Roles of Soil Moisture–Temperature Coupling in Three Types of Heatwaves over the Great Bay Area of China

1.   Introduction

In the context of global warming, heatwaves (HWs) have occurred frequently throughout the world (Scher and Messori, 2019; IPCC, 2023). Both observations and future projections indicate rapid increases in the intensity, frequency, and duration of HWs, exerting serious impacts on ecosystems, human health, and socioeconomics (Perkins-Kirkpatrick and Lewis, 2020; Van Oldenborgh et al., 2022). Extreme HWs are usually accompanied by significant and unusual changes in the soil, leading to a significant drop in crop yields (Vogel et al., 2019; Lesk et al., 2022). HWs are now recognized as the deadliest weather-related disaster, as they exacerbate human exposure to extreme overheating, resulting in a notable increase in mortality (He et al., 2021). During HWs, persistent extreme heat can also intensify the release of stored and anthropogenic heat, leading to severe economic and energy losses (Callahan and Mankin, 2022).

Recently, HWs have emerged as a critical research topic for both the general public and scientists. It is found that the HWs become more complex and extreme. In addition to the HWs traditionally defined based on daytime maximum temperatures, nighttime extreme high temperature events are also becoming more frequent (Su and Dong, 2019(指代不明); Zhang et al., 2020). Furthermore, there is a growing interest in the changes and impacts of “day–night” compound HWs (Chen and Zhai, 2017; Wang et al., 2021(指代不明)), e.g., on human health, energy use, and associated fire risks (Su and Dong, 2019(指代不明); Balch et al., 2022; Xie et al., 2022; Wang et al., 2023). Additionally, different types of HWs vary significantly in the associated large-scale atmospheric conditions and in possible mechanisms during their occurrence. Analyses in southern China and globally (Luo et al., 2022(指代不明)) have found that daytime HWs are often accompanied by enhanced solar shortwave radiation and clear, cloudless weather. In contrast, the atmosphere becomes more humid and clouds increase during nighttime HWs, while compound HWs are associated with the combined conditions of the above two types of HWs.

During HWs, the role of land surface is non-negligible (Chen et al., 2022). The land surface–atmosphere interactions promote the development of HWs and prolong high-temperature events through intensified heat exchanges (Fu and Wang, 2023; Domeisen et al., 2023). For instance, during a positive soil moisture–temperature feedback process, drier soil results in a decrease in latent heat flux and an increase in sensible heat flux that heats the atmosphere, contributing to the persistence of HWs. Moreover, the soil moisture feedback also affects large-scale atmospheric circulations, thus intensifying HWs (Miralles et al., 2014; Benson and Dirmeyer, 2021). Additionally, soil moisture has a significantly larger contribution to the global heatwave risk in humid areas, such as the eastern coastal regions of China (Wang et al., 2024). Soil moisture can sometimes have a greater impact than atmospheric circulation. For example, in the case of the extreme heatwave in eastern China in 2022, the monthly mean atmospheric circulation anomalies failed to explain the cause of the heatwave, while the soil moisture–temperature feedbacks explained part of the temperature anomalies (Jiang et al., 2023(指代不明)). Under the SSP585 high emissions scenario, by 2040–2070, soil moisture effects combined with global warming are projected to lengthen the duration of annual HWs, increase average temperatures, and increase the frequency of HWs in most midlatitude land areas (Zhou et al., 2024). There are also some studies that conducted simulations with and without soil moisture, demonstrating that the soil moisture–temperature coupling amplifies HWs (Horowitz et al., 2022; Cai et al., 2024). The characteristics of land surface substantially affect extreme high temperatures as well. Studies have found that the urbanization effect, the most typical type of anthropogenic land-use change, can increase the frequency of compound HWs to 2–5 times that of the current level (Wang et al., 2021(指代不明)). The interaction between urbanization and the urban heat island (UHI) effect can increase extreme heat events in cities, exposing residents to higher risks (Liao et al., 2018; Jiang et al., 2019).

