Impact of anthropogenic warming on emergence of extreme precipitation over global land monsoon area

Human activities have led to a global temperature increase, and the primary objective of the Paris Agreement is to limit this rise to 1.5 °C of warming level. Understanding the impact of global warming beyond preindustrial conditions on precipitation intensity is crucial for devising effective adaptation and mitigation strategies, particularly in densely populated global land monsoon (GLM) regions. However, the time of emergence (ToE) of extreme summer monsoon precipitation and its dependency on global warming targets has rarely been investigated. Using large ensemble simulations forced by the SSP3–7.0 scenario, we reveal that the impacts of anthropogenic forcing on extreme precipitation intensity become evident in GLM regions before 2050, accompanied by a sudden expansion in areas where the ToE of extreme precipitation occurs. Furthermore, our study demonstrates that achieving the Paris Agreement goal at 1.5 °C of global warming level can prevent the ToE of extreme precipitation in Asian and African monsoon regions. This, in turn, has the potential to halve the number (over one billion) of individuals exposed to extreme precipitation. These findings highlight the urgent need for action to mitigate the risk associated with anthropogenic warming induced climate change.


Introduction
The warmer climate due to rising greenhouse gas concentrations will lead to noteworthy changes in the redistribution of precipitation (Lee and Wang 2014, Wang et al 2020).Moreover, the rise in global temperature will result in more frequent and intense precipitation events (Moon andHa 2020, Ha et al 2020b), leading to severe climatic disasters and economic losses across the globe (Jiang et al 2021, You et al 2022).With a warmer climate, precipitation-related extreme phenomena, such as floods and drought, are expected to become harsh realities affecting human lives (Min et al 2011, Sano and Oki 2022, Satoh et al 2022, Moon et al 2023).Notably, monsoon precipitation serves as a vital freshwater source for over 60% of the world's population (Zhang et al 2018).Consequently, gaining a comprehensive understanding of precipitation changes across the global land monsoon (GLM) is of utmost importance for effective water resource management, disaster mitigation, and reducing the impact of economic damages (Sillmann et al 2017, Wang et al 2021).
The 21st Conference of the Parties (COP21) of the United Nations Climate Change Conference in Paris has proposed the climate target for holding global mean temperature increase to well below 2 • C compared to the preindustrial level with efforts to limit the increase to 1.5 • C. The main concern of society is how precipitation intensity is changing, and how these changes can be attributed to different levels of global warming (Seneviratne and Hauser 2020).
Assessment of the impacts of reducing precipitationrelated hazards by controlling global warming to 1.5 • C compared to 2 • C is essential for mitigation and adaptation plans.Studies have already discussed how limiting global warming to 1.5 • C can reduce the effect of climate change on precipitation (Park et al 2018, Madakumbura et al 2019, Maharana et al 2020, Monerie et al 2021).However, the alterations in both mean and extreme precipitation across GLM in response to anthropogenic climate change remain uncertain within the specified target warming levels.To gain a comprehensive understanding, it is imperative to quantify these changes on the regional monsoon scale (figure S1), which is essential for informed decision-making and climate adaptation strategies.
Here, we investigate the Time of Emergence (ToE) of changes in precipitation intensity over GLM domain (Hawkins and Sutton 2012, Gaetani et al 2020, Dong et al 2022).Identifying the ToE, which is the time when an anthropogenic signal will emerge from the natural variability, provides valuable insight into the detection of regional changes in the hydrological cycle (Giorgi andBi 2009, Gaetani et al 2020).Multi-model simulations from the Coupled Model intercomparison Project (CMIP) or large ensemble simulations indicate that mean and extremely heavy precipitation have already deviated from their quasinatural state in many regions or is likely to do so in the near-future due to anthropogenic forcing (King et al 2015, Martel et al 2018, Nguyen et al 2018, Ha et al 2020a).However, there has been a lack of consideration for ToE of precipitation response to anthropogenic forcing, with a focus on the level of global warming at the regional monsoon domain scale.
In this study, the Community Earth System Model version 2 large ensemble (CESM2-LE) (Danabasoglu et al 2020, Rodgers et al 2021) simulations were mainly used to investigate ToE of mean and extreme precipitation over GLM.We demonstrate that, over most of the GLM regions, the effects of anthropogenic climate change are likely to occur by 2100.This study shows that there are benefits to limiting global warming to 1.5 • C, as this can reduce exposure to the emergence of mean and extreme precipitation over GLM, except the Northern African monsoon.In addition, by examining the exposure of a population to anthropogenic climate impacts, we investigated the social impact of the ToE of summer monsoon precipitation over highly populated GLM regions.The results of this study highlight the urgency for early action to achieve the 1.5 • C global warming target to reduce the likelihood that it is facing abnormal precipitation intensity and related hazards in the summertime.

