1 Introduction

High humidity plays a crucial role in exacerbating detrimental effects of high temperatures by impeding evaporative heat dissipation, thereby limiting the body’s ability to cool down effectively (Sherwood 2018). The accumulation of heat within the human body can result in varying degrees of heat stress, potentially leading to fatal outcomes (Davis et al. 2016; Dunne et al. 2013; Hanna and Tait 2015; Mora et al. 2017; Ramond et al. 2020; Rogers et al. 2021). Noteworthy instances, such as the heatwaves in Chicago (1995) and China (2003), demonstrate that humidity significantly amplifies the severity of heatwaves (Russo et al. 2017). Moreover, incorporating both temperature and humidity in health-related metrics yields more robust projections compared to using individual variable due to the negative correlation between temperature and relative humidity in models under different scenarios (Fischer and Knutti 2012). Recent studies have predicted a global trend where wet heatwaves are expected to surpass dry heatwaves in terms of intensity, frequency, and duration, indicating a heightened severity of wet heatwaves (Meng et al. 2022; Wang et al. 2021; Xu et al. 2020). Nevertheless, our understanding of the development mechanisms of extreme heat stress, particularly at a regional scale, remains limited. The mechanism analysis of extreme heat stress events in North China has not yet been seen.

Based on recent decades’ observations, North China (Fig. S1, 34°–41° N, 112°–121 °E), characterized by the largest alluvial plain in China and one of the world’s most densely populated regions (Kang and Eltahir 2018; Liu et al. 2008; Zheng et al. 2010), emerges as a hotspot for extreme heat stress (Bai et al. 2020; Kang and Eltahir 2018). Considering the important political, economic, and population status of North China and one of the climate hotspots, it is imperative to identify the key factors that govern extreme summertime heat stress in North China because it is helpful to improve the prediction skill of extreme heat stress in North China.

While the global response to global warming suggests an overall increase in atmospheric moisture content, regional variations in this trend are expected due to distinct patterns of moisture-holding capacity (Ha et al. 2022; Held and Soden 2006). North China is an arid agricultural region where human activities, particularly irrigation, significantly contribute to local humidity levels primarily during the crop-growing season spanning late spring to early summer (Bu et al. 2023; Kang and Eltahir 2018). However, extreme heat stress occurrences are most prevalent during mid- to late summer when both temperature and humidity levels are high. The inconsistency between the occurrence of summertime extreme heat stress and irrigation period in North China indicates that irrigation is not the main factor causing summertime extreme heat stress in North China. So, what is the main factor? Is it local or non-local impacts that has a greater impact on extreme heat stress than irrigation?

Most studies on heat stress have been attributed to the effect of large-scale circulation, with subsidence caused by adiabatic warming (An and Zuo 2021; Luo et al. 2022; Yeo et al. 2019). Besides, local land–atmosphere coupling also has significant influence on extremes (Gibson et al. 2017; Koster et al. 2016; Li et al. 2019; Ukkola et al. 2018; Zhang et al. 2020). With the decrease of soil moisture, latent heat flux decreases and sensible heat flux increases. So, surface temperature increases, and then heat stress intensifies. It has also been shown that one kind of extreme event downstream (such as heatwaves) may be caused by another extreme event (such as droughts) upstream (Zhou and Yuan 2022). Therefore, we speculate that extreme heat stress in North China may also be related to upstream climate state, which will be further analyzed in this paper.

A range of heat stress indices has been developed to capture different aspects and levels of heat stress, including wet-bulb globe temperature (Brouillet and Joussaume 2019; Kjellstrom et al. 2016; Li et al. 2020; Parsons 2006; Stull 2011), apparent temperature (Davis et al. 2016; Fischer and Schär 2010; Rastogi et al. 2020), heat index (Diffenbaugh et al. 2007; Matthews et al. 2017; Schwingshackl et al. 2021), excess heat factor (An and Zuo 2021; Nairn and Fawcett 2014), and wet-bulb temperature (TW) (Coffel et al. 2019; Freychet et al. 2020, 2022). Among above indices, TW is commonly used due to its adherence to the second law of thermodynamics, which governs the heat flux in and out of a wet object (Raymond et al. 2017). Severe, dangerous, and deadly levels of heat stress are typically defined by TW thresholds of 27.5 °C, 31.0 °C, and 35.0 °C, respectively (Freychet et al. 2022).

