New Approaches in Dynamic Metrics for Lighting Control Systems: A Critical Review

Abstract

The growing number of daylighting metrics—often overlapping in scope or terminology—combined with the need for prior familiarization to interpret and apply them effectively, has created a barrier to their adoption beyond academic settings. Consequently, this study analyzes a representative set of established and emerging daylighting metrics to evaluate applicability, synergies, and limitations. Particular attention is given to their implications for occupant health, well-being, performance, and energy use, especially within the context of sensorless smart control systems. A virtual room model was simulated using DaySim 3.1 in two contrasting climates—Seville and London—with varying window-to-wall ratios, orientations, and occupancy schedules. The results show that no single metric provides a comprehensive daylighting assessment, highlighting the need for combined approaches. Daylighting Autonomy (DA) proved useful for task illumination, while Useful Daylight Illuminance (UDI) helped identify areas prone to excessive solar exposure. Spatial metrics such as Spatial Daylight Autonomy (sDA) and Annual Sunlight Exposure (ASE) offer an overview but lack necessary granularity. Circadian Stimulus Autonomy (CSA) appears promising for evaluating circadian entrainment, though its underlying models remain under refinement. Continuous Overcast Daylight Autonomy (DAo.con) shows the potential for sensorless lighting control when adjusted for orientation. A nuanced, multi-metric approach is therefore recommended.

1. Introduction

1.1. Background

Over the past few decades, increasing emphasis on energy efficiency and occupant well-being has driven significant innovation in architectural daylighting strategies. One of the most notable outcomes of this shift has been the development of a wide range of daylighting metrics designed to assess the availability and effective use of natural light within buildings. Many of these metrics were developed with the explicit aim of assessing and promoting the effective use of natural daylight in buildings, thereby reducing the reliance on electric lighting as a logical consequence. This has led to their integration into more advanced applications, such as intelligent lighting control systems. A particularly relevant development in this area is the emergence of sensorless control strategies [1], where simulation-based metrics act as predictive inputs for lighting regulation, enabling cost-effective, sensor-free solutions based on architectural and climatic parameters.

What follows is a brief overview of the role daylighting plays in architectural design, energy efficiency, and occupant well-being, serving as a foundation for the critical analysis of a selection of daylight metrics—both well-established and emerging—that exemplify the current landscape of tools available for evaluating and guiding daylight-responsive design.

Effective daylighting strategies offer significant energy-saving potential [2,3,4,5], reducing reliance on electric lighting [6,7,8] and contributing to a building’s overall environmental performance. Lighting design in educational spaces is paramount, requiring attention to visual comfort for learning activities [9] and appropriate thermal conditions [10,11]. Notably, daylight significantly contributes to students’ perceptual capacity [12], a relationship closely linked to the configuration of fenestration [13].

While lighting affects visual and thermal comfort [14,15], its impact extends to non-visual biological processes, particularly the regulation of melatonin, a hormone synthesized by the pineal gland that is strongly modulated by light exposure [16]. The suppression of melatonin by light—particularly short-wavelength (blue) light—quantified as circadian stimulus (CS), is essential for the synchronization of the circadian rhythm, influencing sleep, alertness, and other physiological functions [17].

The human circadian system’s reaction to light diverges significantly from that of the visual system. While visual acuity is most responsive to middle-wavelength light, peaking around 555 nm, circadian entrainment, as measured by melatonin suppression, exhibits heightened sensitivity to short-wavelength (blue) light, with a peak near 460 nm [18]. The circadian system is also critically dependent on the timing of light exposure; light before the minimum core body temperature delays the circadian clock, whereas light after advances it. This system exhibits a temporal dependency, requiring minutes of light exposure to elicit a response, contrasting with the milliseconds of visual response. Empirical models, such as the one developed by Rea et al. in 2005 [16], aim to provide a more nuanced understanding of how light influences circadian rhythms. The sensitivity of the circadian system to light is modulated by prior light exposure, with higher daytime exposure reducing nocturnal sensitivity. Daylight, characterized by its optimal spectrum, intensity, duration, and timing, stands as the ideal synchronizing agent for the human circadian system, a role it has fulfilled for millennia.

Despite daylight’s idealness for synchronizing the human circadian system, in contemporary indoor environments, electric lighting often blurs the distinction between day and night, disrupting natural circadian rhythms and potentially compromising human health. Research has linked circadian disruption to various health issues, including increased morbidity [19], depression [20], and multiple sclerosis [21], with adolescents demonstrating heightened vulnerability [22].

Consequently, the study of daylight in interior spaces is essential for a myriad of reasons. However, daylighting analysis is inherently complex due to the dynamic nature of sunlight. The variability of solar irradiance—affected by factors such as time of day, season, and weather conditions—presents a challenge for designers attempting to accurately predict and optimize daylight performance [23]. Traditional static daylighting maps, which offer a snapshot of illumination at a fixed point in time, fail to capture these temporal variations, limiting their utility in comprehensive daylighting assessments.

Advancements in computational capacity over the past decades have facilitated the development of new indices that allow for a more precise analysis of daylighting in interior spaces. Additionally, these new indices support the implementation of Dynamic Lighting Control Systems (DLCSs) whose effectiveness in maximizing daylight utilization and reducing energy consumption for electric lighting is well-established [24,25,26,27].

Previous reviews have analyzed the impact of dynamic daylighting metrics on building energy performance, such as those by Reinhart et al. [28], and on visual comfort, as discussed by Carlucci et al. [29]; however, they have not explored the implications of these metrics on occupant health and well-being in-depth. Therefore, the new focus of this paper on the health and well-being impacts of dynamic daylighting metrics is both timely and pertinent.

1.2. Objectives

The main objective of this study is to conduct a critical review of key static and dynamic daylighting metrics, with a particular focus on their implications for occupant health and well-being, as well as their role in achieving energy efficiency through reduced electric lighting and its potential impact on thermal loads through solar gains.

Building upon this review, the research further aims to address several key aspects: first, to determine the sensitivity of each metric to various design parameters; second, to explore the correlation, utility, and complementarity among different metrics applied to a singular case study; and finally, to identify the inherent limitations associated with the information each metric provides.

2. Review

2.1. Static Metrics

The Daylight Factor (DF) is defined as the ratio of the illuminance value at a point within an interior space to the exterior illuminance value under a specific overcast-sky condition [30]. Said ratio can be expressed as follows (1):

$$\mathrm{D}\mathrm{F}={\displaystyle \frac{\mathrm{E}\mathrm{i}}{\mathrm{E}\mathrm{e}}}$$

(1)

This metric is considered one of the oldest daylighting indicators—dating back to the early 20th century—developed within the hygienist movement to improve lighting conditions in densely populated urban housing and industrial buildings. For decades, it served as the de facto standard for justifying access to natural light. Conversely, while DF maintains a foothold in some regulatory frameworks, there has been a clear shift in certifications such as LEED v4.1 [31] and WELL v2 [32], which emphasize dynamic, climate-based metrics like sDA or mEDI.

Its calculation considers the window and room geometry and properties such as the surface reflectivity of the materials. However, the window orientation, solar position, and weather conditions are not taken into account, since the calculation scenario considers only an ideal overcast sky [33]. A static luminance distribution allows designers to determine the potential illuminance at a given point in relation to an overcast sky, focusing on the characterization of architecture. The Daylight Factor remains useful for analyzing daylight availability in various architectural elements, including windows [8,34,35], atriums [36], courtyards [37], and skylights [38].

