Optimizing High-Rise Residential Form for Multi-Source Landscape View Access: A Target-Based Visibility Analysis Under Performance Constraints

Abstract

In high-density urban environments, residential design often faces a conflict between maximizing landscape access and maintaining energy-oriented compactness. This study proposes a target-based visibility analysis framework to optimize high-rise forms under strict performance constraints. Utilizing a Quad-mesh reconstruction strategy and Inverse Targeted Ray-Casting, the method accurately quantifies visibility via the cumulative Landscape Visible Surface (LVS) on the target building and Viewpoint-Specific Surface Visibility Rate (R_v) for precise verification against specific landscape targets. The framework is applied to evaluate three morphological prototypes: Compact Tower, Dispersed Tower, and Slab–Tower Hybrid. Quantitative simulations identified the Slab–Tower Hybrid as the optimal solution, demonstrating superior “Visual Morphological Efficiency.” While maintaining a moderate Shape Coefficient (SC = 0.326) to satisfy energy standards, the Hybrid achieved a cumulative Park-View LVS approximately 1.8 times that of the Compact Tower. Furthermore, environmental simulations indicated the Hybrid fosters stable wind environments (0.4–0.7 m/s) and equitable sunlight distribution. The research concluded that through differentiated massing, high-rise architecture can achieve a synergistic balance between visual openness and physical compactness, transforming view analysis from a passive check into an active design driver.

1. Introduction

Rapid urbanization has driven the proliferation of high-rise residential buildings in high-density cities, effectively addressing housing shortages but simultaneously severing the connection between inhabitants and the natural environment [1,2]. In the context of the “concrete jungle,” access to natural landscapes (e.g., rivers, mountains, and parks) is no longer merely an esthetic preference but a critical determinant of urban livability and public health [3,4]. Studies based on attention restoration theory have confirmed that visual exposure to natural elements significantly reduced stress [5,6,7], improves mental well-being [8,9], and enhances the economic value of real estate [10]. Consequently, maximizing Landscape View Access—particularly to multi-source landscapes like rivers, mountains, and parks—has shifted from an aesthetic luxury to a fundamental requirement for healthy urban living.

Recognizing this imperative, architects and scholars have made significant efforts to bridge the gap between the built environment and nature. These contributions primarily focus on visual perception preferences, architectural design philosophy, and the optimization of specific building components. Extensive research has defined what constitutes a “good view.” For instance, Kristl et al. [2] examined how demographic factors influence window view preferences, while Lin et al. [11] confirmed that a higher proportion of natural elements significantly enhances subjective preference. On a macro level, architects have sought a harmonious “self-consistency” between construction logic and nature, exemplified by Antonin Raymond’s work [12] and traditional Chinese architectural order [13]. On a micro level, research has focused on the geometric variables of openings; Ko et al. [14] proposed the “View Access Index” to optimize occupant satisfaction. Similarly, regarding semi-open spaces, Molaei et al. [15] highlighted that balconies facing natural landscapes significantly increase usage frequency. However, a critical gap remains: the existing literature extensively covers subjective perception and local component optimization, yet few studies directly address the adaptability of overall residential building morphology to the natural landscape environment [16]. Current studies often isolate “view” from “building form,” making it difficult to provide explicit guidance on how different morphological prototypes fundamentally determine landscape accessibility.

Furthermore, integrating precise visibility analysis into morphological design remains technically challenging. Current research on visibility analysis largely stems from geography and urban planning [17,18,19]. The foundational concepts of “Isovist” [20] and “Viewshed” [21] are widely used to measure visibility on 2D planes. Building upon these foundations, research has evolved significantly towards complex 3D simulations. Fisher-Gewirtzman and Natapov [22] expanded the definition of visibility analysis, comparing Voxel-based approaches with 3D line-of-sight approaches to analyze complex urban topographies. Subsequent studies have further enhanced these models by incorporating human perception and environmental targets. Fisher-Gewirtzman [23] integrated subjective “weighted views” into quantitative line-of-sight models to better predict spatial perception.

However, most existing methods adopt an “observer-based” perspective, simulating what a person sees from a specific point outward. While effective for evaluating a single viewpoint, this approach is computationally inefficient for optimizing an entire building’s form. These limitations have motivated a shift towards “Object-Visibility Computation,” which reconstructs which parts of a specific target are visible under complex occluders. Recent studies have demonstrated the superiority of this approach in various scales: Kim and Yi [24] advanced this logic by treating the sky hemisphere as a target object to reconstruct precise obstruction boundaries for sky-view factors. In parallel, Salimi et al. [25] employed NURBS-based u, v parameterization to decompose landmark surfaces, enabling the precise evaluation of cumulative visual impacts. Building on this object-oriented logic, Guo et al. [26] formalized an “Inverse Visual Analysis” framework specifically to quantify the river-view accessibility of individual dwellings. By constructing a “River Surface Curve” to delimit the maximum theoretical viewshed, they established a reciprocal calculation model that links the landscape target back to the building envelope. This framework effectively evaluates the spatial relationship between architectural group morphology and natural landscapes, and has been recognized for its utility in residential planning practice [27].

Nevertheless, a critical technical gap persists. While divergent ray-casting methods are highly effective for evaluating spatial openness or landmark visibility within a certain range, they suffer from the “divergence error” when applied to inverse analysis over long distances [28,29,30]. These methods operate on a principle where rays are dense near the viewpoint but spread out as distance increases (the “near-dense, far-sparse” phenomenon). When evaluating visibility from a landscape viewpoint towards a distant building, this leads to sparse sampling on the target surface, causing small architectural features to be missed. Therefore, there is an urgent need for a high-precision method to accurately quantify visibility on complex building forms without being compromised by distance.

Optimizing for views does not happen in a vacuum; it is strictly constrained by physical and climatic factors. In the architectural domain, the trade-off between “Form” and “Performance” is a well-documented dilemma. To maximize view access, building forms tend to become dispersed or articulated [31,32,33]. However, in regions with challenging climatic conditions, such as China’s hot summer and cold winter zone, energy efficiency regulations mandate a low Shape Coefficient (SC)—a ratio of surface area to volume—favoring compact forms to minimize heat transfer [34,35]. Additionally, high-rise typology must strictly adhere to sunlight standards and wind comfort requirements [36,37]. Prior studies have mostly focused on single-objective optimization, rarely establishing a unified framework that couples precise visibility analysis with multi-dimensional performance constraints [38].