The Greater Bay Area (GBA) in South China is one of the most economically developed and active regions in the world. The GBA has a complex geography, where urban agglomerations are surrounded by mountains, forests, farmlands, and rivers, and this area faces the sea to its southeast (Figs. 1a, b). In recent decades, the weather, climate and environment of the GBA have undergone remarkable changes due to global warming and high-frequency human activities, coupled with the specific geography in this region (Hui et al., 2020; Yang et al., 2020), among which the increase in temperature is the most noteworthy (Chao et al., 2020). Li et al. (2021) emphasized that the increasing nighttime temperature in the GBA leads to reduced diurnal temperature ranges, contributing to hotter days and nights. It has also been noted that the increasing trend of extreme heat events in the GBA in recent years is more significant than in other regions of South China, especially in the central GBA (Huang et al., 2023). Moreover, the changes in extreme heat events in the GBA demonstrate that the trends of hot days, hot nights, and HW durations have all increased significantly from 1961 to 2019; and by 2030, the duration of exposure to hazardous discomforts for the Greater Bay Area residents will increase by at least 8.87% (Qing and Wang, 2021; Wang et al., 2021(指代不明)). As the global temperature rises by 1.5°C, the frequency, duration, and intensity of summer HWs in the urban agglomeration of the GBA are projected to respectively increase by 76%, 110%, and 130% compared with the historical period of 1996–2005 (Ma et al., 2022). In both relative and absolute HWs, the greatest increase in heat wave intensity/days is projected over the GBA, with the 2081–2100 increase being 1.5 times that of 2041–2060 (Xu and Zhang, 2022).

While existing research on extreme HWs in the GBA has primarily emphasized the role of the large-scale dynamical background, there is still limited exploration of how land–atmosphere interactions—particularly changes in soil moisture—might influence HWs within the complex underlying subsurface of GBA. There may also be differences in the characteristics and driving mechanisms of different types of HWs in the GBA. Besides, how do various land–atmosphere factors evolve during heatwave evolution remains unknown. All these issues need to be further explored, which will help deepen the understanding of the formation mechanisms and the evolution characteristics of HWs to provide references for predicting and responding to extreme heat events. Therefore, this study firstly classifies the summer HWs in GBA from 1961 to 2020 into daytime HWs, nighttime HWs, and compound HWs. The multi-dimensional spatiotemporal distribution characteristics of these three types of HWs are further analyzed. Subsequently, the characteristics and interactions of land–atmosphere factors during the three types of HW processes in the GBA are explored.

2.   Data and methods

2.1   Data

Daily minimum and maximum temperatures (${\mathrm{T}}_{min}$ and ${\mathrm{T}}_{max}$) from the CN05.1 gridded observational dataset with a horizontal resolution of 0.25°×0.25° are used to define HWs and calculate corresponding heat wave indices. This dataset, provided by the Climate Change Research Center of the Institute of Atmospheric Physics, Chinese Academy of Sciences and available through Wu and Gao (2013), has been widely applied in the studies on extreme temperature events in China (Luo et al., 2020).

Due to the lack of soil data in the CN05.1 dataset, ERA5 hourly reanalysis data are used to analyze the changes of land-atmosphere factors and calculate land-atmosphere coupling indices. The ERA5 reanalysis products provided by the (ECMWF are publicly available from the websites (Hersbach et al., 2023). The specific variables used are 2-m temperature, 2-m dew point temperature, surface pressure, and temperature and humidity of the soil in layer 1 (0–7cm) for the same period which has been used by other researchers to explore its relationship with extreme hot events (García-García et al., 2023; Lyu et al., 2024; Ni et al., 2024). The shallow soil moisture layer directly interacts with the surface energy balance through evaporation and soil moisture-temperature coupling, and its effects tend to be more prolonged and direct (Seneviratne et al., 2010). Compared with the Global Land Data Assimilation System 2.1, ERA5 provides more realistic soil moisture and temperature data in South China (Wu et al., 2021).

2.2   Definition of HWs

As relative thresholds can better reveal the local features of extreme events across different regions (Wang et al., 2020), they are utilized in this study to define the HWs over the GBA, so as to explore the possible driving mechanisms. The specific procedures are as followings. The relative threshold of ${\mathrm{T}}_{max}$ and ${\mathrm{T}}_{min}$ for a certain calendar day at a specific grid point is calculated as the 90th percentile of daily ${\mathrm{T}}_{max}$ and ${\mathrm{T}}_{min}$ (denoted as ${\mathrm{T}}_{max}^{90p}$ and ${\mathrm{T}}_{min}^{90p}$, respectively) in a window of 15 days centered at this day over the period from 1961 to 2020 (60 years ∈ 15 days) at this grid point. It is found that the ${\mathrm{T}}_{max}^{90p}$ are all above 33.5°C on all grids in the GBA, while the ${\mathrm{T}}_{min}^{90p}$ are above 27.0°C in urban areas (supplementary figure S1). On this basis, HWs are classified as daytime, nighttime and compound HWs (Chen and Li, 2017).

