Efficacy of Korea’s Diagonal Plane Restriction for Daylight: Focusing on South-North Direction Blocks within Seoul’s General Residential Districts

Abstract

Since the beginning of the 20th century, daylight access-related regulations have been an integral part of urban development. The Diagonal Plane Restriction in South Korea is one such regulation. This paper evaluates the effectiveness of the policy by investigating buildings in General Residential Districts in Seoul and the factors that have influenced its success. First, selected buildings were analyzed based on their relative core locations, massing morphology, and the glazing proportions on the south elevations. The initial results showed that nearly 50% of the selected buildings had cores on the south side, thus limiting daylight infiltration to the habitable spaces with a glazing ratio under 30%. Secondly, binary logistic regression was performed between identified types and variables to identify the primary factors affecting the ineffective arrangement. The year of a building’s construction and its primary use can significantly predict the effective applicability of a restriction. Lastly, further descriptive analysis shows how policy changes directly contributed to the growth of low-performing types and commercial uses in residential districts. The results of this study point to the shortcomings of urban policy-making without considering the particulars of each urban context. Our findings could assist planners and policy-makers in redesigning regulations for ensuring optimal daylight access in residential districts.

1. Introduction

Natural light has always been considered to be an essential component of architectural design since it has significant economic, ecological, psychological, and well-being benefits for residents [1]. The fact that exposure to direct sunlight significantly impacts human health has been confirmed in numerous previous studies. The direct sunlight and views from inside a room have been proven to reduce stress and improve mood, happiness, and work satisfaction [2,3]. Sun exposure also affects the body’s physiological activity and gene expression. It provides people with vitamin D, and sensible amount of sun reduces the risk of chronic illness and infectious diseases [4].

However, providing sufficient natural light to residents has become an immense challenge for many cities due to their growing densities. Although formal policies on access to light for habitable spaces began to be implemented in the U.S. as early as the late 19th century [5,6,7], policies in most other cities around the world started to be enacted in the mid-twentieth century [8,9,10]. Legislation on light access varies from country to country due to the different regional climate conditions, planning approaches, and urban density needs [11]. Historically, policies securing daylight for buildings have been based on either controlling urban forms for equitable access to sunlight or requiring sufficient window sizes and certain quantities of illumination [12]. The first type of legislation is often part of zoning regulations and requires that specific parts of buildings receive direct sunlight for sufficiently long periods during the day. The second type, which is most common in building codes, requires rooms to be equipped with sufficiently large window openings, which serves as a proxy for the actual illumination level in the room [13].

However, these zoning regulations, which impose floor area limitations to regulate the extent of urban density and daylighting, often cause public disputes. For example, maximizing floor area to get the most economical return from a new building often conflicts with the design goal of providing optimum interior daylight [14]. Further, controlling the urban form is often imperfect and unpredictable in many cities. Zoning codes were mainly designed to prevent specific outcomes—overcrowding of lots and streets—but they did not establish maximum lot sizes, build-to-lines, or minimum heights, which would have promoted a more uniform and predictable design [15]. With these shortfalls, continuous demand for and iterations of policy changes became inevitable.

As can be seen in Korea’s urban planning regulations, these conflicts have become a means of controlling the housing market beyond ensuring a healthy and sound environment. The first diagonal plane restrictions, the subject of this study, were imposed in the early 1970s to ensure that residents had adequate daylight access. Essentially, the purpose of these restrictions was to prevent the city’s living environment from deteriorating due to overcrowding. In 1980, the regulations became more specific for residential districts, and they required buildings to be built at a distance greater than the distance specified by the ordinance from the north side of the site boundary [9]. However, regardless of the regulation’s positive impact on improving the quality of urban living, the issues of ineffective land use and applicability for diverse land conditions have continuously been raised. In 2015, Korea’s Ministry of Land, Infrastructure, and Transport finally agreed to abolish the diagonal plane restriction from the roadside. However, the distance requirement from the north side still remains, and it remains a point of controversy.

Some site geometries and conditions make the measures ineffective, which is one of the various issues that have been raised. Despite the intent of the regulation, many buildings fail to utilize the south side of their buildings as fully habitable spaces, as shown in Figure 1. Consequently, buildings are often built with the south side higher than the north side, and the cores are inevitably located on the south side to facilitate access to the building’s top floor. This inevitably leads to situations in which the public portion of the building occupies the best-lit central space. This led to the research question of how widespread these problems are in urban areas and under what conditions these prescriptive uniform regulations fail to work.