Although originally conceived to ensure minimum hygienic and sanitary standards for natural lighting in occupied spaces (such as homes, schools, and offices), the DF can also serve to estimate the peak energy consumption resulting from electric lighting, considering an overcast sky as one of the most conservative scenarios. While this assumption generally holds true, certain scenarios such as north-facing windows under clear-sky conditions have been proven to be more adverse [39], making it inadequate to set a threshold for the minimum electric lighting requirement. Moreover, as it ignores the dynamic luminous distribution of skies beyond overcast conditions and the actual illuminance required for visual tasks [40], the Daylight Factor (DF) does not accurately reflect electric lighting energy usage, especially in locations with predominantly sunny skies.

While the DF remained the most widespread metric for assessing the potential of natural light inside buildings until 2006, since then it has progressively been replaced by dynamic metrics that more accurately account for the variability of natural lighting, driven by improvements in computational power and its increased accessibility.

2.2. Dynamic Daylighting Metrics for Energy Efficiency

Created in 1989 by the Association Suisse des Electriciens [41] and refined by Reinhart et al. [28,42] in 2001, Daylight Autonomy (DA) measures the proportion of occupied time each year in which daylight meets a specified illuminance level. Inversely, increased DA leads to a decrease in electric lighting usage. It can be expressed as follows (2):

$$\mathrm{D}\mathrm{A}={\displaystyle \frac{{\sum }_{\mathrm{i}}\left({\mathrm{w}}_{\mathrm{f}}·{\mathrm{t}}_{\mathrm{i}}\right)}{{\sum }_{\mathrm{i}}{\mathrm{t}}_{\mathrm{i}}}}\in \left[\mathrm{0,1}\right];{\mathrm{w}}_{\mathrm{f}}=\left\{\begin{array}{c}1 \mathrm{if} {\mathrm{E}}_{\mathrm{D}\mathrm{a}\mathrm{y}\mathrm{l}\mathrm{i}\mathrm{g}\mathrm{h}\mathrm{t}}\ge {\mathrm{E}}_{\mathrm{l}\mathrm{i}\mathrm{m}}\\ 0 \mathrm{if} {\mathrm{E}}_{\mathrm{D}\mathrm{a}\mathrm{y}\mathrm{l}\mathrm{i}\mathrm{g}\mathrm{h}\mathrm{t}}<{\mathrm{E}}_{\mathrm{l}\mathrm{i}\mathrm{m}}\end{array}\right.$$

(2)

where t_i refers to the occupied time throughout the year in hours, and w_f functions as a weighting factor. It is a binary variable that equals 1 if the horizontal illuminance at a given point (E_Daylight) is greater than or equal to a predefined threshold, and 0 otherwise. E_lim refers to the illuminance threshold chosen in accordance with the task. It is generally agreed that a minimum illuminance of 300 lx is required for most tasks, with 500 lx recommended to minimize eye strain [43,44].

The Daylight Autonomy (DA) index offers enhanced precision in determining the annual illuminance of an interior space by accounting for factors relevant to the study of illumination, including diverse sky types and location-specific climatic data. Furthermore, it has served as a foundational metric for other energy efficiency indicators developed over the years. However, DA presents several limitations that constrain its comprehensive applicability in real-world contexts:

-

It does not account for the presence or behavior of electric lighting control systems.

-

It overlooks seasonal and temporal variability, such as the lack of occupancy during summer breaks in educational buildings.

-

It involves a relatively complex process for analyzing and comparing performance across a grid of calculation points, each with its own DA value. This spatial dispersion of results can make it difficult to draw immediate conclusions about the overall daylighting performance of a space, often requiring additional post-processing, visualization, or aggregation techniques to support design decisions.

-

It fails to capture the risk of overexposure or glare, which can negatively impact visual comfort.

-

It ignores the spectral distribution of daylight, which varies throughout the day and plays a crucial role in circadian stimulation, thereby affecting health and well-being.

To address these limitations, several enhanced metrics have been proposed. One of the earliest was Continuous Daylight Autonomy (cDA, DAC, or more commonly DAcon), introduced by Rogers [28]. This calculates the time fraction throughout the year when a minimum illuminance threshold is met by daylight. This metric refines the binary nature of DA by assigning partial credit linearly for illuminance levels below the threshold, and full credit for those that exceed it. In this way, DAcon offers a more nuanced representation of annual daylight availability. An example given by Reinhart et al. in his article [28] explains how, if a work plane requires 600 lx and the provided daylight level is 500 lx, a partial credit of 500/600 (0.8) is attributed. DAcon can be expressed as follows (3):

$$\mathrm{D}\mathrm{A}\mathrm{c}\mathrm{o}\mathrm{n}={\displaystyle \frac{{\sum }_{\mathrm{i}}({\mathrm{w}}_{\mathrm{f}\mathrm{i}}·{\mathrm{t}}_{\mathrm{i}})}{{\sum }_{\mathrm{i}}{\mathrm{t}}_{\mathrm{i}}}}\in \left[\mathrm{0,1}\right];{\mathrm{w}}_{\mathrm{f}\mathrm{i}}=\left\{\begin{array}{c}1 \mathrm{if} {\mathrm{E}}_{\mathrm{D}}\ge {\mathrm{E}}_{\mathrm{L}}\\ {\mathrm{E}}_{\mathrm{D}}/{\mathrm{E}}_{\mathrm{L}} \mathrm{if} {\mathrm{E}}_{\mathrm{D}}<{\mathrm{E}}_{\mathrm{L}}\end{array}\right.$$

(3)

where t_i is the occupied time in a year, and w_fi is the weighting factor dependent on the illuminance threshold. It is a non-binary variable that equals 1 if the horizontal illuminance at a given point (E_D) is greater than or equal to a predefined illuminance threshold (E_L) and 0 if the horizontal illuminance at that given point (E_D) is 0. As long as the value of E_D stays between E_L and 0, the w_fi is assigned a value following the aforementioned example given by Reinhart et al. [28] E_L as mentioned is the illuminance threshold, generally set at 300 lux, for the same reasons previously detailed in the DA explanation.

Despite its utility in determining energy savings from dimmer controls [45,46], DAcon has been underutilized in previous studies. However, recent developments have sparked a renewed interest in this metric, particularly with the emergence of Continuous Overcast Daylight Autonomy (Dao.con) [47]. This new metric is derived both from DAcon and from Overcast Daylight Autonomy (DAo). This last metric was renamed by the authors from Minimum Daylight Autonomy [48] to its current, more informative name.

Overcast Daylight Autonomy (DAo) provides a bridge between static and dynamic metrics by calculating DA with the parameters of DF, more specifically with a fixed overcast sky. Consequently, its equation resembles that of DA (4):

$$\mathrm{D}\mathrm{A}\mathrm{o}={\displaystyle \frac{{\sum }_{\mathrm{i}}({\mathrm{w}}_{\mathrm{f}\mathrm{i}}·{\mathrm{t}}_{\mathrm{i}})}{{\sum }_{\mathrm{i}}{\mathrm{t}}_{\mathrm{i}}}}\in \left[\mathrm{0,1}\right];{\mathrm{w}}_{\mathrm{f}\mathrm{i}}=\left\{\begin{array}{c}1 \mathrm{if} {\mathrm{E}}_{\mathrm{D}\mathrm{O}}\ge {\mathrm{E}}_{\mathrm{L}}\\ 0 \mathrm{if} {\mathrm{E}}_{\mathrm{D}\mathrm{O}}<{\mathrm{E}}_{\mathrm{L}}\end{array}\right.$$

(4)

where the value of the horizontal illuminance (E_Daylight in the original formula) is replaced by E_DO, the daylight illuminance measured at a given point under overcast-sky conditions. Consequently, the calculation procedure follows the same principles as the DA, but considering only overcast skies. While the DAo metric is useful for calculating the percentage of time annually in which a minimum illuminance threshold is met under overcast-sky conditions, it inherits certain limitations from the Daylight Factor (DF) and Daylight Autonomy (DA). Specifically, like the DF, DAo does not consider fenestration orientation.

