Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2024 Jun 4;58(24):10524–10535. doi: 10.1021/acs.est.4c00424

Moderating AC Usage Can Reduce Thermal Disparity between Indoor and Outdoor Environments

Hong Wei †, Bin Chen ‡,§,∥,*, Kangning Huang ⊥, Meng Gao #, Bin Fan †, Tao Zhang †, Ying Tu †, Bing Xu †,¶,*
PMCID: PMC11192031  PMID: 38832650

Abstract

graphic file with name es4c00424_0010.jpg

In the context of escalating urban heat events due to climate change, air conditioning (AC) has become a critical factor in maintaining indoor thermal comfort. Yet the usage of AC can also exacerbate outdoor heat stress and burden the electricity system, and there is little scientific knowledge regarding how to balance these conflicting goals. To address this issue, we established a coupled modeling approach, integrating the Weather Research and Forecasting model with the building energy model (WRF_BEP + BEM), and designed multiple AC usage scenarios. We selected Chongqing, China’s fourth-largest megacity, as our study area due to its significant socioeconomic importance, the severity of extreme heat events, and the uniqueness of its energy infrastructure. Our analysis reveals that AC systems can substantially reduce indoor temperatures by up to 18 °C; however, it also identifies substantial nighttime warming (2–2.5 °C) and a decline in thermal comfort. Particularly for high-density neighborhoods, when we increase 2 °C indoors, the outdoor temperature can be alleviated by up to 1 °C. Besides, despite the limited capacity to regulate peak electricity demand, we identified that reducing the spatial cooled fraction, increasing targeted indoor temperature by 2 °C, and implementing temporal AC schedules can effectively lower energy consumption in high-density neighborhoods, especially the reduction of spatial cooled fraction (up to 50%). Considering the substantial demand for cooling energy, it is imperative to carefully assess the adequacy and continuity of backup energy sources. The study underscores the urgency of reassessing energy resilience and advocates for addressing the thermal equity between indoor and outdoor environments, contributing to the development of a sustainable and just urban climate strategy in an era of intensifying heat events.

Keywords: air conditioning (AC), scenarios simulation, thermal comfort, energy consumption, equality

Short abstract

Minimal research exists on the impact of air conditioning (AC) on thermal comfort equity between indoor and outdoor environments. This study reports the significance of moderating AC spatial and temporal settings on thermal equity, with implications for heat exposure to human health.

1. Introduction

Severe heat events have manifested in cities across the globe in recent years, resulting in significant and unequal socio-economic losses and adverse impacts on human health.1,2 Recent research has indicated that the annual mortality attributed to heat-related illnesses exceeds that associated with any other category of extreme weather events.3 Notably, cases of heatstroke have been widely reported in China, particularly in hot and humid regions.4 One of the most effective methods for enhancing indoor thermal comfort is the use of air conditioning (AC) systems to attain a target cooling temperature within buildings. However, studies have demonstrated that AC usage can exacerbate outdoor thermal discomfort.5−7 For instance, a case study conducted in a district of Tokyo revealed that AC units contribute to a 1.3 °C increase in air temperature.8 Similarly, in Tokyo office areas during summer weekdays, AC operation results in temperature rises of 1–2 °C.9 Limited investigations conducted in Chinese cities have reported similar findings, such as temperature increases of 1–1.5 °C during the daytime and 1.5–2.4 °C during the nighttime in Beijing,7 while minimal daytime increases of 0.6 °C and nighttime increases in Jiangsu Province.10 These studies have estimated the substantial contribution of anthropogenic heating from AC systems by employing a two-way coupling approach between a building energy model (BEM) and an atmospheric model. This approach effectively captures the anthropogenic heating resulting from internal building heat and the release of waste heat by AC systems, which accounts for approximately 20% of the total. It is widely recognized that indoor thermal comfort significantly influences human well-being;11−13 however, recent incidents of heatstroke have highlighted that outdoor workers, such as roof workers and delivery workers, are particularly vulnerable to heat stress during periods of extreme heat.14 Additionally, research indicates that the diversity in indoor thermal environments across buildings, influenced by AC installation and usage, is closely linked to socioeconomic factors.12,15 The prevalent use of ACs exacerbates disparities, as buildings without AC are disproportionately affected by rising outdoor temperatures,16 increasing thermal exposure risks for low-income residents. While numerous studies have focused on achieving equity in human settlements across different income levels,17−20 the specific impact of AC installation and usage on thermal exposure equity remains underexplored. For instance, little research has explored the magnitude of outdoor temperature increase due to indoor temperature reductions. This issue, which is expected to intensify in the context of climate change, warrants further attention.15

In addition to its impact on the urban microclimate, careful consideration must also be given to the energy demands and consumption associated with AC usage. As per the projections by the International Energy Agency, the energy demands for indoor cooling are expected to surge 3-fold, rising from 2020 TW h in 2016 to 6200 TW h in 2050. This surge in demand will constitute approximately 30% of the total global building electricity consumption.21 Mesoscale meteorological models offer the capability to estimate cooling-related energy consumption by incorporating both building and urban climate characteristics into their modeling framework.5,6,22,23 Nevertheless, due to the scarcity of high-temporal and high-spatial-resolution energy consumption data, AC simulation models possess a distinct advantage in describing spatial and temporal energy consumption patterns. Furthermore, these models enable the exploration of AC regulation measures aimed at reducing and adjusting peak electricity demand burdens.

The use of AC exhibits a high degree of temporal and spatial consistency with urban heat events.24,25 Studies have provided compelling evidence of the occurrence of compound events involving heatwaves and disruptions in electricity generation and transmission.26 An illustrative case of such an event unfolded in Chongqing in 2022, resulting in substantial economic and health losses. Furthermore, it is concerning that the cities most susceptible to these compound events are typically megacities characterized by dense populations and developed socioeconomic activities. In such urban centers, a lack of contingency plans for electricity failures renders them highly exposed and vulnerable to heat-related consequences. Climate change-induced warming is projected to persistently increase the incidence of heat-related illnesses and mortality by midcentury. Compound events combining heatwaves with infrastructure failures will exacerbate heat stress. Therefore, it is imperative to quantitatively assess the spatial and temporal impacts of AC usage on urban microclimates and energy consumption.

To gain a deeper understanding of the aforementioned challenges, the primary objective of this study is to employ scenario-based simulations using the Weather Research and Forecasting (WRF) regional climate model. The WRF model, known for its systematic development and widespread utilization, serves as the foundation for quantifying and enhancing our comprehension of the spatial and temporal patterns of AC usage’s impact on urban thermal comfort and energy consumption. Furthermore, this research aims to provide valuable policy recommendations for rational AC usage during extreme heat events.

2. Materials and Methods

2.1. Study Area

Chongqing, situated in the eastern Sichuan region of China, holds pivotal significance for socioeconomic development in the central-western region, and as the fourth largest megacity in China, it accommodates a total population of 32.05 million according to the seventh census (National Bureau of Statistics, 2021). Characterized by a subtropical monsoon climate, Chongqing is renowned as one of China’s “three furnace cities” due to its hot and humid summers.27 The urbanization rate, exceeding 70%, contributes to elevated summertime temperatures within the city. In August 2022, Chongqing recorded its highest temperature since 1961. Compared to the same period, there were 15 additional days with extreme heat (>35 °C), and the average temperature rose significantly by 2.4 °C (China Meteorological Administration, 2022). The extreme demand for AC to provide indoor cooling was unattainable due to electrical failures occurring during heat extremes. This shortage can be attributed to a deficit in hydroelectric power, which typically accounts for over 80% of electricity generation in Sichuan Province, the main region for importing electricity to Chongqing. Consequently, the convergence of extremely high summer temperatures and an energy shortage has rendered Chongqing a vulnerable city with a heightened risk of heat exposure. It is imperative to comprehensively analyze the spatial and temporal patterns of heat and energy consumption in Chongqing and identify optimized strategies to respond effectively to extreme heat events.