Daytime HWs: at least 3 consecutive hot days without any accompanying hot nights (i.e., T_max ≥ ${\mathrm{T}}_{max}^{90p}$and T_min < ${\mathrm{T}}_{min}^{90p}$).

Nighttime HWs: at least 3 consecutive hot nights without any accompanying hot days (i.e., T_max < ${\mathrm{T}}_{max}^{90p}$ and T_min ≥ ${\mathrm{T}}_{min}^{90p}$).

Compound HWs: at least 3 consecutive days with both hot days and hot nights (i.e., T_max ≥ ${\mathrm{T}}_{max}^{90p}$ and T_min ≥ ${\mathrm{T}}_{min}^{90p}$).

In this study, the following indices are employed to depict HWs (Russo et al., 2016).

Heat wave numbers (HWN): the number of HWs.

Heat wave intensity (HWI): the cumulative heat during a heat wave event, given by the sum of the anomalies of ${\mathrm{T}}_{\mathrm{m}\mathrm{a}\mathrm{x}}$ and ${\mathrm{T}}_{\mathrm{m}\mathrm{i}\mathrm{n}}$ exceeding corresponding thresholds during the whole process. Calculating the cumulative value allows for better comparison of heatwaves with varying durations (Russo and Domeisen, 2023).

2.3   Land–atmosphere coupling index

In this paper, the energy balance based on actual evaporation and potential evaporation is used to estimate the soil moisture-temperature coupling intensity, thus quantifying the interaction between land surface and atmosphere (Miralles et al., 2012). The calculation formula is as follows.

where $\mathrm{T}$ denotes the daily 2 m temperature, $\mathrm{H}$ denotes the surface sensible heat flux, ${\mathrm{R}}_{n}$ denotes the net surface radiation, $\mathrm{E}$ and ${\mathrm{E}}_{p}$ respectively denote the actual and potential evapotranspiration, $\mathrm{\lambda }$ denotes the latent heat of vaporization (with a value of 2.5∈10⁶ J/kg), and ${\mathrm{T}}'$, ${\mathrm{H}}'$ and ${\mathrm{H}}_{p}'$ are respectively expressed as their standardized variables. ${e}'$ is expressed by ${\mathrm{H}}'-{\mathrm{H}}_{p}'$ which denotes the energy term, representing the contribution of soil moisture deficit to $\mathrm{H}$. When the soil moisture is sufficient to meet atmospheric demand, ${e}'$ equals 0, while it is positive under dry conditions. The coupling intensity index ${\text{π}}$ is the product of ${\mathrm{T}}'$ and ${e}'$. The larger the value of ${\text{π}}$, the stronger the soil moisture-temperature coupling, and values of ${\text{π}}$ smaller than or equal to zero denote no coupling.

2.4   Analysis methods

The long-term trends in heatwave indices were assessed using the ordinary least squares (OLS) regression model. Additionally, composite analysis was conducted to investigate the changes in land-atmosphere factors and interactions during the occurrence and evolution of three types of HWs. The 60-year climatological average was subtracted to obtain the anomalies. We also used Student’s t-test to examine whether the changes in trends and the anomalies derived from the composite analysis were statistically significant.

3.   Results

3.1   Spatiotemporal distribution characteristics of HWs

Figure 1 illustrates that daytime HWs predominantly occur in the forested areas of the northern GBA, while nighttime HWs are mainly observed in the croplands and forests surrounding the urban areas of the GBA. Compound HWs are concentrated in highly urbanized areas with urban and built-up as the dominant land cover, and in the southern coastal regions. For Compound HWs, the intensity index consists of the cumulative heat of both ${\mathrm{T}}_{min}$ and ${\mathrm{T}}_{max}$ exceeding their respective thresholds, resulting in significantly higher intensities compared with the other two types of HWs, with an annual average HWI reaching up to 12.79°C.

