A Study on Nonvisual Effects of Natural Light Environment in a Maternity Ward of a Hospital in Cold Area

Thanks to the discovery of human eye photoreceptor cells, ipGRC and human nonvisual channels, the nonvisual effects of light have gradually come to our vision and been rationally utilized. Along with this trend, people have expanded their demand for the light environment to both visual and nonvisual needs from only visual needs. With a good natural daylighting, parturients will find their psychological pressure and physiological pain relieved, their rehabilitation rate increased, and they get rehabilitated more quickly. This study was carried out in a representative maternity ward in Harbin based on the latest research results on nonvisual effects at home and abroad. Specifically, the nonvisual effects on the natural light environment in the ward throughout the year were simulated and analyzed from the aspects of equivalent melanopic lux (EML), stimulus frequency (Stim.freq), and circadian effective area (CEA). During the study, natural light in the ward was measured on site, and the evaluation tool and workflow of nonvisual effects were created with the aid of Grasshopper modeling platform, Ladybug+Honeybee, and VB script editor. The results show that the nonvisual effects of natural light on the body of parturients gradually weaken as they further go inside the ward. What is worse, in the most unfavorable all-overcast condition, daylighting on beds far away from the window does not meet the stimulus of human circadian rhythm from April to August. Therefore, additional light is required. The wards have the best nonvisual natural light environment when they are south facing and have a window to floor ratio of 0.3.


Introduction
Natural light is the main light source used by people. Besides its visual functions, people find that it can also have an impact on our circadian rhythm and psychological state, although no scientific explanation demonstrated. Until the retina's third kind of photoreceptor cells-intrinsically photosensitive ganglion cells (ipRGCs)-and nonvisual channels were found in 2002 [1], a rational explanation was delivered. Follow-up studies were carried out on the mechanism of such cells, allowing further verification of the nonvisual biological effects of light stimulation on the body [2].
More and more studies have shown that light's nonvisual effects can directly affect our mood, working efficiency, and physical health [3,4]. This leads to a transition of the evaluation standard for a light environment from a single visual effect to a double evaluation standard of visual and nonvisual. As for parturients awaiting delivery in a hospital, ideal natural light not only alleviates their tension and fear but also has significant effects on their physiological changes, shortening their rehabilitation and reducing the incidence rate of postpartum depression [5,6]. Therefore, an optimal natural light environment with both visual and nonvisual standards acts as an important factor concerning the overall environmental comfort of a maternity ward.
During pregnancy, due to their own physiological needs and the needs of fetal growth, the BMR of parturients is 20%~30% higher than that of ordinary people [7], with flood pressure and heart rate rising and hormone disrupting. The parturients become more sensitive to environmental changes, such as the color, light, and temperature, showing mood swing and a high risk of depression. According to statistics, about 5%-10% parturients suffer from depression [8]. Parturients who have given birth need a good rest to recuperate and stay in an enclosed space for a long time. Apart from the company of their family, the space environment is of great importance. A comfortable environment has significantly positive effect on the regulation of their own emotions. Among a variety of environmental factors constituting the ward space, light environment is regarded as one of the most crucial factors [9]. Studies have shown that nonvisual effects of light affect the secretion of melatonin and the associated circadian rhythm, as well as the body temperature, heart rate, blood pressure, and sleep quality [10][11][12], all of which are related to loneliness, anxiety, depression, and other emotions [13]. In the case of parturients cannot take corresponding drugs, optimal natural light has certain phototherapy effect. Because of its high safety and quick effect, it can remarkably improve the physiological and psychological conditions of parturients.
Since the "Two-child Policy" and "Three-child Policy" have been released one after another, the birth rate of newborns is gradually rising. People's demand for maternity hospitals is increasing, and higher requirements and standards are put forward for the hospital environment. As an important part of the hospital physical environment, the impact of light environment on users' physiology and psychology cannot be ignored. Moreover, natural light is of great study significance based on its strengths such as nocost, complete and continuous spectrum, energy saving, and environmental friendliness.
Previous studies on the function of natural light within the building space environment were limited to two aspects: visual effects and efficient utilization of energy. However, in 2012, The Society of Light and Lighting pointed out that natural light's nonvisual effects should also be taken into consideration [14]. According to the findings of Boubekri et al., optimal natural light and illumination may have positive effects on people's happiness within a building space. People who were less exposed to natural light might show less and poor sleep, less physical activity, and declining living quality [15]. Adequate sleep is of vital importance to parturients. It can not only improve their physical condition but also promote calcium absorption and the development of fetal brain. Hoof et al. studied on the illuminance range that met nonvisual effects. The results showed that in the daytime, a high-level light exposure (over 1000~2000 lx) was needed for human eyes to fully meet nonvisual needs, while excessive light exposure at night should be prevented from affecting sleep quality [16]. Kessel et al. found that photoreceptor cells of different age groups varied in their sensitivity to the light of the same wave band and circadian rhythm stimulation. Such sensitivity to light declined with the increase of age [17]. Therefore, elderly parturients needed more light stimulus compared to those of the right age. In their studies, Andersen et al. took people's activity patterns and viewing directions into consideration and proposed an evaluation model combining visual and nonvisual effects, further promoting the development of studies on nonvisual effects of natural light [18]. According to the current research trend, nonvisual functions of natural light will gradually surpass visual effects and energy utilization, and the therapeutic function of natural light environment will also be rediscovered [19]. Thus, a ward with sufficient natural lighting and nonvisual effects is essential.
Based on available research background, a double obstetrics ward in Harbin Maternity Hospital is taken as the experimental object. As Harbin is located in a harsh region of China, the winters are cold, dry, and long-lasting, receiving less sunlight, which has a greater impact on maternity and needs to be enhanced to benefit from natural light resources. Therefore, the winter time period was selected, and the ward was firstly analyzed with real measurements of natural light, and then, the dynamic simulation analysis of natural daylighting was carried out with the aid of Grasshopper modeling platform and VB editor. Moreover, the results are presented in Excel. On such a basis, the nonvisual effects of the natural light environment in the ward were evaluated in a scientific and quantitative way. By doing so, this paper will provide effective data support for researches on nonvisual effects of medical space in maternity hospitals and fill the gap in research on light environment in maternity wards in China.