To bridge these gaps, this study proposes a comprehensive morphological optimization framework. The primary objective is to establish a quantitative design pathway that couples architectural form with environmental quality. Specifically, this research aims to:

• Propose a Quad-mesh based inverse visibility analysis method using 1 m grids to ensure high-precision visibility quantification;
• Integrate strict Performance Constraints, including the Household Sunlight Perimeter Ratio (HSPR) and SC;
• Identify the optimal morphological strategy through a comparative case study of three prototypes (Compact, Dispersed, and Slab-Tower Hybrid) in Changsha.

This study provides a scientific, evidence-based tool for architects to navigate the complex trade-offs in high-rise residential design.

2. Materials and Methods

To address these technical challenges—particularly the lack of precise visibility quantification for complex building skins—a methodological framework is established. This approach integrates a parametric workflow within the Rhinoceros 3D platform, employing custom algorithms to synchronize morphological design with environmental performance verification.

2.1. Research Framework

This study establishes a quantitative design optimization framework aimed at maximizing the acquisition of external natural landscape resources through architectural form, while strictly adhering to environmental performance standards as boundary constraints. The workflow is structured into three logical phases (Figure 1):

Phase 1: Environmental Context Modeling (Invariant Input).

The first phase involves establishing the digital representation of the site’s fixed environmental context. All surrounding landscape elements (rivers, mountains, parks) are modeled in the Rhinoceros platform to accurately reflect spatial occlusion relationships. Crucially, specific visual boundary curves are defined along the edges of these landscapes to serve as the emission sources for the inverse visual analysis.

Phase 2: Morphological Variable Analysis (Variable Input and Response).

The proposed architectural form serves as the independent variable. By inputting a specific morphological prototype, the framework performs two evaluations: it utilizes an innovative method to identify the ‘Visible Mesh’ that establishes a visual connection with the landscape, while simultaneously calculating key physical indicators—such as Shape Coefficient (Energy), Sunlight hours, and Wind Environment—to optimize performance.

Phase 3: Decision-Making and Optimization (Evaluation Loop).

This is the critical decision-making stage. The logic follows a “Check-and-Refine” process: validation is achieved only if the design maximizes landscape view access while simultaneously satisfying all performance constraints (Energy, Sun, and Wind).

The optimization process is defined as an evolutionary trajectory guided by design intelligence, rather than algorithmic randomization. It initiates with a foundational prototype. Guided by the architect’s design experience and the quantitative feedback from the analysis, the form is iteratively modified—evolving through intermediate morphological variations to a final optimized solution—until a satisfactory balance between landscape maximization and performance compliance is achieved.

2.2. Target-Based Visibility Analysis

The core of the morphological optimization lies in the precise quantification of view access. Target-based visibility analysis method is specifically designed to identify and extract the Landscape Visible Surface (LVS) on the proposed building. By treating the building envelope as the target of visual calculation, this method aims to pinpoint the exact mesh faces that possess a direct line of sight to the surrounding natural landscape resources, thereby quantifying the “visual potential” of the architectural form.

2.2.1. Principle of Inverse Visibility and Visual Boundary Definition

The method is based on the principle of Visual Reciprocity, derived from the methodology proposed by Guo et al. [26]. A critical prerequisite for this analysis is the scientific definition of the “Visual Boundary Curves”—specific linear sources from which visual rays are cast towards the building.

Guo et al. [26] demonstrated that for a neighborhood block, the maximum potential visual scope is defined by the union of visual fields from an ‘Optimal Viewpoint Combination’—specifically, viewpoints located at the site boundary closest to the landscape and at the maximum permissible building height. However, this model assumes a vertically continuous visual corridor; it may become invalid in the presence of intervening overhead obstructions (e.g., viaducts) that segment the view, thereby rendering the simplified landscape boundary curve insufficient.

Assuming unobstructed conditions, the viewshed boundary generated by this optimal combination represents the theoretical maximum visibility range for the entire site. Geometrically, if a viewer can see the nearest edge of a landscape (such as a riverbank), they would naturally have visual access to the landscape’s broader expanse. Consequently, the visual field from any point within the residential cluster constitutes a subset of this maximum scope.

Therefore, this study adopts the boundary lines of this maximum viewshed as the Visual Boundary Curves. For instance, for a river, these curves correspond to the line along the opposite riverbank; for a mountain, they correspond to the ridge line. These curves serve as the valid emission sources for the inverse ray-casting process. If rays emitted from points on these boundary curves can reach the faces of the target building mesh without obstruction, those faces are recognized as having access to the landscape view and are thus categorized as part of the LVS.

2.2.2. Computational Workflow for Visibility Analysis

In the Rhinoceros 3D platform, geometric objects exist in various structural formats, primarily: NURBS Surfaces, Boundary Representations (Brep), Meshes, and SubD objects [39].

Brep: This format defines geometry using precise mathematical boundaries (NURBS surfaces) combined with topological data (vertices, edges, and faces). It is the standard for high-precision modeling of solids and surfaces.

Mesh: This format approximates geometry using a discrete collection of polyhedral faces (typically triangles or quadrilaterals) defined by vertices. Unlike Breps, meshes allow for discrete data processing, making them ideal for computational analysis.

In this study, the Obstruction Environment—including terrain, water bodies, Street trees, Roadside shrubs, and existing buildings—is constructed using Brep models to maintain geometric precision (Figure 2). Conversely, the Target Object (the proposed residential building) is prepared in two formats at the same location: a reference Brep and a computational Quad-Mesh. The Quad-mesh is generated using the “Quad Remesh” algorithm to discretize the continuous building skin into uniform computational units [40]. To illustrate this setup, a standard cube with dimensions of 40 m (L) × 30 m (W) × 50 m (H) was used as a target prototype. When the quad-mesh edge length is set to 10 m, the surface of this cube is discretized into 94 mesh faces (F ₁₋₉₄), as shown in Figure 2.

Visibility calculations are implemented using the Grasshopper visual programming platform (Figure 3). The input parameters for the algorithm include the Target Quad-Mesh, the Obstruction Breps, and the Viewpoint.