In summary, considering the significance of extreme heat stress in North China, this study aims to solve the following scientific questions. First, how has heat stress changed in North China in recent decades? Second, what dominates extreme heat stress in North China, temperature or humidity? Last but not the least, how do upstream variables affect extreme heat stress in North China? The remainder of this paper is arranged as follows: Section 2 describes data and methods. Results of scientific problems are analyzed in Sect. 3. The conclusions and the discussion are given in Sects.  4 and 5.

2 Materials and Methods

2.1 Observations and Reanalysis Data

Daily surface air temperature (T) and relative humidity (RH) are from the National Meteorological Information Center at the China Meteorological Administration under strict quality control, including error correction, continuity detection, and site migration correction (Zhihua et al. 2012). In order to separate temperature and humidity effects in causing extreme TW, we further calculated specific humidity (q) since RH is negatively correlated with T. Evaporation data were obtained from two sources: the European Centre for Medium-Range Weather Forecasts reanalysis version 5 (ERA5) (Hersbach et al. 2020) and the Global Land Evaporation Amsterdam Model (GLEAM) (Martens et al. 2017; Miralles et al. 2011). The ERA5 data combines satellite and reanalysis information, while GLEAM is based on satellite data. For ERA5, hourly evaporation values were accumulated to obtain daily values, while GLEAM data already had a daily temporal resolution. Other atmospheric variables, such as air temperature at 925 hPa, q, horizontal wind at 925 hPa, surface pressure (p), soil moisture, and geopotential height at 500 hPa, were obtained from ERA5 with a temporal resolution of six hours (00:00, 06:00, 12:00, and 18:00 UTC each day). Daily values for these variables were obtained by averaging above four time points. Soil moisture data were accumulated from all four layers (0–7 cm, 7–28 cm, 28–100 cm, and 100–289 cm) provided in ERA5, considering that plant roots have the ability to absorb water from greater depths in arid regions than wet regions.

To calculate the occurrence probabilities of summertime maximum T or q, we divided the number of times when the maximum T or q occurred on each 10-day interval (from June 2 to August 30, nine intervals in total) across all grids in North China by the total sample size, which is the product of number of grids and number of intervals. In Fig. 4.2a–h, we highlighted the 5880-gpm isobar at 500 hPa. The 5880-gpm isobar is commonly used to indicate the area and intensity of NPSH, which is relevant to various indices related to NPSH (Liu et al. 2019). All data used in this study had a spatial resolution of 0.25 × 0.25°. The daily q was calculated from T, RH, and p using the Bolton equation (Bolton 1980):

$$\begin{gathered} e_{s} = 6.112\exp^{17.67 \times T/(T + 243.5)} , \hfill \\ e = e_{s} \times {\text{RH}}\% , \hfill \\ q = \frac{0.622e}{{p - 0.378e}}. \hfill \\ \end{gathered}$$
(1)

where es is saturated vapor pressure (hPa), T is surface air temperature (℃), e is actual vapor pressure (hPa), RH is relative humidity (%), q is specific humidity (kg/kg) and p is surface pressure (hPa).

2.2 Definition of TW, Its Threshold and Definition of Regional Extreme TW Days

TW is a combination of T and RH (Stull 2011; Yu et al. 2021):

$$\begin{gathered} {\text{TW}} = T\tan [A({\text{RH}} + B)^{1/2} ] + a\tan (T + {\text{RH}}) - a\tan ({\text{RH}} - C) + D({\text{RH}})^{3/2} a\tan (E \times {\text{RH}}) - F, \hfill \\ A = 0.152,B = 8.314,C = 1.676,D = 0.004,E = 0.023,F = 4.686. \hfill \\ \end{gathered}$$
(2)

where TW is wet-bulb temperature (℃), T is surface air temperature (℃), tan is a trigonometric function and atan is an inverse trigonometric function, RH is relative humidity (%), A, B, C, D, E, F are 6 empirical parameters.

As mentioned earlier, a severe level of heat stress in North China is indicated by a TW over 27.5 ℃, which coincidentally aligns with the approximate value of the 99th percentile of North China summertime TW considering all grid point values. 27.5 ℃ is sufficiently extreme for heat stress in North China, considering both the physiological impact and adaptability. Therefore, we adopted 27.5 ℃ as the threshold when defining extreme heat stress across all grids in our study.