2.2. WRF Model Configuration

The WRF-Urban 4.1 model was employed to simulate AC energy consumption and its influence on the microclimate in Chongqing during a heatwave period spanning from August 10th to August 30th, 2022, with a 48 h spin-up time. Within this model, the Noah-MP scheme of WRF version 4.1 is coupled with the multilayer building effect parameterization and BEM, as outlined by ref (28). This integrated framework enables the estimation of time-varying indoor temperatures for each floor of buildings and accounts for the energy interactions between buildings and the outdoor atmosphere due to AC usage.

The modeling domain of WRF is centered at coordinates (29.32°N, 106.31°E) and encompasses three two-way nested domains with grid spacings of 9 km (99 × 99 grid cells), 3 km (79 × 103 grid cells), and 1 km (73 × 97 grid cells), respectively (Figure 1). The vertical coordinate system includes 50 sigma levels, with a focus on 13 levels within the lowest 1 km to enhance the resolution of urban boundary layer processes. The physical parametrization schemes used in WRF simulations are as follows: (1) Purdue Lin Scheme;29 (2) Yonsei University Scheme (YSU);30 (3) Grell 3D Ensemble Scheme31,32 only for d01; (4) RRTMG Shortwave Scheme33 and RRTM Longwave Scheme;34 (5) Noah-MP Land Surface Model;35 and (6) Revised MM5 Scheme.36 The initial and boundary conditions are obtained from the National Centers for Environmental Prediction Global Forecast System final gridded analysis data sets with a spatial resolution of 1° × 1° and a temporal resolution of 6 h.

Figure 1.

Figure 1

Open in a new tab

Study area and the domain settings of WRF. (A) Chongqing location in China. (B–D) Land cover and land use of WRF domains d01, d02, and d03. The green triangle in (C) is the validation station.

2.3. Experiment Settings

To gain insights into the spatiotemporal impact of AC on local climate, we established a verified baseline as a control group for comparison with various AC scenarios. The primary distinction between the AC scenarios and the baseline lies in the incorporation of AC applications. Specifically, we implemented five parametrization schemes to characterize the multifaceted effects of AC (Table 1).

Table 1. Design of WRF-Urban Numerical Experiments.

AC scenarios model setup
24 h-AC (Sce1) 24 h use of AC
48 h-electricity-failure (Sce2) same as 24 h-AC, electricity failure happens on two of the hottest days (8.17–8.19)
AC-cooled-fraction (Sce3) same as 24 h-AC, but considering cooled fraction for three urban classes
AC-schedule (Sce4) same as 24 h-AC, but with spatially varying AC working schedules for commercial/industrial (8:00–18:00) and residential (0:00–8:00 and 18:00–24:00)
AC-indoor-temperature (Sce5) same as 24 h-AC, but increase 2 K of indoor temperature, 27°C for residential area, 26°C for commercial area
Open in a new tab

In the 24 h AC scenario (Sce1), we assumed that all areas within buildings operated AC systems continuously throughout the day. Recognizing that not all indoor spaces are cooled and some buildings lack AC altogether, we introduced the concept of a “cooled fraction” as a function of urban classes (AC-cooled-fraction, Sce3). According to previous studies,7 we set cooled fraction to 50% for low-density residential areas, 64% for high-density residential areas, and 95% for commercial areas, close to the fraction setting for Tokyo,9 Madrid,37 and Phoenix.6 Furthermore, to capture the diurnal AC usage patterns more effectively, we introduced distinct working schedules for AC operation in the AC-schedule scenario. In this scheme, the working hours for commercial areas were set from 8:00 to 18:00 local standard time, while residential areas operated AC systems from 0:00 to 8:00 and 18:00 to 24:00. This schedule was validated through observation data, demonstrating its ability to faithfully reproduce the observed double-peak features of AC electric loads.7 Additionally, we conducted a scenario similar to the 24 h AC scenario but with a 2 °C increase in indoor temperature, resulting in 27 °C for residential areas and 26 °C for commercial/industrial areas (AC-indoor-temperature, Sce5).

2.4. Input Data

We replaced the default geographical input data with the 2020 land cover data set from the moderate resolution imaging spectroradiometer (MODIS) (MCD12Q1) at a resolution of 500 m. Furthermore, we reclassified the MODIS land cover category “Urban and Built-up Lands” (value 13) into three distinct urban classes by downscaling the 30 m Global Impervious Surface Area (GISA) data38 to 500 m and calculating the impervious land percentages as follows: low-density residential (30–70%, category 31), high-density residential (70–90%, category 32), and commercial/industrial (>90%, category 33) (Figure 1). Our analysis revealed consistent spatial patterns between GISA and the Global Artificial Impervious Area (GAIA) data set39 for the year 2020. Both data sets accurately captured impervious surface details, and GISA reduced impervious land fragmentation through spatial smoothing (Figure S1). Additionally, we departed from solely representing land use through impervious cover percentages. Instead, we utilized the Essential Urban Land Use Categories (EULUC) land use mapping and corresponded its classification system to our specific land use categories (categories 31, 32, and 33) (Table S1). To accurately represent the thermal characteristics of urban areas in Chinese cities, we adopted values for urban thermal conductivity (1.5 W/mK), heat capacity (1.75106 J/m3 K), fraction of urban landscape (0.5, 0.8, 0.95), and surface albedo (0.2) based on previous studies.7,40

Regarding various urban canopy parameters, we treated them as functions associated with three distinct urban land use categories. To enhance the accuracy of depicting urban vertical morphology, we calculated building height using a newly available data set with a 30 m resolution41 (Table S2). Additional urban morphology attributes, including roughness length (0.15, 0.05, 0.05), street direction (0°, 90°), street width (15 m), and building width (15 m), were set in alignment with a previous urban modeling study conducted for Chengdu, Sichuan.40 Given our specific focus on AC operations, we sourced AC-related parameters from a building energy modeling study. These parameters encompassed the coefficient of performance of AC systems (4.5),6 target temperatures for AC systems across the three urban land use categories (25, 25, and 24 °C), and peak heat generation from equipment (20, 25, and 44.3 W/m2). Furthermore, we utilized population grid data from WorldPop in conjunction with building height data to estimate the peak number of occupants per unit floor area (0.04, 0.03, and 0.13). These estimates closely align with findings from previous research.7

2.5. Model Validation and Robustness Analysis

We established a baseline model without AC systems to assess the model’s reliability by comparing simulation results with weather observations from two National Climatic Data Center (NCDC) stations within our study area, available during the study period. Due to the unavailability of Chongqing’s daily electricity usage data during the study period, we were unable to verify the specific AC usage patterns and thus could not choose a particular scenario result for comparison with validation data. Consequently, we adhered to the methodology of prior studies, validating the spatiotemporal consistency of our model simulations through a baseline scenario.42,43

Given that various parametrization schemes can influence model performance, we conducted a careful selection of an optimized parametrization scheme prior to the scenario modeling. Specifically, we compared the following: (1) the shortwave radiation option of Goddard44 and RRTMG33 and (2) planetary boundary layer (PBL) option of Yonsei University Scheme (YSU)30 and Mellor–Yamada–Janjic Scheme (MYJ).45 Following the selection of the most suitable parametrization scheme, we applied it in various scenario models, as discussed in Section 2.3.