Reanalysis datasets and model simulations
We distinguish the response of anthropogenic forcing and the internal variability of the climate system using the 100-ensemble simulations of the CESM2-LE (Danabasoglu et al 2020, Rodgers et al 2021) under the CMIP6 historical  and Shared Socioeconomic Pathway (SSP) 3-7.0 scenario (SSP3-7.0;2015-2100).To show robustness of our results, we further utilized 50-ensemble simulations of the Canadian Earth System Model version 5 (CanESM5) (Swart et al 2019) and 11-ensemble simulations of IPSL-CM6A-LR developed at the Institut Pierre-Simon Laplace Climate Modeling Center (Boucher et al 2020), which applied the same greenhouse gas emission scenario as CESM2-LE.Additionally, we conducted a ToE analysis using the model simulations forced by the SSP1-2.6 scenario.For validation of large ensemble simulation performance, the daily data obtained from the 55 yr Japanese Reanalysis Project (JRA55) (Ebita et al 2011) with 1.25 × 1.25 horizontal resolutions was analyzed for the period 1958-2021.We also utilized observed daily precipitation datasets from the Global Precipitation Climatology Centre (GPCC) (Schamm et al 2014) for 1982-2020, and the Rainfall Estimates on a Gridded Network (REGEN) (Contractor et al 2020) for 1958-2016.Since all datasets have different spatial resolutions, they were interpolated into 1.25 • 0.9375 • in longitude and latitude.Considering the strong seasonality of monsoon precipitation, our analysis focused only on summertime (from May to September (MJJAS) for the Northern Hemisphere and from November to March (NDJFM) for the Southern Hemisphere).

Definition of the GLM domain
Following Ha et al (2020b), the GLM domain was defined as the region where (1) the annual amplitude of precipitation was larger than 2 mm day -1 , and (2) a greater summertime precipitation intensity resulting from the annual cycle of precipitation compared to the semi-annual cycles (supplementary note 1).The first criterion emphasizes the annual cycle of precipitation, and the second criterion represented the seasonal contrast of the precipitation amount in the harmonic analysis concept.This definition aligns well with the previous GLM definitions ( Wang andDing 2006, Wang et al 2012).In addition, due to the relative insensitivity of harmonic analysis to noise, we could capture useful signals regardless of noise interference (Ha et al 2020b).The global summertime monsoon domain was further divided into nine distinct regional monsoon domains, each character-ized by specific boundaries (figure S1).The regional monsoon domains include the North African (NAF; 60 and South American (SAM; 280

ToE of mean and RX1day
The ToE represents the time when the response to anthropogenic climate change becomes distinguishable from the inherent variability of the climate system.To explore the ToE for both mean and extreme precipitation, we employed the signal-to-noise ratio (S/N).Based on the large ensemble simulations, we defined the signal (S) as the trend observed in the 100-ensemble mean of a given variable, and the noise (N) as the standard deviation across the trend of individual ensemble members (Chung et al 2019, Ha et al 2020a).We selected the year 1958 as the reference year of the precipitation trend because JRA55 reanalysis datasets have been available since that year.The S/N analysis was conducted at each grid point, starting from 1958, and subsequently extending the target trend period by one year.For instance, the analysis periods progressed from 1958-1987-1958-1988, extending finally to 1958-2100.To determine the ToE, we identified the first year in which the S/N surpasses a threshold value, and the S/N remains consistently above this threshold.When employing a threshold value of S/N set at one, ToE patterns for the summertime mean precipitation (hereafter Mean) closely aligned with findings from previous studies (Martel et al 2018, Ha et al 2020a).Consequently, we adopted the threshold (S/N = 1) for the RX1day analysis.ToE of RX1day represents the year in which the index first exceeds the natural variability, and the exceedance remains steady.Additionally, the S/N value serves as a valuable metric to assess the degree of climate deviation, ranging from familiar conditions to those classified as 'unusual (S/N > 1)' , 'unfamiliar (S/N > 2)' , 'unknown (S/N > 3)' , and 'inconceivable (S/N > 5)' (Frame et al 2017, Hawkins et al 2020).

