The features of regional flash droughts in four typical areas over China and the possible mechanisms

Understanding the effects of flash drought on vegetation photosynthesis and potential drivers over China

Flash droughts characterized by rapid onset and intensification are expected to be a new normal under climate change and potentially affect vegetation photosynthesis and terrestrial carbon sink. However, the effects of flash drought on vegetation photosynthesis and their potential dominant driving factors remain uncertain. Here, we quantify the susceptibility and response magnitude of vegetation photosynthesis to flash drought across different ecosystems (i.e., forest, shrubland, grassland, and cropland) in China based on reanalysis and satellite observations. By employing the extreme gradient boosting model, we also identify the dominant factors that influence these flash drought-photosynthesis relationships. We show that over 51.46 % of ecosystems across China are susceptible to flash drought, and grasslands are substantially suppressed, as reflected in both sensitivity and response magnitude (with median gross primary productivity anomalies of -0.13). We further demonstrate that background climate differences (e.g., mean annual temperature and aridity) predominantly regulate the response variation in forest and shrubland, with hotter/colder or drier ecosystems being more severely suppressed by flash drought. However, in grasslands and croplands, the differential vegetation responses are attributed to the intensity of abnormal hydro-meteorological conditions during flash drought (e.g., vapor pressure deficit (VPD) and temperature anomalies). The effects of flash droughts intensify with increasing VPD and nonmonotonically relate to temperature, with colder or hotter temperatures leading to more severe vegetation loss. Our results identify the vulnerable ecological regions under flash drought and enable a better understanding of vegetation photosynthesis response to climate extremes, which may be useful for developing effective management strategies.

Enhancing drought resilience: machine learning-based vulnerability assessment in Uttar Pradesh, India

Drought is a natural and complex climatic hazard. It has both natural and social connotations. The purpose of this study is to use machine learning methods (MLAs) for drought vulnerability (DVM) in Uttar Pradesh, India. There were 18 factors used to determine drought vulnerability, separated into two groups: physical drought and meteorological drought. The study found that the eastern part of Uttar Pradesh is high to very highly prone to drought, which is approximately 31.38% of the area of Uttar Pradesh. The receiver operating characteristic curve (ROC) was then used to evaluate the machine learning models (artificial neural networks). According to the findings, the ANN functioned with AUC values of 0.843. For policy actions to lessen drought sensitivity, DVMs may be valuable. Future exploration may involve refining machine learning algorithms, integrating real-time data sources, and assessing the socio-economic impacts to continually enhance the efficacy of drought resilience strategies in Uttar Pradesh.

Global eight drought types: Spatio-temporal characteristics and vegetation response

The traditional classification of drought events into seasonal and flash types oversimplified the complexity and variability of global drought phenomena, limiting a deeper understanding of drought characteristics and their impacts on vegetation. To address this issue, soil moisture percentile methods and the Soil Moisture Anomaly Percentage Index (SMAPI) were employed to create time series for flash drought (FD) and seasonal drought (SD) events globally from 1981 to 2020. A novel categorization framework was proposed to subdivide the two basic drought categories into eight distinct drought types using a set relationship identification method. The results showed fluctuating trends in the frequencies of Independent FD and Inclusion FD, which declined rapidly after 2011 at rates of 0.05 and 0.04 times/year, respectively. Independent FD frequency was highest in humid areas and decreased with increasing aridity. The spatial distributions of Inclusion FD and SD were similar, with both frequencies highest in extremely arid areas and decreasing with increasing humidity. The frequency of Independent SD, which peaked in semi-arid areas, increased significantly after 2011 at a rate of 0.01 times/year. The occurrence of FD evolving into SD or emerging at the end of SD was rare, with a global average of 0.46 events/decade and little spatial variation. Between 1981 and 2020, FD showed a U-shaped trend in drought duration, while SD showed no clear pattern. The duration of FD showed little difference across arid and humid zones, but the duration of SD decreased significantly with increasing humidity. Vegetation responses to drought varied, with arid regions showing longer response time compared to humid regions. A positive correlation between temperature and solar-induced chlorophyll fluorescence (SIF) during droughts was observed, while precipitation generally showed a negative correlation with SIF. Radiation had a minimal effect on SIF during droughts. The study offered a comprehensive categorization of drought events, enhancing our understanding of their spatiotemporal characteristics and vegetation responses on a global scale.

Re-intensification of flash drought in western China over the past decade: Implications of fluctuating wetting trend

Climate changes and human activities have led to a rise of frequency and intensity of the global flash droughts, resulting in severe consequences for ecosystems, agriculture, and human societies. However, research dedicated to flash droughts in the dryland of western China is relatively limited, leaving their evolutionary characteristics and development processes of these phenomena unclear. To bridge this gap, this study analyzed the spatiotemporal characteristics of flash droughts in western China from 1981 to 2020, based on the standardized evapotranspiration stress index. Additionally, we investigated the development mechanisms by taking meteorological conditions and soil moisture into account. The findings revealed that the northern Qinghai-Tibet Plateau, western Qilian Mountains, and western and southern Loess Plateau are hotspots of flash droughts, characterized by rapid development rates. Across most of the study area, flash drought events persisted between 25 and 30 days. Adequate precipitation is necessary before the onset of flash droughts in western China, while water scarcity and high temperatures played crucial roles in driving the mid-stage of flash droughts. Within the context of the observed "warming and wetting" trend, the average flash droughts occurrence from 2011 to 2020 was approximately 16 % lower than that from 1981 to 1990, and there was a significant annual decrease in spatial coverage of 0.01 % per year. However, in the "wetting in west, drying in east" trend, the spatial coverage of flash droughts has shifted from a declining trend to an insignificant increasing trend since 2000 in the study area, with significant regional differences between the western and eastern regions. Over the past decade, flash droughts had once again intensified in the central Qinghai-Tibet Plateau and the Loess Plateau due to warming and fluctuating wetting trends, raising significant concerns for future ecosystem and agricultural water management in these regions.

Characteristics of flash droughts and their association with compound meteorological extremes in China: Observations and model simulations

Flash droughts have gained considerable public attention due to the imminent threats they pose to food security, ecological safety, and human health. Currently, there has been little research exploring the projected changes in flash droughts and their association with compound meteorological extremes (CMEs). In this study, we applied the pentad-mean water deficit index to investigate the characteristics of flash droughts and their association with CMEs based on observational data and downscaled model simulations. Our analysis reveals an increasing trend in flash drought frequency in China based on historical observations and model simulations. Specifically, the proportion of flash drought frequency with a one-pentad onset time showed a consistent upward trend, with the southern parts of China experiencing a high average proportion during the historical period. Furthermore, the onset dates of the first (last) flash droughts during year are projected to shift earlier (later) in a warmer world. Flash droughts become significantly more frequent in the future, with a growth rate approximately 1.3 times higher in the high emission scenario than in the medium emission scenario. The frequency of flash droughts with a one-pentad onset time also exhibits a significant upward trend, indicating that flash droughts will occur more rapidly in the future. CMEs in southern regions of China were found to be more likely to trigger flash droughts in the historical period. The probability of CMEs triggering flash droughts is expected to increase with the magnitude of warming, particularly in the far-future under the high emissions scenario.