Field Measurement of Natural Light
2.1.1. Selection of Sample Wards. After field visits and online data collection, considering the representativeness of the ward and the operability of field measurement, we chose general double-occupancy maternity wards within Harbin obstetrics and gynecology hospitals as experimental subjects, which meets the requirements of one-way side daylighting in Standard for Daylighting Design of Buildings (GB5003-2013) (daylight factor of side lighting shall not be less than 3% and indoor natural light illuminance shall not be less than 450 lx). Finally, two vacant wards on the 9 th floor of the inpatient department of Harbin Red Cross Hospital were selected as sample wards (Figure 1(a)). These two wards had basically the same parameters, such as indoor arrangement and spatial layout, with built-in toilets and no shelter outside. The difference was that one ward faces south, while the other faces north (Figure 1(b)).

Layout of Measuring Points.
One of the biggest differences between a maternity ward and a general ward is that the former is also equipped with a crib, apart from general beds for parturients to get rest. Due to the low ability of autonomic nerve and self-adjustment, newborns are quite vulnerable to the strong light. This may affect the physiological conditions of newborns such as mood fluctuation, sleep duration, and heart rate [20]. Compared to adults, photoreceptor cells in the eyes of newborns are more sensitive to the light stimulus [17]. Despite their sleep duration of 16~17 h a day, the nonvisual effects of natural light on their bodies cannot be ignored. Hence, apart from the ordinary hospital bed, it is also necessary to set up measuring points on the surface of the crib for illumination measurement.
For in-depth study of the nonvisual effects of natural light in a maternity ward, the behavior observation method was adopted to observe the daily activities of parturients. As such, their daily activity routines were obtained. According to the measured statistics, parturients showed three  3 BioMed Research International kinds of daily activities, namely, lying supine, leaning, and walking. Specifically, they lie supine for about 75% of their time, lean against the bed, and walk around for about 10% of their time, respectively, and do other things for about 5% of their time. According to parturients' activity routines and demand for light environment, the measuring points of nonvisual effects were arranged at different heights in the middle of the surface of ward beds and baby's cots. The measuring points A and B denote the leaning state, 1,000 mm away from the ground; a and b denote lying supine, 550 mm away from the ground; and c and d denote the selected measuring points on the crib, 450 mm away from the ground ( Figure 2).