While DAcon and DAo in isolation have seen limited adoption, DAo.con offers great utility for sensorless lighting smart control systems [47]. By utilizing measured or simulated Daylight Factors and user requirements, the algorithm modifies luminaire’s luminous flux without having to rely on internal light sensors. Input device installation and effective control systems in relation to daylight are among the aspects that make daylight-linked controls (DLCs) difficult to implement [49]. The issue is compounded by the stochastic behavior of occupants [50], which can make work plane input devices less effective. DAo.con could therefore help to mitigate the gap between specialized research and common design practice for DLCs. Continuous Overcast Daylight Autonomy (Dao.con) can be expressed as follows (5):

$$\mathrm{D}\mathrm{A}\mathrm{o}.\mathrm{c}\mathrm{o}\mathrm{n}={\displaystyle \frac{{\sum }_{\mathrm{i}}({\mathrm{w}}_{\mathrm{f}\mathrm{i}}·{\mathrm{t}}_{\mathrm{i}})}{{\sum }_{\mathrm{i}}{\mathrm{t}}_{\mathrm{i}}}}\in \left[\mathrm{0,1}\right];{\mathrm{w}}_{\mathrm{f}\mathrm{i}}=\left\{\begin{array}{c}1 \mathrm{if} {\mathrm{E}}_{\mathrm{i}.\mathrm{o}}\ge {\mathrm{E}}_{\mathrm{t}}\\ {\mathrm{E}}_{\mathrm{i}.\mathrm{o}}/{\mathrm{E}}_{\mathrm{t}} \mathrm{if} {\mathrm{E}}_{\mathrm{i}.\mathrm{o}}<{\mathrm{E}}_{\mathrm{t}}\end{array}\right.$$

(5)

where t_i is the occupied time in a year, and w_fi is the weighting factor dependent on the illuminance threshold, with us assigning values between 1 and 0 while following the same procedure as the DAcon equation. E_i.o is the daylight illuminance defined at a given point under overcast-sky conditions, and E_t is the illuminance threshold defined by lighting requirements.

It must also be noted that DAo.con, while facilitating implementation as a control algorithm for Dynamic Lighting Control Systems (DLCSs) due to its calculability from the Daylight Factor (DF), inherits an intrinsic limitation: the assumption that overcast conditions represent the worst-case scenario for electric lighting demand. As previously discussed, this is not always accurate. Specifically, a north-facing space under clear skies often requires a higher electric lighting demand than under overcast conditions [39]. Consequently, for north-facing spaces, DAo.con may slightly underestimate the electric lighting demand on clear days, necessitating adjustments to avoid under-compensation.

Given that DA does not account for seasonal and temporal variability, Ruiz et al. [51] developed the Partial Daylight Autonomy (DAp) metric. This indicator is derived from DA, and it represents the fraction of the annual occupied time during which a specific daylight illuminance threshold is met, while only considering the actual seasonal use of the space. This indicator can be expressed as follows (6):

$$\mathrm{D}\mathrm{A}\mathrm{p}={\displaystyle \frac{{\sum }_{\mathrm{i}}\left({\mathrm{w}}_{\mathrm{f}}·{\mathrm{t}}_{\mathrm{i},\mathrm{s}}\right)}{{\sum }_{\mathrm{i}}{\mathrm{t}}_{\mathrm{i},\mathrm{s}}}}\in \left[\mathrm{0,1}\right];{\mathrm{w}}_{\mathrm{f}}=\left\{\begin{array}{c}1 \mathrm{if} {\mathrm{E}}_{\mathrm{D}\mathrm{a}\mathrm{y}\mathrm{l}\mathrm{i}\mathrm{g}\mathrm{h}\mathrm{t}}\ge {\mathrm{E}}_{\mathrm{l}\mathrm{i}\mathrm{m}}\\ 0 \mathrm{if} {\mathrm{E}}_{\mathrm{D}\mathrm{a}\mathrm{y}\mathrm{l}\mathrm{i}\mathrm{g}\mathrm{h}\mathrm{t}}<{\mathrm{E}}_{\mathrm{l}\mathrm{i}\mathrm{m}}\end{array}\right.$$

(6)

where t_i,s refers to the occupied time throughout the year including seasonal use in hours, w_f functions as a weighting factor, and E_Daylight is the horizontal illuminance at a given point. E_lim refers to the illuminance threshold chosen in accordance with the task.

As such, DAp excludes non-occupied periods—such as the summer break in educational facilities, when classrooms are not in use. This adjustment allows for a more precise assessment of daylight performance and space design. In fact, differences of up to 10% have been observed between DA and DAp results in the case of classrooms [51], highlighting the importance of considering real occupancy patterns when designing window openings and daylighting strategies.

Finally, and with the aim of generating a more intuitive and easily interpretable indicator—particularly useful for the evaluation of large buildings or urban-scale studies—in 2012, the Illuminating Engineering Society of North America (IES) published the Spatial Daylight Autonomy (sDA), a metric that shifts from point-based daylight analysis to an area-based approach. sDA quantifies the percentage of floor area meeting or exceeding 300 lx for at least 50% of the annual occupied hours, offering a unified percentage for the entire room [52]. It is therefore relatively easy to understand and calculate, facilitating its adoption in design practice. It provides a clear measure of the percentage of space that is adequately illuminated by daylight, simplifying the process of ensuring that designs meet minimum lighting standards. Nevertheless, the sDA area-based approach prevents it from providing information regarding the distribution of daylight across the evaluated area. Thus, it does not account for potential issues such as uneven lighting or areas of excessive contrast, which can impact occupant comfort and visual performance.

2.3. Dynamic Daylighting Metrics for Visual Comfort

Complementary to these dynamic metrics, which focus on achieving adequate illuminance levels for task performance and energy efficiency, there is a set of dynamic metrics that, while based on the same principles, aim to ensure visual comfort by avoiding both excessive illuminance levels (which can cause glare) and insufficient levels (which can lead to visual fatigue)

The first of these metrics is Useful Daylight Illuminance (UDI), developed by Nabil and Mardaljevic [53,54]. UDI complements Daylight Autonomy (DA) by assessing daylight suitability through defined illuminance ranges. It quantifies the time fraction when indoor daylight levels align with occupant visual comfort, specifically between 100 lx and 2000 lx. UDI categorizes illuminance into three thresholds: UDI fell short (UDI-s, <100 lx), UDI achieved (UDI-a, 100–2000 lx), and UDI exceeded (UDI-e, >2000 lx). Like the DA, this approach offers a detailed interpretation of daylight levels based on actual weather data. While DA and UDI both consider the annual occurrence of target illuminance values, UDI extends beyond DA by incorporating an upper threshold, thus accounting for instances of excessive illuminance. Consequently, UDI-a values are typically lower than DA values calculated at the same minimum threshold, such as 300 lx.