The core calculation logic employs a ‘Targeted Ray-Casting’ method (Figure 4). The algorithm constructs lines of sight connecting the Viewpoint (P_view) to the centroid (C_i) of each mesh face on the target object. For the 94-face cube example, this generates 94 initial target rays, which are then categorized into three types based on intersection analysis:

• Type A (Environmentally Occluded): Rays that intersect with any object in the obstruction environment (terrain, surrounding buildings, etc.). These are marked as blocked.
• Type B (Self-Occluded): Rays that do not hit the environment but intersect the target object itself at a point closer than the target centroid (i.e., intersection distance < original ray length). This occurs when a face is on the back side of the building relative to the viewer.
• Type C (Effective Lines of Sight): The remaining rays that encounter no obstructions. These rays maintain their original length and successfully reach the target faces.

Based on this classification, the mesh faces corresponding to Type C rays are identified as visible. The algorithm then reconstructs these valid faces into a new geometric entity called the “Visible Mesh.” In the cube example, the result identifies 27 effective rays, corresponding to a Visible Mesh area of 2700 m² (Figure 4).

The precision of the Quad-Remesh conversion directly influences the calculation results. As the target edge length decreases, geometric fidelity increases, but computational load rises.

Table 1. Sensitivity analysis of visible surface area calculation across different Quad-mesh resolutions.

Target Edge Length of Quad-Mesh	Vertices of Quad-Mesh	Faces of Quad-Mesh	Vertices of Visible Mesh	Faces of Visible Mesh	Area of Visible Mesh	Relative Accuracy (%)
10 m	148	94	45	27	2700	95.59%
5 m	478	376	145	104	2600	99.46%
2.5 m	1702	1504	498	418	2612	98.98%
1 m (Selected)	9886	9400	2781	2584	2584	99.92%
0.5 m (Benchmark)	38,566	37,600	10,739	10,346	2586	100.00%

compares the results for the cube prototype with edge lengths ranging from 10 m to 0.5 m.

Using the 0.5 m grid result as the benchmark ground truth, the analysis shows that a 1.0 m grid size achieves a relative accuracy of 99.92%. Since this setting significantly reduces the face count (and thus calculation time) while maintaining near-perfect accuracy, 1.0 m is selected as the standard precision for this study.

This workflow demonstrates high adaptability to complex architectural forms, such as residential towers with curved balconies. Figure 5 applies this method to a proposed design with 23,207 total quad faces (1 m grid). The calculation successfully filters out occluded areas, resulting in a Visible Mesh of 3992 faces. As shown in the figure, the generated visible surface perfectly matches the visual perspective, validating the algorithm’s reliability for complex morphological combinations.

2.3. Environmental Performance Simulation and Constraints

Architectural form optimization requires a holistic approach where visual esthetics and physical environmental performance are inextricably linked. Following the visibility analysis, this study extends the simulation workflow to assess sunlight, ventilation, and energy efficiency. To ensure methodological consistency, these environmental simulations are conducted within the same high-precision 3D obstruction environment (comprising terrain, surrounding buildings, and vegetation) as defined in Section 2.2. Furthermore, the simulation grid for the target building maintains the 1.0 m precision standard, ensuring that data points for visibility, sunlight, and wind pressure are topologically aligned on the building skin.

The environmental analysis was conducted using Glodon CONCETTO, an AI-driven simulation platform that integrates local regulatory standards with the EnergyPlus Weather database [41]. Unlike simplified analytical formulas (e.g., Bernoulli’s principle), the platform’s AI simulation engine is designed to reconstruct complex environmental interactions within high-density urban settings. To ensure simulation transparency and reproducibility, the boundary conditions were set as follows:

• Sunlight Analysis: Simulations were calibrated to the “Great Cold Day” (January 20), compliant with the Chinese National Standard. The simulation covered the effective time window from 8:00 to 16:00, utilizing a temporal sampling step of 5 min. The calculation plane was set at 0.9 m above the floor level, with a rigorous mesh size of 1.0 × 1.0 m.
• Wind Environment: The wind analysis module was configured to utilize the EnergyPlus meteorological database to establish accurate inlet velocity profiles for the AI simulation. Based on the statistical analysis of the local climate (Changsha, 112.93° E, 28.23° N), the boundary conditions were set to 3.0 m/s from the Northwest (NW) for winter and 1.7 m/s from the Northwest (NW) for summer. The simulation directly outputs the Wind Velocity Magnitude distribution, enabling the identification of stagnant zones (<0.2 m/s) and effective ventilation corridors without relying on manual pressure-difference calculations.

This integrated toolchain allows for rapid iteration and precise quantification of how the building form interacts with natural forces in a complex occlusion environment.

2.3.1. Sunlight Environment: Mandatory Constraints and Health Metrics

Sunlight is a critical factor for residential health. The assessment is divided into two parts: a mandatory regulatory constraint and a quantitative performance metric.

Mandatory Constraint (Chinese National Standard): According to the “Standard for Urban Residential Area Planning” (GB 50180-2018) in China [42], at least one habitable room (such as a bedroom or living room) in every household unit must receive a minimum of 2 h of effective sunlight on the Great Cold Day (January 20th). Any design that fails to meet this threshold is immediately discarded.

Performance Metric (HSPR): In addition to the minimum standard, this study introduces the HSPR to quantify the overall quality of sunlight exposure. Unlike traditional point-based analysis, HSPR measures the extent of the building envelope that functions as a “solar collector” for the unit. It is defined as the percentage of the household’s total perimeter that meets the 2 h sunlight requirement:

$$HSPR={\displaystyle \frac{L_{sun\ge 2h}}{L_{total}}}\times 100\%$$

(1)

where L_sun ≥ 2 h is the length of the exterior wall meeting the sunlight standard, and L_total is the total wall perimeter of the household unit. A higher HSPR indicates a unit with superior natural light penetration and better potential for promoting health.

2.3.2. Wind Environment: Natural Ventilation Potential

The building’s aerodynamic form significantly influences both the outdoor pedestrian comfort and the indoor hygrothermal environment.

Outdoor Environment: The building massing is analyzed to ensure it does not create excessive wind turbulence or stagnation zones at the pedestrian level, facilitating outdoor recreational activities.

Indoor Natural Ventilation: In the “Hot Summer and Cold Winter” climate zone, effective natural ventilation is essential for cooling and maintaining air quality. The optimization goal is to maximize the Wind Pressure Difference between the windward (South) and leeward (North) faces of the building. A significant pressure gradient facilitates the “cross-ventilation” effect, which helps remove indoor heat and moisture. The CONCETTO AI simulation calculates the wind speed distribution around the building at a height of 1.5 m. The simulation directly outputs the Wind Velocity Magnitude distribution, enabling the identification of stagnant zones (<0.2 m/s) and effective ventilation corridors.