Even though the phenomenon is particular to Korea, many other cities face similar dilemmas between as-of-right access to light, diversity in architectural design, and socio-economic growth. Consequently, previous kinds of research have pointed out limitations of daylight-related zoning regulations, especially for frontier cities like New York City but primarily focus on the public experiences on the streets or the urban morphological forms [15,16]. Furthermore, other research efforts regarding daylight-related zoning regulations raise critical concerns regarding the impact of urban development needs and the policy decision-making procedures. For example, cities like Tokyo, which have significantly influenced Korea’s initial urban legislation, have seen a massive building boom in recent decades partially by waiving similar diagonal plane restrictions for economic growth purposes and against livability and public well-being [17]. Recently, in the context of sustainable urban development, some researchers also discussed how daylight-related and density-control regulations affect energy consumption [18]. Similar studies are found in Korea, and previous studies have explored the urban morphologies derived from such regulations, highlighting the limitations of accommodating atypical urban tissues and unique characteristics [19,20]. Other studies have attempted to improve daylight performance by proposing optimum diagonal angles and offset distances through computer-generated simulations [21] or land use efficiency depending on various conditions, such as comparing the allowable floor area ratio (FAR) and achievable based on such height regulations [22]. However, these great research efforts have yet to provide a synthesized process that formalizes the consequential problems through an extensive survey and quantitative analysis with multiple constructed building samples. No study has been undertaken to determine how the building uses or other dynamic urban contexts and geometry reflect the problem and to investigate critical factors influencing it. Through a comprehensive survey and analysis of similar urban blocks in Seoul city, this research aims to classify urban morphologies spawned by the Diagonal Plane Restriction and determine the effectiveness of the regulations on mixed-use buildings in general residential districts and the factors that adversely affect them.

1.1. Zoning Regulations and Daylight

Daylight access is an essential architectural factor that substantially determines building shape, orientation, geometry, and interior space quality [23]. Therefore, legislation regarding daylighting has represented an integral part of modern society. Although such legislation varies from country to country, it tends to fall into three categories. The first relates to the access that buildings have to sunlight. This type of legislation attempts to guarantee buildings and their occupants’ access to sunlight for a predetermined time, typically by establishing local zoning ordinances stipulating how high buildings can be and their setbacks from property lines. The second type of legislation relates to the requirements for windows and sizes and is usually found in building codes. Finally, the third type relates to the quantity of indoor illumination inside a room [24]. Among these various aspects of daylight-related policies, this study focuses on zoning ordinances in Korea, which are considered to be an example of the first type.

Building height is one of the essential factors determining the streetscape and the lighting conditions of space, similar to building setback [25], and cities traditionally rely on zoning codes to control such conditions. Several cities primarily regulate height according to street widths, including Paris, New York, and Japan. In other cities, such as Singapore, maximum heights are restricted by special zoning districts, while in other cities, FAR is the only limiting factor, and the consequential height is controlled indirectly [26]. As shown in the example of New York City [14], daylight access is associated with the light conditions within the streetscape, not with the interior spaces of the buildings. In most other U.S. cities, FAR indirectly controls the heights of buildings to ensure sufficient light and ventilation on the street and nearby buildings [27]. New York City’s Zoning Guidelines represent an evolution of the original document from 1916, which regulated the shape of skyscrapers by requiring “setbacks” above certain heights. The current solution reflects changes from the early 1980s; the “sunlight provision” requires that new constructions obstruct no more than about 75 percent of the sky surrounding them. The field of view of a pedestrian is represented by the ‘daylight evaluation chart: D.E.C.’ (Figure 2) [28].

Historically, Korea’s zoning regulations related to daylight had two distinct focuses: the streetscape and the building’s interior spaces. ‘Diagonal Plane Restriction from the Roadside’ [29], which was abolished in 2015, was intended to maintain a healthy urban environment with appropriate wind and light access by securing sufficient distance between the roads and properties. Even though the abolition of this restriction led to positive effects for a less uniform urban landscape, no alternative regulations substituted the need for sufficient light for the urban environment [30]. A few major active commercial areas within Seoul City are subject to the Street Block Height Limit Restriction, which was introduced in 1999 but did not apply to most residential districts [31]. This regulation was introduced to complement the limitations of the Diagonal Plane Restriction based on street width for commercial areas. However, many studies indicate that urban conditions have worsened since the height requirements were alleviated [32]. The remaining ‘Diagonal Plane Restriction from the Northside [33]’, which is specific to the General Residential Districts, focuses on the access that individual buildings have to daylight rather than the streetscape. Regarding the height of buildings to be built, it is necessary to keep a distance from the adjacent property’s boundary in the north-north direction to secure sunlight, and the height must be less than or equal to the height prescribed by the ordinance.

Figure 3 illustrates how these regulations affect buildings from earlier years and more recent buildings when the road is on the south side of the building. The diagram ‘before 2012’ shows the building form under the influence of both ‘Diagonal Plane Restriction from the Roadside’ and ‘Diagonal Plane Restriction from the Northside’. The first floor must be 1 m from the site boundary and at least 2 m from the second floor. Further, the height up to the third floor must not exceed 8 M. The diagram ‘2012~2015’ demonstrates the change in ‘Diagonal Plane Restriction from the Northside’; specifically, aside from the case of a height of 9 m or less, the diagonal plane must have 1.5 m or more in the north-south direction. The regulatory requirement from 2012 to 2015 mandated that a part with a height exceeding 9 m had to be at least 1/2 of each part of the relevant building. The diagram ‘after 2015’ shows how the building form has changed since the ‘Diagonal Plane Restriction from the Roadside’ was abolished after 2015.

