1. Introduction
There is no doubt that global climate change is causing serious urban warming (Limaye et al. 2018) and due to the rapid urbanization, cities often inadvertently create warmer and drier urban climate conditions than their surrounding areas, which is known as the urban heat island (UHI) phenomenon. These thermal environments cause heat-related health problems [
1,
2]. Meanwhile, heat stress, which can be assessed by human energy budget models, plays an important role in causing heat-related health. According to Epstein and Moran [
3], various age groups, including infants, children, and adults over 65, are most vulnerable to heat-related death since they are more sensitive to excessive heat stress. Heat stress can also reduce workers’ productivity and cause more serious heat-health problems, especially in vulnerable groups [
4].
While individuals have little control over the impact of global issues on heat-related health, there is evidence that urban design ameliorates health problems by increasing human thermal comfort and contributing to reduced heat stress at the urban level [
5,
6]. In this regard, researchers suggest microclimatic urban design (MUD), which is the process of determining the physical organization of buildings and open spaces in urban areas to mitigate thermal environments [
7]. Many studies are underway on MUD and they have found that MUD has positive impacts on urban microclimate and heat-related health [
8]. Their strategies focus on green areas [
9,
10], surface of building and pavement [
5,
11], water bodies [
12,
13,
14], and structure of streets and buildings [
15,
16] at the micro- and mesoscale.
According to the MUD-related research, terrestrial radiation plays an important role in outdoor human thermal comfort. Brown and Gillespie [
17] demonstrated the effect of terrestrial radiation on human thermal comfort through the COMFA model. They found that terrestrial radiation absorbed by a person has negative effects on human thermal comfort during the hot summer weather. All of the materials on Earth emit invisible terrestrial radiation as a function of their surface temperature [
18], and the warmer the object is, the more terrestrial radiation it emits. This invisible energy has a significant effect on outdoor human thermal comfort. Moreover, Shahidan et al. [
18] demonstrated that the solar radiation filtering capacity of trees reduces the terrestrial radiation by cooling the ground, and it is related to improve outdoor thermal comfort in tropical open spaces.
Among the MUD elements, including green, surface, water body, and urban structure, terrestrial radiation is related to the surface. However, surface-related studies have mostly focused on the energy that varies with the albedo of different ground materials, and has rarely considered the amount of terrestrial radiation emitted. Moreover, previous studies have used thermal comfort models to calculate energy budgets that can identify the thermal stress level [
19], and there have been 165 thermal indices developed for estimating indoor and outdoor thermal comfort levels [
20]. However, the impact of terrestrial radiation is not considered in most of the existing models. The existing models typically use air temperature, solar radiation, relative humidity, and wind speed as the main microclimate parameters. For example, Coccolo et al. [
21] conducted a comprehensive review of models and standards of outdoor thermal comfort and stress. A total of 21 thermal comfort models were analyzed in their study. Among them, 11 models consider solar radiation, 21 models consider air temperature, 19 models consider relative humidity, and 15 models consider wind speed. Only one model, the COMFA, considers the terrestrial radiation absorbed by a person, but it does not take into account the surrounding ground conditions at the point where the subject is standing. It only considers the terrestrial radiation of the point where the subject is standing.
By developing a model which can consider the various ground conditions, the ground ratio (GR) of each material, and their terrestrial radiation, this study can contribute to creating MUD strategies by updating existing energy budget models, which can lead to the measurement of more accurate human thermal comfort. Therefore, the goals of the study are to (1) present the need to consider terrestrial radiation for human energy budgets, (2) develop a ground ratio factor (GRF) model that can estimate the difference in terrestrial radiation according to different ground conditions, (3) validate the GRF model through field testing, and (4) propose how the model can be used in microclimatic urban design to ameliorate urban heat islands.
4. Discussion
This study suggests the need to consider terrestrial radiation in thermal comfort measures to create MUD strategies. The findings indicate that terrestrial radiation is emitted differently depending on the proportion of different ground conditions. To recognize the need of terrestrial radiation, it is important to understand why terrestrial radiation is important and how terrestrial radiation affects heat-related human health. Terrestrial radiation has a huge effect on human heat stress. As mentioned in the introduction, terrestrial radiation plays an important role in the outdoor human thermal comfort, which is defined as a state of mind that shows satisfaction with the thermal environment [
28]. Previous studies have found that reducing the terrestrial radiation absorbed by a person improves the human thermal comfort [
18,
29,
30].
High temperatures cause heat stress by making a person thermally uncomfortable. These thermal stresses cause heat-related health problems. Rohat et al. [
31] showed that the heat stress risk, which is linked to the exposure to vulnerable people and extreme temperature events, such as heat hazard, resulted in heat-related health impacts. Moreover, Claudia et al. [
32] assessed the heat stress effects on mortality. Their results showed that there is an increase in the number of outpatients and deaths on hot days by comparing days without thermal stress. Moreover, Kovats and Hajat [
33] conducted a critical review on the relationship between heat stress and public health. Their results showed that heat stress causes heat-related illnesses (clinical signs), which eventually lead to heat death (
Figure 7).
This study simulated the energy budgets to demonstrate the effect of terrestrial radiation on human thermal comfort using the COMFA model proposed by Brown and Gillespie [
17]. An example can be used to illustrate the importance of terrestrial radiation on a person’s thermal comfort. Assume that the air temperature is 25 °C, the ground temperature is 35 °C, the human surface temperature is 25 °C, the ground material is grass, and a person is standing. The energy budget of terrestrial radiation emitted from the ground and absorbed by a person is 510 W/
. The energy budget refers to the degree of thermal comfort and the results showed that terrestrial radiation accounts for a significant portion of the total energy budget (
Figure 8).