Download

Larger image

PowerPoint slide

Fig  1.  (a–b) The location of the GBA and the characteristics of the land surface in the GBA. (c, e, g) Spatial distributions of annual average HWN (units: times) for three types of HWs in the GBA during the summers from 1961 to 2020, namely the daytime HWs (left-hand panels), nighttime HWs (middle panels) and compound HWs (right-hand panels), based on the actual occurrence year. Panels (d, f, h) are the same as panels (c, e, g), but for HWI (units:°C) for three types of HWs.

Figure 2 offers the spatial distribution of annual linear trends in HWN and HWI for three types of HWs, displaying only areas where anomalies are significant at the 0.05 level. The results show that, compared to the other two types of heatwaves, the changes of trends for daytime HWs are smaller, with increases in frequency and intensity mainly concentrated in the northern forested areas. Nighttime HWs exhibit a more obvious increase in occurrence in the whole region of the GBA, though their intensity trends are less marked. Compound HWs exhibit the most significant changes, with a notable increase in both frequency and intensity in the central GBA. The trend in HWI for compound HWs can reach up to 0.29°C year⁻¹ (figure 2f), suggesting a close linkage between compound HWs and urbanization.

Download

Larger image

PowerPoint slide

Fig  2.  The spatial distributions of the annual linear trends of HWN (units: times year⁻¹) for three types of HWs in the GBA during the summers from 1961 to 2020, namely the daytime HWs (left-hand panels), nighttime HWs (middle panels) and compound HWs (right-hand panels). Panels (d, e, f) are the same as panels (a, b, c), but for HWI (units:°C year⁻¹) for three types of HWs.

Globally, the middle and high latitude regions of the Northern Hemisphere exhibit more intense compound HWs, with most of the increase in compound HWs occurring in the second half of 1983-2012 (Jiang et al., 2023(指代不明)). Significant changes in daytime HWs and nighttime HWs are primarily concentrated in northeastern and central Asia (Wu et al., 2023). Both globally and regionally, these conclusions are similar to those we obtained. Chen et al. (2017) (无参考文献) found that most compound HWs occur in eastern China, especially along the southern coastal regions, which are economically developed and densely populated, where the frequency of increase is the highest in China. Recent studies have shown that urbanization contributes 36-58% to compound HWs in the Yangtze River Delta and Pearl River Delta, while its contribution in the Beijing-Tianjin-Hebei region is relatively smaller, ranging from 16-29% (Lin et al., 2024).

3.2   Characteristics of land-atmosphere factors during HWs

3.2.1   Statistical characteristics of land-atmosphere factors

Strong soil moisture-atmosphere coupling plays a critical role during HWs. To verify this effect in the GBA, this section firstly calculates the correlations of the temperature and moisture of soil in the layer (0–7 cm) with the heat wave indices during the same period. The results indicate a strong correlation between extreme high-temperature events and surface soil temperature or moisture in the GBA, with all correlation coefficients statistically significant at the 0.05 significance level.

To intuitively illustrate the soil and atmospheric characteristics during heat wave events, the land-atmosphere factors during each event of the three types on each grid of the GBA during 60 years are statistically analyzed. Results show that nighttime HWs are the most frequent, with the number of frequencies reaching 3,062, followed by daytime HWs (2,681) and compound HWs (2,127) from 1961 to 2020.

Histograms and kernel density estimation curves at the edges (figure 3a) reveal that 2 m temperature during HWs ranges between 25.46°C and 33.28°C in the GBA. Moreover, the average temperatures during daytime HWs and compound HWs are higher than those during nighttime HWs, consistent with the classification criteria for HWs. The near-surface specific humidity during nighttime HWs ranges from 16.84 g kg⁻¹ to 23.05 g kg⁻¹, while it ranges from 17.12 g kg⁻¹ to 23.07 g kg⁻¹ during compound HWs. 73% of nighttime HWs and 55% of compound HWs occur with higher specific humidity than the climatological mean in summer (20.35 g kg⁻¹). In contrast, daytime HWs tend to occur under drier conditions, with 74% events accompanying with specific humidity lower than the climatological mean in summer.

Figure 3(b) displays that the surface soil temperature during three types of HWs in the GBA all ranges from 23.73°C to 34.63°C. Besides, most HWs occur with surface soil moisture ranging from 1.01 m³ m⁻³ to 1.50 m³ m⁻³, while 74% of HWs in the GBA occur under drier conditions than the climatology mean. However, nighttime HWs tend to occur under wetter soil conditions than the other two types, with 45% of these events occurring when soil moisture exceeds the climatology mean.