Furthermore, considering that both indoor and outdoor temperatures are influenced by building attributes, as noted in refs (43 and 46), collecting comprehensive building data on a city-wide scale presents a significant challenge. To address this, we conducted a robustness analysis by testing various building attributes. Specifically, we adjusted the heat capacity values for roofs, buildings, and the ground (AKSR, AKSB, and AKSG) to 0.75 and 1.5 (the default value) and adjusted the thermal efficiency of heat exchangers to 0.5 and 0.75 (the default value). Leveraging robust simulations, we then proceeded to examine the impacts of the aforementioned scenarios in greater detail.

2.6. Heat Index and Energy Consumption Calculation

The heat index (HI), also referred to as the apparent temperature, reflects how the temperature is perceived by the human body when combined with relative humidity (RH), making it a crucial parameter for assessing human comfort.47 When exposed to excessive heat, the human body initiates perspiration or sweating as a cooling mechanism, highlighting the significant relevance of HI to human health.48,49 The National Oceanic and Atmospheric Administration (NOAA) defines HI as a composite measure of air temperature (T) and relative humidity (R), as represented in eq 1 below. To account for the fact that WRF does not directly provide RH values, we derive them using eq 2.

2.6. 1
2.6. 2

where psfc is the surface pressure, and Q2 is the specific humidity.

Furthermore, taking into account human physiological responses to varying HI levels, NOAA has developed an HI classification system that aids in assessing health risks and informing urban climate policies. In this system, HI values below 80 are considered safe, those between 80 and 90 warrant caution, and readings above 90 indicate extreme caution. NOAA further designates danger and extreme danger thresholds at values exceeding 103 and 125, respectively. However, it is noteworthy that in the course of our simulation study, we were unable to model temperature maxima high enough to observe any areas with HI values exceeding 100.

In terms of energy consumption, we need to estimate the magnitude of electricity used in the report of the Chongqing Statistical Yearbook and compare it with the model results. Drawing on observations from Beijing’s monthly mean electricity loads,7 we note that peak months exhibit electricity consumption 1.6 times greater than low-peak months. Thus, we conservatively estimate that AC-related electricity consumption is approximately 60% of that used for other daily purposes.

3. Results

3.1. Model Validation

To validate both the model’s reliability and the chosen parametrization scheme, we initiated a comparative analysis by examining the results obtained using different parametrizations, as discussed in Section 2.6. The outcomes, illustrated in Figure 2, demonstrate that the shortwave radiation parametrization schemes of Goddard and RRTM yield similar levels of accuracy, with RRTM outperforming in modeling wind speed. In the realm of PBL parametrizations, YSU exhibits a slight advantage over MYJ in replicating atmospheric temperature and wind speed patterns. Consequently, we adopted RRTM as the shortwave radiation scheme and YSU as the PBL scheme for subsequent simulations. Furthermore, given the diverse land use characteristics within urban areas, we conducted a comparative assessment of model results using two different land use parameterizations: EULUC and impervious land percentage based on GISA. The analysis revealed that the latter, the GISA-based approach exhibited slightly superior performance. Hence, all subsequent simulations incorporated the GISA-based land use parametrization. Figure 2 illustrates a minor deviation in the model’s representation of maximum daily temperatures, specifically a cold bias. However, considering the model’s overall performance and the fact that differences among scenarios can offset simulation errors for extreme values within a day, the model’s performance is deemed acceptable.

Figure 2.

Figure 2

Open in a new tab

Model parameterization validation. (A–C) Shortwave test between RRTM and Goddard. (D–F) PBL test between MYJ and YSU. (G–I) Land use test between GISA and EULUC. Difference between model and observation of AT (A,D,G), U10 (B,E,H), and V10 (C,F,I).

3.2. AC Spatiotemporal Contribution on 2 m Air Temperature

Across all scenarios, significant differences in 2 m air temperature (T2) compared to the baseline were evident, as illustrated in Figure 3. Given the similarity in temperature difference patterns on a daily scale, we chose the 19th of August as a representative day for a more detailed examination of the daily profile. Figure 3B reveals minimal temperature variation during daytime hours (8:00–18:00), while a distinct urban warming effect of more than 2 °C was observed during nighttime (19:00–7:00), with maximum temperature increases exceeding 5 °C. Although similar temperature changes were observed, variations emerged among scenarios. Specifically, the 24 h AC (Sce1) and AC-cooled-fraction (Sce3) scenarios exhibited similar temporal patterns, whereas AC-schedule (Sce4) displayed a different pattern. Regarding temperature warming, the 24 h AC scenario registered the highest increase, slightly surpassing AC-cooled-fraction, while AC-schedule resulted in the lowest temperature rise. To ascertain that the differences in T2 were attributable to AC usage, we conducted an analysis to estimate the temporal relationship between sensible heat flux changes due to AC and T2 differences, as presented in Figure S2. This analysis revealed a significant temporal correlation between sensible heat flux and T2 differences, with a noticeable 2 h lag between the two variables.

Figure 3.

Figure 3

Open in a new tab

(A) T2 temporal difference between scenarios and baseline during the study period and (B) on 19th August. The gray area in (A) highlights the period corresponding to (B).

We proceeded by mapping the spatial differences in T2 to observe the areas experiencing warming at 19:00 in the 24 h AC scenario (Figure 4A) and the AC-schedule scenario (Figure 4C), which corresponds to the time when the warming effect is most pronounced, as indicated in Figures 3. Figure 4A,C illustrates that the increases in T2 are closely linked to urban land use. Consequently, we conducted a zonal statistical analysis based on land use types, and the results of which are presented in Figure 4B,D. In the case of the 24 h AC scenario (Sce1), it is apparent that commercial and industrial areas serve as hotspots where T2 rises are attributed to AC usage. In contrast, for the AC-schedule scenario (Sce4), commercial and industrial areas deactivate their AC systems at 18:00, resulting in limited T2 changes thereafter (Figure 4D). Additionally, the spatial distribution of T2 warming from 0:00 to 24:00 on the 19th of August is depicted in Figure S3, while the zonal statistics of urban land use at 7:00 and 19:00 are presented in Figure S4. These Supporting Information Figures illustrate that the changes in T2 according to land use types are consistent at both time points in the 24 h AC scenario.

Figure 4.

Figure 4

Open in a new tab

T2 difference among different scenarios at 19:00 of 19th August. Spatial T2 difference between Sce1 (A), Sce4 (B), and baseline. Zonal statistics of T2 difference on urban land use (low-density residential area, high-density residential area, and commercial/industrial area) between Sce1 (C), Sce4 (D), and baseline.