1.2. Building Codes and Daylight

In addition to zoning regulations, various building codes also play an essential role in securing adequate daylight for residential buildings. Residents can enjoy both light and visual contact at a particular latitude, where sunlight primarily enters through windows [34]. Therefore, the window size and the illumination quantity have always been crucial criteria in developing relevant regulations for daylighting [10]. The most preferred direction in the northern hemisphere is the south because it gives residents the best possible daylight [35], which is the reason behind the diagonal plane restriction.

In Sweden, the general recommendation is either to ensure a minimum room glass-to-floor ratio of 10% or to meet a daylight factor threshold of 1% on any specific point in a room [36]. Japanese building code stipulates that habitable rooms in continuous occupancy buildings should have window sizes that are not less than 14% or 1/7th of the total floor area of the building and between 20% and 40% of the floor area in other types of buildings. In Hong Kong, for rooms, at least one window is required: (i) the area of the glass must be not less than 1/10th of the floor area of the room, and (ii) the windows must, to an extent, be at least equal in the aggregate to 1/16th of the floor area of the room, and be opened in such manner that the top of the opening of each window is at least 2 m above the floor [37]. Korea’s building code also requires the window opening of a room to be not less than 1/10th of the floor area of the room [38].

Zoning regulations are the primary focus of the present research; however, the relative proportions of windows on the south elevation are used to measure how effective the regulation is. Therefore, the importance of this study is further reinforced by the building codes about windows and illuminance mentioned above.

1.3. Seoul’s Residential Districts

This paper evaluates the effectiveness of the ‘Diagonal Plane Restriction’ specifically for mixed-used buildings in parts of General Residential Districts in Seoul. To this end, it is necessary to discuss the distinctive characteristics of Korea’s zoning regulations for specific districts, which differ from those of many other countries.

Japan adopted zoning in 1919 and subsequently imposed its system on the Korean peninsula through the Chosun Urban District Planning Decree of 1934. Under the decree, Seoul began implementing a zoning system with three categories of zoning districts—Residential, Commercial, and Manufacturing—which expanded to 12 in 1988 [39]. Residential districts can be divided into three categories: Residential Only, General Residential, and Semi-Residential. General Residential Districts (G.R.D.) can be divided into three sections of differing densities: G.R.D. 1 refers to low-density residential areas, G.R.D. 2 to medium-density areas, and G.R.D. 3 to high-density areas. G.R.D. With some exceptions, G.R.D. 3 mainly consists of apartment housing complexes. The scope area of this study is primarily G.R.D. 2 and G.R.D. 3.

The city’s zoning system legally allows for a variety of commercial and business uses in residential districts, often creating areas with almost full commercial use (Figure 4). Some see this provision as offering greater flexibility than cities in the U.S. and Europe and providing the capability to produce diverse mixed-use neighborhoods [40]. However, the infrastructure of many of these areas, such as their roads and waste management systems, are not designed to support a high ratio of commercial activities. The phenomenon is most visible closer to subway stations with a large floating population. Although these commercialized areas are one of the most “universal” areas in Korea, there are many difficulties involved in creating a sound neighborhood living environment due to the uniform application and indiscriminate development of urban building regulations [41].

This study includes urban blocks in these unique parts of Seoul’s Residential Districts to investigate how the building’s programmatic uses relate to the Diagonal Plane Regulation’s effectiveness.

2. Materials and Methods

2.1. Research Scope and Data Preparation

This study analyzes north-south directional blocks within the Type 2 and Type 3 G.R.D. near the Sadang Station and the Anam Station (Figure 5). Since the Diagonal Plane Regulations for securing the right to daylight are based on the distance from the boundary line of adjacent land in the north direction, the properties within the north-south directional blocks are typically more constrained than those within the east-west directional blocks. To satisfy the offset distance requirements, east-west directional blocks can utilize the road width, located either on the north side or when the road is located on the south side, which comprises a certain distance that is guaranteed in the urban context. By contrast, north-south directional blocks cannot benefit from such contextual advantages. Moreover, the blocks in this present analysis include buildings with a high proportion of commercial and business use as their primary function compared to other buildings in General Residential Districts. In this study, 109 buildings near the Sadang Station and 101 buildings near the Anam Station—i.e., 210 buildings in total—are analyzed.

The data for this research are primarily obtained from the ‘Toji-Eum’ website, which is a comprehensive portal service provided by Korea’s Ministry of Land, Infrastructure, and Transport. The portal includes land use and city-wide planning information for all areas within the peninsula. Moreover, all other morphological and geometrical data are collected from GIS-based web platforms and physical measurements during the site visits (

Table 1. Summary of the data sources for each variable.