2.3.3. Energy Efficiency: SC

The SC is a fundamental indicator of thermal efficiency. It represents the ratio of the building’s external surface area in contact with the atmosphere (F_target) to its enclosed volume (V_target).

$$SC={\displaystyle \frac{F_{target}}{V_{target}}}$$

(2)

A lower SC implies a more compact form with less heat loss. According to the General Specifications for Building Energy Conservation and Renewable Energy Utilization (Chinese National Standard GB 55015-2021) [35], the SC for high-rise residential build-ings in this climate zone is subject to a strict limit:

$$SC\le 0.40$$

(3)

This value serves as a rigid constraint in the optimization process. While lower values are preferred for energy saving, any design exceeding 0.40 is considered non-compliant with national energy codes and is eliminated.

2.4. Case Study: A Residential Project in a Complex Landscape Context

To verify the effectiveness of the proposed coupling mechanism, this study applies the optimization workflow to a real-world residential project, detailing the specific site constraints, the construction of the large-scale simulation environment, and the logic behind the generation of three morphological prototypes for comparative analysis.

2.4.1. Site Context and Constraints-Driven Solution Space

The empirical case study is a residential development project located in Changsha, China. Changsha is a prominent example of a “Landscape City,” characterized by the harmonious integration of mountains, rivers, and the urban fabric. The project site is strategically positioned at the convergence of three primary natural landscape resources (Figure 6).

• River (Water): The site overlooks the expansive Xiang River to the east, including a prominent river island.
• Mountain: A continuous mountain range (the southern extension of the Yuelu Mountain system) provides a scenic backdrop to the south.
• Park: A large-scale urban ecological park is located immediately to the west.

However, architectural design in high-density contexts is a process of “Design Possibility Contraction” [43]. While the initial design space may appear vast, the plot is shaped by a set of rigid urban constraints that strictly confined the morphological exploration to a limited feasible set: On the western side, an underground road tunnel and a subway line impose strict structural protection requirements, leaving only a narrow, north–south-oriented strip as the final developable area. Meanwhile, the eastern section is already occupied by three high-rise residential towers (130–150 m) and several four-story auxiliary buildings. Additionally, a residential project to the northwest is currently under construction. Due to updated zoning regulations and a prolonged development timeline, the proposed building faces a strict height limit of 80 m. These combined constraints—the narrow buildable footprint, significant visual occlusion from existing eastern towers, and the strict height cap—create a highly restricted ‘gap.’ Within this confined spatial envelope, optimizing the architectural form becomes critical to maximizing visual access to the surrounding landscape, as traditional design strategies are rendered ineffective by such dense boundary conditions.

2.4.2. Simulation Environment Setup

To accurately quantify the visual interaction between the building and its surroundings, a comprehensive 3D Site Model covering an area of 4500 m × 4500 m was constructed within the Rhinoceros platform (Figure 7). The proposed building is situated at the geometric center of this environment. The model incorporates high-fidelity Brep geometries representing the complex terrain, all existing surrounding buildings, and critical road infrastructure. While the proposed methodology is fully capable of vegetation modeling (as verified in Section 2.2), this case study prioritizes the morphological optimization of high-rise massing. Given that the visual field of high-rise units (typically situated above the third floor) naturally clears the height of ground-level vegetation, tree obstruction was deemed negligible for this macro-scale form generation. Consequently, vegetation was omitted to enhance computational efficiency.

Under this configuration—where major architectural and terrain obstacles are rigorously accounted for—the simulation operates on a definitive geometric principle: a clear line of sight to the distant shoreline or ridgeline indicates a valid visual connection to the landscape body. Accordingly, by applying the “Visual Boundary Definition” principles established in Section 2.2.1, three distinct Landscape Boundary Curves were extracted to drive the simulation. As shown in Figure 7, these curves do not merely trace geometric edges but serve as the source trajectories for the inverse visibility calculation:

• River Surface Curve: Rather than using a simple geometric centerline, this curve is extracted to represent the maximum river-view extent relevant to the site, thereby ensuring the accuracy of the visibility calculation.
• Mountain Ridgeline Curve: Defined along the crest of the southern mountain range. As the highest visible contour, this curve dictates the vertical silhouette of the landscape background. Any line of sight reaching the ridgeline implies visual access to the mountain mass below.
• Park Boundary Curve: Defined along the interface between the urban fabric and the ecological park, representing the near-field visual boundary of the western green space.

A critical divergence from previous methodologies lies in the analysis range. While Guo et al. [26] limited their visual structure analysis to an 800 m radius, the scale of the Xiang River and Yuelu Mountain landscape necessitates an extended range of 2250 m. At such a magnitude, traditional “diverging ray-casting” methods become ineffective, as the angular spread would cause the spatial gap between rays to exceed the dimensions of the target building elements. This limitation fundamentally justifies the adoption of the “Targeted Ray-Casting” method (Section 2.2.2). Unlike traditional divergent ray-casting where the sampling gap expands with distance, this method anchors each visual ray to the centroid of a specific 1 m² Quad-mesh unit. Consequently, the simulation precision is solely determined by the mesh resolution and remains invariant to the visual distance. This mechanism maintains a constant sampling density on the building skin across the entire 2250 m range, ensuring that the visibility of every architectural unit is accurately calculated without the resolution degradation inherent in traditional viewshed methods.

2.4.3. Design Evolution and Forward Optimization Process

Unlike “Inverse Optimization” methods that rely solely on algorithmic generation, this study adopts a “Forward Design Optimization” framework [44], which integrates architectural heuristics with quantitative performance feedback. As emphasized by De Luca et al. [45], identifying the appropriate morphological typology is a prerequisite for high-performance design. Consequently, under the rigid constraints detailed in Section 2.4.1, the design team employed a “Representative Selection” strategy. Rather than randomly sampling the design space, the process involved an iterative optimization within each viable topological category to derive the “Best-in-Class” candidate for comparative analysis (

Table 2. Iterative design evolution and optimization process of the three representative typologies.

	Compact Typology				Dispersed Typology			Hybrid Typology
Optimal Representative *
Iterative Variations *
Legend of Functional Zones		Core		Interior Staircase			Study Room			Cloakroom
		Corridor		Dining and Living Room			Kitchen			Balcony
		Foyer		Bedroom			Bathroom			Terrace Garden

Note: ‘Optimal Representative’ * displays the final prototypes selected for comparative simulation based on expert scoring and stakeholder feedback; ‘Iterative Variations’ * shows the intermediate schemes generated during the forward optimization process.

