Concurrent drought and heatwave events over the Asian monsoon region: insights from a statistically downscaling CMIP6 dataset

Concurrent drought and heatwave (CONDH) can cause tremendous ramifications on socioeconomic activities and human health, and the drought-heatwave (D-H) dependence was revealed to be one of the major factors of the CONDH across most global land regions. However, insufficient attention has been paid on the CONDH over the Asian monsoon region, and the impact of the D-H dependence is even more rarely considered. Based on a statistically downscaling the Coupled Model Intercomparison Project Phase 6 (CMIP6) dataset, we explore the spatial distributions of the intensity, frequency, and duration of the CONDH. In addition, the qualitative impacts of D-H dependence on the intensity, frequency, and duration of the CONDH have been further investigated via comparing these three metrics of the CONDH with those of heatwave. In the period of 1961–2014, the spatial pattern of intensity and duration of the CONDH bear a large resemblance to that of heatwave, with more severe CONDH (heatwave) occurring over South Asia, and relatively long-lasting CONDH (heatwave) occupying over the low latitudes of the Asian monsoon region. The frequency of the CONDH presents large discrepancies with that of heatwave. East Asia (South Asia) is hit by more frequent CONDH (heatwave) than other sub-regions. The D-H dependence is conducive to the intensity, frequency, and duration of the CONDH, especially over the East Asia, Eastern Siberia, and Western Siberia. In the period of 2015–2100, the intensity and duration of the CONDH and heatwave maintain the historical spatial structures. The frequency of the CONDH and heatwave change remarkably relative to 1961–2014, with largest value over eastern central Asia. The D-H dependence is conducive to the three features of the CONDH, and such positive contributions would weaken in response to global warming, especially under higher emission scenario.


Introduction
Drought and heatwave can directly and indirectly reduce water availability and increase energy consumption, causing tremendous ramifications on the environmental and social systems (Perkins et al 2012, Dai 2013, Trenberth et al 2014, Zhai et al 2018, Chen et al 2020, Wang et al 2021, 2023, Dong et al 2023).Such adverse effects would be exacerbated when these two extreme events occur simultaneously, which is referred to as concurrent drought and heatwave (CONDH) event (Mueller and Seneviratne 2012, Mazdiyasni and AghaKouchak 2015, Hao et al 2018, Kong et al 2020, Mukherjee and Mishra 2021, 2022, Zhang et al 2023).The Asian monsoon region is extremely vulnerable to climate extremes, given that more than 60% of the world population live here and mostly depend on agriculture for their livelihoods (Kim and Bae 2020, Kim et al 2020, Ge et al 2021).However, numerous studies have been focused on drought and heatwave independently over this region (Son and Bae 2015, Zscheischler et al 2018, Sutanto et al 2020, Ge et al 2021, Yang et al 2021, Zhang et al 2022, Yin et al 2023), which might have grossly underestimated the impact of the CONDH.A comprehensive understanding of the CONDH is in an urgent need for the disaster prevention, mitigation, and relief over the Asian monsoon region.
Drought and heatwave are not independent to each other due to the thermodynamic relationship between precipitation and temperature.Previous studies constructed the concept of drought-heatwave (D-H) dependence to depict the tendency for drought and heatwave facilitate or inhibit each other when they occurred simultaneously, and signified that stronger D-H dependence is often associated with more frequent CONDH (Lyon 2009, Zscheischler and Seneviratne 2017, Hao et al 2020, Kong et al 2020).For instance, Zscheischler and Seneviratne (2017) elucidated that the D-H dependence contributed to the frequency of the CONDH in warm seasons over most continental areas of the Northern Hemisphere.Kong et al (2020) revealed that the D-H dependence is conducive to the intensity, frequency, and duration of the summer CONDH over Eastern China during 1962-2015.On the other hand, observation evidences and model simulations illustrated that the soil moisture-temperature feedback was one of the main drivers of the D-H dependence, particularly for the transitional regimes between wet and dry climate zones (Seneviratne et al 2006, 2010, Muller and Seneviratne 2012, Holmes et al 2017, Vogel et al 2017).Specifically, high temperature can induce soil moisture deficiency through accelerating evapotranspiration, moreover, sustained soil moisture deficit leads to increase in surface temperature by partitioning surface energy flux into more sensible heating and less latent heating.Since the Asian monsoon region consists of tropical, warm temperate, arid, cold, and polar climate zones with different soil moisture-temperature interactions, the sixth assessment report of the intergovernmental panel on climate change divides this region into eight subregions, including East Asia (EAS), eastern central Asia (ECA), Eastern Siberia (ESB), Western Siberia (WSB), Russian Far East (RFE), the South Asia (SAS), Southeast Asia (SEA), and the Tibetan Plateau (TIB) (Iturbide et al 2020, figure 1(c)).However, which region possess the strongest D-H dependence remains unknow.This issue is essential for identifying hotspots of the CONDH over the Asian monsoon region.