On each grid point, if TW exceeded 27.5 ℃, we defined the day on this grid point as an extreme TW day. To identify regional extreme TW days, we defined them as days when more than 50% of North China’s land area (lower than 925 hPa) exceeded 27.5 ℃ (An and Zuo 2021). However, when doing the composition analysis, we excluded any regional extreme TW days that occurred within five days of a previous regional extreme TW day. The exclusion was implemented to prevent the characteristics of the development of regional extreme TW days from being confounded. The selection of a 5-day period was based on the observation that evident development (anomaly) typically began three or four days before a regional extreme TW day, and the peak intensity of TW usually lasted for 2 days. The period from the onset of development to decline to normal levels is regarded as regional extreme TW period.

3 Results

3.1 High Sensitivity of Heat Stress to Humidity

An analysis of data from recent decades (1981–2020) reveals a significant (P < 0.1) increase in mean and maximum TW intensity across almost all of North China during boreal summer period (June to August) (Fig. 1a, b). The total number of days surpassing the TW threshold (27.5 °C) and the maximum duration of consecutive extreme TW days also experienced notable increases (Fig. 1c, d). While some isolated regions exhibited negative trends, they were mostly insignificant. The eastern part of North China displayed the most substantial positive trend in mean and maximum TW intensity, with a maximum value exceeding 0.4 °C/decade (Fig. 1a, b). Additionally, the southern part of North China exhibited the greatest rise in total number of extreme TW days, with a positive trend of 0.4 days/decade (Fig. 1c). Notably, areas in northwestern North China, including Tai Hang Mountain region, and central Shandong Province, which are characterized by higher altitudes, rarely surpassed the TW threshold. Because air temperature decreases with the increase of altitude, air temperature is lower in high mountains than other regions in North China. TW is a function of air temperature and humidity, so low temperature will inevitably lead to low TW. In this study, the definition of extreme heat stress is based on absolute thresholds. So, according to the definition, the number of days when TW exceeds the threshold is always zero. As a result, certain grid points in these regions show no trend in total number of days exceeding the TW threshold or the maximum duration of consecutive extreme TW days (Fig. 1c, d).

Fig. 1
figure 1

a Trend (℃/decade) of the mean wet-bulb temperature (TW, ℃). b Trend (℃/decade) of the maximum TW. c Trend (days/decade) of total extreme TW days. d Trend (days/decade) of the maximum consecutive extreme TW days. All trends are calculated in each grid point in North China from 1981 to 2020 during the boreal summer (June to August). In regions where slashes are displayed in a–d, the trends are statistically significant at the 90% confidence level. The analysis encompasses the North China region, which spans from 34° N to 41° N and 112° E to 121° E. Extreme TW day is defined as a day when the TW exceeds the designated threshold

Full size image

The days when TW exceeded the threshold were consistently observed in late summer, specifically between July 10 and August 20 (Fig. 2a). The temporal pattern aligns closely with the occurrence dates of maximum summertime q during summer across all grid points. However, the consistency between TW and q does not hold true between TW and T. In other words, extreme TW days always coincide with the presence of maximum q, whereas extreme TW days may not necessarily coincide with the occurrence of maximum T. The disparity between T and q can be explained by the fact that before August 10, T had already reached its peak and remained high for several days, while q was still in the process of rising (Fig. 2b). Only when q also reached a high level (after July 10) did the days exceeding the TW threshold occur. In summary, the occurrence of days surpassing the TW threshold is contingent upon both T and q being elevated, but the decisive factor in triggering extreme TW days is the level of q.

Fig. 2
figure 2

a The left axis represents the number of extreme wet-bulb temperature (TW) days within different 10-day intervals (shown in gray), while the right axis represents the occurrence probabilities of maximum T (shown in blue) and q (shown in red). These counts and probabilities are calculated for each grid in North China during the boreal summer from 1981 to 2020. b The intraseasonal variability of North China-averaged surface air temperature (T) (shown in blue) and surface specific humidity (q) (shown in red) is depicted, with dashed lines representing each year and a solid line representing the multi-year mean from 1981 to 2020. The possible date range for extreme TW days is indicated between the black dashed lines. c A scatter plot illustrates the relationship between T, q, and TW in North China during the boreal summer from 1981 to 2020. d Same as c, but for extreme TW days

