Abstract
The increasing concurrences of heatwaves and droughts in the context of global warming have attracted much attention from the scientific community given their devastating social and environmental impacts. In this study, the effects of heatwaves in each adjacent week of flash drought onset on the intensification rate of soil moisture were quantified through a meta-Gaussian-based conditional probability model. Results showed that both heatwaves and flash droughts have become more frequent since the middle of the 1990s. For the seasonal distributions, except for the southwestern region where flash droughts lagged behind heatwaves, there was a good synchronization between the two climate extremes. Strong correlations between heatwaves and flash droughts were found in the northeastern, northern, and southwestern regions. Heatwaves with varied timing of emergence behave differently on the formation of flash droughts, along with significant regional differences. Short-term impending hot conditions were crucial for the breakout of flash droughts, especially for the week when flash droughts were initiated, the emergence of heatwaves was likely to increase the intensification rate of soil moisture by 20% compared to those with no heatwaves in their development stage.
HIGHLIGHTS
The composite patterns of concurrent heatwaves and flash droughts are investigated.
Conditional probability is adopted to quantify the effect of heatwaves on drought.
Heatwaves increase the intensification rate of drought events by about 20%.
INTRODUCTION
In the context of global warming, recent years have witnessed growing frequencies of dry and hot events which brought great threats to the ecological environment and human society (Yuan et al. 2015; Naumann et al. 2018; Seneviratne et al. 2021). In addition to the occurrence of each isolated climate extremes alone, the probability of concurrent multiple extremes (e.g., droughts and heatwaves) also increases as a result of globally rising temperatures (AghaKouchak et al. 2014; Ridder et al. 2020; Hao 2022). Numerous observational and modeling studies have shown the enhanced trends of such compound events both in severity and magnitude in many regions around the world (Hao et al. 2018; Zscheischler et al. 2018; Manning et al. 2019; Alizadeh et al. 2020; Geirinhas et al. 2021; Mukherjee & Mishra 2021).
Reasons for the chain process of cascaded climate extremes are complicated, which involve both atmospheric dynamics mechanisms and land–atmosphere feedbacks. For instance, large-scale atmospheric circulation and sea surface temperature anomalies are commonly recognized as the early diagnostic signals for the initiation of droughts and heatwaves (Zhou et al. 2019). Atmospheric blocking which induces significant precipitation and temperature anomalies (Sousa et al. 2017; Lenggenhager & Martius 2019; Zhu et al. 2021) creates favorable weather conditions for the occurrence of heatwaves and drought. Meanwhile, its effects can be spatially variable given the location of blocks occurring in different geographical regions (Röthlisberger & Martius 2019). In addition to synoptic circulation anomalies, local and remote land–atmosphere feedbacks also play a vital role in the simultaneous occurrences of hot and dry extremes, as well as for the intensification of both extremes. For instance, based on observations and atmospheric general circulation model products, the co-variability of temperature and precipitation has been investigated both at global and regional scales with negative correlations detected over land (e.g., Trenberth & Shea 2005; Koster et al. 2009). Soil moisture also strongly modulates near-surface heat and aridity deficits through soil moisture-temperature coupling and soil-precipitation coupling (Seneviratne et al. 2010; Schumacher et al. 2022). The linkages between droughts and heatwaves are still under study, in particular, the influences of heatwaves on the formation of flash droughts have rarely been discussed.
In the absence of a universal definition, flash drought typically refers to a drought with sudden onset and rapid rate of intensification (RI). Given the close relation between hot conditions and flash droughts, the temperature has been adopted by previous studies in their ways of defining and identifying flash drought events. For instance, Mo & Lettenmaier (2015) considered the anomalies of temperature both in precipitation deficits driven and heatwave driven flash droughts, and explicitly illustrated the role of high temperatures in the latter type which leads to evapotranspiration increases and therefore decreases of soil moisture. In addition, the temperature was also implicitly incorporated in some monitoring indices (e.g., the evaporative stress index) for flash drought prediction and forecasting purposes (Otkin et al. 2014; Lorenz et al. 2021). Focused on the physical drivers of flash drought, several studies analyzed the impacts of antecedent meteorological conditions on the depletion of soil moisture (Ford & Labosier 2017; Qing et al. 2022). However, the specific role of hot extremes on flash droughts, with varied timing of emergence during the drying process, is still unclear.