Regarding the impact of a 48 h electricity failure (Sce2), we examined its influence on T2 in comparison to the 24 h AC scenario (Sce1). Interestingly, our analysis revealed that if electricity fails for 48 h, the effects persist for an additional day (Figure 5A). During the electricity failure, a decrease in T2 is observed, followed by a rapid increase in T2 as extensive AC usage ensues, with the highest increase occurring at 12:00, approximately 4 h after power restoration (Figure 5B). Further investigation into the T2 increase across different land use types, as depicted in Figure 5C, highlights that the most significant T2 increase occurs in commercial and industrial areas. This finding is corroborated by a similar result observed in the case of a 72 h electricity failure, confirming the reliability of our findings and presented in Figure S5.

Figure 5.

Figure 5

Open in a new tab

T2 difference between Sce1 and Sce2. (A)T2 temporal patterns of Sce1 and Sce2 during the study period. (B) T2 difference between Sce1 and Sce2 from 17th August to 21st August, corresponding to the gray time span in (A). (C) T2 difference between Sce1 and Sce2 at 12:00 on 19th August, corresponding to the red point in (B).

3.3. AC Spatiotemporal Contribution on Thermal Comfort

In addition to examining differences in T2, we conducted a comparative analysis of HI variations between the baseline and scenarios Sce1, Sce3, and Sce4. Upon assessing the temporal patterns of HI differences alongside T2 differences, we observed a temporal alignment between them (Figure S6). Given that HI is influenced by both T2 and RH (eq 1), we investigated whether specific humidity (Q2) follows a regular daily pattern akin to T2. As depicted in Figure S7, Q2 differences exhibited irregular fluctuations near zero, and the association between Q2 differences and latent heat flux from AC was found to be insignificant. Furthermore, the spatial patterns of HI differences closely mirrored those of T2 differences (Figures S8 and 6); however, the impact of AC usage on HI increase was found to be more pronounced than on T2 increase (Figure 6).

Figure 6.

Figure 6

Open in a new tab

Linear association between T2 difference and HI difference in (A) Sce1and (B) Sce4 at 19:00 on 19th August.

In accordance with NOAA’s HI classifications of “safe” (HI < 80), “caution” (80 < HI < 90), and “extreme caution” (HI > 90), we categorized HI into these three classes and quantified the impact of AC use on the transformation of HI levels. As depicted in Figure 7, the diurnal profiles of 24 h AC (Sce1) and AC-cooled-fraction (Sce3) exhibit similarities, with the latter showing a lower transition percentage. The results highlight the hours of 3:00 and 19:00 as critical times when transitions to “caution” and “extreme caution” occur, respectively. While AC-schedule (Sce4) displays similar daytime patterns, notable differences emerge during nighttime. The decision to switch off AC in commercial areas significantly enhances heat comfort, particularly during nighttime hours. Furthermore, spatial representations of HI level transitions are presented in Figure S9. It is observed that 24 h AC (Sce1) extends the area of “extreme caution” by 28% (at 15:00) and 32% (at 18:00), AC-cooled-fraction (Sce3) expands this area by 23% (at 15:00) and 19% (at 18:00), while AC-schedule (Sce4) results in an 18% increase in the “extreme caution” area at both 15:00 and 18:00.

Figure 7.

Figure 7

Open in a new tab

HI level transition in 24 h of 19th August between baseline and (A) Sce1, (B) Sce3, and (C) Sce4.

3.4. AC Spatiotemporal Contribution on Indoor Temperature

The temporal patterns of indoor temperature differences between AC scenarios (Sce1, Sce3, and Sce4) and the baseline are depicted in Figure S10. Both 24 h AC and AC-cooled-fraction exhibit lower indoor temperatures than the AC schedule, the latter showing a distinctive double peak attributed to the AC working schedule. Additionally, Figure S11 illustrates the spatial patterns on the 19th of August, providing a direct representation of significant indoor temperature decreases occurring from 12:00 to 22:00, which coincides with the essential time span of AC operation. Further exploration of indoor temperature differences under the influence of 24 h AC (Sce1) in three urban land use types (Figure S12) reveals a consistent trend of substantial improvement in indoor thermal conditions from 12:00 to 22:00 due to AC usage. When comparing indoor temperatures between 24 h AC (Sce1) and AC-schedule (Sce4), it is noteworthy that a time lag exists between the activation of AC and the point at which indoor temperatures reach the target level (Figure 8). For instance, there is a 4 h delay between AC activation (18:00) and the achievement of a 25 °C indoor temperature (22:00) in residential areas, with a longer 6 h lag in commercial/industrial areas due to larger temperature differentials between indoor and outdoor environments.

Figure 8.

Figure 8

Open in a new tab

Indoor temperature difference between Sce1 and Sce4 in 24 h of 19th August in (A) low-density residential area, (B) high-density residential area, and (C) commercial/industrial area.

3.5. AC Spatiotemporal Contribution on Energy Demands

We computed the electricity consumption of AC in various scenarios (Figure S13) for a typical summer day, resulting in values of 335.56 × 106 kW (24 h-AC), 200.61 × 106 kW (AC-cooled-fraction), and 226.13 × 106 kW (AC-schedule). Remarkably, these values are consistent in magnitude with the estimations obtained from the Statistical Yearbook (792 × 106 kW). Our findings underscore that the daytime hours (10:00–20:00) represent a substantial period of heightened electricity demand, and adjusting the AC operating schedule effectively postpones the onset of peak electricity consumption periods by 2 h (Figure S13). Additionally, our analysis reveals that AC energy consumption is predominantly influenced by T2 rather than Q2 (Figure 9). Specifically, for every 1 °C increase in T2, the energy consumption increases by approximately 106 kW.

Figure 9.

Figure 9

Open in a new tab

Linear association between AC consumption and T2, Q2.

4. Discussion

4.1. Model Bias

Based on the model results, it effectively captures the overall diurnal climate patterns, albeit with a noticeable underestimation of daily maximum temperatures, as evident in Figure 2. Several factors may contribute to this discrepancy: (1) inadequate parametric scheme. The current parametric scheme may not be the optimal choice for describing a city characterized by complex terrain. Chongqing’s unique topography, nestled amidst mountainous terrain, coupled with scattered impervious land, presents a challenging environment for precise simulation. (2) Soil parameter limitations. The absence of soil observation data hampers our ability to refine soil parameters. The default soil parameters within the model could potentially overestimate actual soil thermal conductivity, particularly during the midday hours, where the impact might be more pronounced. (3) Crop-related parameters. It is worth noting that the station’s proximity to cropland introduces an additional layer of complexity. In mid-August, Chongqing typically witnesses crop harvesting activities, exposing bare land. This bare land has the propensity to absorb shortwave radiation, potentially elevating temperatures in the vicinity. Consequently, this could contribute to the observed higher temperatures at noon in the station data.