Variable Type	Data Source
Massing Arrangement; Site Depth-D(m); Site Width-W(m); Window-Wall Ratio	Site Visit and Physical Survey; https://map.naver.com (accessed on 2 September 2022) https://www.landbook.net (accessed on 2 September 2022) https://www.valueupmap.com/ (accessed on 2 September 2022) https://smap.seoul.go.kr/ (accessed on 2 September 2022)
Primary Use; Building FAR (%); Building Coverage Ratio (%); Built Year; No. of Floors	http://www.eum.go.kr/web/am/amMain.jsp (accessed on 2 September 2022)

).

2.2. Methods and Process

The study consists of four main stages: (1) Selecting sample buildings and gathering data, (2) Defining massing types according to the relative location of the building cores, (3) Analyzing the different massing types and daylighting effectiveness levels of the south elevation, and (4) Processing the data by using statistical analysis methods. Figure 6 shows the framework of the research methods and process.

By statistically analyzing the different massing types and the factors associated with their formation, this study attempts to identify the factors that significantly affect massing types and limit the effectiveness of the diagonal plane restriction. Binary logistic regression analysis was conducted using IBM SPSS Statistics 26 to identify the most significant common factors affecting the sample massing types. The ‘Statistical Package for the Social Sciences (SPSS)’ is a package of statistical programs from I.B.M. for solving research problems through analysis, hypothesis testing, geospatial analysis, and predictive analytics [42].

Binary regression analysis is a type of logistic regression analysis, and unlike linear regression analysis, it does not limit the dependent variables to continuous numbers. Binary logistic regression is a form of regression that is used when the dependent variable is a true or forced dichotomy and when the independent variables are of any type. Multinomial logistic regression was designed to handle the case of dependent variables with more classes than two, though it is sometimes also used for binary dependent variables since it generates somewhat different output [43]. With the ineffective type being a dichotomous variable, the binary logistic regression model is considered to be an appropriate model in which both continuous and categorical independent variables are applicable. The factors were evaluated in detail while utilizing statistical graphs to determine whether there was any congruity in the trends and to investigate further the potential cause of the patterns found in the regression models.

2.3. Defining Massing Types and Variables

This section describes the data in detail, along with a few variables that require further explanation (

Table 2. Characteristics of massing types.

Type	Plan Arrangement	Massing Types		Primary Use	Building FAR	Building Coverage Ratio	Built Year	Building Area	Site Depth (D)	Site Width (W)
S1			-	R: 47.6% B: 0.0% NC1: 4.8% NC2: 47.6%	155.54%~ 276.8% (Mean: 208.8%)	45.85%~ 59.98% (Mean: 56.2%)	1979~ 2021	319 m2~ 1582 m2 (Mean: 599 m2)	10.5 m~ 22.4 m (Mean: 14.5 m)	10.7 m~ 20.3 m (Mean: 14.4 m)
S2				R: 60.0% B: 10.0% NC1: 0.0% NC2: 30.0%	135%~ 245.7% (Mean: 190%)	49.8%~ 59.93% (Mean: 56.7%)	1985~ 2015	97 m2~ 938 m2 (Mean: 390 m2)	9.6 m~ 15.9 m (Mean: 13.3 m)	11.7 m~ 23.9 m (Mean: 13.3 m)
S3				R: 29.7% B: 1.6% NC1: 18.8% NC2: 50.0%	127.97%~ 292.39% (Mean: 205%)	45.03%~ 63.25% (Mean: 55.9%)	1980~ 2019	144 m2~ 2230 m2 (Mean: 625 m2)	9.9 m~ 22.6 m (Mean: 15.4 m)	8.6 m~ 21.7 m (Mean: 14.2 m)
M1			-	R: 60.0% B: 0.0% NC1: 6.7% NC2: 33.3%	137.85%~ 258.21% (Mean: 193%)	40.2%~ 59.94% (Mean: 56%)	1987~ 2021	292 m2~ 2332 m2 (Mean: 678.5 m2)	11.3 m~ 22.7 m (Mean: 14.2)	9.4 m~ 20 m (Mean: 17 m)
M2				R: 68.8% B: 0.0% NC1: 6.3% NC2: 25.0%	125.19%~ 288.96% (Mean: 202%)	45.88%~ 59.96% (Mean: 56%)	1974~ 2019	250 m2~ 2230 m2 (Mean: 613.9 m2)	10.1 m~ 23.4 m (Mean: 12.5 m)	9.3 m~ 25.5 m (Mean: 15.8 m)
N1			-	R: 45.5% B: 0.0% NC1: 0.0% NC2: 54.5%	97.67%~ 242.53% (Mean: 180%)	49.46%~ 59.87% (Mean: 56%)	1968~ 2018	128 m2~ 960 m2 (Mean: 575.2 m2)	9.6 m~ 28.7 m (Mean: 17.7 m)	2 m~ 28.2 m (Mean: 13.3 m)
N2				R: 57.1% B: 28.6% NC1: 14.3% NC2: 0.0%	132.5%~ 269.35% (Mean: 210%)	57.69%~ 59.81% (Mean: 59%)	1986~ 2003	303 m2~ 889 m2 (Mean: 528.6 m2)	13.4 m~ 16.7 m (Mean: 15 m)	10.5 m~ 18.3 m (Mean: 14 m)
N3				R: 42.1% B: 0.0% NC1: 31.6% NC2: 26.3%	49%~ 298.56% (Mean: 201.43%)	39.37%~ 143.6% (Mean: 205%)	1967~ 2017	98 m2~ 2230 m2 (Mean: 510.8 m2)	9.1 m~ 19.5 m (Mean: 13.4 m)	8.7 m~ 20.8 m (Mean: 14.3 m)
EX			-	R: 73.1% B: 7.7% NC1: 7.7% NC2: 11.5%	68.66%~ 205% (Mean: 157%)	43.07%~ 57.04% (Mean: 56%)	1968~ 1989	115 m2~ 2644 m2 (Mean: 479.4 m2)	8.9 m~ 23.4 m (Mean: 14.5 m)	10.3 m~ 24.5 m (Mean: 14.6 m)