Field Measurement.
In the study, we adopted TES1332A illuminometer (resolution: 0.1 lx; measuring range: 0:1~2 × 10 5 lx; margin of error: ±3%) to measure the illuminance at each measuring point. SNDWAYSW-M50 handheld infrared laser rangefinder (measuring range: 0-50 m; accuracy: ±1.5 mm) was used to measure window size, room depth, and bay net value. At the same time, a tape measure was also used for auxiliary measurement. The ward was measured on January 9-10 (two sunny days) and 12-13 (two completely cloudy days) in 2021. During measurement, the curtains were pulled back, and all indoor lights were turned off, reducing the influence of artificial light on measurement. According to the timetable of sunrise and sunset

BioMed Research International
in Harbin, the illuminance of the measuring points in the two wards is measured from 6 : 00 to 17 : 00 every day. The measurement was performed once every hour for three times at an interval of 5 s. Take the mean value of the three measurements as the final result.
Different from the previous evaluation of static daylighting and illumination, the evaluation of natural daylighting with nonvisual effects is badly sensitive to external factors. It seems that the illumination intensity and spectral distribution of natural light are invariably in a fluctuating state, but there are also certain inherent laws to follow [21]. In order to make the evaluation more accurate, natural light can be simplified to some extent. For example, D65, D75, D55, and other CIE standard light sources can substitute the sky in the south, north, northeast, and other directions, respectively. On account of the research conducted by Cajochen et al. on illumination intensity and human body vigilance [22], it is summarized that the starting threshold of natural light illuminance to stimulate the circadian rhythm is as follows. Among CIE standard light sources, D65 is 190 lux, D55 21 0lux, and D75 170 lux. Therefore, it can be concluded that the natural daylighting illuminance value of a room facing north, which can receive natural light from the north, reaches 170 lux, that of a room facing east or west reaches 210 lux, and that of a room facing south reaches 190 lux. Although there is no specific requirement for the daily lighting time that should meet the nonvisual effects of the human body, the WELL Building Standard defines that the minimum light stimuli standard for nonvisual effects is 250EML and lasts not less than 4 hours per day. As a result, in the actual measurement of this paper, the natural light of not less than 190 lux for at least 4 hours every day is regarded as the standard, on the basis of which the human body is believed to have an effective circadian rhythm on this day.

Simulation Technology Platform.
Grasshopper is the most commonly used parametric design and visual pro-gramming platform for architects thanks to its great openness and efficiency and compatibility with many plug-ins. At present, DIVA and Ladybug+Honeybee, two mature daylighting computing plug-ins, are running on Grasshopper platform. Roudsari et al. found that the combination of Honeybee and Ladybug works better than DIVA in meteorological data analysis and visualization and also natural daylighting simulation. For this reason, Ladybug+Honeybee is selected for related daylighting calculation in the development of nonvisual evaluation tools.
Honeybee, mainly developed by Mostapha Sadeghipour Roudsari of the University of Pennsylvania, is a powerful port plug-in in Grasshopper toolkit for daylighting simulation and energy consumption analysis. With Honeybee, daylighting was mainly analyzed in line with the following procedures: setting weather conditions, importing models, assigning materials, setting simulation parameters, making simulated calculation, and visualizing results. In terms of the simulation type, users can analyze illuminance, brightness, shade, glare, etc. and then use Run daylight simulation, a module, to call Radiance or Daysim for daylighting calculation.