UDI’s upper threshold attempts to address excessive illuminance, which can contribute to overheating and glare. However, it does not directly assess glare phenomena. Glare is influenced by factors beyond simple illuminance levels, such as the size and luminance of light sources, and their position relative to the observer [55,56]. The upper illuminance threshold can serve as a guideline value in the characterization of architecture, but it cannot replace a more comprehensive study of glare using specific metrics such as Unified Glare Rating (UGR) [43,57] or Daylight Glare Probability (DGP) [58,59,60].

To complement sDA, IES also introduced another visual comfort-oriented dynamic lighting metric: Annual Sunlight Exposure (ASE). This metric quantifies the percentage of an area that receives direct sunlight above a specified direct sunlight illuminance threshold for more than a defined number of hours per year [52]. While sDA serves to indicate whether a minimum level of illuminance is maintained throughout the year in an area, ASE focuses on excessive levels of illuminance due to direct sun, which can result in unwanted glare, discomfort, or excessive heat gain. Being a spatial metric, its strengths and weaknesses are akin to those of sDA. While it offers a clear and actionable estimate for designers, it lacks the granularity offered by other non-spatial metrics. Additionally, ASE suffers from the same shortcomings as UDI’s upper threshold. Issues such as glare or discomfort are linked to but not limited to an excessive illuminance value [55,56,61]. Therefore, while a high ASE value indicates excessive illuminance within the study area, it does not provide a precise assessment of phenomena such as glare or overheating, for which more specific and better-qualified metrics already exist.

2.4. Dynamic Daylighting Metrics for Health and Well-Being

Despite their usefulness in addressing energy efficiency through daylight utilization, the metrics previously discussed do not fully describe the impact of daylight on occupant health and well-being—particularly with respect to the spectral composition of daylight. Their focus is placed on the visual system and therefore they do not account for the influence of light on circadian entrainment. To address this deficiency, the Circadian Stimulus Autonomy (CSA) is defined by its authors as “the percentage of days when a threshold for adequate melatonin suppression is met solely by daylight during a specific time of the day, typically in the morning” [61,62,63,64].

However, it is important to note that CSA does not account for the timing of light exposure or individual variations in circadian sensitivity, which are critical factors in real-world applications of circadian lighting. These limitations should be considered when interpreting CSA values in design contexts.

CSA was developed from the calculation of DA and the illuminance required to achieve an acceptable value of circadian stimulus (CS). To achieve appropriate circadian entrainment, Acosta et al. argue that a CS value between 0.3 and 0.4 is adequate [61,62], based on previous studies conducted on office workers [20]. Circadian stimulus can be expressed as follows (7):

$$\mathrm{C}\mathrm{S}=0.7·\left(1-{\displaystyle \frac{1}{1+{\left(\frac{{\mathrm{C}\mathrm{L}}_{\mathrm{A}}}{355.7}\right)}^{1.1026}}}\right)$$

(7)

where CL_A is the circadian light. CL_A is estimated from the spectral power distribution (SPD) and illuminance perceived by the study subject through Rea et al.’s model of human circadian phototransduction [16,65], considering an exposure time of 1 h with a fixed 2.3 mm diameter pupil.

Thus, CSA must be calculated using a CS value of 0.4—based on the light flux received and its combined SPD, considering the inner reflections of the room analyzed. It can be expressed as follows (8):

$$\mathrm{C}\mathrm{S}\mathrm{A}={\displaystyle \frac{{\sum }_{\mathrm{i}}\left({\mathrm{w}}_{\mathrm{f}\mathrm{i}}{\mathrm{t}}_{\mathrm{i}}\right)}{{\sum }_{\mathrm{i}}{\mathrm{t}}_{\mathrm{i}}}}\in \left[\mathrm{0,1}\right];{\mathrm{w}}_{\mathrm{f}\mathrm{i}}=\left\{\begin{array}{c}1 \mathrm{if} {\mathrm{E}}_{\mathrm{D}}\ge {\mathrm{E}}_{\mathrm{L},\mathrm{C}\mathrm{S}}\\ 0 \mathrm{if} {\mathrm{E}}_{\mathrm{D}}<{\mathrm{E}}_{\mathrm{L},\mathrm{C}\mathrm{S}}\end{array}\right.$$

(8)

where t_i is the occupied time in a year, w_fi is the weighting factor dependent on the illuminance threshold, E_D is the daylight illuminance defined at a given point, and E_L,CS is the illuminance threshold required to achieve a CS value of 0.4.

Despite its promising potential, the Circadian Stimulus Autonomy (CSA) metric currently faces several challenges. The established threshold fails to account for individual variability, including factors such as age [66,67], chronotype [68], and exposure history [69], as well as differences in circadian sensitivity among individuals [70]. Furthermore, CSA is based on the circadian stimulus (CS) model developed by Rea et al., which is still undergoing evaluation and refinement by the scientific community, rendering its premature use in regulatory or certification applications.

Additionally, a significant challenge in the application of CSA lies in its inherent variability across locations and times. Firstly, natural daylight’s spectral power distribution (SPD) changes throughout the day and year due to solar position—angle of incidence and solar elevation—and atmospheric conditions—cloud cover, pollution, and humidity. The spectral composition of daylight, particularly the presence of blue light, crucial for circadian stimulus, varies significantly from morning to afternoon and between seasons. Secondly, geographical location plays a crucial role. Regions closer to the equator experience more consistent day lengths and therefore more stable daylight exposure, while higher latitudes exhibit dramatic seasonal differences in daylight availability and spectral quality. Thirdly, the circadian efficacy of light is dependent on the timing of exposure. Morning exposure to blue-enriched light has a stronger impact compared to later times of day [18]. Consequently, the same illuminance level can yield different circadian stimuli if the spectral composition varies. While the metric’s developers have attempted to address these dynamic factors by calculating CSA during early morning hours using an average spectral power distribution (SPD) from studied locations [64], this approach still introduces a degree of uncertainty. Therefore, further research is required to ensure that CSA accurately assesses circadian entrainment.

In addition to the circadian stimulus (CS), a more widespread metric for evaluating circadian-effective lighting is the melanopic Equivalent Daylight Illuminance (mEDI), as defined by the International Commission on Illumination in CIE S 026:2018 [71]. mEDI expresses the melanopic impact of a given light source by referencing it to the illuminance produced by a standardized daylight spectrum (D65), thereby offering a biologically relevant, yet easily interpretable, metric. Its standardization and spectral basis have facilitated its adoption in a range of regulatory frameworks and design tools, including the WELL Building Standard, DIN/TS 67600:2022-08, and emerging ISO/CIE 8995-1:2025 norms [32,72,73,74,75].

Unlike CS, which simulates non-linear melatonin suppression based on intensity, spectrum, timing, and exposure duration, mEDI is a spectrally weighted static illuminance metric, originally conceived for snapshot assessments. It does not model temporal dynamics or circadian phase sensitivity. However, its relatively simple formulation enables efficient integration into simulation workflows, and recent studies have extended its use to hourly or annual analyses using dynamic climate data and radiance-based simulation engines [76,77]. This trend has significantly increased its utility in daylight-centric lighting design and building performance assessments.