4.2. Heat Stress Increase from AC

The findings presented in Sections 3.2 and 3.3 demonstrate that various scenarios of AC use exacerbate outdoor heat stress. In the 24 h AC simulation, we observe a nighttime warming of approximately 2–2.5 °C in urban districts, along with a daytime warming of over 0.5 °C. Similar conclusions were proposed by other studies,7 however, our results showed a lower warming in daytime and higher warming in nighttime compared to simulation results of Beijing (1–1.5 °C in daytime and 1.5–2.4 °C in nighttime). This discrepancy can be attributed to variations in background climates and disparities in urban land use spatial distribution and morphology between Beijing and Chongqing. The consensus regarding AC’s stronger impact on evening urban temperatures stems from the restricted depth of the urban boundary layer.5 During the day, the warming effect of AC scenarios spreads uniformly throughout a deeper boundary layer, facilitated by enhanced vertical mixing. Conversely, at night, a stable atmospheric layer inhibits the dissipation of waste heat generated by AC systems into the higher atmosphere. Consequently, when AC is extensively used at night, heat accumulates near the surface, resulting in warming. Specifically, between 17:00 and 8:00, 19:00 emerges as the pivotal time when AC usage profoundly affects the outdoor environment. This phenomenon arises from the substantial temperature disparity between indoor and outdoor environments, leading to heightened AC power consumption and elevated waste heat emission, all while the boundary layer depth diminishes. Consequently, at 19:00, when AC is widely employed in urban areas, more than half of the regions categorized under cautionary HI transform into the extreme caution category. Furthermore, the impact of AC-induced warming varies across different land use types, with commercial areas experiencing an average temperature increase of 0.5 °C compared to other residential areas during nighttime AC usage. Consequently, the implementation of nighttime AC usage restrictions holds the potential to significantly enhance the overall outdoor heat comfort levels within the city (Figures 7 and 4D).

Regarding electricity failures, we observed that while indoor environments may not maintain a comfortable temperature during such events, the outdoor conditions experience a relief of approximately 2 °C compared to the conditions in the 24 h AC scenario. However, it is essential to note that outdoor heat stress experiences a sudden increase following the restoration of electricity. This is attributed to the accumulation of excess heat indoors over the period when cooling systems were unavailable. When the AC systems are reactivated, this accumulated heat must be expelled outdoors, resulting in a warming effect that persists for several hours.

Indoor temperatures remain relatively stable based on land use types. Consequently, the usage of AC effectively reduces indoor temperatures, with reductions of up to 18 °C achieved during periods of high outdoor temperatures (12:00–20:00) (Figure S12). The time required for indoor temperatures to reach the target temperature after the activation of AC systems is 6 h for commercial areas and 4 h for residential areas (Figure 8). This 2 h difference is a consequence of divergent AC operational times. During the daytime, when the temperature differential between indoor and outdoor environments is significant, the 6 h cooling period substantially lowers indoor temperatures in commercial buildings. Conversely, during nighttime, when the temperature differential is comparatively smaller, residential buildings achieve the target indoor temperature after 4 h of cooling.

4.3. Indoor and Outdoor Thermal Disparity

The trade-off intrinsic to AC use is vividly presented in our results, which demonstrate a notable disparity between indoor comfort and outdoor heat stress. From a spatial perspective, this trade-off manifests as thermal inequity: while indoor inhabitants enjoy cooler temperatures, outdoor temperatures rise, disproportionately impacting those who cannot retreat indoors, such as street vendors or delivery workers. From a temporal perspective, we devote particular attention to the temperature increase observed at 19:00–20:00 in densely built-up areas. This focus is 2-fold: first, because this time and location coincide with a significant deterioration of the thermal environment, largely attributable to AC usage (Figure 7); and second, because it corresponds to the period when individuals conclude their workday and congregate in these areas. To quantify this inequality spatially and temporally, we conducted a simulation identical to the 24 h AC scenario but increased indoor temperatures by 2 °C (27 °C for residential areas and 26 °C for commercial areas) (Sce5). The results demonstrated that the maximum increase in the average outdoor T2 was approximately 0.3 °C at 19:00, with increases of up to 1 °C observed in high-density built-up urban areas (Figure S14). Compared to the conditions providing indoor thermal comfort, the 1 °C rise in T2 has adverse economic and health effects on the outdoor thermal environment. Since there is a significant positive correlation between thermal comfort and labor productivity, human health in extreme hot weather has been observed in different regions around the world by previous studies.47,49−51 It is also noted that the physical health hazards of thermal deterioration are disproportionally related to demographic structure, which is more harmful to the elderly,52,53 people of lower socioeconomic status and education,53 and those with heat-related diseases.54−56 Despite the lack of detailed demographic data, the finding of disproportional heat stress increase between indoor and outdoor environments is pivotal, underscoring that modest indoor temperature increases can markedly reduce outdoor heat stress, thereby narrowing the thermal comfort gap across indoor and outdoor spaces.

4.4. Energy Consumption from AC

The findings from Section 3.5 indicate that the energy consumption of 24 h AC is the highest, followed by AC-schedule, with the lowest consumption observed in AC-cooled-fraction scenarios. An important insight emerges that a 2 °C increase in indoor temperature (Sce5) can yield energy savings of up to 12%. The spatial distribution of energy consumption closely aligns with land use patterns, consistent with previous studies.7 Altering the AC operating schedule results in a distinct diurnal energy consumption profile for the AC-schedule compared to the other two AC scenarios (Figure S13). Specifically, peak consumption occurs between 14:00 and 19:00, aligning AC-schedule (Sce4) more closely with electricity consumption observations reported in previous studies.6,7 However, during the period of highest AC demand (12:00–20:00), AC-schedule’s ability to manage peak electricity demand is limited. Therefore, additional management strategies should be considered to ensure a stable electricity supply during these critical hours.

Notably, we found that AC consumption is strongly related to T2 (Figure 9). The heat released by AC units contributes to a warming effect, potentially leading to increased AC energy consumption. This positive feedback loop has significant implications and should not be overlooked in urban climate projections, particularly in hot cities with widespread AC usage.5,57 Given the association between T2 and AC energy consumption, it is worth noting that a 2 °C temperature rise resulting from AC usage could lead to a 2 × 106 kW increase in energy consumption. This positive feedback effect will exacerbate inequality in the living environment. The additional investigation of energy consumption to achieve indoor thermal comfort targets will increase the proportion of people who cannot afford AC uses.

4.5. Implications

Numerous policy implications arise from the conclusions drawn in our study. First, urban residents are advised to avoid outdoor activities in areas with widespread AC usage during the nighttime, particularly bustling commercial zones and high-density residential areas. Additionally, following an electricity outage, individuals should remain indoors for up to 6 h after power is restored to mitigate the heightened outdoor heat exposure resulting from intensive AC usage. We advocate for a recalibration of indoor temperatures exceeding the standard 2 °C increase to optimize energy conservation while improving the outdoor thermal environment. For commercial entities, precooling strategies could be effective, considering the time lag for temperature equilibration. Policymakers must consider AC usage schedules that reflect occupation patterns across different zones to efficiently manage energy use and enhance thermal comfort. It has been demonstrated that deactivating AC in commercial zones can effectively regulate outdoor thermal comfort. Furthermore, efforts should be directed toward enhancing emergency response capabilities to ensure energy security and continuous public infrastructure services, particularly considering the limited effectiveness of modifying AC schedules in alleviating peak electricity consumption. In the context of developed first-tier or tourist cities, where land use schedules may not be fixed, strategies to reduce AC intensity or cooling fraction should be explored to manage energy consumption and enhance heat comfort. Moreover, improving ventilation in high-density buildings and enhancing AC efficiency are valuable steps toward reducing heat accumulation. These adaptations can reduce energy demand and temper the urban heat island effect, fostering a more thermally equitable urban landscape.