Evaluation
Method. At present, there are three kinds of simulation evaluation methods for nonvisual effects of dynamic natural light in the world, namely, weighted DA value evaluation, cap model, and circadian effective area (CEA) evaluation. Among them, CEA evaluation takes into account the influencing factor of personal photosensitive history for the first time. It is convenient to calculate and has stronger graphical expression ability of the analysis results. Therefore, CEA evaluation is mainly adopted here for simulation analysis and evaluation of nonvisual effects of natural light.

Workflow Establishment of Simulation Evaluation.
The simulation analysis of nonvisual effects mainly includes daylighting simulated calculation and visualization analysis. With the aid of plug-in Honeybee, the annual dynamic natural lighting is simulated, and data are processed with the Then, the results can be visualized with the combination of Grasshopper's built-in components and Excel. According to the analysis of the measured data, it is found that the illuminance value of the ward facing the north is less sensitive to the weather change, and the illuminance value is low. Therefore, the ward facing the south is taken as the sample ward for computer simulation analysis.
(1) Daylighting Simulated Calculation. Based on the wards' measurement data, a 3D model was built in Rhino. The model was converted to the Honeybee Zone that can be identified by Honeybee, and then it was endowed with materials. The color, roughness, highlights, and other factors of materials will exert an impact on the indoor natural daylighting. Honeybee has a calling component of the Radiance material library, through which plastic and glass materials in Radiance are directly available. The reflectivity and transmittance of the indoor materials used in this paper are acquired with field measurements (Table 1).
Studies have shown that the human body can respond differently when stimulated by natural light at different times of a day, among which the light stimuli from 6 : 00 to 10 : 00 can affect the circadian rhythm at most and effectively advance the circadian rhythm cycle for most of the popula-tion [23]. Nearly 75% of the population, whose circadian rhythm cycle exceeds 24 hours, needs to advance their rhythm cycle every day to better adapt to the light and dark cycle in the natural environment. However, the natural light stimulation during the hours of 10 : 00 -18 : 00 cannot change the circadian rhythm cycle but improve the arousal level of the human body [22]. The nonvisual illuminance requirement of EML_250 for 4 hours every day is proposed in the latest WELL Building Standard 2018, and it is stipulated that the requirement should be met from 9 : 00 to 13 : 00 every day to improve the circadian rhythm of users. However, considering the ward round beginning at 8 : 00 in the hospital and the rest of parturients, the daily analysis period was set at 8 : 00 -12 : 00. Equivalent melanopic lux (EML) thresholds are at 250.
In this study, Annual Daylight Simulation was adopted, where Daysim was called to calculate the annual dynamic natural daylighting and to convert the visual illuminance into equivalent melanopic lux through setting conversion coefficient. By setting the time range and EML threshold for daily calculation, the whole-year hourly EML as well as the whole-year daily stimulus frequency (integer between 0 and 7) meeting EML threshold requirements can be obtained. It can be used as the basis for the follow-up analysis of year-round nonvisual effects in these wards. See Figure 3 for a detailed flow chart of the simulation process. (2) Visualization Analysis. In this paper, some evaluation components were compiled based on VB and combined with Grasshopper's built-in components for data processing. Then, Excel was used to present data, including the annual equivalent melanopic lux (EML), stimulus frequency (Stim.freq), and percentage of circadian effective area (CEA). The specific simulation process is shown in Figure 4.
The analysis chart of whole-year nonvisual illuminance can directly show the nonvisual illumination level of 8760 hours throughout the year, which can be taken as a basis for other analysis indicators. If the EML average of natural light received by a certain point in a certain direction reached 250EML, i.e., the minimal light stimulus standard for nonvisual effects, we deemed that circadian effective (CE) was generated at this measuring point this day. Moreover, in order to quantify the changes of availability of CE stimulus over a period of time (e.g., one week), the concept of stimulus frequency (Stim.freq for short) was proposed [24]. According to the number of days reaching effective CE on the current day and the following 6 days, the effect of CE maintenance on the current day was determined and