). Even though the morphological types can be classified in several different ways, we have categorized representative buildings based on the core locations, which significantly impact the elevations at which daylight infiltration is most prominent.

2.3.1. Massing Types

Upon investigating the buildings near Sadang Station and Anam Station, similarities and differences—in particular, those involving the arrangement of the plans—gradually emerged and became critical elements defining the different massing types. Depending on the orientation and location of the core, all buildings were classified as S, M, or N. EX represents buildings with two or fewer floors, which are considered to be minimally affected by Diagonal Plane Restriction. In Type S, the core is located on the south side; in Type M, it is located on the middle part; and in Type N, it is located on the north side. Next, massing types S, M, and N were subdivided per more specific locations of cores. The core is located near the road on the south side of the building in S1. S2 is for buildings with the core located in the middle, and S3 is for those with the core away from the road on the south side. M1 is when the core is located near the road in the middle part of the plan, whereas M2 is when the core is in the central part, away from the road. N1 is for buildings with the core near the road on the north side, N2 is for those with the core located in the middle, and N3 is for those with the core away from the road on the north side.

2.3.2. Primary Use

Building permits and registrations in Korea require owners to list a single primary use, even for mixed-use buildings [44]. Korea’s zoning regulations allow for a greater variety of commercial and business uses within the General Residential Districts to facilitate the creation of convenient residential areas. This represents a clear difference from other countries, which strictly limit commercial or business use in residential zoning districts. Korean law has adopted the concept of Neighborhood Commercial Uses to manage commercial uses in residential districts [40]. Even though the selected buildings are located within the residential districts, the main uses vary from residential to various commercial uses. For this reason, we have included ‘Primary Use’ as one of the critical variables. R is the abbreviation for Residential; B is the abbreviation for Business; NC1 stands for Type 1 Neighborhood Commercial Facilities, and NC2 stands for Type 2 Neighborhood Commercial Facilities.

2.3.3. Building FAR

Floor area ratio (FAR) is the ratio of the above-ground floor area to the site area of the building, which is often lower than the legal limit. In Korea, over the last few decades, multiple changes have been made to the allowable FAR policy, reflecting social demand. Figure 7 illustrates the range of FAR for the selected buildings based on their construction years.

2.3.4. Building Areas

The building area includes the above and below-ground floor areas, excluding rooftop structures. We have also separated areas into Residential and Non-Residential purposes. Residential uses include apartments and single-family housing; Non-Residential purposes include business facilities, education, and research facilities, warehouses, and various neighborhood commercial uses. Figure 8 depicts the distribution of geometrical characteristics of the samples; site area, building area, and site width and depth. The box ranges from the first quartile to the third quartile of the distribution and the range represents the interquartile range. A line across the box indicates the median, and outliers are shown as open circles. Figure 9 depicts the variations in the building area of each sample.

2.4. Analysis of Southern Elevations for Daylighting

Based on the massing types listed in Table 2, we analyzed the south elevations of the sample buildings. The windows for the habitable spaces are directly affected by the daylight, and they are most critical compared to the windows for the core or corridor. To calculate the ratio of the rooms that can benefit from the Diagonal Plane Restriction, both ‘Core Exterior Glazing Ratio’ and ‘Habitable Space Glazing Ratio’ were analyzed. As presented in

Table 3. Analysis of potential daylight access on south elevations.