4.6. Uncertainty and Prospect

Several uncertainties should be acknowledged in this study. First, the absence of available electricity consumption data related to AC use hinders the verification of the reliability of scenario simulation results. Although we have optimized the simulation’s performance by adopting parameters demonstrated in previous studies, it is important to note that the study’s primary objective is to compare spatiotemporal differences among scenarios, enhancing our understanding of AC usage impact patterns rather than providing precise quantitative values. Future research endeavors can achieve more accurate AC and anthropogenic heat simulations with access to electricity consumption data. Second, the scenarios employed in this study may not entirely mirror reality. For instance, the AC usage time schedule may not precisely align with that of the AC-schedule scenario (Sce4). Due to challenges in data acquisition, we did not account for the varying thermal properties of different building categories, such as thermal conductivity, thermal efficiency of heat exchangers, and heat capacity, among others. These properties are believed to correlate with factors such as building age and construction materials.43,46 To affirm the robustness of our findings, we modified the heat capacity values for roofs, buildings, and the ground (AKSR, AKSB, and AKSG) and thermal efficiency of the heat exchangers to evaluate the reliability of our results. The results in Figures S15 and S16 demonstrated that the spatiotemporal patterns of modifying the heat capacity and thermal efficiency of the heat exchangers are robust. Despite these simplifications, our study aims to shed light on spatiotemporal patterns through these simplified simulations, raising awareness of this issue and laying the groundwork for further investigations. Lastly, this study did not conduct a detailed calculation or discussion of heat exposure due to a lack of demographic data. Details of population distribution and structural data such as age, occupation, education level, etc. are essential to further understanding equity in heat exposure. Further research efforts could estimate heat-related socioeconomic benefits by association analysis between heat comfort and working efficiency and human health (hospitalization and mortality).50,51,58

Acknowledgments

This study was supported by the Open Research Program of the International Research Center of Big Data for Sustainable Development Goals (CBAS2022ORP02), the National Key Research and Development Program of China (2022YFB3903703 and 2022YFE0209300), the National Natural Science Foundation of China (grant number 42090015, 72091514), the University of Hong Kong Seed Fund for Strategic Interdisciplinary Research Scheme, and the University Grants Committee (UGC) Collaborative Research Fund (CRF) Young Collaborative Research Grant (C2002-22Y).

Data Availability Statement

All codes and models are available on the WRF tutorial Web site (https://www.mmm.ucar.edu/models/wrf). The land cover data of MODIS, GAIA is downloaded from the Google Earth Engine (GEE) data platform (https://developers.google.com/earth-engine/datasets/). GISA is downloaded from the authors’ Web site (http://irsip.whu.edu.cn/resources/resources_en_v2.php), and EULUC is downloaded from the Web site (https://data-starcloud.pcl.ac.cn/zh). The station observation data is found on the National Climatic Data Center (NCDC) (https://www.ncei.noaa.gov/cdo-web/).

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.est.4c00424.

  • EULUCclassification system and WRF land use classification system; urban morphology parameters of different land use; impervious land cover in GAIA and GISA; additional model parameters setting; spatiotemporal patterns of 2 m air temperature, HI, indoor temperature, and air conditioner consumption of different scenarios; and robustness test of models’ parameters (PDF)

The authors declare no competing financial interest.

Supplementary Material

es4c00424_si_001.pdf (3.1MB, pdf)