While the two approaches are conceptually distinct—mEDI offering a normative spectral benchmark, and CS providing a biologically grounded dynamic model—they are not mutually exclusive. When used together, they can provide complementary insights: mEDI can ensure compliance with regulatory thresholds, while CS can guide personalized or task-specific lighting strategies aligned with human physiological responses.

A major limitation of mEDI stems from its static formulation, which may not fully capture the dynamic nature of circadian-effective lighting. It does not incorporate critical temporal variables such as timing, duration, or pattern of exposure, all of which significantly influence circadian responses. Additionally, mEDI assumes spectral constancy during a given exposure, which can be problematic under natural daylight, whose spectral power distribution (SPD) varies with time of day, weather conditions, and solar position. This variability can lead to mismatches between predicted and actual physiological responses. Moreover, although mEDI values can be computed at regular intervals (e.g., hourly) to approximate dynamic behavior, such simulations still treat each timestep independently, lacking the ability to model cumulative or non-linear effects on the circadian system. Lastly, mEDI, by focusing exclusively on spectral content, does not distinguish between light exposures occurring at different times of day, which can result in overestimations or underestimations of circadian impact depending on context. While its simplicity facilitates implementations in performance simulations and building standards, further developments are needed to improve its alignment with biologically dynamic processes such as circadian entrainment. A brief summary and comparison of both metrics are shown in

Table 1. Comparative summary: mEDI vs. CSA.

Aspect	mEDI	CSA
Origin	CIE S 026:2018	Acosta et al. [61,62]
Basis	Spectrally weighted equivalent to D65 daylight	Non-linear physiological model of melatonin suppression
Input Parameters	Spectral power distribution (SPD)	SPD, intensity, exposure duration, circadian timing
Temporal Dimension	Static (but adaptable to dynamic simulations)	Explicitly time-based and dynamic
Physiological Accuracy	Indirect (based on ipRGC sensitivity)	Direct modeling of biological response (but with the model still in development)
Simplicity of Use	High—requires only SPD	Medium—requires multiple biological inputs
Standardization and Adoption	High—used in WELL, DIN/TS 67600:2022-08, ISO standards	Lower—cited, but not part of international standards
Main Limitation	No modeling of exposure duration or adaptation	Model complexity and fewer tools available

.

3. Methods

Having thoroughly reviewed the principal static and dynamic daylighting metrics, this section outlines the methodology employed to offer a comparative analysis of them. The aim is to critically examine the correlation, utility, and potential symbiosis among these diverse metrics when applied under a consistent study situation. By doing so, this research seeks to not only understand their interrelationships and sensitivities to specific design parameters but also to uncover the inherent limitations in the information they provide, thereby contributing to a more nuanced understanding of their overall practical applicability in architectural design and building performance assessment. The methodology involved the development of a virtual room model, serving as the basis for calculating each of the previously discussed metrics. These calculations were performed on a 50-point analysis grid, with the study systematically varying parameters such as the room’s geographical location, orientation, window-to-wall ratio, and typical occupied hours. The specifics of the method employed will be detailed in the subsequent sections.

3.1. Properties of the Virtual Model

The computational model for this study is based on a simplified rectangular room geometry, rather than based on specific real-world case studies. The space measures 3.0 m in height, 6.0 m in width, and 3.0 m in depth, resulting in a floor surface area of 18.0 m² and a volume of 54.0 m³. Reflectance values were set at 40% for the floor, 60% for the walls, and 80% for the ceiling across all models. Regarding fenestration, a single square window centered on the façade with a 0.7 transmittance was adopted for all virtual models. This configuration was chosen to isolate the impact of window size variations, intentionally excluding the influence of window position and transmittance. While the present study primarily focuses on the comparative analysis of dynamic metrics rather than the impact of various window configurations, future research could fruitfully explore the effects of diverse window placements and surface reflectances, assessing the metrics’ performance under such varied conditions as highlighted by previous studies [62,64].

The window dimensions were varied in both width and height to achieve window-to-wall ratios (WWRs) of 20%, 30%, and 40%. While not being based on any national or European regulation, this specific range was selected based on feasibility considerations for common use environments, balancing minimum illuminance requirements with overall energy performance. While a fully glazed façade might theoretically yield optimal results across all metrics—along with minimized electric lighting energy consumption—it would lead to a drastic increase in other energy demands for heating and cooling. Fifty calculation points positioned 0.8 m above the floor, distributed in 5 lines and 10 rows at 0.6 m intervals, formed the study grid across the analysis plane—following the point layout required for sDA calculations. This configuration enabled more precise spatial metric calculations, and it allowed for simultaneously accelerating the computational process by avoiding an excessively dense array of points. For the non-spatial dynamic metrics, the authors decided to focus on the central section displayed in red in Figure 1; since the central areas of the room are typically occupied for the majority of the time, its results are more relevant for the comparative analysis of metrics. In future studies, it would be of great interest to evaluate the behavior of the metrics in the lateral sections as well.

The calculation process incorporated two specific occupancy profiles: a full-day schedule from 08:00 to 18:00—a timetable commonly used for the calculation of sDA—and a half-day schedule spanning 08:00 to 14:00. This dual approach allows for a comprehensive analysis of lighting performance under varying usage patterns, capturing both the full spectrum of a typical workday, including critical peak solar hours and late afternoon conditions, and the distinct implications of morning-focused occupancy by concentrating on the period of highest natural light availability. Daylight Saving Time (DST) was deliberately excluded, reflecting the anticipated discontinuation of DST within the European Union, offering results uninfluenced by seasonal clock shifts for long-term applicability.

3.2. Location and Orientation of the Model

The spectral power distribution (SPD) of light incident on the occupant’s eye is determined by the average sky conditions and subsequent reflections within the surrounding environment. For this analysis, two distinct geographical locations were selected: Seville, Spain; and London, UK. These were selected as representative locations due to their contrasting daylight availability conditions. Seville, located in southern Spain, features a Mediterranean climate with high annual solar radiation levels and abundant daylight, especially during summer months. In contrast, London, situated in northern Europe, presents significantly lower daylight availability, characterized by frequent overcast conditions and shorter daylight hours, particularly in winter. These differences allow for the assessment of daylight metrics under diverse climatic scenarios, providing a more comprehensive comparison of their performance. The virtual model was oriented to both north and south in each location. This approach is fundamental for a comprehensive lighting study, as these orientations represent the extremes of natural light availability: the south maximizes direct solar exposure, while the north offers more consistent diffuse daylight and minimal direct solar impact, allowing for a thorough analysis of metric performance under diverse conditions.

Figure 2 illustrates the average SPD for each location, displaying the spectral distribution of each sky as derived from statistical data on the annual percentage of overcast skies.

The analysis was conducted using DaySim 3.1, based on climate files derived from long-term meteorological data (typically spanning 5–10 years). The SPD presented does not correspond to a single theoretical sky type (e.g., those defined in ISO/CIE 15469:2004(E)/CIE S 011/E:2003 [33]), but rather to a weighted average between clear and overcast-sky conditions, reflecting their relative frequency in the local climate data. For CS calculations, the spectral input was obtained by combining typical SPDs for each sky condition. Therefore, the resulting SPD represents a location-specific annual average, not an idealized or fixed sky model [78], and the prevailing climatic conditions of each site. The standard CIE D65 spectral power distribution (SPD), commonly representing a typical sky vault, was included for comparative purposes. This D65 SPD curve closely resembles the spectral distribution observed under intermediate sky conditions.