Type 1		Type 2		Type 3		Type 4
Massing type: S1		Massing type: S2		Massing type: S3		Massing type: M/N
Core Exterior Glazing Ratio: 1.6%~39.2%		Core Exterior Glazing Ratio: 1.5%~9.6%		Core Exterior Glazing Ratio: 3.2%~23.3%		Core Exterior Glazing Ratio: 0%
Habitable Space Gazing Ratio: 6.5%~27.8%		Habitable Space Gazing Ratio: 16%~43.7%		Habitable Space Gazing Ratio: 12.3%~29.3%		Habitable Space Gazing Ratio: 42.3~66.4%

, the glazing ratio is calculated as a percentage of the total wall area. In the case of the S1 and S3 types, where the core accounts for a significant portion of the elevation, the proportion of lighted windows receiving substantial sunlight was less than 30% of the total area, which was much lower than that of the S2 Type M and Type N, whose proportions range from 43.7% to 66.4%. The results are primarily due to the core interfering with the habitable rooms in the south, where daylight is essential. As can be seen in the examined cases, despite the intended purpose of the regulations, it is difficult for smaller buildings to fully utilize the benefits provided by the regulations due to various constraints.

2.5. Effectiveness of Regulation on Selected Buildings

As opposed to Types M and N, where the south elevation is entirely devoted to habitable space, Type S inevitably has a significant area occupied by public space (core). Based on the southern elevation analysis, in particular, types S1 and S3 display the lowest proportion of windows in habitable rooms, thus indicating the weakest performers for the Diagonal Plane Restrictions. Compared to Types S1 and S3, Type S2, in which the short side of the core interferes with the south elevation, appears to be less problematic. Figure 10 shows the quantities of the different massing types, and Types S1 and S3 together account for 40% (85) of the total sample quantities. Further, Figure 11 shows the actual distribution map of all types, and Figure 12 shows the distribution of S1 and S3, where no significant pattern is found in the distribution.

3. Results

Binary Logistic Regression Analysis

This study uses statistical analysis to identify factors that are significantly correlated with specific massing types when the diagonal plane restriction is applied in an ineffective manner. The earlier observational research shows that, among the nine massing types (Type S1~3, Type M1~2, and Type N1~3, Type EX), types S1 and S3 show significantly lower glazing proportions at the south elevations of buildings, yet the highest among the 210 samples. Binary Logistic Regression was used to test the research objective.

The dependent variable was taken in terms of whether the building was classified as type S1 or S3 (0- is Type S1 or S3; 1- is NOT Type S1 or S3). The predictor variables included zoning, primary use, site area, width, depth, building coverage ratio, residential and residential use percentage, and the number of floors. The zoning period and primary uses were further divided into separate variables per categorical options.

Based on the ‘Binary Logistic Regression Analysis,’

Table 4. Binary logistic regression analysis result: S1S3 or NOT.

Variables in Equation S1S3 or NOT
		B (Coefficients)	Wald	df	p-Value (Sig.)	Odds Ratio: Exp(B)
Step 1a	Built Year	−0.067	10.013	1	0.002 ***	0.935
	Main Program: Residential vs. Non-residential	−0.392	5.446	1	0.020 **	0.676
	Site Area	0.004	1.286	1	0.257	1.004
	Building Coverage Ratio (%)	0.002	0.013	1	0.908	1.002
	Site Width-North South Direction	0.082	1.533	1	0.216	1.086
	Site Depth-West East Direction	−0.043	0.458	1	0.499	0.958
	Building Area	0.006	1.389	1	0.239	1.006
	Building FAR	−0.003	0.380	1	0.538	0.997
	Residential Use	0.092	0.984	1	0.321	1.097
	Non-residential Use	0.078	0.688	1	0.407	1.081
	No. of Floors	−0.112	0.294	1	0.588	0.894

*** Significant at the 99% confidence level, ** Significant at the 95% confidence level.

provides the regression coefficients, odds ratios, and statistical significance. The p-values indicate that two of the eleven variables significantly predict the classification outcome at a level above 5%. The results show that the most significant predictors for massing types with the ineffective application (S1, S3) of the Diagonal Plane Restrictions are the ‘Built Year’ and ‘Primary Use’. For example, the probabilities of S1 and S3 buildings being constructed yearly decreased by a factor of 0.945 (p-value < 0.01). Further, they are 0.676 (p-value < 0.01) times less likely to be used for residential purposes. Despite the need for a more detailed analysis, we can say that buildings with ineffective regulation applications (S1, S3) have been affected by the construction years and are primarily used for non-residential purposes.

The influencing factors for effective buildings (S2, M, N) were also examined using Binary Regression Analysis, even though Type M and N buildings are not as crucial to the objectives of this research.

Table 5. Binary logistic regression analysis result: S or NOT.