Full size image

The scatter plot in Fig. 2c provides a visual representation of the relationships among TW, T and q across all grid points in North China. It is evident that TW increases with both T and q; however, the rates of increase differ. As TW rises, it is influenced more significantly by changes in q compared to changes in T. On average, when TW increases from 5 to 15 ℃, T changes by 10 ℃ (from 10 to 20 ℃), while q only changes by 5 g/kg (from 5 to 10 g/kg). However, when TW increases from 15 to 30 ℃, T still only changes by 10 ℃ (from 20 to 30 ℃), but q experiences a larger change of approximately 15 g/kg (from approximately 10–25 g/kg). For values of TW exceeding the threshold, TW is minimally affected by variations in T and is more strongly controlled by changes in q (Fig. 2d). In this scenario, if q remains constant, alterations in T will only have a limited impact on TW. Conversely, if T is fixed, TW will increase significantly with rising q.

3.2 The Development and Maintenance of Extreme Heat Stress

The spatial distribution and the intensity of TW in North China are closely linked to the patterns observed in q and T, as illustrated in Fig. 3. Approximately three to four days prior to regional extreme TW days, anomalies in TW, q, and T began to manifest. All of the anomalies typically intensified over a two-day period and then gradually returned to their normal levels. Two days before the occurrence of regional extreme TW days, abnormally high TW values emerged in the southern region of North China (Fig. S2d). On the following day, TW anomaly expanded and strengthened continuously, extending the anomaly to entire North China. Finally, on regional extreme TW days, TW anomaly extension and strengthening reached their peaks, with TW values surpassing 27 ℃ in more than half of North China (particularly the southern part) and even exceeding 28 ℃ in certain areas (Fig. 3.1f). Although TW weakened slightly one day after extreme TW days, some areas in southern North China still experienced TW values above 27 ℃, aligning with the occurrence of consecutive regional extreme TW days that typically persisted for an average of 2.1 days (Fig. 3.1g). Two days after regional extreme TW days, the majority of North China exhibited TW values below 27 ℃, resembling the conditions observed two or more days before regional extreme TW days (Fig. 3.1h). Notably, TW intensity remained consistently high in southern North China and gradually decreased with increasing latitude (Fig. 3.1e–g).

Fig. 3
figure 3

The composition for day-to-day intensity variations of TW (1a–h; ℃), q (2a–h; g/kg), and T (3a–h; ℃) from 5 days before (a) to 2 days after (h) regional extreme TW day (f). Regional extreme TW day is defined as a day when the TW exceeds the designated threshold in more than 50% of the North China area, excluding mountainous regions above 925 hPa

Full size image

When comparing the developments of q and T with TW, it is evident that the changes in q align more closely with observed TW than T (Fig. 3.2a–h). Typically, TW is 1–2 ℃ lower than T due to the cooling effect of q. Three days before a regional extreme TW day, the spatial extent of q at a level of 20–22 g/kg began to expand toward southern regions of North China (Fig. S2.2c). One day later, the level of q at 20–22 g/kg extended to cover areas within North China, although q values to south of North China remained higher than those within North China (Fig. S2.2d). Another day later, q in southern North China increased to exceed 22 g/kg (Fig. 3.2e), marking the first instance where q in North China reached the same level as in the southern regions. Finally, on regional extreme TW days, the region with q > 22 g/kg reached its maximum, with q exceeding 20 g/kg in most parts of North China (Fig. 3.2f). On the first day after regional extreme TW days, q experienced a rapid decline, and only a small area in southern North China maintained q > 22 g/kg (Fig. 3.2g). Importantly, the spatial pattern of q on regional extreme TW days and the subsequent day closely resemble the corresponding TW pattern (Fig. 3.1f–h, 2f–h). The consistent patterns between TW and q further confirm that q is the decisive factor in the occurrence of regional extreme TW days.

The development and the maintenance of T differ from those of TW and q from 5 days before to 2 days after regional extreme TW days. Overall, T intensity developed earlier and lasted longer than that of TW and q (Fig. 3.3b). It is noted that the fairly constant temperature difference of 2–4 ℃ between North China and a region to southern North China is conducive to transporting heat to North China through temperature advection (Figs. S1.3a–h).