As the 'dry gets drier, and wet gets wetter' paradigm is widely recognized (Dai 2013, Trenberth et al 2014), the shifting of climate zones can definitely alter the D-H dependence.For example, Zscheischler and Seneviratne (2017) showed that the D-H dependence during 2001-2100 would increase over northern extratropic regions, eastern Asia, and some parts of western South America relative to 1871-1969. Lyon (2009) ) elucidated that the D-H dependence did not substantially change over South Africa under global warming.In the Asian monsoon region, climate zones are shifting to arid climate zone under global warming (Dai 2013, Trenberth et al 2014, Son and Bae 2015).Nevertheless, a limited number of studies have been carried out to investigate how the D-H dependence changes in the future over this region.
Based on the above premises, the primary goals of this study are: (1) what is the spatial characteristics of the CONDH over the Asian monsoon region?(2) How and to what extent D-H dependence can modulate the CONDH over different regimes of this region?
(3) How the D-H dependence changes in the future?The remainder of this paper is organized as follows: section 2 depicts the data, the definition of summer heatwave and CONDH, and the D-H dependence.The results obtained are discussed in section 3. The conclusions and discussion are supplied in section 4.

Data
In this study, we mainly focus on summer heatwave and CONDH.Daily precipitation and maximum temperature in summer are obtained from the first ensemble member ('r1i1p1f1') of 20 models (table S1) in the NASA Earth Exchange Global Daily Downscaled Projections (NEX-GDDP), including historical simulations  and future projections (2015-2100) under the SSP126 and SSP585 scenarios (Thrasher et al 2012, Maurer andPierce 2014).The periods of 1961-2014 and 2015-2100 are the reference time slices of the CMIP6 dataset in historical simulations and future projections (Eyring et al 2016).The NEX-GDDP dataset is generated via correcting bias and downscaling daily variables from the CMIP5 and CMIP6 projects (NEX-GDDP-CMIP5 and NEX-GDDP-CMIP6).
As the predecessor of the NEX-GDDP-CMIP6, the NEX-GDDP-CMIP5 has showed considerable improvements in simulating climate extremes at regional scale compared with the raw CMIP5 output, and has been widely used in assessment climate change impacts at regional scale (Chen et al 2017, Bao and Wen 2017, Xu and Wang 2019).In this study, the performance of the NEX-GDDP-CMIP6 over the Asian monsoon region is assessed by the global meteorological forcing dataset (GMFD), which merges information from reanalysis, in situ observations, and satellite remote sensing products (Sheffield et al 2006).The GMFD with the horizontal resolution of 0.25 • × 0.25 • has been widely employed as observational validation in the field of hydrometeorology (Wood et al 2002, 2004, Raghavan et al 2018, Xu and Wang 2019).The mean values of summer precipitation (maximum temperature) during 1961-2014 averaged over the Asian monsoon region obtained from the GMFD is 4.00 mm d −1 (26.72 • C), which is 3.98 ± 0.16 mm d −1 (26.75 ± 0.03 • C) in the NEX-GDDP-CMIP6.The observed spatial patterns of mean values of summer precipitation (r = 0.99, p < 0.01, figures S1(a) and (e)) and maximum temperature (r = 0.99, p < 0.01, figures S1(c) and (g)) are highly reproduced by the NEX-GDDP-CMIP6.Similar result can be obtained for the standard deviations of summer precipitation (figures S1(b) and (f)) and maximum temperature (figures S1(d) and (h)).Thus, the NEX-GDDP-CMIP6 dataset shows a good reproducibility for the observed first-and secondorder moments of these two variables in summer, which is a relatively reliable dataset that can be applied for assessment of extreme climate change over the Asian monsoon region.