References

  1. Thompson V.; Kennedy-Asser A. T.; Vosper E.; Lo Y. T. E.; Huntingford C.; Andrews O.; Collins M.; Hegerl G. C.; Mitchell D. The 2021 western North America heat wave among the most extreme events ever recorded globally. Sci. Adv 2022, 8 (18), eabm6860 10.1126/sciadv.abm6860. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Callahan C. W.; Mankin J. S. Globally unequal effect of extreme heat on economic growth. Sci. Adv. 2022, 8 (43), eadd3726 10.1126/sciadv.add3726. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Shindell D.; Zhang Y.; Scott M.; Ru M.; Stark K.; Ebi K. L. The effects of heat exposure on human mortality throughout the United States. GeoHealth 2020, 4 (4), e2019GH000234 10.1029/2019gh000234. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Li Y.; Li C.; Luo S.; He J.; Cheng Y.; Jin Y. Impacts of extremely high temperature and heatwave on heatstroke in Chongqing, China. Environ. Sci. Pollut. Res. 2017, 24, 8534–8540. 10.1007/s11356-017-8457-z. [DOI] [PubMed] [Google Scholar]
  5. Salamanca F.; Georgescu M.; Mahalov A.; Moustaoui M.; Wang M. Anthropogenic heating of the urban environment due to air conditioning. J. Geophys. Res. Atmos. 2014, 119 (10), 5949–5965. 10.1002/2013JD021225. [DOI] [Google Scholar]
  6. Salamanca F.; Georgescu M.; Mahalov A.; Moustaoui M.; Wang M.; Svoma B. M. Assessing summertime urban air conditioning consumption in a semiarid environment. Environ. Res. Lett. 2013, 8 (3), 034022. 10.1088/1748-9326/8/3/034022. [DOI] [Google Scholar]
  7. Xu X.; Chen F.; Shen S.; Miao S.; Barlage M.; Guo W.; Mahalov A. Using WRF-urban to assess summertime air conditioning electric loads and their impacts on urban weather in Beijing. J. Geophys. Res. Atmos. 2018, 123 (5), 2475–2490. 10.1002/2017JD028168. [DOI] [Google Scholar]
  8. Kikegawa Y.; Genchi Y.; Yoshikado H.; Kondo H. Development of a numerical simulation system toward comprehensive assessments of urban warming countermeasures including their impacts upon the urban buildings’ energy-demands. Appl. Energy 2003, 76 (4), 449–466. 10.1016/S0306-2619(03)00009-6. [DOI] [Google Scholar]
  9. Ohashi Y.; Genchi Y.; Kondo H.; Kikegawa Y.; Yoshikado H.; Hirano Y. Influence of air-conditioning waste heat on air temperature in Tokyo during summer: Numerical experiments using an urban canopy model coupled with a building energy model. J. Appl. Meteorol. Clim. 2007, 46 (1), 66–81. 10.1175/JAM2441.1. [DOI] [Google Scholar]
  10. Liu K.; Du Y.; Chen W.; Wu X. Simulation of interaction between high-temperature process and heat emission from electricity system in summer. Glob. Energy Interconnect. 2022, 5 (6), 692–702. 10.1016/j.gloei.2022.12.010. [DOI] [Google Scholar]
  11. Pan J.; Tang J.; Caniza M.; Heraud J.; Koay E.; Lee H. K.; Lee C. K.; Li Y.; Nava Ruiz A.; Santillan-Salas C. F.; et al. Correlating indoor and outdoor temperature and humidity in a sample of buildings in tropical climates. Indoor air 2021, 31 (6), 2281–2295. 10.1111/ina.12876. [DOI] [PubMed] [Google Scholar]
  12. Tamerius J.; Perzanowski M. S.; Acosta L. M.; Jacobson J. S.; Goldstein I. F.; Quinn J. W.; Rundle A. G.; Shaman J. Socioeconomic and outdoor meteorological determinants of indoor temperature and humidity in New York City dwellings. Weather Clim Soc 2013, 5 (2), 168–179. 10.1175/WCAS-D-12-00030.1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Nguyen J. L.; Schwartz J.; Dockery D. W. The relationship between indoor and outdoor temperature, apparent temperature, relative humidity, and absolute humidity. Indoor air 2014, 24 (1), 103–112. 10.1111/ina.12052. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Occupational Safety and Health Administration . Worker’s rights. Available from: https://www.osha.gov/heat-exposure/case-studies (accessed 2024-5-8).
  15. Quinn A.; Tamerius J. D.; Perzanowski M.; Jacobson J. S.; Goldstein I.; Acosta L.; Shaman J. Predicting indoor heat exposure risk during extreme heat events. Sci. Total Environ. 2014, 490, 686–693. 10.1016/j.scitotenv.2014.05.039. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Zuurbier M.; van Loenhout J. A. F.; le Grand A.; Greven F.; Duijm F.; Hoek G. Street temperature and building characteristics as determinants of indoor heat exposure. Sci. Total Environ. 2021, 766, 144376. 10.1016/j.scitotenv.2020.144376. [DOI] [PubMed] [Google Scholar]
  17. Wu S.; Yu W.; Chen B. Observed inequality in thermal comfort exposure and its multifaceted associations with greenspace in United States cities. Landsc. Urban Plan. 2023, 233, 104701. 10.1016/j.landurbplan.2023.104701. [DOI] [Google Scholar]
  18. Zhou W.; Troy A.; Morgan Grove J.; Jenkins J. C. Can money buy green? Demographic and socioeconomic predictors of lawn-care expenditures and lawn greenness in urban residential areas. Soc. Nat. Resour. 2009, 22 (8), 744–760. 10.1080/08941920802074330. [DOI] [Google Scholar]
  19. Zhou W.; Huang G.; Pickett S. T.; Wang J.; Cadenasso M.; McPhearson T.; Grove J. M.; Wang J. Urban tree canopy has greater cooling effects in socially vulnerable communities in the US. One Earth 2021, 4 (12), 1764–1775. 10.1016/j.oneear.2021.11.010. [DOI] [Google Scholar]
  20. Dong C.; Yan Y.; Guo J.; Lin K.; Chen X.; Okin G. S.; Gillespie T. W.; Dialesandro J.; MacDonald G. M. Drought-vulnerable vegetation increases exposure of disadvantaged populations to heatwaves under global warming: A case study from Los Angeles. Sustain. Cities Soc 2023, 93, 104488. 10.1016/j.scs.2023.104488. [DOI] [Google Scholar]
  21. Huang K.; Lee X.; Stone B.; Knievel J.; Bell M. L.; Seto K. C. Persistent increases in nighttime heat stress from urban expansion despite heat island mitigation. J. Geophys. Res.: Atmos. 2021, 126 (4), e2020JD033831 10.1029/2020jd033831. [DOI] [Google Scholar]
  22. Ciancio V.; Falasca S.; Golasi I.; Curci G.; Coppi M.; Salata F. Influence of input climatic data on simulations of annual energy needs of a building: EnergyPlus and WRF modeling for a case study in Rome (Italy). Energies 2018, 11 (10), 2835. 10.3390/en11102835. [DOI] [Google Scholar]
  23. Pappaccogli G.; Giovannini L.; Zardi D.; Martilli A. Assessing the Ability of WRF-BEP+ BEM in Reproducing the Wintertime Building Energy Consumption of an Italian Alpine City. J. Geophys. Res.: Atmos. 2021, 126 (8), e2020JD033652 10.1029/2020jd033652. [DOI] [Google Scholar]
  24. US Department of Energy . US Energy Sector Vulnerabilities to Climate Change and Extreme Weather; US Department of Energy: Washington DC, USA, 2013. [Google Scholar]
  25. Sailor D. J.; Pavlova A. Air conditioning market saturation and long-term response of residential cooling energy demand to climate change. Energy 2003, 28 (9), 941–951. 10.1016/S0360-5442(03)00033-1. [DOI] [Google Scholar]
  26. Stone B.; Mallen E.; Rajput M.; Gronlund C. J.; Broadbent A. M.; Krayenhoff E. S.; Augenbroe G.; O’Neill M. S.; Georgescu M. Compound climate and infrastructure events: how electrical grid failure alters heat wave risk. Environ. Sci. Technol. 2021, 55 (10), 6957–6964. 10.1021/acs.est.1c00024. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Wang Y.; Li G.. Analysis of “furnace cities″ in China using MODIS/LST product (MOD11A2). 2014 IEEE Geoscience and Remote Sensing Symposium; IEEE. 2014.
  28. Salamanca F.; Martilli A. A new building energy model coupled with an urban canopy parameterization for urban climate simulations—Part II. Validation with one dimension off-line simulations. Theor. Appl. Climatol. 2010, 99, 345–356. 10.1007/s00704-009-0143-8. [DOI] [Google Scholar]
  29. Chen S.-H.; Sun W.-Y. A one-dimensional time dependent cloud model. JMSJ Ser.II. 2002, 80 (1), 99–118. 10.2151/jmsj.80.99. [DOI] [Google Scholar]
  30. Hong S.-Y.; Noh Y.; Dudhia J. A new vertical diffusion package with an explicit treatment of entrainment processes. Mon. Weather Rev. 2006, 134 (9), 2318–2341. 10.1175/MWR3199.1. [DOI] [Google Scholar]