Variables in Equation S or NOT
		B (Coefficients)	Wald	df	p-Value (Sig.)	Odds Ratio: Exp(B)
Step 1a	Built Year	−0.056	7.363	1	0.007 ***	0.945
	Main Program: Residential vs. Non-residential	−0.498	8.138	1	0.004 ***	0.608
	Site Area	0.004	1.698	1	0.193	1.004
	Building Coverage Ratio (%)	0.000	0.000	1	0.982	1.000
	Site Width-North South Direction	0.090	1.889	1	0.169	1.094
	Site Depth-West East Direction	−0.102	2.890	1	0.089	0.903
	Building Area	0.007	1.683	1	0.195	1.007
	Building FAR	0.000	0.006	1	0.936	1.000
	Residential Use	0.129	1.273	1	0.259	1.138
	Non-residential Use	0.128	1.227	1	0.268	1.136
	No. of Floors	−0.240	1.383	1	0.240	0.787

*** Significant at the 99% confidence level.

,

Table 6. Binary logistic regression analysis result: M or NOT.

Variables in Equation M or NOT
		B (Coefficients)	Wald	df	p-Value (Sig.)	Odds Ratio: Exp(B)
Step 1a	Built Year	−0.048	2.437	1	0.118	0.953
	Main Program: Residential vs. Non-residential	0.066	0.065	1	0.799	1.068
	Site Area	0.000	0.003	1	0.954	1.000
	Building Coverage Ratio (%)	0.030	0.337	1	0.561	1.030
	Site Width-North South Direction	−0.266	9.300	1	0.002 ***	0.766
	Site Depth-West East Direction	0.154	3.273	1	0.070	1.166
	Building Area	−0.018	0.944	1	0.331	0.982
	Building FAR	−0.001	0.029	1	0.866	0.999
	Residential Use	−0.349	0.677	1	0.411	0.705
	Non-residential Use	−0.326	0.594	1	0.441	0.722
	No. of Floors	0.093	0.098	1	0.755	1.098

*** Significant at the 99% confidence level.

and

Table 7. Binary logistic regression analysis result: N or NOT.

Variables in Equation N or NOT
		B (Coefficients)	Wald	df	p-Value (Sig.)	Odds Ratio: Exp(B)
Step 1a	Built Year	0.097	16.344	1	0.000 ***	1.102
	Main Program: Residential vs. Non-residential	0.411	4.966	1	0.026 ***	1.508
	Site Area	−0.001	0.127	1	0.722	0.999
	Building Coverage Ratio (%)	−0.009	0.178	1	0.673	0.991
	Site Width-North South Direction	0.108	2.523	1	0.112	1.114
	Site Depth-West East Direction	0.010	0.031	1	0.860	1.010
	Building Area	0.001	0.040	1	0.841	1.001
	Building FAR	0.004	0.511	1	0.475	1.004
	Residential Use	0.024	0.035	1	0.852	1.024
	Non-residential Use	0.024	0.036	1	0.850	1.024
	No. of Floors	−0.579	6.283	1	0.012 ***	0.560

*** Significant at the 99% confidence level.

present the results. For Type M, only one variable—’Site Width’—has a significant association with the specific type. Coefficients of the site width and the type are negatively associated with the odds of 0.768 (p-value < 0.01). The results show that, for each unit increase in width, the probability of being Type M decreases by 0.768 times. For Type N, ‘Zoning Period,’ ‘Primary Use,’ and ‘Number of floors’ are all significantly correlated with the specific types. The coefficients of all three predictors are negatively associated with type N.

Even though the site width and the number of floors affect some types, the ‘Built Year’ and the ‘Primary Use’ of the buildings appear to be the most significant predictors for types S1 and S3, and the next section will examine how each variable is related to the proposed types.

4. Discussion

Unlike many previous studies discussing similar policies, which focus on the consequential urban morphology or effects on the public realm beyond the buildings, this study aims to quantify the interaction between the policy efficacy and the complex social contextual factors. The most influential factors affecting the policy’s effectiveness were identified through comprehensive qualitative observation and categorization combined with quantitative statistical analysis.

4.1. Built Year and Effectiveness of the Regulation

The results of the previous binary analysis showed that ‘Built Year’ is one of the critical factors to consider. To better understand the correlation between the built year and type, we examined the regulatory changes that may have influenced the morphology of the sample buildings at a particular period. This includes but is not limited to changes in building FAR, site coverage ratio, parking requirements, etc. Following this, the built years were categorized into five ‘Zoning Periods’ depending on the essential FAR changes in the relevant regulations (

Table 8. Zoning period categorization.

Zoning Period	Years	Floor Area Ratio (%)	Diagonal Plane Restriction Changes
P1	Before 1979	200	Enactment of the relevant law (1971)
P2	1980~1985	180
P3	1986~1999	250 > 300 > 400 (1994)
P4	2000~2015	250	9m Offset from North Side, (2012, See Figure 3)
P5	2015-	250	Abolishment of Road Side Restriction (2015)

, Figure 13). Results were obtained based on descriptive analyses of the samples grouped according to the predetermined zoning periods.