Download

Larger image

PowerPoint slide

Fig  3.  (a) The joint distribution of 2-m temperature versus 2-m specific humidity during each heat wave event in the GBA from 1961 to 2020, with scatter plots at the center and kernel density estimation curves at the edges (y-axis at the edges in units of numbers). Panel (b) is the same as Panel (a), but for the joint distribution of surface soil temperature versus soil moisture, the gray dotted lines represent the climatological means, and the numbers in the upper right corner denote the total number of events for three types of HWs, respectively. (c) Specific locations of the three representative regions in the GBA. Panels (d, e) are the same as Panels (a, b), but for representative regions.

Based on the conclusions in section 3.1, three representative regions with highly-frequent HWs are selected, where the ratio of HWN in representative regions to the total HWN in the GBA exceeds 65% for all three types (figure 3c). The representative region for daytime HWs is denoted as AREA1, located in the northern part of the GBA. Nighttime HWs are more frequent around the periphery of the GBA, with the northwest, northeast, and southwest regions denoted as AREA2. Compound HWs are most frequent in the central and southwestern coastal regions of the GBA, denoted as AREA3. Statistical analysis is also performed in the representative areas and similar conclusions are obtained. Daytime HWs occur under the conditions of low surface specific humidity and low surface soil moisture, indicating dry atmosphere and soil conditions. Nighttime HWs occur under relatively wetter atmosphere and soil conditions compared to other types of HWs. Compound HWs exhibit comprehensive characteristics, which undergo the conditions of the other two types. What roles do land and atmosphere play during these HWs, and what changes have taken place?

3.2.2   Spatial distribution of land–air factors

Since the classification for HWs in the GBA region take into account the temperature relative to the thresholds at daytime and nighttime, the characteristics of each land-atmosphere factor during HWs are also discussed for both daytime (0600 UTC, 1400 BST) and nighttime (1800 UTC, 0200 BST). The hourly data have been detrended to remove the effect of long-term climate change on the final results, highlighting the spatial distribution of land-atmosphere anomalies of three types of HWs. The spatial distributions of 2-m temperature, near-surface specific humidity, as well as soil temperature and soil moisture are generally consistent with the statistical characteristics discussed in Section 3.2.1. Here, we investigate changes in cloud cover, wind speed and surface flux during heat waves.

For total cloud cover (figures 4a–f), daytime HWs are associated with an overall negative cloud cover anomaly, while nighttime HWs show a positive anomaly in cloud cover. At daytime, Compound HWs result in a negative cloud cover anomaly, though some regions show positive anomalies at nighttime

Considering both wind speed (figures 4g–l) and specific humidity (supplementary figure S2), at daytime, the wind speed is negatively anomalous, dominated by anomalous northwesterly winds, and near-surface specific humidity showing a negative anomaly during daytime HWs. For nighttime HWs, there is an anomalous strengthening of southwesterly winds, which carries the water vapor, leading to a positive specific humidity anomaly in GBA. During compound HWs, anomalous northwesterly winds and a negative wind speed anomaly occur in the center of the GBA, accompanied by an abnormally low specific humidity.

At nighttime, the wind speed of all three types of HWs is generally lower than at daytime. Daytime HWs are dominated by anomalously weak northwesterly winds, while nighttime HWs still show a relatively wet atmospheric condition with strengthened southwesterly winds, which bring water vapor and contribute to maintaining high temperature and humidity during the night. For compound HWs, anomalous westerly winds are dominant, especially in the southwestern GBA, where wind speeds are significantly enhanced.

Download

Larger image

PowerPoint slide

Fig  4.  The average total cloud cover (in %) ,10-m wind speed (in m s⁻¹) during daytime HWs (left-hand panels), nighttime HWs (middle panels), and compound HWs (right-hand panels) in the GBA from 1961 to 2020 are presented. The composite anomaly spatial distributions at the daytime (0600 UTC, 14:00 BJS; panels a-c, g-i, m-o) and at nighttime (1800 UTC, 02:00 BJS; panels d-f, j-l, p-r) are shown, with the dotted regions and solid red lines indicating anomalies significant at the 0.05 level.