  31. Grell G. A. Prognostic evaluation of assumptions used by cumulus parameterizations. Mon. Weather Rev. 1993, 121 (3), 764–787. . [DOI] [Google Scholar]
  32. Grell G. A.; Dévényi D. A generalized approach to parameterizing convection combining ensemble and data assimilation techniques. Geophys. Res. Lett. 2002, 29 (14), 38-1–38-4. 10.1029/2002gl015311. [DOI] [Google Scholar]
  33. Iacono M. J.; Delamere J. S.; Mlawer E. J.; Shephard M. W.; Clough S. A.; Collins W. D. Radiative forcing by long-lived greenhouse gases: Calculations with the AER radiative transfer models. J. Geophys. Res.: Atmos. 2008, 113, D13. 10.1029/2008jd009944. [DOI] [Google Scholar]
  34. Mlawer E. J.; Taubman S. J.; Brown P. D.; Iacono M. J.; Clough S. A. Radiative transfer for inhomogeneous atmospheres: RRTM, a validated correlated-k model for the longwave. J. Geophys. Res.: Atmos. 1997, 102 (D14), 16663–16682. 10.1029/97JD00237. [DOI] [Google Scholar]
  35. Yang Z. L.; Niu G. Y.; Mitchell K. E.; Chen F.; Ek M. B.; Barlage M.; Longuevergne L.; Manning K.; Niyogi D.; Tewari M.; et al. The community Noah land surface model with multiparameterization options (Noah-MP): 2. Evaluation over global river basins. J. Geophys. Res.: Atmos. 2011, 116 (D12), D12110. 10.1029/2010jd015140. [DOI] [Google Scholar]
  36. Jiménez P. A.; Dudhia J.; González-Rouco J. F.; Navarro J.; Montávez J. P.; García-Bustamante E. A revised scheme for the WRF surface layer formulation. Mon. Weather Rev. 2012, 140 (3), 898–918. 10.1175/MWR-D-11-00056.1. [DOI] [Google Scholar]
  37. Izquierdo M.; Moreno-Rodríguez A.; González-Gil A.; García-Hernando N. Air conditioning in the region of Madrid, Spain: An approach to electricity consumption, economics and CO2 emissions. Energy 2011, 36 (3), 1630–1639. 10.1016/j.energy.2010.12.068. [DOI] [Google Scholar]
  38. Huang X.; Yang J.; Wang W.; Liu Z. Mapping 10 m global impervious surface area (GISA-10m) using multi-source geospatial data. Earth Syst. Sci. Data 2022, 14 (8), 3649–3672. 10.5194/essd-14-3649-2022. [DOI] [Google Scholar]
  39. Gong P.; Li X.; Wang J.; Bai Y.; Chen B.; Hu T.; Liu X.; Xu B.; Yang J.; Zhang W.; et al. Annual maps of global artificial impervious area (GAIA) between 1985 and 2018. Remote Sens. Environ 2020, 236, 111510. 10.1016/j.rse.2019.111510. [DOI] [Google Scholar]
  40. Wang H.; Liu Z.; Zhang Y.; Yu Z.; Chen C. Impact of different urban canopy models on air quality simulation in Chengdu, southwestern China. Atmos. Environ. 2021, 267, 118775. 10.1016/j.atmosenv.2021.118775. [DOI] [Google Scholar]
  41. Huang H.; Chen P.; Xu X.; Liu C.; Wang J.; Liu C.; Clinton N.; Gong P. Estimating building height in China from ALOS AW3D30. ISPRS J. Photogramm. Remote Sens 2022, 185, 146–157. 10.1016/j.isprsjprs.2022.01.022. [DOI] [Google Scholar]
  42. Adilkhanova I.; Santamouris M.; Yun G. Y. Coupling urban climate modeling and city-scale building energy simulations with the statistical analysis: Climate and energy implications of high albedo materials in Seoul. Energy Build 2023, 290, 113092. 10.1016/j.enbuild.2023.113092. [DOI] [Google Scholar]
  43. Adilkhanova I.; Santamouris M.; Yun G. Y. Green roofs save energy in cities and fight regional climate change. Nat. Cities. 2024, 1, 238–249. 10.1038/s44284-024-00035-7. [DOI] [Google Scholar]
  44. Matsui T.; Zhang S. Q.; Lang S. E.; Tao W. K.; Ichoku C.; Peters-Lidard C. D. Impact of radiation frequency, precipitation radiative forcing, and radiation column aggregation on convection-permitting West African monsoon simulations. Clim. Dyn. 2020, 55, 193–213. 10.1007/s00382-018-4187-2. [DOI] [Google Scholar]
  45. Janjić Z. I. The step-mountain eta coordinate model: Further developments of the convection, viscous sublayer, and turbulence closure schemes. Mon. Weather Rev. 1994, 122 (5), 927–945. . [DOI] [Google Scholar]
  46. Aksoezen M.; Daniel M.; Hassler U.; Kohler N. Building age as an indicator for energy consumption. Energy Build 2015, 87, 74–86. 10.1016/j.enbuild.2014.10.074. [DOI] [Google Scholar]
  47. Anderson G. B.; Bell M. L.; Peng R. D. Methods to calculate the heat index as an exposure metric in environmental health research. Environ. Health Perspect. 2013, 121 (10), 1111–1119. 10.1289/ehp.1206273. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Lee D.; Brenner T. Perceived temperature in the course of climate change: an analysis of global heat index from 1979 to 2013. Earth Syst. Sci. Data 2015, 7 (2), 193–202. 10.5194/essd-7-193-2015. [DOI] [Google Scholar]
  49. Kim H.; Ha J.-S.; Park J. High temperature, heat index, and mortality in 6 major cities in South Korea. Arch. Environ. Occup. Health 2006, 61 (6), 265–270. 10.3200/AEOH.61.6.265-270. [DOI] [PubMed] [Google Scholar]
  50. Reid C. E.; Mann J. K.; Alfasso R.; English P. B.; King G. C.; Lincoln R. A.; Margolis H. G.; Rubado D. J.; Sabato J. E.; West N. L.; et al. Evaluation of a heat vulnerability index on abnormally hot days: an environmental public health tracking study. Environ. Health Perspect. 2012, 120 (5), 715–720. 10.1289/ehp.1103766. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Zhao M.; Lee J. K. W.; Kjellstrom T.; Cai W. Assessment of the economic impact of heat-related labor productivity loss: a systematic review. Clim. Change 2021, 167, 22. 10.1007/s10584-021-03160-7. [DOI] [Google Scholar]
  52. Fouillet A.; Rey G.; Laurent F.; Pavillon G.; Bellec S.; Guihenneuc-Jouyaux C.; Clavel J.; Jougla E.; Hémon D. Excess mortality related to the August 2003 heat wave in France. Int. Arch. Occup. Environ. Health 2006, 80, 16–24. 10.1007/s00420-006-0089-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Medina-Ramón M.; Zanobetti A.; Cavanagh D. P.; Schwartz J. Extreme temperatures and mortality: assessing effect modification by personal characteristics and specific cause of death in a multi-city case-only analysis. Environ. Health Perspect. 2006, 114 (9), 1331–1336. 10.1289/ehp.9074. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Schwartz J. Who is sensitive to extremes of temperature? A case-only analysis. Epidemiology 2005, 16, 67–72. 10.1097/01.ede.0000147114.25957.71. [DOI] [PubMed] [Google Scholar]
  55. Semenza J. C.; et al. Excess hospital admissions during the July 1995 heat wave in Chicago. Am. J. Prev. Med. 1999, 16 (4), 269–277. 10.1016/S0749-3797(99)00025-2. [DOI] [PubMed] [Google Scholar]
  56. Stafoggia M.; Forastiere F.; Agostini D.; Biggeri A.; Bisanti L.; Cadum E.; Caranci N.; de’Donato F.; De Lisio S.; De Maria M.; et al. Vulnerability to heat-related mortality: a multicity, population-based, case-crossover analysis. Epidemiology 2006, 17, 315–323. 10.1097/01.ede.0000208477.36665.34. [DOI] [PubMed] [Google Scholar]
  57. Takane Y.; Kikegawa Y.; Hara M.; Grimmond C. S. B. Urban warming and future air-conditioning use in an Asian megacity: importance of positive feedback. npj Clim. Atmos. Sci. 2019, 2 (1), 39. 10.1038/s41612-019-0096-2. [DOI] [Google Scholar]
  58. Parsons L. A.; Masuda Y. J.; Kroeger T.; Shindell D.; Wolff N. H.; Spector J. T. Global labor loss due to humid heat exposure underestimated for outdoor workers. Environ. Res. Lett. 2022, 17 (1), 014050. 10.1088/1748-9326/ac3dae. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

es4c00424_si_001.pdf (3.1MB, pdf)

Data Availability Statement

All codes and models are available on the WRF tutorial Web site (https://www.mmm.ucar.edu/models/wrf). The land cover data of MODIS, GAIA is downloaded from the Google Earth Engine (GEE) data platform (https://developers.google.com/earth-engine/datasets/). GISA is downloaded from the authors’ Web site (http://irsip.whu.edu.cn/resources/resources_en_v2.php), and EULUC is downloaded from the Web site (https://data-starcloud.pcl.ac.cn/zh). The station observation data is found on the National Climatic Data Center (NCDC) (https://www.ncei.noaa.gov/cdo-web/).

Articles from Environmental Science & Technology are provided here courtesy of American Chemical Society

RESOURCES