10
BioMed Research International divided into 5 levels (A: 7 days; B: 5-6 days; C: 3-4 days; D: 1-2 days; and F: 0 days). The higher the level, the better. The Stim.freq analysis chart mainly presented daily stimulus level at certain points throughout the year, only expressed in time dimension rather than spatial dimension. According to the WELL standard, the EML threshold of 200 and 150, respectively, required 75% and 100% working areas to meet the conditions. Hence, an indicator should be introduced to manifest the spatial condition during evaluation of nonvisual effects. In CEA evaluation, a CEA index has been put forward to represent the percentage of areas that met specific Stim.freq within the space. The CEA analysis chart clearly shows the percentage of measuring points to be calculated which met or exceeded the set Stim.freq within the space. Figure 5 shows the mean value of all day illuminance of each measuring point in both wards, respectively, facing south and north in sunny and cloudy days. In terms of the ward facing south, by comparing the 190 lx reference line, it can be seen that the natural light stimulus of each measuring point can reach the nonvisual threshold during 8 : 00-15 : 00 in sunny days ( Figure 5(a)). Among them, an optimal illuminance value was determined at the bed near the window (test points B and b), which can meet the daily nonvisual needs of 10 hours. On the condition of cloudy day ( Figure 5(b)), the average illuminance value at each measuring point in a ward

11
BioMed Research International was lower than that of sunny day as a whole. The average illuminance value of measuring points close to the window was better than that far away from the window, basically reaching the illumination standard of more than 190 lux for 4 h a day, satisfying the demands of parturients' nonvisual effects. The illuminance values measured at those beds far away from the window (test point c) reached the threshold of nonvisual effects only 2 h daily, while those measured deeper from the room (test points A and a) failed to exceed 190 lx for an hour daily. Therefore, it can be considered that far away from the window (test points A, a, and C), light stimulus failed to reach the stimulus threshold of nonvisual effects on cloudy day, which is not conducive to maintaining the physical health of parturients. In terms of the ward facing north, there was no significant difference in the average illuminance value at each measuring point in the ward in sunny days ( Figure 5(c)). Except for the two time periods of 6 : 00-7 : 00 a.m. and 16 : 00-17 : 00 p.m., the illuminance change range of each measuring point in time ranges is small. Compared with 170 lx reference line, it is clear that only the illumination values measured near the window (test points B, b, c, and d) between 10 : 00 and 13 : 00 meet the standard but did not exceed the standard by much. Hence, the ward was deemed with less ideal natural light environment and could not meet nonvisual needs of human body. On the condition of cloudy day ( Figure 5(d)), it can be found that with the increase of room depth, the illuminance value of the measuring point decreased gradually, and the range was stable. Compared with the 170 lx reference line, in the whole ward, only two test points (test points b and d) near the window met the 4 h illuminance needs daily merely on the condition that parturients were lying. Basically, it cannot meet the nonvisual needs of parturients, which can be considered to be harmful to health. Figure 6 shows the annual time map of the measuring points representing the leaning and lying supine status of parturients as well as those on the surface of baby's cots. As illus-trated, the all-year EML level of the bed (test points B, b, and d) near the window is quite high. Irrespective of a cot or a ward bed and the status of parturient, the natural daylighting of these three measuring points can meet the needs of nonvisual effects of the human body; that of the cot (test point c) in the middle can basically meet the nonvisual needs. In contrast, parturients, either lying supine or leaning against the bed, will expect variable and low equivalent EML levels at the measuring points in ward beds (test point A and a) away from the window and especially a lower level from April to August than that in other months. Therefore, it is hard to meet their physiological rhythm stimuli.