Definitions of summer heatwave and CONDH
For each grid cell, a heatwave event is identified when the daily maximum temperature exceeds a relative threshold for at least three consecutive days.The threshold is calculated as the 90th percentile of daily maximum temperature of all summer days during 1961-2014.For each month, the frequency of heatwave refers to the total number of heatwave events.The intensity of a heatwave event is defined as the mean value of daily maximum temperature in all days during this event.The duration of a heatwave event represents the total number of days during this event.
If there is a heatwave event at the end of previous month and another heatwave event at the beginning of the following month, it is recorded as two heatwave events.The definition of heatwave in the current study has been widely adopted in different regions of the globe (Mueller and Seneviratne 2012, Perkins et al 2012, Mazdiyasni and AghaKouchak 2015, Chiang et al 2018).
The CONDH is defined as a heatwave event occurring in a meteorological drought context, which can be depicted by the standardized precipitation index (SPI), the standardized precipitation evapotranspiration index (SPEI), and indices associated with the palmer drought severity index (PDSI).
Except for the SPI, potential evapotranspiration is required in the SPEI and the PDSI, which is not available in the NEX-GDDP-CMIP6 dataset.Thus, the SPI is employed to represent soil moisture condition (Guttman 1998(Guttman , 1999)).Previous studies signified that the seasonal drought can be well represented by the 3 m SPI, which is calculated by fitting a gamma distribution to 3 m precipitation (Seneviratne et al 2006, 2010, Hirschi et al 2011, Mueller and Seneviratne 2012).In the current study, the drought month is defined as a month whose 3 m SPI value is smaller than −0.8, as adopted in previous studies (Mueller and Seneviratne 2012, Holmes et al 2017, Vogel et al 2017).The definitions of frequency, intensity, and duration of the CONDH are the same as the above heatwave, but for the meteorological drought month.The spatial characteristics of three metrics of the CONDH defined based on 3 m SPI bear large resemblance to those defined based on 1 m SPI (figure not shown).

The D-H dependence
According to previous studies (Lyon 2009, Kong et al 2020), the D-H dependence can be measured as the ratio of heatwave in drought months to that in all months, denoted as Taking the period of 1961-2014 as an illustrative example, DM (TM) refers to the number of drought (all) summer months.HI dry (HI) represents heatwave intensity in drought (all) condition, which is calculated as cumulative intensity for heatwave events drought (all) summer months.HF dry (HF) denotes heatwave frequency in drought (all) condition, which is calculated as cumulative frequency for heatwave events in drought (all) summer months.HD dry (HD) represents heatwave duration in drought (all) condition, which is calculated as cumulative days for heatwave events in drought (all) summer months.The ratio is greater than 1, suggesting that intensity, frequency, and duration of heatwave in drought condition are larger than those in all condition.That is, the D-H dependence is conducive to intensity, frequency, and duration of the heatwave develop toward larger value.The ratio is smaller than 1, indicating that the D-H dependence is not conducive to intensity, frequency, and duration of heatwave development.If the ratio is equal to 1, the CONDH shows no differences from heatwave, qualitatively indicating that drought and heatwave are independent of each other.Therefore, this measurement of the D-H dependence can qualitatively demonstrate that drought provide a favorable/unfavorable condition for heatwave development when drought and heatwave occur simultaneously.