The aim of this study, therefore, is to evaluate to what extent the drying process may be accelerated under heatwaves before and during the onset of flash drought. Investigations were conducted over China on the basis of reanalysis of soil moisture products and temperature observational data. The effects of heatwaves on flash drought were quantified by using both multi-regression model and conditional probability model. The remainder of this paper is organized as follows: Section 2 describes the data used in this study, and the methods for quantifying the effects of heatwaves. Section 3 presents the results, Section 4 discusses the findings, with conclusions drawn at the end.
DATA AND METHODS
Data
ERA-Interim is a global atmospheric reanalysis produced by the European Centre for Medium-Range Weather Forecasts (ECWMF; https://apps.ecmwf.int/datasets/data/interim-full-daily/levtype=sfc/). Following a sequential data assimilation scheme, the ERA-Interim reanalysis uses available observations combined with prior information to produce a forecasting and estimates physical parameters such as precipitation and soil moisture as well (Dee et al. 2011). The soil moisture data from ERA-Interim were employed to identify drought conditions in China. Daily soil moisture estimations of four layers (0–7 cm, 7–28 cm, 28–100 cm, 100–289 cm) are provided at a spatial resolution of 0.75° and are updated on a monthly basis. Other optional interpolated data with different spatial resolutions are also available. In this study, records of the top three layers at 0.25° spatial resolution from 1979 to 2018 were collected, and they were converted into volumetric moisture contents at a depth of 1 m in units of m3/m3. To eliminate the seasonality impact, original daily series were aggregated into weekly values, then converted into soil moisture percentile for every 52 weeks.
The daily maximum temperature was employed to identify heatwaves. Records of 756 meteorological stations evenly distributed in China were obtained from the National Meteorological Science Data Center (http://data.cma.cn/), covering a temporal span from 1961 to 2016. This dataset is released with strict quality control, and is recommended as a priority option for analyzing climate extremes in China. To match the spatial resolution of soil moisture data, the site-based observations were interpolated to gridded data at a spatial resolution of 0.25° by using the Kriging method, as its interpolation accuracy is higher than other interpolation methods such as ordinary nearest neighbor and inverse distance weighting (Lin et al. 2002; Chen et al. 2010).
Methods
Heatwaves identification
Heatwaves broadly refer to periods of days with anomalously warmer temperatures than normal. The variables related to measuring such extreme hot weather can be maximum, minimum or apparent temperature. The critical threshold for representing the normal condition can be fixed thresholds or percentiles, and the warm spell of heatwaves can be consecutive days or cumulative counts of single days above a predetermined threshold (Perkins et al. 2012). All these constitute the diversity of heatwave definitions. This study focused on the impacts of heatwaves on soil moisture depletion, and the extreme hot condition in the daytime is more important for revealing such heat effects. Therefore, the daily maximum temperature (Tmax), combined with the relative thresholds was employed for identifying heatwaves given the vast territory of the research domain. Following the common methodological definition, heatwave periods of at least three consecutive days are detected when anomalies of maximum temperature exceed the 90th percentile of Tmax for each calendar day. The 90th percentile is derived from the probability density function computed by using the maximum temperature data averaged on a 15-day moving window during the 30-year (1961–1990) climatological period. The thresholds were calculated for each grid cell separately, which allows for comparative analysis between grids in different regions.