3.3. Assessment and Comparison of Metrics

The 300 lx and 500 thresholds for DA calculations were chosen to provide a comprehensive evaluation of the indoor lighting environment. The 300 lx threshold represents the widely accepted minimum illuminance level required for comfortable execution of common office and educational tasks [43]. Its inclusion ensures an assessment of basic visual sufficiency and functional compliance. Conversely, the 500 lx threshold signifies a more optimal and desirable illuminance level, particularly suitable for tasks demanding higher visual acuity, prolonged focus, or enhanced productivity [79]. By evaluating DA at that higher threshold, this study can quantify the potential for achieving a superior indoor environmental quality, moving beyond mere compliance to foster improved occupant well-being and reduced reliance on electric lighting. In the case of DAo, DAcon, and DAo.con, the authors considered it appropriate to use the same thresholds applied to DAs of 300 lx and 500 lx to facilitate a comparative analysis between the various illuminance metrics.

While Useful Daylight Illuminance (UDI) was originally conceived with a lower threshold of 100 lx to quantify illuminance levels below the 300 or 500 lx minimums, common in other daylighting metrics [53,54], a lower threshold of 300 lx has gained traction over the years [28,80]. For the present study, it was decided to employ 300 lx as a lower bound for UDI and 2000 lx for the upper one. Since the main objective of this article is to establish a comparison between the different illuminance metrics, the authors considered that a more similar upper threshold for UDI and ASE would facilitate the comparative analysis. Thus, the illuminance values will be divided between the following:

• Values that fall short of the accepted range: UDI_<300.
• Values that fall within the accepted range: UDI_300–2000.
• Values that surpass the accepted range: UDI_>2000.

For Spatial Daylight Autonomy (sDA), the IES established a threshold of 300 lx, based on the aforementioned criteria. Its complementary metric, Annual Sunlight Exposure (ASE), was concurrently defined with a 1000 lx threshold [52]. Therefore, the thresholds selected for the study at hand are sDA₃₀₀, ASE_1000–250.

The calculation of the Circadian Stimulus Autonomy (CSA) in this study followed the methodology established in previous research [61,62,81]. This approach integrates the average spectral autonomy from both locations (Figure 2), the orientation of the models, and Equation (7) to derive the final CS values. For London, a consistent threshold of 320 lx was determined for both north and south orientations. In Seville, the thresholds varied by orientation, with 290 lx established for the north-facing facade and 360 lx for the south-facing facade. These illuminance thresholds were derived by setting a target CS value of 0.4 in Equation (7), then calculating the required CL_A to achieve that CS value, based on the average SPD of the sky in each location (Figure 2) and considering orientation.

These thresholds serve as a valuable reference for evaluating circadian entrainment in the specified scenarios, given that, in conventional lighting applications, natural light’s SPD variation—attributable to location and climate—exerts only a very limited impact on melatonin suppression throughout the day.

3.4. Combination of Variables for the Calculation Models

Table 2. Variables considered for the development of the calculation models.

Location	WWR	Orientation	DA Threshold	Occupied Hours
Seville (SEV)	20% (20)	North (N)	300 lx (300)	08:00 to 18:00 (0818)
London (LON)	30% (30)	South (S)	500 lx (500)	08:00 to 14:00 (0814)
	40% (40)

presents the variables used to generate the 48 calculation models in this study. Each variable has been assigned a bracketed code for subsequent differentiation of each hypothesis.

3.5. Calculation Engine

This study was conducted using DaySim 3.1, a daylighting simulation program built upon the Radiance ray-tracing engine. This engine was developed by the Building Technologies Department at Lawrence Berkeley National Laboratory and validated by several studies [82,83]. The accuracy of DaySim 3.1 has also been tested in various research works [42,84,85].

4. Results

Following the methodology outlined, DA, DAcon, DAo, DAo.con, CSA, UDI, sDA, ASE, and DA_avg values were calculated based on the variables established in Table 2 and the preceding sections. The graphical analysis presents the results corresponding to the middle segment of the calculation grid. This segment was selected for visual clarity and representativeness; however, the complete 50-point grid was fully utilized to ensure the precise calculation of comprehensive spatial metrics such as sDA, ASE, and DA_avg, as detailed in Section 3.1. The results obtained for each metric are presented graphically as follows: Seville, 20% WWR in Figure 3; Seville, 30% WWR in Figure 4; Seville, 40% WWR in Figure 5; London, 20% WWR in Figure 6; London, 30% WWR in Figure 7; and London, 40% WWR in Figure 8.

In the upper-left corner of each model, its unique identification code, based on the variables described in Table 2, is displayed. For all cases, models with south-facing (S) glazing are arranged on the left, while those with north-facing (N) glazing are on the right. It can also be observed that the top four figures correspond to a Daylight Autonomy (DA) with an illuminance threshold of 300 lx, and the bottom four to an illuminance threshold of 500 lx. Beneath each figure, a legend is provided indicating the various metrics evaluated in each section of the models. Additionally, each figure’s caption includes a reminder of the identification system.

5. Discussion of Results

5.1. DA₃₀₀ and DA₅₀₀ and Non-Spatial Metrics’ General Patterns

The results obtained from DA highlight notable differences between thresholds. The variation between DA calculated with 300 lx (DA₃₀₀) and at 500 lx (DA₅₀₀) thresholds is particularly pronounced, with the latter yielding values up to 46% higher in certain configurations, such as in the SEV40N_500 case (Figure 5).

All dynamic metrics evaluated—including DA, DAo, DAcon, DAo.con, and Circadian Stimulus Autonomy (CSA)—exhibit a consistent spatial pattern: values tend to decrease with an increasing distance from the window. However, the rate of decline and the final values reached within the test space vary across metrics, reflecting their distinct sensitivities to daylight distribution.

The key aspects to highlight from this section are as follows:

• Significant differences observed between DA₃₀₀ and DA₅₀₀, with DA₅₀₀ showing up to 46% higher values in some cases.
• All dynamic metrics show a spatial decrease with distance from the window.
• The rate of decrease and final values differ between metrics, indicating a varying sensitivity to daylight distribution.

5.2. DAo.con, Dacon, and DAo

A clear pattern emerges when comparing the performance of DA and DAo.con under varying orientations. In north-facing fenestrations, DAo.con frequently yields higher values than Daylight Autonomy (DA), while in south-facing orientations, DAo.con generally falls below DA, except near the window. This trend aligns with the metric’s underlying assumptions and formulation: the diffuse illuminance provided by overcast skies is advantageous in north-facing spaces, enhancing daylight availability, whereas in south-facing rooms, overcast conditions reduce direct solar gains, increasing the need for electric lighting.

These results, consistent with the observations in the review section, suggest the potential value of a contextual adjustment to the DAo.con algorithm in north-facing spaces. Such a modification could help improve the accuracy of electric lighting predictions by avoiding the overestimation of daylight availability under predominantly diffuse conditions.

As a hybrid metric combining DAo and DAcon, DAo.con consistently produces intermediate values between them across all models. While DAo and DAcon have seen limited application individually, their integration in DAo.con offers promising utility for Dynamic Lighting Control Systems—particularly if further refined to account for orientation-specific daylight characteristics.