Two-phase linear regression
To identify the abrupt expansion time of ToE (circle symbol in figure 3), we applied a two-phase linear regression analysis to the cumulative ToE fraction (figure S2).This analysis utilized two linear regression models, representing the cumulative fraction of ToE before and after the change point, respectively.The α and β indicated the regression coefficients, t was the years of the study period, and n indicated the length of the time series.As we investigated the changes in ToE from 2000 to 2100, n was 101 in this study.The optimal change point (m), which indicates when ToE begins to expand rapidly, was determined by identifying the statistically significant maximum value of α 2 − α 1 at a confidence level of 95%.For each regression model utilized to calculate the optimal change point, the length of the time series exceeded 30 years to ensure robustness and reliability.

Global mean temperature rise (∆GMT) and exposed population
The global mean temperature (GMT) was calculated using a 31 yr moving average centered on the analyzed year, based on individual ensemble members.The rise in GMT (∆GMT) was obtained from the difference between the 31 yr moving averaged GMT and the average of GMT during the preindustrial period (1850-1900), as illustrated in figure S3.The ensemble mean of the ∆GMT was used when the ToE is converted to ∆GMT.According to CESM2-LE under the SSP3-7.0scenario, ∆GMT is projected to reach 1.5 • C in 2030 and 2 • C in 2044.Note that the year corresponding to specific global warming levels depends on the emission scenarios.
For population data, we used the gridded population count dataset from the Global 1-km Downscaled Population Base Year and Projection Grids based on the SSP3 (regional rivalry) v1.01 (Gao 2017(Gao , 2020)), interpolated to 1.25 • × 0.9375 • resolution.The exposed population is defined as the sum of individuals in each grid exposed to ToE of precipitation at different ∆GMT levels.We aggregated these results for each regional monsoon domain.Since the dataset is provided at decade intervals, this study extrapolated population figures for 2030 and 2040 to represent the exposure levels at 1.5 • C and 2 • C global warming levels, respectively.

Projection of changes in GLM precipitation
We focus on examining the mean and extreme precipitation intensity during the summertime over GLM domains due to seasonality of monsoon precipitation.Specifically, the extreme precipitation index was derived as the daily maximum precipitation intensity (RX1day) observed during the summertime.To assess the performance of model projections, we compared simulated precipitation with observed and reanalysis datasets, particularly evaluating the climatology (figure S4-S6) and interannual deviation (figure S7) of Mean and RX1day from 1958-2021.In CESM2-LE simulations, both Mean and RX1day exhibit a pattern correlation coefficient (PCC) of up to approximately 0.7 with the GPCC, REGEN, and JRA55 datasets, although RX1day is slightly overestimated.Moreover, CESM2-LE demonstrates favorable simulations of the standard deviation of precipitation indices in comparison to other large ensemble simulations (figure S7).The precipitation bias has been improved in CESM2 as compared to its previous version (Danabasoglu et al 2020).Given this improvement, this study primarily relies on CESM2-LE to demonstrate the anthropogenic impacts on changes in precipitation intensity over GLM domains.
Changes in both Mean and RX1day between the historical  period and the end of 21st century (2071-2100) were examined (figure 1).A significant increase in Mean intensity is projected for the NAF and Asia monsoon domains, whereas a decrease is projected for the American monsoon domains (figures 1(a) and S8).The enhanced precipitation in warm climate can primarily be attributed to thermodynamic factors within the moisture budget equation (Endo and Kitoh 2014, Moon and Ha 2020).On the other hand, the weakened intensity of Mean over the American monsoon domains can be linked to the gradient of sea surface temperature between the El Nino-like eastern Pacific and relatively cool Atlantic oceans (Lee and Wang 2014, Wang et al 2020), as well as the increase in evapotranspiration (Satoh et al 2022).
A substantial increase in RX1day is projected across the overall monsoon domains, except for NAM (figure 1(b)).The time series of the regionally averaged RX1day show a clear intensification of extreme precipitation compared to the historical period  (figure S8).In the far future (2071-2100), the GLM domain-averaged intensity of RX1day is projected to increase by 23%.This projected increase in extreme precipitation across GLM raises concerns about the possibility of encountering unprecedentedly heavy precipitation.In addition, it is important to recognize that the impacts of this projected intensity increase and its associated timing will differ across various monsoon domains shaped by factors such as population density, vulnerability, and resilience (Byers et al 2018, Zhang et al 2018).Thus, adopting a tailored regional approach is crucial for effective climate mitigation policies.

