Unraveling the impact of green infrastructure and urban forms in 3D perspective on land surface temperature using remote sensing

The rapid growth of the global urban population causes cities' dynamic development and densification, and hence the re-composition of urban green spaces, which are increasingly under pressure. Therefore, urban space must be appropriately shaped to be resident-friendly. Particular attention must be paid to green infrastructure (GI), an essential component of a suitable quality of life. In spatial planning, analyses were often based on two-dimensional (2D) mapping of GI using satellite or aerial imagery and 2D indicators, where vegetation showed a uniform spatial distribution, regardless of volume. Meanwhile, the vertical structure of vegetation is heterogeneous, especially in urban environments. Similarly, buildings are characterized by different forms and volumes. Increasingly widespread LiDAR (Light Detection and Ranging) technology, particularly Airborne Laser Scanning (ALS), provides precise three-dimensional (3D) point cloud data for vast areas and can help analyze the complex structure of built-up and green spaces. The main goal of the research was to investigate the spatial relationship between the volume of vegetation, urban forms and land surface temperature (LST) in the rapidly developing Luxembourg City. The objectives are to (1) estimate the volumes of vegetation (VV) and buildings (BV) and, based on this, to create 3D indices; (2) investigate the characteristics of the seasonal distribution of LST; (3) analyze the relationships between 3D indices and LST. We proposed Vegetation 3D Density Index (V3DI), Building 3D Density Index (B3DI) and Vegetation Volume to Building Volume (VV2BV) Index (Figure 1). The VV was calculated using ALS (2019) point clouds processed into voxels (0.5 m), which define a cube surface area, and VB as 3D solid models. The LST maps were generated from Landsat 8 OLI/TIRS satellite imagery for January, April, August, and October. We compared the results of 3D GIS spatial analyses with population statistics to determine the residents' environmental situation and indicate possible vegetation deficits in urban space. Indices were estimated for districts, cadastral parcels, a grid of 100 m and for each building individually using a 100 buffer. We found that GI covered 1689 ha of Luxembourg City, accounting for 33% of the entire administrative area. The 3D GIS analyses show that the total volume of vegetation (> 1.0 m above ground) was about 40 million m3, equating to 328 m3 of greenery per resident. The overall VV2BV(%) index calculated for Luxembourg was 41.6%. Only five districts were characterized by a high value for the VV2BV index, which indicates areas well supplied with GI in relation to built-up areas. The results show that the VV is weakly negatively correlated with the BV (coefficients equal to -0.39); VV had a high relative influence on LST. There was a significant negative correlation between VV and LST and a positive correlation between urban forms and LST. This research study shows the potential transformation of the urban planning process by offering precise strategies based on 3D indices, including green infrastructure, to address urban heat island challenges effectively and urban sustainability.