3.3 Impact of Upstream Conditions on Extreme Heat Stress

The development and the maintenance characteristics of regional extreme TW days, as illustrated in Fig. 3, demonstrate a consistent pattern regardless of whether the parameter is TW, q, or T. The anomalies first emerged in southern regions of North China and then progressed toward North China (Fig. S2). It indicates that the influence of upstream factors is critical.

In the case of q development in North China, it is primarily influenced by the intensification of evaporation in southern regions of North China and subsequent transportation through low-level humidity flux (Fig. 4.1a–h). Three days before a regional extreme TW day, the area with anomalous evaporation exceeding 4 mm/day expanded in southern regions of North China (Fig. 4.1c). The evaporation anomaly strengthened the following day, with evaporation exceeding 5 mm/day at certain grid points, but the spatial extent of anomaly remained unchanged (Fig. 4.1d, 1e). Evaporation in southern regions of North China continued to intensify, reaching its maximum one day before and on regional extreme TW days (Fig. 4.1e, 1f). Concurrently, starting from 2 days before regional extreme TW days, the humidity flux into North China also strengthened in tandem with evaporation (Fig. 4.1d). One day before and on regional extreme TW days, the humidity flux into North China peaked, transporting a substantial amount of q to North China (Fig. 4.1e, 1f). One day after regional extreme TW days, the intensity of evaporation decreased slightly, while the humidity flux into North China significantly weakened (Fig. 4.1g).

Fig. 4
figure 4

1a–h, The composition for day-to-day variations of ERA5 evaporation (shading; mm/day) and humidity flux at 925 hPa (vector; g s−1 hPa−1 m−1) from 5 days before (1a) to 2 days (1h) after regional extreme TW day (1f) in East China (14°–42° N, 100° E–125° E), with the black box representing North China. The blue lines in 1a–h are the 5880-gpm isobar at 500 hPa. 2a–h, Day-to-day variation of temperature advection (℃/day) in North China from 5 days before (2a) to 2 days (2h) after regional extreme TW day (2f)

Full size image

In North China, there were minimal noticeable changes in evaporation around regional extreme TW days, except for a slight increase observed specifically on regional extreme TW days. However, the slight increase in evaporation was significantly lower compared to the anomalously higher evaporation occurring in southern regions of North China. The temporal development characteristics and the spatial patterns of evaporation, as indicated by GLEAM dataset, were found to be consistent with those observed in ERA5 dataset described earlier (Fig. S3), which means that above finding is robust.

The development of T in North China was influenced by the temperature difference between northern and southern regions, which facilitated a constant positive temperature advection toward North China. The positive temperature advection was observed to first appear three days before regional extreme TW days and continued to intensify until one day before regional extreme TW days. The advection intensity reached over 2 ℃/day across most of North China, with some grid points experiencing up to 4 ℃/day (Fig. 2. 4.2c, 2e). The increase in T (2–4 ℃) between normal situations and regional extreme TW days was predominantly governed by the positive temperature advection (Figs. 3.3a–f and 4.2a–f).

Notably, there were no abnormal sinking motions associated with high pressure in North China before and on regional extreme TW days (Fig. 4.1a–f). Instead, during the eight-day period, the Northwest Pacific Subtropical High (NPSH) weather system developed and persisted to the south of North China, leading to increased evaporation in Yangtze River Basin (YRB) (Fig. 4.1a–f).

The positive evaporation anomaly observed before regional extreme TW days is attributed to the western extension of the NPSH (Fig. 4.1d–f). The NPSH extends to YRB three days before regional extreme TW days (Fig. 4.1c). One day before and during regional extreme TW days, NPSH reached its peak, accompanied by strong sinking motion, which leads to an increase in T and acceleration of evaporation in YRB (Fig. 4.1e, 1f). As a result, soil moisture in YRB experienced a notable reduction, with a final deficit of over 0.05 m3/m3 two days after regional extreme TW day (Fig. S4l). In contrast, soil moisture in North China did not decrease even after regional extreme TW days (Figs. S3h–l). Three days before regional extreme TW days, soil moisture in YRB began to decrease due to enhanced evaporation (Fig. S4d). Just one day after regional extreme TW days, the soil moisture difference compared to regional extreme TW days becomes negative, with a level of exceeding 0.05 m3/m3 (Fig. S4h).