Flash drought identification
In the absence of a universal definition, flash drought is frequently characterized by a sudden onset and rapid intensification which has distinguished features from traditionally slowly evolving droughts. Definitions focusing on the rapid soil drying process have been mostly used in previous research (e.g., Hunt et al. 2014; Ford & Labosier 2017; Osman et al. 2021). Following this notion, a quantitative method of measuring the intensification rate was employed to identify flash drought events. According to Otkin et al. (2018), there are two key requirements to recognize a flash drought. One refers to the events that should actually fall into drought and also contain some periods of soil moisture conditions approaching the critical level of vegetation moisture stress. This can be controlled by two thresholds associated with soil moisture percentiles, i.e., the upper limits of 40th percentile and the lower limits of 20th percentile, respectively (Svoboda et al. 2002; Liu et al. 2020a). To eliminate the seasonality effect, soil moisture percentiles were estimated separately for each week. Thirteen candidate theoretical probability distributions (including BETA, GAMMA, LOGIST (LOG), Loglog (LLG), Generalized extreme value (GEV), Weibull (WBL), Exponential distribution (EXP), Generalized pareto (GP), NAKAGAMI (NAKA), Birnbaum-Saunders(BIRN), Normal distribution (NORM), RAY, and RICI) were employed to fit soil moisture series, and the optimal distribution was chosen as the one that can pass the Kolmogorov–Smirnov test at the 95% significance level, along with the minimum root-mean-square error (RMSE).
Multivariate linear regression
Conditional probability
Framework of copula-based analysis for estimating the intensification rate of soil moisture conditioned on heatwaves in adjacent weeks under varying scenarios.
Framework of copula-based analysis for estimating the intensification rate of soil moisture conditioned on heatwaves in adjacent weeks under varying scenarios.
RESULTS
Spatial and temporal distribution of heatwaves and flash droughts
Spatiotemporal distributions of heatwaves and flash droughts in China during 1961–2016. (a) Spatial distribution of heatwave frequencies, (b) annual and decadal mean values of heatwave days, (c) spatial distribution of flash drought frequencies, and (d) annual and decadal means of the proportion of flash drought area.
Spatiotemporal distributions of heatwaves and flash droughts in China during 1961–2016. (a) Spatial distribution of heatwave frequencies, (b) annual and decadal mean values of heatwave days, (c) spatial distribution of flash drought frequencies, and (d) annual and decadal means of the proportion of flash drought area.
Weekly distributions of heatwave (HW) and flash drought (FD) occurrences (i.e., the ratio of heatwave/flash drought events in each week to annual average heatwave/flash drought events) in four regions of China during 1961–2016.
Weekly distributions of heatwave (HW) and flash drought (FD) occurrences (i.e., the ratio of heatwave/flash drought events in each week to annual average heatwave/flash drought events) in four regions of China during 1961–2016.
Variation of soil moisture in response to heatwaves
Variations of soil moisture in response to heatwaves: (a) depletion curves of soil moisture during the dry spells between two adjacent precipitation events (t1 represents the first day after the termination of precipitation events) in summer. Light red shadows represent the cases during heatwaves and light blue shadows for non-heatwave days. Blue and red solid lines represent the mean scores of soil moisture. Data were from the grid cell where the Beijing city is located. (b) Average changes of soil moisture in response to heatwaves from 1979 to 2016. Negative values indicate increments of soil moisture during heatwaves, while positive values represent the reduction of soil moisture during heatwaves. Please refer to the online version of this paper to see this figure in colour: http://dx.doi.org/10.2166/nh.2023.022.
Variations of soil moisture in response to heatwaves: (a) depletion curves of soil moisture during the dry spells between two adjacent precipitation events (t1 represents the first day after the termination of precipitation events) in summer. Light red shadows represent the cases during heatwaves and light blue shadows for non-heatwave days. Blue and red solid lines represent the mean scores of soil moisture. Data were from the grid cell where the Beijing city is located. (b) Average changes of soil moisture in response to heatwaves from 1979 to 2016. Negative values indicate increments of soil moisture during heatwaves, while positive values represent the reduction of soil moisture during heatwaves. Please refer to the online version of this paper to see this figure in colour: http://dx.doi.org/10.2166/nh.2023.022.