The key aspects to highlight from this section are as follows:

• DAo.con typically exceeds DA in north-facing fenestrations but tends to be lower with south-facing orientations, especially away from the window.
• To better serve as a dynamic metric for sensorless lighting control, DAo.con requires contextual adjustments in north-facing rooms to avoid overestimating daylight availability.

5.3. Circadian Stimulus Autonomy

In some models—such as LON20S_300_0814 in Figure 6—CSA and DA₃₀₀ show similar results; however, the results generally diverge markedly. This highlights the distinct criteria and specific applications of both metrics. Although CSA presents inherent limitations, as discussed in the introduction, their joint evaluation offers complementary insights across a wide variety of case studies. In contrast, DA₅₀₀ shows minimal alignment with DA, consistently returning substantially lower values throughout the space.

Both metrics display a clear spatial gradient, with higher percentages near the window. Only in the Seville models with a South orientation and a WWR above 30% are high percentages maintained at the rear of the room, as depicted in Figure 4 and Figure 5. It is also noted that achieving high CSA values may correlate with illuminance levels exceeding the upper UDI threshold (>2000 lx) for a significant portion of the time. However, this does not occur towards the back of the room in the Seville-based models with a WWR above 30% and a southward orientation, as can be seen in Figure 4 and especially Figure 5. Further research is warranted to determine whether the illumination levels required to achieve adequate circadian entrainment could result in discomfort or glare for some occupants, particularly those located near window openings.

The key aspects to highlight from this section are as follows:

• CSA and DA₃₀₀ occasionally align, but generally show distinct results, reflecting their different purposes and offering complementary insights, despite CSA’s known limitations. DA₅₀₀ consistently yields lower values, showing minimal correlation with CSA.
• High CSA values may coincide with exceeding upper UDI thresholds (>2000 lx) near windows, raising concerns about potential discomfort or glare, warranting further investigation.
• The models located in Seville with a WWR > 30% and south-facing orientation exhibit relatively high values deeper into the room, despite UDI showing no major signs of potential discomfort.

5.4. Useful Daylight Illuminance

UDI clearly illustrates how the indoor illumination distribution varies significantly with the distance from the window. The distribution also varies significantly depending on the geographic location and room orientation. For example, models located in Seville with a south-facing orientation—such as the one depicted in Figure 5—show acceptable daylight levels that extend toward the rear of the room. This contrasts with other configurations where illuminance levels decline more sharply, underscoring how the “effective depth” of daylight penetration is inherently influenced by both climatic and orientational factors.

Moreover, UDI effectively complements ASE by identifying specific points within the space where daylight levels exceed comfort thresholds. It allows for a more detailed analysis of the spatial distribution of daylight, providing a nuanced understanding of where illuminance levels may become excessive.

The key aspects to highlight from this section are as follows:

• UDI highlights the strong influence of distance from the window, location, and orientation on both indoor daylight distribution and potential glare.
• In Seville’s south-facing models, acceptable UDI values (within the lower and upper thresholds) extend deeper into the room compared to other cases.
• UDI complements ASE by identifying concrete points where illuminance values exceed the upper threshold, supporting a more precise analysis of potential glare or discomfort.

5.5. sDA, ASE, and DA_avg

The results for sDA and DA_avg tend to be comparable—or nearly identical—in several London models, particularly during the 08:00 to 18:00 timeframe. In Seville with a 500 lx threshold, the gap between them begins to widen, although in Seville with a 300 lx threshold, the similarity remains notable. Despite the absence of a perfect correlation between sDA and DA₃₀₀—some models like SEV40_300 (Figure 5) exhibit differences of up to 26%—both metrics appear to provide similar information in many cases. As expected, DA₅₀₀ presents significantly greater discrepancies, reaching up to 46% in the case of SEV40_500.

However, the relationships among lighting metrics are not consistent across all geographic locations and orientations. An example of this variability is observed in London with a north-facing orientation (Figure 6, Figure 7 and Figure 8), where DA_avg tends to slightly exceed sDA. In contrast, for most other analyzed orientations and locations, sDA typically shows higher values than DA_avg. This variability underscores that the interpretation of the relative performance between DA_avg and sDA is highly context-dependent, shaped by both geographic location and room orientation.

In south-facing orientations, ASE also shows a close correlation with the other two spatial metrics, given the direct solar exposure characteristic of unshaded south-facing apertures. Nevertheless, the commonly used ASE threshold of 10% as an upper limit appears questionable. Firstly, in almost all models, achieving sDA values recommended by IES LM-83-12 [52]—between 50% and 75%—also entails exceeding this 10% ASE threshold. Additionally, in the London models with a northward orientation (Figure 6, Figure 7 and Figure 8), ASE consistently exceeds 10%, yet according to UDI, the results are not especially concerning in terms of excessive illuminance across the room, except for those points closest to the window.

As was mentioned in the introduction, spatial metrics lack the granularity inherent to other metrics. UDI, by contrast, reveals that high illuminance values above 2000 lx are typically concentrated mostly in the first two or three points near the window. Beyond this zone, values within an “acceptable” UDI range become predominant, while the rear areas of the space more often fall below minimum thresholds. An exception is found in Seville with a south-facing orientation, where acceptable illuminance levels are maintained even at the very back of the room.

This lack of spatial precision can be observed in the comparative analysis of SEV30S_300_0814 (Figure 4) and LON40S_300_0814 (Figure 8). Although both models display nearly identical sDA (100% and 100%) and ASE (64% and 66%, respectively) values, the UDI results differ markedly: in the London case, the rear half of the room falls below the minimum UDI threshold between 30% and 44% of the time, whereas in Seville, most of the space remains within acceptable daylight levels. This suggests that while spatial metrics like sDA and ASE can provide a useful indicative assessment, they may obscure important spatial variations in lighting quality that are better captured by more granular metrics such as UDI.

While simulation-based analysis provides valuable insights, future research involving empirical validation in real buildings will be essential to confirm the predictive reliability of these metrics under actual use conditions.

The key aspects to highlight from this section are as follows:

• sDA and DAavg show similar or nearly identical results in several London models during the 08:00–18:00 period. In Seville, with a 300 lx threshold, they remain closely aligned in various models; however, there are discrepancies of up to 26%.
• In London with a north-facing orientation, DAavg slightly exceeds sDA, though in most other cases, sDA exceeds DAavg.
• ASE shows a stronger correlation with sDA and DAavg in many of the south-facing models due to increased direct solar exposure.
• In most models, achieving IES-recommended sDA levels (50–75%) necessitates exceeding the 10% ASE limit recommendation. Additionally, in London north-facing models, ASE surpasses 10%, but the UDI values suggest that excessive illuminance is confined to areas near the window.
• Spatial metrics like sDA and ASE lack granularity; they overlook variations that UDI captures more precisely. Models with identical sDA (100%) and ASE values (64–66%) show major differences in daylight distribution when compared with the UDI results.

5.6. Complementarity and Utility of Daylighting Metrics

DA₃₀₀ remains the primary reference metric for daylighting due to its ease of interpretation and proven utility in assessing whether a space is adequately lit for a specific task. Its variant, DA₅₀₀, introduces a stricter threshold that can better ascertain the luminous quality of spaces where tasks requiring higher visual perception are performed. However, by definition, neither DA300 nor DA500 accounts for potential discomfort caused by glare, making it necessary to rely on complementary metrics.