To analyze the future changes in the D-H dependence, we compare the DH In the current study, the change between two periods is denoted as the value in the latter period divided by that in the former period.The signal-to-noise ratio is employed to assess the robustness of the multi-model result, which is constructed as the ratio of the NEX-GDDP-CMIP6 multimodel ensemble mean (MME) to the corresponding inter-model standard deviation.The ratio greater than 1 means a high robustness of the result (Giorgi et al 2018, Xu et al 2021).The EAS, ECA, ESB, WSB, RFE, SAS, SEA, and TIB denotes the domain of 19.5

Spatial characteristics of summer heatwave and CONDH during 1961-2014 and the impact of the D-H dependence
The spatial characteristics of intensity, frequency, and duration of summer heatwave and CONDH during 1961-2014 over the Asian monsoon region are investigated based on the GMFD and the NEX-GDDP-CMIP6 (figure 1).The intensity, frequency, and duration of the CONDH (heatwave) obtained from the NEX-GDDP-CMIP6 MME shares a similar spatial pattern with that derived from the GMFD (figures 1(a)-(l)).The area-averaged intensity and duration of the CONDH (heatwave) over each subregion obtained in the two datasets are nearly the same (figures 1(m), (o), (p) and (r)).Slightly differences exist between the GMFD and the NEX-GDDP-CMIP6 dataset in depicting the frequency and duration of the CONDH (heatwave) over each subregion.The area-averaged frequency of the CONDH and heatwave over the whole domain in the GMFD are 58.5 times and 16.6 times, which are slightly smaller than those (60.2times and 19.2 times) in the NEX-GDDP-CMIP6 MME, particularly in SEA (figures 1(n) and (q)).Therefore, the NEX-GDDP-CMIP6 dataset can well capture the observed characteristics of summer heatwave and CONDH over the Asian monsoon region.Taking the NEX-GDDP-CMIP6 MME as an illustrative example, the characteristics of the CONDH and heatwave are described detailly below.
The spatial and zonal distributions of the CONDH intensity bear a large resemblance to those of heatwave intensity (figures 1(d) and (j)), and the area-averaged intensity of the CONDH over each sub-region is nearly the same as that of heatwave (figure 1(p)).Heatwave and CONDH tend to occur severely over SAS, whereas they exhibit relatively low intensity over TIB.The spatial pattern of the CONDH frequency presents large discrepancies with that of heatwave frequency (figures 1(e), (k) and (q)).Overall, the frequency of heatwave (60.0 ± 2.58 times) is about three times as much as that of the CONDH (19.0 ± 1.85 times) over the whole domain.
Heatwave occurs more frequently over SAS, while the frequency of the CONDH shows relatively lower values over this region.The CONDH duration shows an analogues spatial structure to heatwave duration, and the former is about one day larger than the latter over the entire domain (figures 1(f), (l), and (r)).The long-lasting CONDH/heatwave is located mainly at the mid-low latitudes, especially over SEA, SAS, TIB, and EAS.Therefore, the intensity and duration of summer CONDH exhibit analogues spatial structures to those of heatwave in the historical period, suggesting that drought can exert little impacts on the two features of the CONDH over the Asian monsoon region.On the other hand, the drought can affect the occurrence of summer CONDH, given the large discrepancies in the frequency of the CONDH and heatwave.
The signal-to-noise ratios for the three metrics of heatwave and CONDH are greater than 1 over almost the entire region, except for western central Asia where the CONDH is absent due to the absence of drought month (SPI > −0.8, figure not shown), indicating that the spatial characteristics of the heatwave and CONDH derived from the NEX-GDDP-CMIP6 are robust over most Asian monsoon region.Besides, the NEX-GDDP-CMIP6 shows good consistencies with the GMFD in depicting the characteristics of the CONDH and heatwave.In summary, the above characteristics based on the NEX-GDDP-CMIP6 dataset are reliable and robust.
The D-H ratio is further examined to investigate the qualitative impact of the D-H dependence on summer CONDH over different parts of the Asian monsoon region during 1961-2014 (figure 2).The DH i , DH f , and DH d share a similar spatial pattern, with values greater than 1 over almost the entire region, indicating that the D-H dependence is inducive to the intensity, frequency, and duration of the CONDH except for western central Asia.The DH i over the SEA, SAS, TIB, ECA, EAS, RFE, ESB, and WSB are about 1.44, 1.30, 1.47, 1.36, 1.59, 1.41, 1.66, and 1.61, indicating that the intensity of heatwave in drought condition is about 1.3-1.7 times of heatwave in all condition (figure 2(d)).The DH f over the SEA, SAS, TIB, ECA, EAS, RFE, ESB, and WSB are about 1.57, 1.38, 1.56, 1.47, 1.74, 1.52, 1.85, and 1.68, suggesting that the frequency of heatwave in drought condition is about 1.4-1.9times of heatwave in all condition (figure 2(e)).Moreover, the DH f shares a quite similar spatial structure with the CONDH frequency, and large values appear over EAS, ESB, and WSB (figures 1(e), 2(b) and (e)).That is, more frequent CONDH tends to occur over the region with stronger D-H dependence.The duration of heatwave in drought condition is about 1.4-1.9times of heatwave in all condition (figure 2(f)).