There are no thermal comfort-related studies considering the different terrestrial radiation according to the different ground conditions. Therefore, this study is significant enough in conducting leading research on related issues through model development. This study developed two versions of the GRF model: The GRF-G model that can use ground temperature and the GRF-A model that can use air temperature. The models were developed with the theories and compared with actual measurements for validation. The result that both models have a statistically significant correlation between PTR and ATR (correlation coefficients = about 0.8 ~ 0.9, p < 0.01) shows that it is theoretically possible to estimate terrestrial radiation through the models.
This study also found that when considering two different ground materials, the correlation between the PTR and ATR was higher than considering three ground materials. For the GRF-G model, the average coefficient of correlation analysis is 0.903 for two different materials and 0.810 for three different materials. For the GRF-A model, the average coefficient of correlation analysis is 0.814 for two different materials and 0.779 for three different materials at the 0.01 significant level. This indicates that the model is more accurate when considering fewer ground materials.
There are several limitations that need further research. When measuring the terrestrial radiation emitted from the ground using instruments, it is important to know how the instruments theoretically measure the terrestrial radiation and how terrestrial radiation is received by a real person. Ground temperature is usually measured vertically downward using a thermal camera (FLIR E6) or a pyrgeometer. Assuming that a person can be represented by a vertical cylinder, the terrestrial radiation from the ground (
Figure 9A) should be translated into the amount of terrestrial radiation received by a vertical cylinder (
Figure 9B).
The CNR4 upper longwave detector has a view of nearly 180° and the lower detector has a view of 150° (
Figure 10). This indicates that it can measure all of the terrestrial radiation from the sky and ground hemisphere except for the space of 30°. It seems that the missing 30° might have only a minor effect on the amount of terrestrial radiation for a horizontal flat plate, but it might be quite important for a vertical cylinder. Therefore, further research needs to complement this limitation through theoretical and empirical grounds.
Further research should also be conducted on a more accurate conversion using various methods and considering additional microclimate parameters. Ground temperature is used to calculate the terrestrial radiation, which provides the most accurate value. However, ground temperature data have an issue in that data cannot be easily obtained relatively compared to the air temperature data. Therefore, previous studies attempted to estimate the ground temperature using the air temperature. This study used the regression analysis approach using only air temperature to estimate the ground temperature. In this study, the regression analysis was used. The results showed R
2 values of 0.762 for asphalt, 0.663 for concrete, and 0.573 for grass, respectively. This approach provides statistically significant results, but various methods for a more accurate conversion should be considered, which can include additional microclimate parameters. Many methods have been proposed to convert air temperature to ground temperature. Khan et al. [
22] suggested the model for the asphalt ground temperature. In their model, various microclimate parameters, such as wind speed (m/s), air temperature (F), relative humidity, and solar radiation (W/
) were considered to estimate the ground temperature (F). According to their results, the model has an
value of more than 0.9. Moreover, according to Diefenderfer et al. [
34], their model included the daily maximum ambient temperature (°C), calculated daily solar radiation (W/
, and depth from the surface (m) to calculate the predicted daily maximum ground temperature (°C). Their adjusted
value of the model was 0.771 for maximum temperature and 0.798 for minimum temperature. In addition, this study only considered air temperature and sky condition, while relative humidity and wind speed were not considered. Basically, relative humidity and wind speed can be theoretically similar when measuring terrestrial radiation for each measuring point. However, at the same time, they can affect the amount of terrestrial radiation delivered from the ground to the person. Therefore, to conduct a more accurate conversion, a further consideration of various methodologies is needed.
Additional models for various areas should be developed, since climate relationships and characteristics may vary depending on the location. According to the Köppen climate classification, climatic conditions vary depending on the region, continent, country, etc. As a pioneering research, this study was limited to the City of College Station, located in the Cfa climate zone (humid subtropical climate) for the increasing validity and reliability of the results, but additional research for different climatic zones is needed to use the GRF model in more diverse areas. The relationships between microclimate parameters vary depending on the climate zones. Mildrexler et al. [
35] conducted a global analysis to compare air temperatures with land surface temperatures using satellite data of every World Meteorological Organization station on Earth. Their results showed that the relationship varied by latitude and longitude. Emissivity also varies across areas. Brown and Gillespie [
17] found that most of the landscape design elements have a coefficient of 0.9 or more, thus when we are satisfied with the estimate of 10% of reality, we can use uncorrected equations. However, it is necessary to use more accurate emissivity values to measure the accurate terrestrial radiation. According to Wang et al. [
36], even the same material may have different emissivity depending on changes in the temporal or spatial distribution. Therefore, further studies should consider various climates and seasonality considering the relationship between emissivity and terrestrial radiation.
Finally, for a better terrestrial radiation estimation, a wider range of ground materials, the consistent properties of each material, and securing more samples should be considered. This study considered only three materials: Asphalt, concrete, and grass. Soil was excluded due to the difficulty in securing a target site as it is a private issue. Therefore, additional ground materials need to be considered to apply the CRF model to various materials. Moreover, even if it is the same ground material, the characteristics of albedo and emissivity vary slightly depending on the different neighborhood environments. For example, according to the results of this study, grass has slightly different albedo and emissivity depending on the measurement spots since they have different grass conditions depending on their management status. These differences need to be considered for accurate research results. Furthermore, as a pioneering research, we tried to secure the sample number which can be statistically significant to present the need to consider different amounts of terrestrial radiation. More than 30 samples were targeted and 51 for A-C, 46 for A-C, 44 for C-G, and 39 for A-C-G were used. However, further research will need to increase the samples and survey areas to achieve more accurate and statistically significant results.