The land interacts with the atmosphere through energy exchanges, making it necessary to analyze the surface latent heat flux and sensible heat flux. At daytime, for all three types of HWs, the shortwave solar radiation heats the atmosphere, leading to an abnormal increase in temperature and enhanced evapotranspiration, corresponding to an anomalous increase in the latent heat flux directed from the surface to the atmosphere (figures 5a-c). The maximum latent heat flux during daytime HWs can reach 126.92 W/m². Due to partial blocking of shortwave radiation by clouds, the latent heat flux is reduced during nighttime HWs. At nighttime (figures 5d-f), without shortwave radiation, energy fluxes decrease accordingly.

At daytime, all three types of HWs still show an anomalous increase in the sensible heat flux (figures 5g-i), which heats the atmosphere and contributes to extreme high temperatures. The largest positive sensible heat flux anomalies occur over farmland and forests. At nighttime, the atmosphere transfers more sensible heat flux to the land because the GBA has been in a state of high atmospheric temperature relative to land surface temperature during the summers of 1961-2020. As a result, the climatic state of nighttime sensible heat flux shows negative values in GBA. During nighttime HWs and compound HWs, a negative sensible heat flux anomaly occurs, indicating an abnormal increase in the heat flux from the atmosphere to the surface, which amplifies nighttime high temperatures.

Download

Larger image

PowerPoint slide

Fig  5.  The average latent heat flux (in W m⁻²) and sensible heat flux (in W m⁻²) during daytime HWs (left-hand panels), nighttime HWs (middle panels), and compound HWs (right-hand panels) in the GBA from 1961 to 2020 are presented. The composite anomaly spatial distributions at the daytime (0600 UTC, 14:00 BJS; panels a-c, g-i) and at nighttime (1800 UTC, 02:00 BJS; panels d-f, j-l) are shown, with positive values indicating fluxes directed from the surface to the atmosphere. Dotted regions and solid red lines indicate anomalies significant at the 0.05 level.

3.3   Characteristics of land-atmosphere factors and local land-atmosphere interactions during heatwave evolution

Subsequently, the changes in land-atmosphere factors during HWs in the GBA are analyzed. For the selection of the period before and after the occurrence of HWs, we refer to the work of Xie and Zhou (2023), and based on the previous work indicated that the average duration of HWs is 4.56 days. The analysis is based on all heatwave events that occurred in the GBA region at each grid point during the 60-year period from 1961 to 2020.Therefore, we mark the first day of a heat wave event as day 0 and take the period from 5 days before to 5 days after day 0 (denoted as ±5) to explore heat wave evolution features in this study. Then, using the method of composite analysis, we identify the meteorological characteristics of specific days during the evolution of different types of heat waves, and subtract the 60-year climatological average for the corresponding day to obtain composite anomalies.

The results demonstrate that as HWs develop, abnormal increases are observed in both 2-m temperature and surface soil temperature, accompanied by negative anomalies of surface soil moisture (figure 6). For the daily mean temperature, nighttime HWs, compound HWs and daytime HWs peak at day 0, day +1, and day +2, respectively. Compound HWs have the largest positive anomaly of up to +2.6°C. Besides, the evolution in 2-m temperature is generally consistent with that of surface soil temperature. Soil moisture exhibits opposite patterns to soil temperature. In both daytime HWs and compound HWs, soil moisture decreases gradually during the HWs, and significantly rebounds after +2 days. The soil moisture during nighttime HWs gradually dries as the event develops. However, the negative anomaly remains relatively stable in the later stages (figure 6b), contrasting with the other two types of HWs, where soil moisture decreases significantly again in the later stages. Therefore, compared with the other two types, the surface soil during nighttime HWs is relatively wetter.

From the perspective of near-surface specific humidity, nighttime HWs maintain a positive specific humidity anomaly throughout the heat wave, whereas compound HWs sustain a positive anomaly after day 0. Daytime HWs (figure 6a) experience consistently dry atmospheric conditions from day −5 to day +3, with the specific humidity reaching its maximum negative anomaly on day −1. Consequently, the average surface specific humidity is the lowest during daytime HWs.

Download

Larger image

PowerPoint slide

Fig  6.  Evolutions of composite anomalies in 2 m temperature (red line), 2 m specific humidity (blue line), surface soil temperature (yellow line) and surface soil moisture (green line) averaged over the GBA from day −5 to day +5 for (a) daytime HWs, (b) nighttime HWs and (c) compound HWs during 1961–2020, the circles denote the anomalies significant at the 0.05 level.