Timing of emergence of change in precipitation intensity
To demonstrate the anthropogenic impacts on the changes in intensity of precipitation in comparison to natural variability, we conducted a ToE analysis.This approach allowed us to capture the emergence of unusual precipitation intensity compared to the historical period.By the year 2100, the forced signal of Mean induced by the anthropogenic effects is detected in all regional monsoon domains, except for most of the AU (figure 2).ToE of Mean occurs in whole regional monsoon domains within the mid-future (2041-2060).In the case of IND and NAF, ToE of Mean is expected to occur earlier (before 2020) along with the increasing trend of Mean.It can be attributed to amplified land-ocean thermal contrast, which leads to a rise in sea level pressure in the Pacific region A substantial intensification of extreme precipitation is expected across the GLM by 2100.Unlike Mean intensity which shows regional variations, all regional monsoon domains are projected to experience enhanced heavy precipitation intensity in the future.This increase can be attributed to the higher water-holding capacity of the atmosphere in warm climate (Wang et al 2020, Ha et al 2020b), leading to an escalation in the intensity of extremely heavy precipitation (Emori andBrown 2005, Pfahl et al 2017).In several regions such as ICP, EA, SAF, and AU, the ToE for RX1day will occur earlier (by the year 2050) compared to ToE for Mean.ToE of RX1day comes with relatively lower uncertainty than that of the Mean (figure 2(c)).
The response of regional Mean and RX1day intensity to anthropogenic forcing obtained from the CESM2-LE is comparable to the spatial patterns of ToE for precipitation intensity in CanESM5-LE and IPSL-CM6A-LR-LE simulations (figures S9 and S10).In CanESM5-LE, the ToE for Mean is projected to occur in all regions of IND and EA by 2100 (figure S9).The ToE for Mean in NAF and EA is projected to occur before 2020.Moreover, the ToE for RX1day exhibits overall earlier emergence compared to other models.This can be attributed to the rapid global temperature rise since the preindustrial periods in CanESM5-LE, reaching 1.5 • C by 2010 (figure S3).The IPSL-CM6A-LR-LE simulations project 1.5 • C warming around 2020, while the CESM2-LE model shows a more gradual approach, reaching it around 2030.These results indicate warmer climate leads to earlier occurrence of ToE induced by anthropogenic forcing.Consequently, in CanESM5-LE, IPSL-CM6A-LR-LE, and CESM2-LE simulations, ToE of extreme precipitation in GLM is projected to manifest in 2025, 2037, and 2045, respectively.Despite slight variations in timing of ToE across the models, the distinction concerning vulnerable areas remains consistent.Furthermore, our analysis on the simulations under the SSP1-2.6 scenario, a sustainable pathway (figure S11 and supplementary note 2) shows a reduction in the occurrence rate of ToE for Mean and RX1day in GLM by about 6.5% and 15.1%, respectively.
The ratio of the area exhibiting sensitive response in precipitation intensity to anthropogenic forcing varies over time within each regional monsoon domain (Moon andHa 2020, Pokhrel et al 2021).We calculated the timing of the abrupt expansion of the sensitive region due to anthropogenic forcing based on a two-phase linear regression (figure 3).In terms of ToE of Mean, EA and SAF are likely to experience a rapid expansion of sensitive areas, even before ∆GMT rises by 1.5 • C. EA is projected to have high precipitation sensitivity due to rising temperature  2023).ToE of RX1day area is expected to expand rapidly in the near-future (2021-2040) across all regional monsoon domains except for ICP.The enhancement of extreme precipitation can be attributed to increased atmospheric moisture content driven by thermodynamics (Donat et al 2016) and is closely linked to changes in GMT (figure S3).As ∆GMT reaches 2 • C, the areas affected by anthropogenic forcing steadily increase.In a warmer climate at 3 • C of ∆GMT, most monsoon regions have already responded to anthropogenic impact.Although the median ToE of precipitation changes is expected to occur in the mid-future (figure 2), precipitation begins to respond rapidly to global warming much earlier, even before ∆GMT increase of 1.5 • C (figure 3).
In terms of GLM, the signals of anthropogenic forcing for Mean and RX1day are projected to reach 72% and 87% of the cumulative fraction by the end of the 21st century, respectively (figure 3(i)).While anthropogenic effects on Mean in GLM gradually expand over time, the areas where robust anthropogenic forcing affects RX1day compared to internal variability have expanded rapidly since 2022.By 2100, more than 75% of the regional monsoons will be affected by the enhanced RX1day due to anthropogenic forcing, except for the NAM.Thus, these results emphasize the importance of efforts to mitigate greenhouse gas emissions to prevent or delay extreme precipitation-related hazards.