Stimulus Frequency.
In order to sustain a healthy circadian rhythm cycle, Konis suggests that at least 5 days of effective physiological rhythmic stimulus within any 7 days are required to avoid upsetting the internal clock of human body [25]. Therefore, a threshold of Stim:freq ≥ 5 d/wk was selected here as the standard of Stim.freq analysis. Figure 7 shows the year-round distribution of Stim.freq at each measuring point, whereas Table 2 shows the yearround time percentage of Stim.freq at each level in each measuring point. As illustrated below, from 8 : 00 to 12 : 00, the measuring points B, b, c, and d near the window basically sustain their respective critical value Stim:freq ≥ 5 d/wk during the whole year, and the percentage of time in which the actual value exceeds the critical value is over 90% throughout the year. In particular, the measuring point d sustains a stimulation frequency Stim:freq ≥ 7 d/wk per day throughout the year, except winter (December, the following January, and February). The measuring point c in the middle baby's cot hits the critical value at all times except for June. So, it follows that natural daylighting basically satisfies the nonvisual effects. Both measuring points A and a are far from the window. Their Stim.freq can reach the acceptable critical value about 70% of the time throughout the year. Unsatisfactorily, from April to August, the natural daylighting of nonvisual effects is quite poor and detrimental to the health of parturients. Comparatively, irrespective of being close to or away from the window, parturients who lie supine receive much more nonvisual natural light stimulus 13 BioMed Research International than those who lean against the bed. It is also evident that natural light in winter has little impact on nonvisual effects, and the distance from the window and parturients' behavior condition can only produce slight changes in the nonvisual natural light stimulus received by the body.

Effective Circadian Rhythm
Percentage. In addition to lying supine and leaning against the bed, parturients tend to walk around. For the purpose of evaluating the nonvisual effects of natural light under such a state, it is necessary to evaluate the nonvisual natural light environment in a whole ward. With an available circadian effective area, the dynamic natural light in a building space can be evaluated. In this study, a factor of personal sensitization history is introduced for the first time. With this factor, the nonvisual effects of natural light in a space throughout the year can be analyzed. Hence, a grid at a certain distance from the ground is simulated and calculated in this daylighting simulation. In the nonvisual simulation, the WELL stipulates that the position of 1,500 mm from the ground denotes the height of human eyes when standing, typically the horizontal viewing direction. There are 522 grid measuring 200 mm × 200 mm.
As shown in Figure 8, the surrounding images show the distribution of the average CEA values of the month and the mapping of the Stim.freq of each measuring point of the grid in space. The image in the middle shows the CEA maps under each Stim.freq after categorization and merging based on the five levels (F-rank not shown). The effect of seasonal changes on the distribution of natural daylighting of nonvisual effects in the building is evident. Among others, the nonvisual stimulus is suboptimal from April to August, when the CEA above the critical value approaches 60% at the best, but only for a very few days. Overall, except for the four months from April to August, the Stim.fre in other months is 5 d/wk or above, a value quite satisfactory. The daylighting of nonvisual effects in the ward decreases as parturients go inside, especially in areas near the toilet and doorway. All these areas have quite low stimulus throughout the year, indicating that they are biologically dark and need to be supplemented with additional lighting.