).

• Optimization of the Compact Typology (Scheme 1.x)

The initial strategy prioritized land-use efficiency within the narrow site setbacks by adopting a central-core layout. While early iterations explored various core placements, they resulted in deep floor plates that compromised daylighting. Through structural refinement and expert consultation, Scheme 1.1 (Compact Tower) was selected as the optimal representative. This prototype minimizes the building footprint but inherently faces challenges regarding self-shading and limited view corridors.

2.

Evolution of the Dispersed Typology (Scheme 2.x)

To overcome the limitations of the compact form and enhance natural ventilation, the massing was fragmented to maximize the external perimeter. Although initial variations generated excessive surface areas leading to poor Shape Coefficients, the design was refined through multiple rounds of iterative adjustment to balance the degree of fragmentation. Scheme 2.1 (Dispersed Tower) was finalized as the representative prototype, offering an optimized trade-off between ventilation benefits and energy conservation limits within this category.

3.

Synthesis of the Hybrid Typology (Scheme 3.x)

Acknowledging the respective trade-offs of the previous strategies, a hybrid approach was developed to synthesize the efficiency of the slab with the multi-orientation potential of the tower. While early attempts failed to meet wind comfort standards, subsequent iterations adjusted the orientation of the slab components. The final Slab-Tower Hybrid (Scheme 3.1) was identified as the optimal solution, featuring a topologically twisted geometry that aligns with both the River and Park view axes while maintaining a moderate Shape Coefficient.

Therefore, the comparative analysis presented in Section 3 is not a random ranking of arbitrary forms, but a validation of these three optimized representatives (Figure 8), ensuring that the conclusions reflect the maximum potential of each design strategy.

3. Results

3.1. Geometric Efficiency Analysis

Geometric efficiency in this study refers to the balanced relationship between a building’s landscape accessibility and its energy-related form performance. Ideally, the architectural form should maximize landscape visibility while maintaining low envelope-driven energy demand. Accordingly, the analysis proceeds in three steps: first, examining the inherent geometric attribute of the form (Shape Coefficient); second, quantifying landscape visibility from both global LVS and target-based perspectives; and finally, synthesizing these metrics to determine “Visual Morphological Efficiency,” which characterizes the balance between landscape acquisition and energy conservation.

3.1.1. Inherent Geometric Attribute: SC

The SC is an inherent geometric attribute of architectural form and is commonly adopted as a proxy for envelope-related energy performance, where a lower SC generally indicates a more compact mass with reduced heat-transfer area. In the present context, however, its implication for landscape visibility is not straightforward because visibility is governed not only by envelope extent but also by self-occlusion and external blockage within a dense urban setting. All three schemes satisfy the regulatory threshold (SC ≤ 0.40) for the “Hot Summer and Cold Winter” climate zone, yet they differ in compactness: the Compact Tower achieves the lowest SC (0.302), the Dispersed Tower exhibits the highest SC (0.348) due to its fragmented massing, and the Slab–Tower Hybrid remains intermediate (0.326), indicating a balanced degree of compactness and articulation.

3.1.2. Landscape Visibility Assessment: Cumulative LVS and Viewpoint-Specific Rates

To precisely quantify the natural landscape acquisition of the proposed project (the Target), we defined specific landscape boundary curves and established a series of “Estimated Effective Viewpoints” along each curve:

• River Landscape: 20 estimated effective viewpoints (No. 0–19).
• Mountain Landscape: 14 estimated effective viewpoints (No. 0–13).
• Park Landscape: 20 estimated effective viewpoints (No. 0–19).

The determination of these viewpoints followed an adaptive sampling strategy based on distance and geometric inflection to ensure analytical robustness.

• Distance-based Spacing: For proximal landscape resources (Park), equidistant viewpoints were set at intervals of less than 50 m. For distant resources (River and Mountain), the interval was set within 150 m.
• Geometric Feature Inclusion: Key geometric turning points along the ridgelines were explicitly included to capture morphological shifts in the landscape boundary.
• Efficiency Filtering: To ensure the validity of the dataset, continuous segments where five consecutive viewpoints lacked any visual connection to the target were excluded.

This resulting density is sufficient to capture the continuous visual interaction between the high-rise massing and the landscape, ensuring that the comparative evaluation remains stable against minor positional variations.

Based on these defined viewpoints, we employed two complementary methods to quantify visibility:

• Cumulative LVS areas:

Landscape accessibility is evaluated through the cumulative LVS areas associated with each landscape category (Figure 9,

Table 3. Quantitative comparison of cumulative LVS areas.

	Building Surface Area	River-View	Mountain-View	Park-View
Compact Tower	11,896.2 m2
Dispersed Tower	15,992.5 m2
Slab-Tower Hybrid	13,385.3 m2
Legend *

Note: The bottom legend illustrates the visualization method using the Slab-Tower Hybrid (River-View) as a representative example. “RSC” denotes the River Surface Curve.

). The green rays in Figure 9 illustrate the inverse targeted ray-casting from the landscape boundary curves to the building envelope, determining the specific surface areas visible from each resource. The results show a consistent hierarchy, with the Slab–Tower Hybrid performing best across view types. Taking Park-View as an illustrative example, the Hybrid reaches a cumulative Park-View LVS of 95,628.6 m², which is approximately 1.8 times that of the Compact Tower (53,723.9 m²) and 1.3 times that of the Dispersed Tower (74,949.2 m²). River-View and Mountain-View exhibit the same overall pattern, indicating that the Hybrid form provides the most effective visual access under the given occlusion conditions.

2.

Viewpoint-Specific Surface Visibility Rate (R_v)

To reveal the distribution of visibility across the landscape boundaries, we calculated the Surface Visibility Rate for each specific estimated effective viewpoint. This is defined as the percentage of the Target’s visible mesh area (seen from a specific viewpoint) relative to the Target’s total envelope surface area:

$$R_v\left(i\right)={\displaystyle \frac{A_{visible}\left(i\right)}{A_{target\_total}}}\times 100\%$$

(4)

where i corresponds to the specific viewpoint number on the landscape curve. Figure 10 integrated these rates into line charts, clearly illustrating the performance differences.

