For decades, urban analysis operated under a "flat-earth" paradigm. Researchers mapped cities using two-dimensional grids, tracking building footprints, horizontal land-use expansion, and surface greenness indices from top-down satellite imagery. However, modern cities do not merely expand outward; they grow upward and inward, forming complex three-dimensional geometries.
When we flatten a city into a 2D plane, we lose the structural context that governs its physical, ecological, and socioeconomic realities. A single 2D pixel cannot distinguish between a one-story warehouse and a fifty-story skyscraper, nor can it capture how wind flows through a street canyon or how solar radiation bounces off a glass facade. To address these limitations, a growing body of research is shifting toward volumetric urban morphology—modeling the built and natural environment as a continuous 3D space.
The Thermodynamics of 3D Urban Canyons
Traditional regional thermal models rely heavily on 2D land-cover data to analyze the Urban Heat Island (UHI) effect. Yet, as Chunbo Li et al. and Yuwen Wang et al. point out, flat models fail to account for how three-dimensional urban geometry traps and reflects heat.
To understand why a city retains heat, we must look at the volumetric interactions within "urban canyons"—the spaces flanked by buildings on both sides. In a flat model, a building's thermal footprint is static. In reality, its height and volume dictate the Sky View Factor (SVF) (the extent of sky visible from the ground) and the ratio of ground area to facade area. As demonstrated by Xiaorui Li et al., "tall and slim" building configurations mitigate surface temperatures far better than "short and fat" structures. This happens because vertical facades provide mutual shading and facilitate wind paths that flush out trapped heat, whereas dense, low-slung blocks trap long-wave radiation.
By incorporating 3D structural data and surface roughness, frameworks like the DeepUrbanModeller (DUM) proposed by Yuan Zang et al. can simulate heat trapping and reflection in urban canyons with ultra-high precision. This volumetric perspective is echoed by Kyungil Lee et al., who emphasize that understanding wind paths and heat retention requires accounting for 3D urban geometry rather than just 2D surface temperatures. Furthermore, Hongchao Xu et al. quantified that while 2D coverage ratios drive baseline temperatures, 3D building patterns provide a substantial, independent contribution to explaining localized temperature variations. Even at a fundamental physics level, Aniket Dashpute et al. show that explicitly modeling heat diffusion in three dimensions prevents the spatial artifacts that typically plague 2D thermal analysis by accounting for how heat moves into the depth of an object.
From Green Cover to Living Vegetation Volume
The limitations of 2D modeling are equally pronounced when evaluating urban ecosystems. Historically, urban green spaces were quantified using two-dimensional metrics like canopy coverage or the Normalized Difference Vegetation Index (NDVI). However, a flat green polygon treats a manicured lawn and a dense, multi-layered urban forest identically.
To accurately calculate ecological benefits like carbon sequestration, pollutant removal, and localized cooling, researchers are transitioning to 3D metrics such as Living Vegetation Volume (LVV). By leveraging Mobile Laser Scanning (MLS) point clouds, Wen Fan et al. demonstrated how fitting geometric primitives (like cylinders and spheres) to segmented tree data captures the true physical volume and biomass of vegetation.
This volumetric approach reveals the hidden mechanics of environmental mitigation. For instance, Byeong-chan Kim et al. found that accurate cooling assessments require accounting for canopy height and vertical density rather than surface greenness alone, as the volume of a forest dictates its capacity for transpiration and shading. This principle extends to natural landscapes as well: Blaise Tymen et al. showed that a 3D volume-based model of canopy architecture significantly outperforms traditional vertical-gap models when predicting micro-environmental variables like light, temperature, and humidity on the forest floor. Even when analyzing social impacts, such as the relationship between tree coverage and property crime investigated by Yifei Chen et al., understanding how greenery affects natural surveillance and concealment ultimately requires looking past flat canopy maps toward the physical density and vertical structure of the environment.
Simulating and Generating the Vertical City
Because cities grow both horizontally and vertically, traditional urban expansion models that focus solely on 2D land-use transitions provide an incomplete picture of metropolitan development. Modern frameworks address this by coupling horizontal land-use changes with vertical building height predictions.
For example, Xiaoping Liu et al. developed an integrated 3D simulation model that synergistically predicts horizontal land conversion alongside vertical growth. Similarly, Chuanglin Fang et al. advanced urban growth modeling by implementing a 3D Cellular Automata (3DCA) framework. By incorporating economic geography factors—such as proximity to transport and distance to the city center—their model successfully simulates the self-organization and emergence of high-rise clusters and central business districts. From a theoretical perspective, Junze Yin argues that capturing the true structural evolution and density of urban environments requires moving from static 2D spatial footprints to time-dependent, 3D volumetric fractal analysis.
This push for three-dimensionality is also transforming automated urban planning and layout generation. Generative models that operate purely on 2D grids often overlook critical architectural constraints and zoning regulations, a limitation highlighted by Dongjie Wang et al.. To resolve this, next-generation layout systems like Text2City, developed by Nanxuan Zhao et al., transition from flat road-and-building maps to 3D outputs that natively account for building heights and terrain topography, ensuring that generated cityscapes are both visually and functionally rational.
Socioeconomic Mapping and Energy Harvesting
Beyond climate and simulation, the vertical dimension is a vital indicator of human activity and resource potential. In high-density metropolitan areas, socioeconomic indicators are deeply tied to vertical volume. The authors of Lightweight and Robust Representation of Economic Scales from Satellite Imagery demonstrated that integrating 3D information from digital surface models is necessary to accurately capture economic scales, as relying solely on 2D imagery underestimates the economic density of vertical growth. This is directly applicable to population mapping: Nando Metzger et al. showed that incorporating building height data allows deep learning models to estimate building occupancy rates rather than raw horizontal density, drastically improving the precision of fine-grained population maps in high-rise environments.
Finally, 3D morphology dictates how cities can harvest renewable energy. Traditional solar cadasters map photovoltaic (PV) potential as flat overlays on roofs. However, Francesco Nex et al. demonstrated that transitioning to full 3D web-based visualizations is essential in complex terrains and urban structures to accurately model how solar radiation interacts with multi-faceted surfaces. In dense street canyons, energy generation shifts from roofs to building facades. To optimize this, Peng Wang et al. introduced a "space aspect ratio index" that accounts for the vertical dimension of installations and height-specific shading effects, proving that traditional 2D street metrics are insufficient for modern photovoltaic integration.
Whether analyzing solar yields, predicting urban heat, or mapping populations, the conclusion across these diverse domains is identical: to understand how a city functions, we must stop treating it as a flat map and begin modeling it as a three-dimensional volume. This spatial context is precisely what enables planners to intuitively evaluate complex interactions, moving past the information silos of traditional 2D geographic systems as noted by Wei Xi et al. and Chen Zhu-Tian et al..
Key Foundational Work
The Three-dimensional Urban Growth Simulating Based on Cellular Automata (2013) — contributed 3 concept elements to this direction.
FACTORS TO FORM DEWELLING AREA IN BUILT-UP HOUSES AREA : from view of building-Use in Setagaya ward (1995) — contributed 2 concept elements to this direction.
3D solarweb : a solar cadaster in the Italian alpine landscape (2013) — contributed 1 concept elements to this direction.