Therefore, the D-H dependence provides positive contributions to the intensity, frequency, and duration of summer CONDH in the historical period over the Asian monsoon region, especially over EAS, ESB, and WSB.It is worthy of notice that the spatial pattern of the CONDH intensity (duration) is different from that of the DH i (DH d ), but consistent with that of heatwave intensity (duration) (figures 2(a) and 1(a); figures 2(c) and 1(c)).These features suggest that the intensity and duration of summer CONDH can be affected by the D-H dependence, but they are largely determined by heatwave.Our result is consistent with previous studies, which signified that heatwaves are extreme events that meet certain thresholds, resulting in small differences in intensity and duration of heatwave between drought condition and all condition (Mazdiyasni and AghaKouchak 2015).On the other hand, the frequency of summer CONDH is dominated by the D-H dependence, which has been revealed previously (Mueller and Seneviratne 2012, Zscheischler and Seneviratne 2017, Hao et al 2020).For example, Mueller and Seneviratne (2012) revealed that regions with stronger soil moisture-temperature feedbacks (D-H dependence) often possesses more frequent CONDH in the warm season.Zscheischler and Seneviratne (2017) elucidated that the increased D-H dependence can exacerbate the increase in the CONDH frequency over the northern extratropic, eastern Asia, and some parts of western South America.The above conclusion obtained from the DH i , DH f , and DH d is robust, because the signal-tonoise ratios are greater than 1 over almost the entire region.

Projected characteristics of summer heatwave and CONDH and changes in the D-H dependence
The projected spatial patterns of summer heatwave and CONDH over the Asian monsoon region during 2015-2100 are displayed (figure 3).Consistent with features observed during 1961-2014, the projected intensity and duration of the CONDH also display similar spatial structures with those of heatwave (figures 3(a) and (d); figures 3(c) and (f)), indicating that these two metrics of the CONDH are mainly determined by heatwave during 2015-2100 under the SSP126 and SSP585 scenarios.Severe CONDH (heatwave) tends to occur over SAS, and long-lasting CONDH (heatwave) occurs primarily at the mid-low latitudes.The CONDH shows a longer duration but a comparable intensity compared with heatwave under both two scenarios.The two types of extreme events over the eight sub-regions exhibit slightly stronger intensities and much longer durations under the higher emission scenario (figures 3(g) and (i)).The frequencies of summer heatwave and CONDH in the future change remarkably relative to the historical period (figures 1(e) and 3(b); figures 1(k) and 3(e)).The frequency of summer heatwave and CONDH share a similar spatial structure with each other during 2015-2100 (figures 3(b) and (e)), suggesting that the frequency of the CONDH is mainly determined by heatwave in the future period.Specifically, SAS experiences the highest frequency of heatwave during 1961-2014, while this region is projected to experience the least occurrence of heatwave during 2015-2100.ECA is projected to experience the most frequent CONDH and heatwave during 2015-2100 among the eight sub-regions, possibly attributed to the spatial heterogeneity of global warming with more rapid increase in temperature over the high latitudes than that over the low latitudes (Taylor et al 2013, Pithan and Mauritsen 2014).
The three D-H ratios during 2015-2100 under the two scenarios share analogous spatial structures with those during 1961-2014, but with smaller magnitudes (figures 2(a) and 4(a); figures 2(b) and 4(b); figures 2(c) and 4(c)), indicating that the D-H dependence over most of the Asian monsoon subregions would weaken in response to global warming.The DH i , DH f , and DH d averaged over the eight sub-regions under the SSP126 and SSP585 scenario are greater than 1, and the value in each sub-region under the SSP126 scenario is larger than that under the SSP585 scenario (figures 4(a) and (b)).That is, the D-H dependence exerts a positive impact on the intensity, frequency, and duration of the CONDH in the future, and the impact is more pronounce in the lower emission scenario.It is worth noticing that the DH f over SEA under the SSP585 scenario is equal to 1, indicating that drought and heatwave are independent of each other.Our result is consistent with that of Vogel et al (2017), who revealed that the soil moisture-temperature interactions exerted little impact on the increase in the hottest days over SEA during 2081-2100 relative to 1951-1970 based on the experiments initialed by CMIP5 output in the RCP585 scenario.