• River View (Figure 10a): The Slab-Tower Hybrid (Gray line) shows a distinct advantage. Notably, at Viewpoint 3, the Hybrid achieves a visibility rate exceeding 40%, whereas the other prototypes remain below 20%. It maintains a consistent lead from Viewpoint 7 to 19.
• Mountain View (Figure 10b): The Hybrid maintains a high-efficiency “plateau,” with visibility rates consistently exceeding 50% for Viewpoints 0–2 and 6–10, significantly outperforming the Compact Tower (~40%) and Dispersed Tower (<38%).
• Park View (Figure 10c): Similarly, the Hybrid peaks near 50% visibility and maintains a substantial margin over the others across the majority of sampling points (No. 0–10 and 16–19).

In summary, the comparative trajectories in Figure 10 demonstrate a high degree of structural stability. The Slab-Tower Hybrid consistently defines the upper bound of performance across the vast majority of sampling points, maintaining a distinct margin over the Compact and Dispersed prototypes. This persistent hierarchy confirms that the identified morphological advantage is robust and intrinsic to the design strategy, indicating that the comparative results are insensitive to minor variations in viewpoint positioning or sampling density.

3.1.3. Synthesis: Balancing Energy Conservation and Landscape Acquisition

A key finding emerged when LVS was interpreted alongside the total building surface area. The Dispersed Tower attained higher visibility largely by expanding the building envelope (15,992.5 m²), implying increased construction exposure and potentially greater envelope-related energy losses. In contrast, the Slab–Tower Hybrid achieved the highest cumulative visibility outcomes with a more moderate envelope area (13,385.3 m²) and a lower SC than the Dispersed Tower (0.326 vs. 0.348). This indicated a higher “visual morphological efficiency”: rather than relying on a brute-force increase in surface area, the Hybrid configuration leverages strategic geometric alignment and mass articulation to reduce self-occlusion and better engage available view corridors. Overall, the comparison suggests that, within the same regulatory SC constraint, architectural form-making can substantially influence landscape visibility, and the Slab–Tower Hybrid achieves superior visual returns per unit envelope area by coupling visual performance with geometric compactness.

3.2. Sunlight Environment and Health Metrics

Following the visibility analysis, the study evaluated sunlight performance to ensure that the optimization of landscape views does not compromise solar accessibility, a critical determinant of resident health. The simulation, conducted using Concetto AI, mapped precise solar hours onto every square meter of the building envelope under identical constraints (Figure 11).

As a baseline, the simulation confirms that all three architectural forms satisfy the mandatory code requirement of providing at least 2 h of sunlight to at least one habitable room per household on the Great Cold Day. However, meeting the minimum code does not equate to uniform living quality. To quantify the disparities in health performance beyond basic compliance, we utilize the HSPR (see

Table 4. Comparative analysis of HSPR.

	Compact Tower Residence			Dispersed Tower Residence			Slab–Tower Hybrid Residence
	South Side (min)	West or East Side (min)	HSPR	South Side (min)	West or East Side (min)	HSPR	Toward South Side (min)	Toward North Side (min)	HSPR
Unit 1	480	285	40.23%	480	234	46.22%	385	94	58.47%
Unit 2	480	159	26.19%	480	219	47.54%	315	30	26.68%
Unit 3	300	270	42.89%	289	234	51.57%	304	0	33.68%
Unit 4	170	144	17.50%	255	189	33.02%	200	0	25.21%
Unit 5	304	230	28.03%	255	225	38.36%	274	0	45.43%

Note: The calculation results presented in this table correspond to the first-floor residential units.

). The HSPR measures the percentage of a household’s external perimeter that receives effective sunlight; a higher value indicates superior solar accessibility, contributing to a healthier and more comfortable indoor environment.

The HSPR results reveal distinct performance characteristics driven by the geometric logic of each prototype:

• Compact Tower: The sunlight distribution exhibits significant unevenness. While some units achieve reasonable values, Unit 4 performs poorly at only 17.50%, and the majority of units are below 30%. This indicates severe self-shading issues inherent to the centralized core layout.
• Dispersed Tower: This form demonstrates the most superior and balanced performance. Three out of five units exceed an HSPR of 46%, with Unit 3 reaching a peak of 51.57%. Even the lowest-performing unit maintains a ratio of 33.02%. The dispersed layout effectively increases the solar exposure perimeter for individual households, eliminating “sunlight dead zones.”
• Slab-Tower Hybrid: The results show a polarized performance. Unit 1 achieves the highest HSPR (58.47%) among all tested units (benefiting from the slab-end position). However, north-facing units (e.g., Units 2 and 4) show a lower metric (approx. 25–26%), relying heavily on south-facing exposure.

3.3. Wind Environment and Natural Ventilation

The assessment of the wind environment addresses two distinct scales: the outdoor pedestrian comfort and the potential for indoor natural ventilation.

At the site level, the computational fluid dynamics simulation reveals negligible differences among the three architectural forms (

Table 5. Comparative simulation of the outdoor wind environment.

	Winter	Summer
Compact Tower
Dispersed Tower
Slab-Tower Hybrid

Note: White blocks represent the building massing; the color gradient represents wind velocity in m/s. Meteorological data source: EnergyPlus Weather Database for Changsha.

). In winter, wind speeds within the residential cluster remain below 2 m/s, preventing uncomfortable drafts, while summer conditions offer a light breeze across the site. Consequently, all three proposals satisfy the standard comfort requirements for pedestrian circulation and outdoor recreational activities.

However, a critical divergence appears when evaluating the building envelope’s potential for indoor ventilation. In the “Hot Summer and Cold Winter” climate zone, the design priority shifts seasonally: while winter strategy focuses on airtightness for heat preservation, maximizing natural ventilation in summer is paramount for passive cooling and thermal comfort. The simulation results regarding the building envelope (Figure 12) highlight significant performance disparities driven by the morphological traits of each iteration:

• Compact Tower: This form exhibits the poorest ventilation potential. Its bulky, centralized mass creates extensive “Wind Shadow” zones on the leeward side, resulting in stagnant airflows (<0.2 m/s) around the envelope. This lack of air movement is detrimental to heat dissipation, significantly increasing the reliance on mechanical cooling during summer.
• Dispersed Tower: Despite its fragmented volume, this form suffers from uneven airflow distribution. While corner units experience higher wind velocities, the complex recesses and setbacks induce vortex formation, leading to severe airflow stagnation in the re-entrant corners. This results in poor ventilation equity, where some households enjoy good airflow while neighbors face stagnation.
• Slab-Tower Hybrid: The hybrid form demonstrates optimal ventilation performance. Its articulated linear geometry effectively acts as a wind guide, channeling the prevailing winds to maintain a consistent speed range of 0.4–0.7 m/s (Soft Breeze) across most facades. This stable wind pressure distribution is crucial for facilitating effective cross-ventilation for the majority of residential units.