Discussion
Based on the measurement data, in terms of the ward facing south, the standard was to meet the natural lighting of no less than 190 lx for at least 4 h every day. In sunny days, the light stimulus at each measuring point in the ward could meet parturients' needs for circadian rhythm stimulation. While in all cloudy days, the natural lighting environment near the window was not ideal. In terms of the ward facing north, the standard was natural lighting of no less than 170 lx for at least 4 h every day. Whether cloudy or sunny, natural daylighting in the ward failed to meet parturients' needs for circadian rhythm stimulation. It can thus be seen that although natural light can meet parturients' visual needs in the ward, it still fails to meet their nonvisual, requiring supplementary illumination. In the Code for Design of General Hospital (GB51039-2014), the orientation of the ward is not clearly specified. It only requires favorable daylighting, ventilation, and matching with terrain. But in fact, depending on the local climate and direction, the houses in some places are livable even they face east or west. Through software simulation, the nonvisual effects of different orientations of sample wards can be optimized and analyzed, to explore the orientation of hospital wards suitable for cold areas (Harbin).
It can be seen from Figures 9 and 10 and that the change of orientation has different effects on different measuring points. The farther away the hospital bed is from the window, the more sensitive the natural light nonvisual effect is to the change of orientation. Hence, when selecting a direction the ward faces, priority should be given to the natural daylighting having nonvisual effects near the entrance. In the ward facing east or south, percentage of test points exceeding the threshold (Stim:freq ≥ 5 d/w) exceeded 70%. In the ward facing west or north, only those measuring points near the entrance had poor performance of natural light's nonvisual effects, and percentage exceeding the threshold (Stim:freq ≥ 5 d/w) was about 50%. Therefore, in terms of natural light's nonvisual effects on parturients, the east and south direction is optimal and less different from each other, the west direction is the worst, and the north direction is on the middle level. Through the aforesaid computer simulation, it can be concluded that the natural daylighting having nonvisual effects is problematic at beds away from the window. Especially from April to August, natural light basically cannot meet parturients' needs of circadian rhythm stimulation. This is because the ward is deep inside the war and the solar altitude angle is high from April to August, so the sunlight cannot directly reach the test points. The area near the entrance requires a supplement of high-intensity artificial illumination. Apart from the supplementary artificial illumination, the ward can also be optimized. However, it is quite complicated and demands a huge amount of work to implement optimization methods, such as changes of the ward's plane layout and proportion, or the combination between wards. In comparison, it is more simple and direct to improve the ward's daylighting by changing the ward's window-to-floor ratio.
According to Figures 11 and 12, the change of windowto-floor ratio had different impacts on the nonvisual light environment at different locations in the ward. It has the greatest impact on test point a near the entrance and the smallest impact on measuring point a near the window. That is, with the increase of room depth, the impact of the change of window-to-floor ratio on the stimulus frequency of test point increases. When the window-to-floor ratio reached 0.35, percentages of Stim:freq ≥ 5 d/w at each measuring point were basically the same, and the ratio rising above 0.35 had the same impacts on each test point. When the ratio was 0.3, Stim.freq values at each measuring point all maintained at a higher level throughout the year, and percentages of Stim:freq ≥ 5 d/w all exceeded 70% at each point. Compared to that of 0.35, the ratio of 0.3 is more energysaving and environmentally friendly, so 0.3 was a more appropriate standard value out of these 6 ratios.

Conclusions
In this paper, we adopted WELL standard and CEA evaluation method that were used to quantify nonvisual effects of natural light in wards. According to the study findings, a ward's natural light environment that meets visual needs does not necessarily meet parturients' nonvisual needs. A ward's nonvisual natural light environment has something to do with parturients' behaviors and ward orientation. From the obtained results, the following conclusions can be drawn: (1) Computer simulation has basically the same results with those of the field measurement-natural light's nonvisual effects on people gradually decrease with the increase of depth perpendicular to the ward's lighting surface width. Especially during April-September or on the whole-day cloudy condition, as for parturients away from the window, no matter lying supine or leaning, the natural light basically cannot meet their needs for circadian rhythm stimulation. In that way, supplementary illumination is required. However, during the field measurement, it was found that illuminance values near the window in the ward rose significantly in the morning. If the illuminance value is too high, it is prone to glare and visual fatigue, resulting in physical discomfort. Shading facilities need to be added appropriately to avoid such situations (2) Natural daylighting of the overall nonvisual effects in the ward can vary significantly with seasonal changes. In summer, nonvisual stimuli received in the area near the window are remarkably better than those near the entrance. While in winter, there are no great changes. Although some areas received a higher value of nonvisual illuminance and a longer duration every day in summer, in terms of meeting the threshold standard and the uniformity, nonvisual natural light environment in the ward is better in winter than that in summer (3) The change of orientation and window-to-floor ratio has a certain effect on the natural lighting of nonvisual effect in maternity ward. According to the field measurement, the ward facing south has better nonvisual effects than that facing south, which maintain at a higher level. Computer simulation analysis shows that the priority of different directions that wards face in the cold area (Harbin) is as follows: south>east>north>west. When the window-to-floor ratio is 0.3, the nonvisual natural light environment is optimal

Data Availability
The original contributions presented in the study are included in the article, and further inquiries can be directed to the corresponding author/s.

Ethical Approval
Ethical review and approval was not required for this study on human participants in accordance with local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

Conflicts of Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.