The land-atmosphere interactions during the evolutions of the three types of HWs are further quantified. As daytime HWs and compound HWs intensify, it is found that the larger the temperature and energy terms become, resulting in stronger soil moisture-temperature coupling compared with nighttime HWs. And the soil condition of compound HWs is relatively drier and soil moisture can significantly influence the sensible heat exchange by adjusting the Bowen ratio. Hence, the soil moisture-temperature coupling is stronger, the value of ${\text{π}}$ during compound HWs reaches a maximum of +2.78 (figure 7c).

However, the soil moisture-temperature coupling for nighttime HWs is rather weaker, fluctuating around zero (figure 7b). This may be due to the fact that the surface soil moisture for nighttime HWs is relatively higher than that for other HWs during their evolution. The sensible heat exchange between soil and near-surface atmosphere is primarily driven by temperature difference rather than soil moisture. Consequently, the value of coupling intensity index ${\text{π}}$ is close to zero. As a result, there is basically no soil moisture-temperature coupling.

Download

Larger image

PowerPoint slide

Fig  7.  Evolutions of composite standardized anomalies in temperature term (${\mathrm{T}}'$, red line), energy term (${\mathrm{H}}'-{\mathrm{H}}_{p}'$, blue shading), and coupling intensity index (${\text{π}}$, purple line) averaged over the GBA from day −5 to day +5 for (a) daytime HWs, (b) nighttime HWs and (c) compound HWs during 1961–2020.

4.   Conclusions and discussion

Our study used the CN05.1 gridded observational dataset and ERA5 reanalysis data to investigate three types of HWs in the GBA during the summers over the past 60 years. The spatial distribution characteristics of the three types of HWs in the GBA are presented, followed by a statistical analysis of the characteristics of land-atmosphere factors for each type. On this basis, representative regions with highly-frequent HWs of three types are respectively determined and spatially verified. Additionally, the changes in land-atmosphere factors and soil moisture-temperature coupling during the evolutions of three types of HWs are further analyzed. The main conclusions are as follows.

Both daytime and nighttime HWs predominantly occur in the areas with croplands and forests, with the former primarily in the forested regions of the northern GBA and the latter on the periphery of the GBA. In contrast, compound HWs are mainly found in the highly-urbanized central GBA and the southern coastal regions.

The statistical characteristics and spatiotemporal evolution of land–atmosphere factors during the three types of HWs reveal that the less (more) cloud cover as well as the low (high) atmospheric humidity affects the high (low) shortwave flux, which favor the warming of daytime (nighttime). Such a warming mechanism has also been found in previous studies (Luo and Lau, 2017). In the GBA, both the 2 m specific humidity and soil moisture are substantially lower during daytime HWs. While during nighttime HWs, the atmosphere and soil conditions are relatively wetter than those in the other two types of HWs. Compound HWs exhibit features that combine the characteristics of the other two types. The results suggest that there are significant differences in the influence of land-atmosphere factors on different types of HWs in different representative regions, further demonstrating the important role of these factors in affecting HWs.

These differences can be further understood by examining the specific mechanisms through which land–atmosphere interactions influence different types of HWs. Due to the high temperature and humidity conditions in the GBA, the soil is wet most of the time, dominated by atmospheric processes predominantly affecting the land. However, soil moisture can somewhat exacerbate the high temperatures of heat wave events in the GBA. For daytime HWs and compound HWs, the soil moisture exhibits larger negative anomalies compared with nighttime HWs (figures 6a and 6c). Drier soil conditions result in relatively lower actual evapotranspiration and thus increase the difference between actual and potential evapotranspiration. The soil moisture can also modify the Bowen ratio, thereby altering sensible heat fluxes and affecting the near-surface air temperature. Therefore, the soil moisture-temperature coupling is relatively stronger (figure 7a and 7c), far beyond the summer climate state, which can exacerbate and sustain the high temperatures of these two types of HWs (supplementary figure S3). However, in the later stages of nighttime HWs, soil moisture does not recover as quickly as it does during daytime or compound HWs, instead remaining in a negative anomaly. The anomalously high nighttime temperatures are primarily maintained by the greenhouse effect of increased water vapor. The relatively adequate soil moisture supply minimizes its impact on temperature changes, resulting in no soil moisture-temperature coupling (figures 6b and 7b).