3.4. Comprehensive Decision

To determine the final architectural prototype, a multi-criteria trade-off analysis was conducted by synthesizing results from the landscape, sunlight, wind, and energy simulations.

Table 6. Multi-criteria performance comparison matrix of the three morphological prototypes.

Performance Metric	Compact Tower	Dispersed Tower	Slab-Tower Hybrid
Landscape Visibility	Low Severe self-occlusion	Medium Gained via surface expansion	High Highest morphological efficiency
Sunlight Equity (HSPR)	Poor Uneven; max unit < 30%	High Balanced; most units > 46%	Moderate Polarized; relies on south units
Ventilation Quality	Poor Stagnant air (<0.2 m/s)	Medium Uneven; vortex issues	Optimal Stable cross-ventilation (0.4–0.7 m/s)
Energy Efficiency (SC)	High SC = 0.302 (Best)	Low SC = 0.348 (Worst)	Balanced SC = 0.326 (Moderate)
Overall Verdict	Rejected Fails living quality standards	Sub-optimal High energy cost for views	Selected (Optimal) Best balance of View, Health and Energy

summarizes the comparative performance of the three schemes.

As indicated in Table 6, the Slab–Tower Hybrid is selected as the most suitable solution under the site-specific constraints. The decision follows a clear hierarchy. First, with landscape performance as the primary objective, the Hybrid achieves the highest visual efficiency by aligning its massing with the river and mountain view axes, rather than relying on a simple increase in envelope area as in the Dispersed Tower. Second, regarding the critical health requirement of ventilation, it provides the most stable cross-ventilation potential—an important contributor to summer thermal comfort in Changsha—while avoiding the localized air stagnation observed in the Compact Tower. Third, in terms of constraint management for energy and sunlight, although the Compact Tower is thermally the most efficient, it underperforms in both habitable sunlight access and landscape visibility. The Hybrid scheme offers a more balanced profile: it maintains a moderate SC (0.326), substantially lower than the Dispersed option, while still securing acceptable sunlight conditions for most units.

Consequently, the Slab–Tower Hybrid is selected as the optimal prototype for this site context, as it statistically satisfies the performance constraints while maximizing the primary objective of landscape view access.

4. Discussion

4.1. Decoupling View from Envelope: The Necessity of Precise Computation and Passive Morphological Control

This study responds to the fundamental challenge of quantifying landscape accessibility in high-density urban environments. A prevalent assumption in conventional design practice is linear: that maximizing the building envelope naturally leads to maximized views. However, the comparative results between the Dispersed Tower and the Slab-Tower Hybrid refuted this linearity. Although the Dispersed Tower possessed the largest surface area, its effective landscape visibility was compromised by self-shading and mutual occlusion. Conversely, the Slab-Tower Hybrid achieved superior visibility with a smaller, more energy-efficient envelope, effectively decoupling “view maximization” from “envelope expansion”.

This decoupling has profound implications for energy efficiency. In the context of passive architectural design, the building envelope serves as the primary interface for thermal exchange. Therefore, reducing the envelope area relative to the building volume—quantified as the Shape Coefficient (SC)—is critical for minimizing energy demand. While absolute energy consumption is influenced by numerous micro-scale variables such as HVAC systems and wall U-values [46], literature confirms that morphological factors are the most cost-effective drivers of energy efficiency during the early design phase [47]. As Shi et al. [48] demonstrated, passive morphological strategies (including SC optimization and orientation) can improve energy efficiency by over 20% before any active technologies are applied.

Consequently, by adopting SC as a key geometric proxy for thermal performance, this study validates a crucial architectural logic: high-performance morphology does not require a trade-off between “opening up for views” and “closing down for energy.” Instead, through precise inverse computation, architects can identify specific morphological configurations—like the Slab-Tower Hybrid—that simultaneously maximize landscape acquisition while maintaining a compact, energy-efficient form. This reinforces the necessity of integrating precise visibility calculation with fundamental geometric energy indicators at the very onset of the design process.

4.2. Site-Specificity and the Scale of Landscape Acquisition

The coupling of architectural form with natural landscape is not merely an esthetic intuition but a quantifiable data description rooted in site-specificity [49]. The essence of this research is to predict the landscape accessibility for future residents realistically. In this specific case study, the target resources are large-scale natural entities: the Xiang River, the Mountain, and the Urban Park. The modeling strategy deliberately omitted vegetation interference for two reasons: first, the permeability of foliage allows for partial visibility [50]; second, trees primarily affect the lower floors, whereas the focus here is on the macro-relationship between the high-rise massing and the distant horizon [51].

However, this method is scale-dependent. If the research object were a small-scale landscape feature—for instance, if the abstract viewpoint used in the algorithm validation (Section 2, Figure 5) were hypothetically assumed to be a garden sculpture—vegetation occlusion would become a critical variable [52]. In such micro-scale scenarios, the model would need to account for the specific canopy dimensions to accurately measure the visible wall area. Thus, the proposed workflow is adaptable: it provides a rigid framework for quantification, but the definition of “obstacles” must be adjusted according to the scale of the landscape target (Macro vs. Micro).

4.3. Evidence-Based Design Strategies: From Morphological Topology to Programmatic Intervention

The core value of the proposed computational workflow lies in its ability to translate simulation data into tangible architectural strategies. Architectural design in this climate zone inevitably faces a “Vectorial Conflict”: the Solar Axis (optimal South orientation for thermal and health performance) opposes the Landscape Axis (East for the River, West for the Park). To address this, the Slab-Tower Hybrid employs a strategy of Topological Differentiation. The massing is twisted to capture the eastern river view and western park view. However, this morphological optimization introduces a secondary challenge: the orientation of the slab components deviates from the strict South, creating a bias towards the Southeast and Southwest. This deviation exposes the facade to excessive solar radiation during summer.