Useful Daylight Illuminance (UDI) might seem an appropriate complement to DA, yet its upper illuminance threshold does not ensure an accurate assessment of glare either. Nevertheless, beyond evaluating whether the analyzed space exhibits illumination deficiencies based on the chosen illuminance threshold, UDI provides insights into zones potentially subject to excessive solar incidence, which could, in turn, affect thermal demand—especially in the case of cooling. This being stated, UDI alone cannot determine excessive solar gain, as there are other more suitable metrics for that purpose. According to the results of this study, it could be argued that UDI proves most effective when used in conjunction with spatial metrics.

Both sDA and ASE provide an overall view of the daylighting availability and spatial distribution in a given space. Their simplicity makes them highly attractive for designers with limited experience in daylight simulation tools. However, as demonstrated in previous sections, their unreflective use without the assistance of other metrics can lead to significant errors. Two spaces may present similar sDA and ASE values while displaying very different luminous conditions. Therefore, particular caution is required when comparing spaces with different locations and orientations. In this regard, a metric like UDI serves as a valuable tool to better understand what is truly occurring within that space. Nevertheless, neither UDI nor ASE accurately assesses visual discomfort due to glare, necessitating the use of a specialized metric designed for such concerns.

While a combination of UDI, sDA, and ASE—or alternatively DA, sDA, and ASE, if more spatial detail is needed—along with appropriate glare metrics might appear sufficient, it is important to emphasize that none of these are explicitly aimed at promoting healthier spaces beyond fulfilling the minimum illuminance required to avoid vision impairment. In this context, CSA can be useful for estimating the percentage of days on which adequate melatonin suppression is achieved to promote circadian entrainment. Even so, the intrinsic limitations of CSA—already outlined in the introductory review—must not be overlooked. The underlying models of circadian entrainment on which this metric is based are still under active development. Therefore, future revisions to the CSA are expected to change as the model is updated based on new research.

Beyond traditional metrics focused on visual comfort and emerging indices related to circadian health, the effective integration of daylight into building operation strategies requires sophisticated tools. In particular, the design of intelligent lighting control systems—particularly those aiming for sensorless operation—demands a nuanced understanding of daylight availability under specific sky conditions. It is precisely to address this need that the Continuous Overcast Daylight Autonomy (DAo.con) metric has been developed. DAo.con is a hybrid dynamic daylighting metric designed to estimate the availability of daylight under overcast-sky conditions. It builds upon DAcon and DAo by combining their respective approaches—temporal continuity and daylight-only analysis—into a unified model that is especially suited to predicting lighting demand in unfavorable daylighting scenarios. Thus, DAo.con integrates core concepts from both DAcon and DAo, combining them for an assessment under overcast-sky conditions with a gradient-based evaluation approach. This enables DAo.con to reliably estimate electricity consumption under suboptimal daylighting conditions without relying on light sensors, which may be affected by the stochastic behavior of occupants. Although placing sensors at the working plane level would theoretically provide the most accurate measurements of available illuminance, this configuration carries the risk of occasional obstructions, potentially reducing the accuracy of the collected data. However, as seen in the previous analysis section, its algorithm requires refinement when applied to north-facing rooms, particularly in climates dominated by clear skies.

Building upon the preceding discussion of results,

Table 3. Assessment of daylighting metrics: Strengths and limitations.

Metric	Scope	Strengths	Limitations
Daylight Factor (DF)	Health and visual comfort	- Simple and quick to calculate - Historically used and widely understood - Yields conservative energy consumption estimates	- Ignores orientation and climate - Only overcast skies are considered
Daylight Autonomy (DA)	Energy efficiency and visual comfort	- Useful for minimum illuminance assessments - Reflects daylight availability based on climate data - Serves as the basis for other dynamic metrics	- Does not consider artificial light control - Ignores seasonal and spectral variation - Binary pass/fail threshold may oversimplify performance
Continuous DA (DAcon)	Energy efficiency	- Introduces gradual scoring for values below the illuminance threshold	- Lack of adoption in design guidelines and regulations
Overcast DA (DAo)	Energy efficiency	- Serves to link DF with dynamic metrics - Useful in diffuse sky conditions	- Inherits many DF limitations
Continuous Overcast DA (DAo.con)	Energy efficiency	- Useful for sensorless DLCS - Can be derived from DF	- Requires further adjustment for north-facing rooms in clear-sky scenarios
Partial Daylight Autonomy (DAp)	Energy efficiency	- Adjusts DA to seasonal use - Useful for schools or buildings with different occupancy patterns	- Tied to specific occupancy profiles - Not widely adopted
Spatial DA (sDA)	Visual comfort	- Intuitive interpretation of the results - Recommended by IES and LEED - Useful for a broad illuminance assessment	- Its lack of granularity can hide illuminance imbalances
Annual Sunlight Exposure (ASE)	Visual comfort	- Intuitive interpretation of the results - Recommended by IES and LEED - Useful to estimate the risk of overexposure	- Its lack of granularity can hide illuminance imbalances - Does not directly assess glare
Useful Daylight Illuminance (UDI)	Visual comfort	- Takes into account both insufficient and excessive illuminance - Useful in tandem with spatial metrics	- Does not directly assess glare - No formal regulatory basis
Circadian Stimulus Autonomy (CSA)	Health and well-being	- Estimates circadian stimulus - Relates spectrum, timing, and health to an illuminance value	- Based on circadian entrainment models still in development - Does not account for relevant factors such as chronotype, age, or prior exposure

has been compiled to present the main advantages and limitations of each of the analyzed metrics. An additional column focused on the main scope of the metric has been added to provide further context.

6. Conclusions

This study aimed to critically evaluate a selection of static and dynamic daylighting metrics to understand their individual roles, limitations, and interrelations in assessing daylight performance in architectural spaces. Through comparative simulations of a standardized room in Seville and London under varying orientations, window-to-wall ratios, and occupancy schedules, the analysis sought to determine the sensitivity, consistency, and relevance of each metric for visual comfort, energy performance, and circadian health.

There is no consistent correlation between the different daylighting metrics when considering different locations, orientations, occupancy hours, or illuminance thresholds. While the goal of achieving one or two indices capable of reliably determining the quality of daylight within a space is desirable, the current reality is that a combination of distinct metrics is needed to achieve a comprehensive and robust evaluation. A more promising direction could involve the development of intuitive tools that simplify both the calculation and graphical interpretation of results. This approach could empower non-specialized designers to draw meaningful conclusions efficiently, sacrificing accuracy or requiring excessive time. Consequently, a clear understanding of the specific strengths and applications of each metric remains essential for a complete daylighting assessment:

• DA remains a valuable and easily interpretable metric for assessing daylight performance, effectively indicating whether a space receives sufficient natural light for specific visual tasks.
• While UDI’s upper threshold is not reliable for accurate glare assessment, it is useful for identifying specific areas within a space where excessive solar incidence may lead to increased cooling loads.
• Identical sDA and ASE values can represent significantly different lighting conditions, especially when comparing spaces across different locations and orientations. Therefore, these metrics should be used in conjunction with others—such as UDI or DA—to gain a more accurate understanding of real luminous performance.
• CSA is a promising tool for estimating the percentage of days on which occupants experience adequate melatonin suppression to support circadian entrainment. However, the underlying biological models are still evolving, meaning future adjustments to the CSA methodology are anticipated.
• DAo.con is a particularly useful metric for the design of sensorless lighting control systems. Nevertheless, its algorithm requires further refinement for spaces with north-facing orientations, particularly in climates dominated by clear-sky conditions.