As shown in Figure 14, over 85% (182) of properties were constructed after the 1990s, i.e., P3, which was marked by radical changes in FAR. The prevalence of Type S3 and Type N3 increased during this period. It should be noted that changes in the overall number of constructions for each period are not necessarily proportional to changes in the composition of the different types. For example, during P3~P5, the proportion of N types significantly decreased, whereas S types, primarily type S1 and S3, increased, accounting for more than 60% of the overall constructions during P4~P5 (Figure 15).

It is evident that the types exhibiting ineffective applicability of the Diagonal Plane Restrictions (S1 and S3) are on the rise. By contrast, there has been a rapid decline in N types that would benefit most from the restrictions, which clearly illustrates the policy’s limitations. Although further research is needed to quantify the causes, the results indicate that fewer buildings have been constructed with south-facing units or habitable spaces in recent years. This appears abnormal, given Korea’s cultural preference for southern exposures. Further, since the sample blocks are located near the boundary of General Residential Districts, which have a high concentration of commercial uses, the results could also suggest changes to the use of buildings ranging from residential to commercial. A detailed discussion of the interrelation will be presented in the following subsection.

4.2. Primary Uses and Effectiveness of the Regulation

The previous binary analysis result showed that primary use is another critical factor to consider. Thus, we continued with further descriptive analysis since the specific uses could provide additional information about the regulations’ applicability. As shown in Figure 16, 52% (109) of the sample buildings’ primary uses were for non-residential purposes, which is unexpectedly high considering that the blocks are located within the General Residential Districts. Moreover, the non-residential use proportion was significantly higher for Types N3 (57%, 22) and S3 (70%, 45).

Some previous research highlights one of the recent phenomena occurring in Seoul city: the commercial gentrification and proliferation of food and beverage stores in regions with high volumes of pedestrians [45,46]. Examples of cases include parts of General Residential Districts, such as our sample blocks adjacent to major transportation stations.

Neighborhood commercial uses within residential districts have also been subject to the most frequent changes in regulations [47] compared to other uses. In the early 90s, the government established NC Type 2 for extensive commercial use, including karaoke bars, massage parlors, car sales, and small offices [48]. Such zoning changes have accelerated the commercialization of residential districts’ boundaries and increased the number of buildings wherein light infiltration is not necessarily a primary design consideration.

Many commercial spaces are less dependent on daylight than residential spaces; hence, benefitting from the Diagonal Plane Restrictions became less critical for increasing property values. Therefore, we can conclude that the increase in Types S1 and S3 in recent years can be attributed to this change in primary uses, as shown in Figure 16.

5. Conclusions

This research analyzed 210 sample buildings within Seoul City’s General Residential Districts near Sadang and Anam Stations to determine whether or not the Diagonal Planes Restrictions are effective for daylight access and what factors influence their success.

The results of the initial descriptive analysis showed that 94 (46.7%) of the samples had cores located at the south side of the building occupying nearly 40% of the south elevation, which is critical for daylight exposure. Consequently, the overall ratio of windows at south elevations for such types was less than 30%, which was attributed to the contradictory situation whereby the shared public core spaces occupied a significant portion of the south elevation, thus limiting the allocation of habitable rooms. Further statistical analysis was conducted for S1 and S3 types, which were identified as the most problematic types. The binary analysis results presented in this work identify two statistically significant common factors affecting the effectiveness of Diagonal Plane Restriction: built year and primary use.

Among the buildings constructed after 1998, the proportion of samples exhibiting ineffective application of the Diagonal Plane Restrictions (Type S1, S3, 62%) is greater than the percentage of effective cases (Type S2, M, N, 38%) and has been rising. The regulation changes to the primary uses appear to be interrelated with the results. The types presenting the ineffective application of the Diagonal Plane Restrictions had higher proportions of non-residential primary uses, where daylight access is considered to be less critical that it is for residential primary uses.

It is possible for regulations such as Diagonal Plane Restriction, which initially intended to provide adequate lighting for south-facing elevations of buildings, to have an adverse effect on internal layout, thus causing contradictory results. Therefore, although there is a need for a comprehensive analysis, the above findings indicate the need for the creation of additional contents that enable the flexible application of the regulations, which are currently applied uniformly without taking into account the characteristics of individual properties, uses, and contexts. This result also provides an exemplary case that illustrates the limitations of Euclidian zoning policies in light of ongoing global discussions regarding form-based or performance-based zoning. Additionally, the convergent research process used in this study, both qualitative and quantitative, provides a methodology that may be useful in evaluating other cities’ policies where local zoning ordinances stipulate the height and setback of buildings for daylight access.

Due to the small number of samples, there are inherent limitations in identifying exact solutions based on the observations presented in this paper. Consequently, we are considering further research to develop alternative standards to ensure daylight in residential districts through simulations and models considering diverse context-specific characteristics.