The simulation results necessitated a compensatory architectural intervention to resolve this thermal drawback. As illustrated in Figure 13, the design introduces Sky Communal Terraces to mitigate the solar gain while enhancing living quality. The specific generation mechanism and multidimensional benefits of this strategy are articulated as follows:

• Programmatic Interlocking (The 9-Meter Logic):

The distinct 9 m clear height of these terraces is derived from the structural rhythm of the unit mix. The design systematically integrates Duplex Units (spanning floors H and H + 1) with Simplex Units (on floor H + 2). By staggering the terrace locations—for instance, positioning the Duplex terrace on the West and the Simplex terrace on the East—a vertical void spanning three structural floors is created above each platform. This rhythmic arrangement articulates the facade and generates voluminous, airy communal pockets within the high-rise density.

2.

Climatic Response (Solar Altitude and Shading):

These terraces function as effective horizontal shading devices. In the “Hot Summer and Cold Winter” zone (Changsha, approx. 28° N), the Solar Altitude Angle is extremely high during the summer solstice, peaking near 85° around noon. The deep overhangs of the terraces are strategically designed to block direct solar radiation during the critical high-intensity window of 11:00–13:00. By intercepting the steep solar rays before they reach the interior, the terraces significantly reduce the cooling load despite the building’s orientation deviation.

3.

Dual Functionality (Social and Viewing):

Beyond their climatic function, these terraces serve as superior observation decks. Compared to indoor windows, these semi-outdoor platforms offer wider, frameless panoramic views of the river and mountains, creating a high-quality social space for residents to engage with the natural environment.

By integrating morphological twists with climatic shading structures, the study demonstrates a holistic Evidence-Based Design process. It proves that architectural form optimization is not about finding a “perfect” shape, but about balancing conflicting vectors (View vs. Sun) through specific spatial interventions.

4.4. Methodological Applicability and Directions for Further Optimization

This study presents a specific optimal solution—the Slab-Tower Hybrid—tailored to the unique landscape constraints of the Changsha site. It is crucial to acknowledge that while the computational workflow possesses universal applicability, the resulting architectural form is not a generalized formula. Every architectural project is essentially a unique innovation responding to distinct loci [53,54]. The value of this research lies not in prescribing a fixed shape, but in providing a rigorous, evidence-based tool that empowers architects to make informed decisions within complex, site-specific constraints.

While the current workflow successfully resolved the macro-morphological trade-offs, the methodology holds significant potential for deeper integration with architectural tectonics and performance analysis.

First, regarding decision-making in complex scenarios, the current optimization benefited from a site that allowed for the simultaneous maximization of all three landscape types. However, in more intricate urban environments, landscape resources often present conflicting availability. Future research should introduce a weighted evaluation system based on resident preferences or real estate valuation, enabling the algorithm to handle scenarios where trade-offs between different view targets are unavoidable.

More importantly, a critical direction for future research lies in the precise correlation between the “Visible Mesh” and the internal residential layout. Since the computational model has already achieved a high-precision resolution (1 m²) for the global envelope, and given that the overall building facade is essentially a composite of the exterior walls of multiple individual residential units, it is computationally feasible to map the global visibility data directly onto specific household envelopes. This data mapping would transform the design process from “Massing Optimization” to “Functional Allocation and Fenestration Strategy.” By identifying the specific facade zones with the highest cumulative visibility for each unit, architects can scientifically align primary habitable rooms (e.g., master bedrooms and living rooms) with these high-value surfaces, while relegating service spaces to low-visibility zones. This data-driven approach effectively prevents the mismatch mentioned in traditional design—such as placing a blank gable wall facing a prime landscape—by ensuring that the ‘visible patches’ on the envelope are activated through appropriate fenestration and functional placement.

Furthermore, this correlation directly dictates the design of the facade openings. Zones with high landscape accessibility warrant maximized glazing to capture the view, whereas zones with negligible visibility should adopt reduced window sizes or solid walls to conserve energy. This inevitably leads to a dialectical conflict between Landscape Acquisition and Thermal Insulation, as larger windows increase solar gain and heat loss. Therefore, a valuable trajectory for future work is to transition from global morphological optimization to local façade articulation. Building upon the optimized form established in this study, future research will investigate the quantitative equilibrium between window-to-wall ratios, visual quality, and energy consumption, thereby achieving a multi-scalar integration of architectural form and detailed fenestration.

5. Conclusions

This study investigated the complex trade-offs between landscape visibility, energy efficiency, and environmental comfort in high-density urban residential design. By comparing three distinct morphological prototypes under identical regulatory constraints, the research identified the Slab–Tower Hybrid as the optimal topological solution for the multi-source landscape context of Changsha.

Quantitative evidence from the simulations confirmed that the Hybrid form successfully resolves the inherent conflict between “View Maximization” and “Energy Conservation.” Unlike the Dispersed Tower, which relies on brute-force envelope expansion to increase visibility (resulting in a higher SC of 0.348), the Hybrid achieves the highest landscape acquisition—exemplified by a cumulative Park-View LVS approximately 1.8 times that of the Compact Tower—while maintaining a moderate SC (0.326). This performance demonstrated a superior “Visual Morphological Efficiency,” effectively delivering “more view with less envelope.” Simultaneously, in terms of environmental health, the Hybrid prototype overcomes the limitations of the other forms by ensuring stable cross-ventilation (0.4–0.7 m/s) and eliminating the “sunlight dead zones” inherent to the centralized core layout, thereby providing a balanced and healthy living environment.

Beyond the specific case study, this research contributes a methodological framework for “Target-Based Visibility Analysis.” Distinct from traditional methods that quantify visibility as a generic percentage of open space, the Inverse Targeted Ray-Casting method developed here shifts the focus to specific landscape resources. By integrating the R_v, the framework allows architects to verify whether the building’s geometric twists accurately align with the vector direction of valuable landscape nodes. This transforms view analysis from a passive post-design check into an active, precision-oriented design driver, ensuring that morphological articulation directly responds to site-specific assets.

Looking forward, while this research established the optimal logic at the morphological massing scale, the comprehensive realization of landscape performance requires a multi-scalar approach. Future research will shift focus from the macro-scale building form to the micro-scale adaptability of unit floor plans and facade articulation. Specifically, subsequent studies will investigate how internal functional layouts, window-to-wall ratios, and balcony configurations can be optimized to further capture the landscape targets defined in this study. The ultimate goal is to integrate these multi-scalar findings into a systematic design framework for high-rise residential architecture that seamlessly bridges the gap between urban morphology, energy constraints, and the human experience of nature.