The above analysis suggests that the D-H dependence provides positive contributions to the intensity, frequency, and duration of summer CONDH in the future period under both two scenarios.Nevertheless, the impacts of the D-H dependence on the three features of summer CONDH in the future are less positive compared with those in the historical period, especially under the SSP585 scenario.
To investigate how the CONDH changes with the weakening of the D-H dependence, the changes in the three metrics of summer heatwave and CONDH, numbers of drought months, and the D-H dependence during 2046-2065 and 2080-2099 with respective to 1995-2014 are displayed (figures 5 and S3).Under the SSP126 scenario, the frequency of summer heatwave (CONDH) over the entire region during 2046-2065 increase, which is about 2.4 (2.0) times as that during 1995-2014.The change magnitude of summer heatwave (CONDH) frequency over each sub-region during 2046-2065 are nearly the same as The decreases over all the sub-regions during the two future periods under both two scenarios, and the decreased magnitude is relatively large under the SS585 scenario.Change in the CONDH can be affected by heatwave, drought, and the D-H dependence.The above feature suggests that a weakened impact of the D-H dependence exists in the future relative to the historical period, especially under higher greenhouse gases concentration.Similar result can be also obtained for the DH i and the DH d (figure S3).Therefore, the increases in the three features of summer CONDH during 2046-2065 and 2080-2099 over the Asian monsoon region under the SSP126 and SSP585 scenarios are mainly determined by the changes in heatwave.
In addition to the ratio of CONDH to heatwave, the D-H dependence can be measured by copula methodology, which has been widely used in modeling dependence between hydroclimatic variables (Nelsen 2006, Genest and Favre 2007, Zscheischler and Seneviratne 2017, Hao et al 2020).Studies based on copula methodology signified that the interannual correlation coefficient between precipitation and temperature was generally a good indicator for the The change between the two periods is defined as the value in the latter period divided by that in the former period.The box-and-whisker plot shows the 10th, 25th, 50th, 75th, and 90th percentiles of the 20 models from the NEX-GDDP-CMIP6 dataset.
influence of the D-H dependence on the likelihood of the CONDH.A larger absolute value of correlation coefficient indicates a greater impact of the D-H dependence on the CONDH.To verify our result, we explore changes in the D-H dependence measured by interannual correlation coefficient between summer temperature and precipitation (figure not shown).Result shows that change in the correlation coefficient is spatially heterogeneous over the Asian monsoon region, and there is decreased correlation coefficient in most sub-regions during 2046-2065 and 2080-2099 with respective to the historical period, which is consistent with our result that the D-H dependence would decrease over almost the whole region in the future.

Conclusions and discussion
The Asian monsoon region is extremely vulnerable to extreme events due to its population density and monsoon climate (Kim and Bae 2020, Kim et al 2020).To date, the severe ramifications caused by the CONDH have been witnessed but largely underestimated by the independent investigations on drought and heatwave over this region (Son and Bae 2015, Ge et al 2021, Yang et al 2021).Providing a comprehensive understanding of the CONDH is necessary to appropriately assess the risks associated with its concurrency.
Previous studies revealed that the D-H dependence can exert significant impacts on the warmseason CONDH across most global land regions, and the impacts exhibit regional diversity due to the soil moisture condition (Lyon 2009 Based on a statistically downscaled CMIP6 dataset, we investigate the historical  and future (2015-2100) spatial characteristics of summer CONDH over the Asian monsoon region, and further explore the D-H dependence through comparing intensity, frequency, and duration of the CONDH with that of heatwave in averaged condition.In the period of 1961-2014, more severe CONDH tends to occur over SAS, and relatively long-lasting CONDH occupies over the low latitudes of the Asian monsoon region, whereas EAS is hit by more frequent CONDH than other sub-regions.The D-H dependence is conducive to the intensity, duration, and frequency of the CONDH over Asian monsoon region, especially over EAS, ESB, and WSB.The D-H dependence is the main factor of the CONDH frequency, while the intensity and duration of the CONDH are primarily controlled by heatwave in average condition.In the period of 2015-2100, the intensity and duration of the CONDH maintain the historical spatial structures.The CONDH frequency change remarkably relative to the historical period, and ECA is projected to experience the most frequent CONDH.The D-H dependence provides positive contributions to the three features of the CONDH, and such positive contributions would weaken relative to the historical period.The intensity, frequency, and duration of the CONDH in the future are dominated by that of heatwave in average condition, respectively.