Correlation between the intensification rate of soil moisture and the proportion of heatwave days in adjacent weeks of flash droughts. Areas with insufficient samples or the correlation failed to pass the significance test were marked blank.
Correlation between the intensification rate of soil moisture and the proportion of heatwave days in adjacent weeks of flash droughts. Areas with insufficient samples or the correlation failed to pass the significance test were marked blank.
Quantify the role of heatwaves on flash droughts
Weighting coefficients of heatwave days in each adjacent week derived from the multivariate regression model in each region. T0 represents the week when flash drought initiates. The blue shadows show the 75th and 25th quantiles of the weighting coefficients in the region.
Weighting coefficients of heatwave days in each adjacent week derived from the multivariate regression model in each region. T0 represents the week when flash drought initiates. The blue shadows show the 75th and 25th quantiles of the weighting coefficients in the region.
(a) Comparisons between empirical and theoretical probabilities for the intensification rate of soil moisture over China. (b) Root mean square errors for 13 candidate probability distributions in four major flash drought-affected regions.
(a) Comparisons between empirical and theoretical probabilities for the intensification rate of soil moisture over China. (b) Root mean square errors for 13 candidate probability distributions in four major flash drought-affected regions.
Increasing ratio of the intensification rate of soil moisture by heatwaves in adjacent weeks of the initiation of flash droughts in four regions. The orange areas represent heatwaves that promote the intensification rate of soil moisture, while the grey areas represent negative effects. The red dots and solid lines for each probability distribution curve show the most probable, and 90% confidence intervals of the increasing ratio values. Please refer to the online version of this paper to see this figure in colour: http://dx.doi.org/10.2166/nh.2023.022.
Increasing ratio of the intensification rate of soil moisture by heatwaves in adjacent weeks of the initiation of flash droughts in four regions. The orange areas represent heatwaves that promote the intensification rate of soil moisture, while the grey areas represent negative effects. The red dots and solid lines for each probability distribution curve show the most probable, and 90% confidence intervals of the increasing ratio values. Please refer to the online version of this paper to see this figure in colour: http://dx.doi.org/10.2166/nh.2023.022.
DISCUSSION AND CONCLUSIONS
Recent studies mostly focused on concurrent climate extremes such as heatwaves and droughts (e.g., Alizadeh et al. 2020; Mukherjee & Mishra 2021; Hao 2022), while limited attention has been paid to the relation between heatwaves and flash droughts. As a new drought type increasingly discussed in the recent 20 years, flash drought is characterized by the sudden onset and rapid intensification, and their relation with hot conditions may be different from traditional slowly-evolving droughts. Trenberth et al. (2014) stated that the increased warming due to climate change may exacerbate the drying process in a quicker and more intense manner. Such reinforcements from the heating condition were also observed in this study. According to Figure 4, it is found that soil moisture under heatwaves decreased more rapidly than in non-heatwave days, meanwhile, the decrement of soil moisture varied over different regions. To investigate the specific roles of heatwaves with the varied timing of emergence, we used two models, i.e., the multivariate regression model and the meta-Gaussian copula model, to quantify such effects. From Figures 6 and 9 we can see that both models revealed similar quantitative results, where heatwaves occurred exactly in the week when flash drought initiated had the strongest forces, and the intensification rate of soil moisture was likely to be accelerated by 20% according to the meta-Gaussian model. This highlights the importance of short-term impending hot conditions for the breakout of flash drought.