We summarize the driving mechanisms of high temperatures during the day and night for the three types of HWs. As shown in figure 8, the daytime HWs in the GBA occurred with clear, cloudless skies, abnormally small northwesterly winds, and unusually high shortwave radiation during the daytime (figures 4a, 4g, 5a and 5g). The excessive shortwave radiation heats the surface, leading to an abnormal increase in sensible heat flux and latent heat flux which further causes a reduction in soil moisture. The soil conditions are relatively drier and soil moisture can significantly influence the sensible heat exchange by adjusting the Bowen ratio (figure 6a). Hence, A strong soil moisture-temperature coupling process further amplifies sensible heat flux release, resulting in persistent high daytime temperatures (figure 7a). However, at night, the lack of clouds allows the surface to emit longwave radiation freely, preventing sustained nighttime high temperatures.

For nighttime HWs, both atmospheric humidity and soil moisture are generally high during these events, accompanied by abnormally increased cloud cover. The clouds block a significant portion of incoming shortwave radiation, and the soil is relatively wetter compared to the other two types of HWs which makes there is almost no coupling between soil moisture and temperature (figure 7b). Under these conditions, the atmosphere and land are not sufficiently heated during the daytime, preventing daytime temperatures from exceeding the heatwave threshold (figures 4b,4h, 5b and 5h). At night, however, the high temperature is mainly driven by the greenhouse effect brought by water vapor and increasing downward longwave radiation, which maintains high and prolonged nighttime temperature.

Compound HWs occur with the combined conditions of the other two types. The reduced cloud cover allows for increased shortwave radiation to heat both the surface, leading to corresponding changes in latent and sensible heat (figures 4c, 4i, 5c and 5i). At the same time, a stronger coupling between soil moisture and temperature exists, the value of ${\text{π}}$ during compound HWs reaches a maximum of +2.78 (figure 7c). This positive coupling enhances the sensible heat flux contributes to the persistence of high daytime temperatures. Besides, the high temperatures during compound HWs are maintained not only by soil moisture but also by the greenhouse effect of water vapor (figure 6c). At night, higher humidity, as the main factor, enhances the insulating effect of water vapor, which further warms the atmosphere, leading to persistent high temperatures during the night.

Download

Larger image

PowerPoint slide

Fig  8.  Conceptual model diagram of the main physical mechanisms of daytime HWs, nighttime HWs and compound HWs.

This study primarily used dry-bulb temperature to define HWs. However, humidity also plays a critical role in influencing the intensity and health impacts of HWs. Different definitions of dry and wet HWs may lead to discrepancies in research results. In East Asia, the dynamics of dry and wet HWs defined by relative humidity differ (Ha et al., 2022), and in South China, dry-wet HWs defined by both wet-bulb and dry-bulb temperatures are related to the north-south movement of the South Asia High and the western North Pacific subtropical high (Luo et al., 2022(指代不明)). Therefore, it is crucial to consider humidity when setting heatwave thresholds.

This study qualitatively analyzed the spatial distribution characteristics of the three types of HWs in the GBA, as well as possible influencing factors and mechanisms. However, the impacts of the highly-urbanized and unique topography of the GBA on HWs have not been considered. During high-temperature events, the UHI effect intensifies, which further increases surface air temperature (Ward et al., 2016). Wind speed and moisture are key factors contributing to the synergistic effects of UHI and HWs (Li and Bou-Zeid, 2013; Hong et al., 2019; Rogers et al., 2019). Zhao et al. (2018) found that differences in urban versus rural evaporation rates are one of the key contributors to the synergistic effects during daytime. And urban surfaces, with their increased heat absorption, enhance sensible heat flux, which exacerbates extreme heat, while latent heat cooling from surface soil and evapotranspiration can help alleviate extreme local temperatures (Pyrgou et al., 2020). Therefore, how the GBA urban heat island effect synergizes with compound HWs concentrated in highly urbanized areas and the other two types of HWs deserves in-depth analysis. It is also worth investigating how the UHI effect in the GBA influences compound HWs, particularly in highly urbanized regions. Moreover, the unique topography of the GBA facilitates the formation and development of foehn winds, which can partially offset the cooling effects of sea breezes (Ramamurthy and Bou-Zeid, 2017; Hirsch et al., 2021). Therefore, future research should employ numerical simulations to determine the relative contributions of various physical processes over the complex underlying surface in the GBA, such as the impact of surface soil moisture on HWs and the roles of valley and sea-land breezes. Furthermore, these simulation results can validate the conclusions of the current study.