Several caveats need to be mentioned in this study.First, in addition to the ratio of CONDH to heatwave, the D-H dependence can be measured by copula methodology (Nelsen 2006, Genest and Favre 2007, Zscheischler and Seneviratne 2017, Hao et al 2020).Studies based on copula methodology signified that factors that can influence the interannual correlation coefficient between precipitation and temperature can affect the D-H dependence, such as the local soil moisture-temperature feedback and remote large-scale oceanic-atmospheric circulations (Deng et al 2019, 2020, Mukherjee et al 2020, Feng and Hao 2021).However, physical processes have rarely been discussed in studies based on the ratio of CONDH to heatwave, which is a limitation of this measurement.Second, this study defines the CONDH based on the monthly SPI index.There are indices representing drought at the synoptic scale, such as, the weekly self-calibrated PDSI, which is calculated based on weekly total precipitation, weekly mean temperature, available water content, and potential evapotranspiration (Wells et al 2004).However, not all variables are available in the NEX-GDDP-CMIP6 dataset to calculate the synoptic drought indices.Third, the impact of the D-H dependence on the CONDH over the Asian monsoon region will weaken in response to global warming, especially under the higher emission scenario.Our preliminary result shows that the interannual correlation coefficient between temperature and precipitation during 2046-2065 and 2080-2099 will decrease with respective to 1995-2014 over most Asian monsoon region, which may partly explain the weakened D-H dependence.More efforts should be devoted onto investigations of physical processes responsible for changes in D-H dependence.Finally, we signify that the future changes in intensity, frequency, and duration of CONDH are largely determined by heatwave.However, the key processes involved in changes in heatwave, droughts, and CONDH remain unsolved.Delving into these issues is essential for deeply and comprehensively understanding of the occurrence of CONDH events, which warrants further investigations.
i , DH f , and DH d during 2046-2065 and 2080-2099 under the SSP126 and SSP585 scenarios with those during 1995-2014.The periods of 2046-2065 and 2080-2099 have been widely used to represent mid-term and long-term changes (Eyring et al 2016, Ge et al 2021, Yang et al 2021, Xu et al 2023).

Figure 2 .
Figure 2. Spatial pattern of the (a) DH i , (b) DH f , and (c) during 1961-2014 over the Asian monsoon region.Area-averaged (d) DH i , (e) DH f , and (f) DH d over the eight sub-regions.The stippling in (a)-(c) denotes that the signal-to-noise ratio is greater than 1.The bars (vertical lines) in (d)-(f) represent the MME (STD) from the NEX-GDDP-CMIP6 dataset.

Figure 4 .
Figure 4. Spatial pattern of (a), (b) DH f , and (c) DH d over the Asian monsoon region during 2015-2100 under the SSP126 scenario.Area-averaged (d) DH i , (e) DH f , and (f) DH d during 2015-2100 over each sub-region under the SSP126 and SSP585 scenarios.The stippling in (a)-(c) denotes that the signal-to-noise ratio is greater than 1.The bars (vertical lines) in (d)-(f) represent the MME (STD) from the NEX-GDDP-CMIP6 dataset.

Figure 5 .
Figure 5. Box-and-whisker plots of the changes in the frequency of summer (a) heatwave and (b) CONDH, (c) numbers of drought months, and (d) the DH f during 2046-2065 (blue boxes) and during 2080-2099 (red boxes) with respective to 1995-2014 under the SSP126 and SSP585 scenarios averaged over each sub-region.The change between the two periods is defined as the value in the latter period divided by that in the former period.The box-and-whisker plot shows the 10th, 25th, 50th, 75th, and 90th percentiles of the 20 models from the NEX-GDDP-CMIP6 dataset.
, Zscheischler and Seneviratne 2017, Hao et al 2020, Kong et al 2020).The Asian monsoon region consists of various climate zones with different soil moisture conditions, which would change into arid climate zone in the context of global warming (Peel et al 2007, Dai 2013, Son and Bae 2015).However, the impacts of the D-H dependence on the CONDH over different parts of the Asian monsoon region has been rarely considered, and how the D-H dependence would change in response to global warming remains as an open question.