4 Discussion

Our study reveals that upstream evaporation plays a crucial role in increasing q in North China, leading to extreme heat stress. We have shown that q is the primary factor influencing extreme summertime heat stress in North China, highlighting a relationship where upstream meteorological variables impact downstream variables. While the relationship has been proposed in previous studies (Yang et al. 2021; Zeng and Yuan 2021; Zeng et al. 2019), its application to extreme heat stress, particularly in North China, has not be done to the best of our knowledge.

Between July 10 and August 20, a period when TW exceeded the defined threshold, the retreat of NPSH toward YRB was observed, influencing evaporation patterns. The influence of NPSH edge resulted in increased precipitation frequency over North China, which had a local cooling effect but hindered evaporation. The increased precipitation limits evaporation and cools the land, thus damping heat stress. Although the ocean serves as a significant source of water vapor, our analysis did not find a notable difference in ocean evaporation between regional extreme TW days and other days (Fig. 4.1a–h). As a result, water vapor from the oceans did not play a major role in the occurrence of regional extreme TW days. One explanation is that a substantial amount of water vapor dissipates during long-distance transportation. The dissipation can be observed in Fig. 4.1d–g, where humidity flux is prominent over the Yellow Sea but significantly reduces when it moves inland due to dissipation, before intensifying again over YRB. The proximity of YRB to North China, combined with the significant upstream evaporation resulting from the extension of NPSH, also played a critical role in minimizing dissipation. As a result, the impact of dissipation on water vapor supply was greatly reduced. Therefore, to accurately predict and project heat stress in North China, it is essential to provide a precise description of upstream evaporation, considering the influence of NPSH and the specific geographical characteristics of North China.

We propose that q is the dominant factor controlling heat stress, particularly during extreme days. The explanation lies in the nonlinearity between TW and q, which clarifies the varying importance of q at different TW levels (Fig. 2c) (Stull 2011). This finding is in line with the established notion that q is more crucial than T in arid areas of China for extreme heat stress (Wang et al. 2019). However, the relative importance of q and T varies across different time scales and regions. For instance, in a study of heat stress in the United States, T is equally or even more dominant than q in some climatologically cool and moist regions (Raymond et al. 2017). Besides, on a global interannual scale, T is the decisive driver of TW changes (Yu et al. 2021).

Our analysis revealed a consistent linear trend of increasing heat stress intensity and frequency in North China, with the majority of grid points showing a positive trend. However, there were a few grid points around Beijing city where an insignificant negative trend was observed. To a large extent, this can be attributed to the frequent artificial landscape irrigation practices employed in the highly developed historic city of Beijing (Yuan et al. 2009). The irrigation process involves the phase change of water, resulting in land cooling and lower temperatures. As a result, the average temperature around Beijing (39.5°–42° N, 115° E–117° E) is 1–2 ℃ lower than that in North China, while still exhibiting similar interannual variability (Fig. S5). On top of that, according to the directions of humidity flux in Fig. 4.1, Beijing city is unaffected by humidity flux from YRB to North China. It proves again that if there is no upstream humidity transport, the trend of heat stress in North China would not increase.

5 Conclusion

In conclusion, our findings emphasize that q is the decisive factor in driving extreme heat stress in North China. Moreover, the unique role of upstream evaporation in YRB is vital in contributing the crucial additional humidity required to create extreme heat stress in North China. The occurrence of high upstream evaporation during regional extreme TW days can be attributed to the westward extension of NPSH.

This study highlights the decisive role of humidity in creating extreme heat stress in North China, despite both humidity and temperature reaching their peak values during TW extremes. Remarkably, the transportation of humidity and the advection of temperature from upstream regions to North China have played a critical role in the formation of extreme heat stress. The upstream influence implies that when predicting extreme heat stress in North China, it is not only crucial to consider local changes but also necessary to account for upstream processes, particularly upstream evaporation. Therefore, a comprehensive understanding of both local and upstream factors is essential for accurate predictions of extreme heat stress in North China.

Future work should be focused on improving the accuracy of evaporation prediction. It has been proved that evaporation prediction based on artificial intelligence method has great scientific potential (Abed et al. 2023). Therefore, it is worth studying the prediction of evaporation by combining physical model with artificial intelligence. Another direction of future research is to focus on the influence of upstream variables on other extreme events in North China, such as heavy rains and droughts. Extreme self-propagation is also a hot research topic. Drought, for example, always spreads itself downwind. (Schumacher et al. 2022).