpret the relationship between two-dimensional (2-D)/3-D landscape patterns and NTL intensity. Two study areas were selected to investigate the landscape–NTL relationship at the parcel and subdistrict levels. The major findings of this study include the following: 1) 2-D and 3-D urban landscape patterns have a close relationship with NTL intensity at the parcel and subdistrict scales; 2) the combinational metric of 2-D and 3-D landscape patterns has a stronger relationship with NTL intensity than either the 2-D or 3-D landscape metrics alone; 3) the correlations between most landscape metrics and NTL intensity are not simply positive or negative but change as metrics grow; and 4) the urban socioeconomic level is not only related to a single landscape metric sometimes but tends to the result of metrics interaction. These findings may help urban planners and government officials make more reasonable urban landscape planning policies under the goal of sustainable development.
Highlights
THE world is becoming an urban planet. 55% of the world’s population has been living in urban areas since 2007, and that share is expected to increase to 68% by 2050 [1]
As we only focused on the 3D landscape metrics of buildings, seven 3D landscape metrics were calculated in ArcGIS 10.4 to measure the 3D properties of buildings: total volume (TV), mean height (MH), height variance (HV), normalized height variance (NHV), 3D percentage of landscape (3DPLand), 3D largest patch index (3DLPI), and cubic index (CI)
We conducted a further analysis with the Shapley additive explanations (SHAP) method to explore how these 2D and 3D landscape metrics are related to nighttime light (NTL) intensity and how they interact with each other, which will advance our knowledge of urban development
Summary
THE world is becoming an urban planet. 55% of the world’s population has been living in urban areas since 2007, and that share is expected to increase to 68% by 2050 [1]. Urbanization is usually defined as population increases in cities versus rural areas, urban development is a complex process and contains two key aspects: spatial and socioeconomic. Previous studies revealed a strong correlation between urban spatial and socio-economic development, and it has been a long-standing notion that urban expansion and socioeconomic development can promote each other [9,10,11,12]. Urban agglomeration, and metropolitan scales, Shi et al [14] argued a slightly different point that there are long-term bidirectional causality and short-term unidirectional causality from socioeconomic development to spatial expansion. These studies have revealed the causality between urban horizontal spatial development and socioeconomic level. Urban spatial development is horizontal but vertical [15]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