Cases in other weeks were complicated, along with significant regional differences. For instance, the positive effects of heatwaves for the southern region covered seven complete preceding weeks. However, for other regions in China, such positive effects of heatwaves were manifested as intermittent signals, combined with negative values at certain intervals. The reasons for the negative values are complicated. From Figure 9 we can see that most negative values before the 2-week lead were close to zero, indicating a weak correlation of heatwaves with flash drought onset. Ford & Labosier (2017) analyzed meteorological conditions at four pentads prior to flash drought onset, and also found a gradually enhanced correlation from 4-pentad to 1-pentad lead. In other words, flash droughts are more relevant to short (no more than 2-week lead) periods of hot conditions in most regions. Besides, negative effects were also found for weeks after the flash drought was initiated. This may be related to the transition between energy-limited and water-limited conditions as a result of persistent reduction of soil moisture under heatwaves. Vegetation also influences the drying process by controlling the rate of evapotranspiration (e.g., Zhang & Zhang 2019; Yin et al. 2021). Liu et al. (2022) found that under water-stressed conditions, vegetation may alleviate droughts, and such alleviating effects differ among different vegetation types. This may in part explain the negative roles of heatwaves in restraining the depletion of soil moisture, and the varied timing of such negative effects of heatwaves after flash droughts initiated in different regions (Figure 9). In the context of global warming, drastic climate change poses great challenges to global security and sustainable development, accelerating the rate of interregional hydrological cycles, leading to significant global extreme events such as droughts or heatwaves (Bian et al. 2022; Sun et al. 2023). Droughts may trigger or exacerbate heatwaves through the effect on surface fluxes, which in turn generate positive feedback to droughts by increasing evaporative demand or reducing precipitation (Tilloy et al. 2019). From the standpoint of land–atmosphere feedback, the potential role of land–atmosphere feedback is in increasing the frequency of compound dry and hot extreme events under climate change (Zscheischler & Seneviratne 2017). The prevailing view is that heatwaves caused by global warming may not cause droughts to occur, but they may make existing droughts faster and more severe (Trenberth et al. 2014). In addition, a modest positive drought-heat effect has been found using the numerical experiments with the Weather Research and Forecasting (WRF) model; however, the evidence of drought-heat-drought-coupled feedback was still lacking (Osman et al. 2022).
The major findings of this study are given as follows: Both heatwaves and flash droughts became more frequent since the middle of the 1990s, and spatially higher frequencies of heatwaves (20–45%) and flash droughts (10–15%) were found in southern China, with low frequencies for both climate extremes in the northwestern region. For the seasonal distributions, except for the southwestern region where flash droughts lagged behind heatwaves, there was a good synchronization between the two climate extremes. Strong correlations between the heatwave days and the intensification rate of soil moisture were found in the northeastern, northern, and southwestern regions, with the values of CC above 0.7. Consistent quantitative results from the multivariate regression model and meta-Gaussian model suggest that short-term impending hot conditions were crucial for the breakout of flash drought, especially for the week when flash drought initiated, the intensification rate of soil moisture under heatwaves was likely to increase by 20% comparing to those with no heatwaves in their development stage. Meanwhile, heatwaves with varied timing of emergence behaved differently on the formation of flash drought, and the persistence of positive effects for accelerating soil moisture decline also differs over different regions. The results highlighted the necessity of recognizing the composite patterns of concurrent heatwaves and flash droughts, which is essential for objective assessment of compound events, and drought adaptation and management strategies.
AUTHOR CONTRIBUTIONS
X.Z. carried out the analyses, prepared the figures and wrote the manuscript. Y.L. prepared the figures and designed the paper. Y.Z. wrote the manuscript and supervised the formulation of this manuscript. Q.M. and G.P. supervised the formulation of this manuscript. Y.Q. and H.Y. prepared the data. All authors discussed the results and contributed to the final paper.
FUNDING
This research has been supported by the National Natural Science Foundation of China (grant nos 41901037, 42171021, and 42071040), the National Natural Science Foundation of Jiangsu Province, China (grant no. BK20220145), and the Central Guidance for Local Science and Technology Development fund projects, under Grant no. 2021ZY0027.
DATA AVAILABILITY STATEMENT
All relevant data are included in the paper or its Supplementary Information.
CONFLICT OF INTEREST
The authors declare there is no conflict.