Global warming level to slow down the ToE of precipitation change
The changes in the intensity of precipitation are mainly caused by temperature rise (Trenberth et al 2003).In other words, we can estimate the potential ∆GMT level from the emergence of unusual conditions in Mean and RX1day (Zhang et al 2018, Satoh et al 2022, Guo et al 2023).Figures 4(a Categorizing the climate target level of ∆GMT associated with ToE of both Mean and RX1day reveals that NAF is likely to suffer changes in the intensity of Mean and RX1day within 1.5 • C of ∆GMT (figure S12).
SAF and IND are projected to experience ToE of both indices in response to anthropogenic impacts below 2 • C level which is the target of the Paris Agreement.To preserve lower ∆GMT is important to reduce the likelihood of risk arising from future changes in summer precipitation intensity over the GLM region.
Due to population growth and the expansion of anthropogenically affected areas, more people in GLM are exposed to ToE of summertime precipitation under the SSP3-6.0scenario (figures 4(c) and (d)) (Gao 2020, Chen and Sun 2021, Guo et al 2023).At the 1.5 • C warming level, the exposed population to ToE of Mean accounts for 1.0 billion (12% of the world's population), while RX1day accounts for 900 million (10% of the world's population).When we set the target level as 2 • C of ∆GMT, 1.3 and 1.9 billion populations are likely to expose the changes in Mean and RX1day in response to the anthropogenic impacts, respectively.Achieving the 1.5 • C goal can prevent approximately 100 million people from being exposed to the intensification of the Mean in SAF caused by anthropogenic forcing.Furthermore, the differential impact between 1.5 • C and 2 • C warming conditions is likely to lead to intensification of RX1day, particularly over SAF, IND, ICP, and EA, resulting in flooding disasters and dramatic economic losses (Shi et al 2021).These results show the vulnerability that approximately 21% of the total population is exposed to unfamiliar wet conditions with ToE of RX1day at 2 • C warming level of ∆GMT.Our findings further imply that achieving the Paris Agreement by limiting the temperature rise below 1.5 • C compared to the preindustrial level can prevent 11% of the world's population from experiencing abnormally enhanced RX1day over GLM.

Discussion
This study unveils that anthropogenic activity has discernible impacts on changes in the intensity of Mean and extremely heavy precipitation during the summertime in most GLM domains by the end of the 21st century.We highlight the emergent expansion of sensitive regions that are likely to experience ToE due to anthropogenic forcing.By 2040, monsoon regions responding to anthropogenic precipitation changes will explosively increase under the SSP3-7.0scenario.Although uncertainties such as internal variability and model uncertainty (figures S9 and S10) still exist (Hawkins and Sutton 2011), our findings shed light on the early emergence of anthropogenic impact on both mean and extreme precipitation intensity over GLM.
Under the SSP3-7.0scenario, changes in precipitation intensity due to anthropogenic forcing are already affecting some parts of the monsoon regions, even before reaching the 1.5 • C warming level of ∆GMT.At a warming level of 2 • C, the population exposed to extreme precipitation intensity is twice that of the population exposed to extreme precipitation in a 1.5 • C warming climate.Around a billion people living in NAF, SAF, and IND are likely to experience both unusual mean and extreme precipitation conditions at the 2 • C level of ∆GMT (figures 4 and S12).NAF is projected to be exposed to enhanced extreme precipitation due to anthropogenic forcing, regardless of the temperature increase.Note that this study only shows an exposed population under the SSP3-7.0scenario.The changes in the exposed population are subject to the emission scenarios (Xu et al 2022).However, limiting global warming to 1.5 • C can lead to a reduction in exposure to precipitation extreme events across GLM under all SSP scenarios (Li et al 2020, Shi et al 2021).Therefore, regionspecific mitigation strategies against climate change are urgently needed.
To project how far the climate is likely to shift from the historical range, we can set the S/N threshold for ToE analysis to higher values (Frame et al 2017, Hawkins et al 2020).Even at a higher S/N threshold of 2, ToE of precipitation remains similar to their ToE with unusual (S/N = 1) (figures 2 and S13).This result suggests an urgent need for mitigation and adaptation strategies in several monsoon domains, including ICP and African monsoons which are expected to suffer inconceivable climate (S/N > 3).
The COP27 in Egypt agreed to create a fund to compensate developing countries for climate change 'loss and damage' .To mitigate the risk of climate change due to human activity, we should also study other indicators of vulnerability, such as water demand and human capital (Mackay 2008).Despite calls to prioritize climate action, current global efforts are insufficient to achieve the 1.5 • C warming target (Matthews and Wynes 2022).This study highlights the importance of adaptation and mitigation strategies for water resources and precipitation-related hazards considering societal aspects.It provides valuable insights into anthropogenic impact on unfamiliar future climate.are accessible at https://sedac.ciesin.columbia.edu/data/set/popdynamics-1-km-downscaled-pop-baseyear-projection-ssp-2000-2100-rev01.

Figure 1 .
Figure 1.Projected changes in mean and extreme precipitation under climate change during the summertime.Spatial distribution of ensemble median values of the changes (%) in (a) Mean and (b) RX1day in the end of the 21st century (2071-2100) compared to the historical period (1958-1987) based on the CESM2 large ensemble simulations.Cross symbol denotes statistically nonsignificant at the 95% confidence levels (Student's t test).Black lines indicate the global monsoon domains.The results were obtained during the summertime (MJJAS in the Northern Hemisphere, and NDJFM in the Southern Hemisphere).

Figure 2 .
Figure 2. Time of Emergence (ToE) for mean and extreme precipitation estimated by 100-ensemble CESM2-LE projections.ToE (year) of (a) Mean and (b) RX1day based on CESM2-LE during the summertime.The colors indicate both ToE and the direction of the trend of precipitation index.Black lines denote the regional monsoon domains.(c) ToE of Mean (black) and RX1day (red) depending on the regional monsoon domains.Marks indicate the ToE based on the 100-ensemble mean and the error ranges present one standard deviation from the individual 100-ensemble simulations.The triangle and inverted triangle denote the increasing and decreasing dominant trend within the domains, respectively.The closed (open) marks denote the rate at which the ToE appears more (less) than two-thirds of the monsoon domain.
and a decline in sea level pressure over the Eurasian continent (Lee and Wang 2014).These spatial patterns prompt the moisture transport from the Pacific Ocean to the Asian monsoon domain.Moreover, the intensified hemispheric thermal contrast characterized by a warmer NH and cooler SH contributes to an increase in Asian monsoon precipitation and weakened AU monsoon precipitation (Sun et al 2010, Wang et al 2014).In addition, the increased atmospheric moisture in warm climate driven by greenhouse gas forcing contributes to the intensification and earlier occurrence of ToE for the Mean (Ha et al 2020a).

Figure 3 .
Figure 3. Cumulative fraction of the total area with Time of Emergence (ToE) for mean and extreme precipitation.The cumulative fraction (%) of the total area with ToE (year) of change signals from the CESM2-LE over regional monsoon domains during the summertime.Black and red lines indicate the Mean and RX1day, respectively.Shaded area denotes one standard deviation of individual 100-ensemble simulations.Circle denotes the time in which the cumulative fraction increases abruptly using the two-phase linear regression method.Dashed vertical lines indicate the +1.5 • C, +2 • C, and +3 • C global mean temperature rise compared to the preindustrial period (1850-1900).

Figure 4 .
Figure 4. Global mean temperature rise and exposed population corresponding to Time of Emergence (ToE) for mean and extreme precipitation.(a), (b) Global mean temperature rise (∆GMT, • C) compared to the preindustrial period (1850-1900) which corresponds to the ToE of (a) Mean and (b) RX1day.The ∆GMT is derived from the average of individual 100-ensemble simulations.Black lines denote the regional monsoon domains.Hatch represents the regional monsoon domain where ToE appears in less than two-thirds of the area.(c, d) Exposed population (10 8 persons) in which the median ToE of (c) Mean and (d) RX1day is earlier than or at the same time as that of ∆GMT < +1.5 • C and +2 • C. Color bars indicate the exposed population in each regional monsoon domain.
) and (b) illustrate the ∆GMT levels corresponding to the median ToE values for Mean and RX1day are shown.Even at relatively low ∆GMT levels (<1.5 • C), some regions may witness changes in precipitation due to the anthropogenic effect.Nevertheless, adhering to climate mitigation targets of 1.5 • C or below 2 • C of ∆GMT can help reduce precipitation-related disasters and mitigate the impacts of anthropogenic forcing on Mean and RX1day.The ToE of RX1day appears to be particularly sensitive to global warming levels.The median ToE values for RX1day in IND, ICP, EA, and SAF correspond to ∆GMT values between 1.5 • C and 2 • C levels under the SSP3-7.0scenario (figure 4(b)).