Drought In China Research Articles

Transcriptome analyses revealed chilling response genes in mango (Mangifera indica L. cv. Keitt) leaf

ABSTRACT Mangifera indica L. cv. Keitt is a cold-stress fruit plant native to China's drought river valley. Chilling stress affects productivity. Understanding the mechanisms of chilling stress is important to increasing chilling resistance in mango. Leaves of Keitti were subjected to 4 °C for 0, 3, 6, and 9 h for RNA-Seq-based transcriptome analysis, respectively. The chlorophyll content, carotenoid content, catalase, and peroxidase activities significantly increased during 9 h. The leaves responded to the stress by enhancing photosynthetic pigment content and antioxidant enzyme activity. After 3 h of chilling, 410 genes were differentially expressed. WRKY70 and PLD1 were significantly up-regulated after 9 h. Compared to 9 and 0 h, there were 1123 DEG. The DEGs are enriched in hormonal signal transduction, secondary metabolites, and the abiotic stress response. Similarly, the transcriptional factor families including NCED2, MYB73, and HLH162 up-regulated. The study will promote research on the development of chill-resistant mangoes.

Satellite-derived aridity index reveals China's drying in recent two decades

The expansion of dryland has caused a huge impact on the natural environment and human society. Aridity index (AI) can effectively reflect the degree of dryness, but spatiotemporally continuous estimation of AI is still challenging. In this study, we develop an ensemble learning algorithm to retrieve AIs from MODIS satellite data in China from 2003 to 2020. The validation proves the high match between these satellite AIs and their corresponding station estimates with a root-mean-square error of 0.21, bias of -0.01, and correlation coefficient of 0.87. The analysis results indicate China has been drying in recent two decades. Moreover, the North China Plain is undergoing an intense drying process, whereas the Southeastern China is becoming significantly more humid. On the national scale, China's dryland area shows a slight expansion, while the hyper arid area has a decreasing trend. These understandings have contributed to China's drought assessment and mitigation.

Bayesian vine copulas improve agricultural drought prediction for long lead times

Drought prediction models generally focus on shorter lead times (1–3-months) as their performance drastically declines at longer lead times (> 3 months). However, reliable agricultural drought prediction model with longer lead times is fundamental for reducing the impacts on agriculture sector, ecosystem, environment, and water resources. We propose a novel agricultural drought prediction model for long lead times by integrating vine copulas with Bayesian model averaging (hereafter, BVC model). Considering the previous meteorological drought, antecedent hot condition, and agricultural drought persistence as three predictors, the BVC model predicts agricultural drought with 1–6-month lead times. Here we focus on summer season (e.g., August) drought in China where agricultural drought impacts are more pronounced. Compared with optimal vine copula (OVC), average vine copula (AVC), and persistence-based models, the proposed BVC model performs better for 1–6-month lead times. Our findings can improve agricultural drought management, food security assessment, and early drought warning.

Spatiotemporal Characteristics of Meteorological and Agricultural Droughts in China: Change Patterns and Causes

Drought is complex and is also one of the main disasters affecting China. Exploring the response of agricultural drought and meteorological drought to climatic factors helps us to understand the causes of drought. In this paper, we evaluated the temporal and spatial characteristics of soil drought and meteorological drought (SMD) and explored their responses to climatic factors and latent heat fluxes (LHF), and then explained their variation from the perspective of atmospheric circulation. The following results were obtained. (1) Meteorological drought has gradually increased in the Liaohe River Basin, on the North China Plain, and on the Loess Plateau while average soil moisture has been maintained at only approximately 25%. The impacts of drought are very serious in these regions. (2) LHF response to short time-scale (3-month scale) drought performance is high in the dry season, and the regions with high correlation coefficients are spatially distributed and concentrated in the monsoon climate zone. The regions with high correlation coefficients between drought and LHFs on long time scales (12-month scale) are concentrated in the coastal basin of southeast China. (3) Short- and long-term SMDs showed highly responsive and significant relationships with PDO, showing variations in the southeast coastal basin, the Pearl River basin, the northwest inland basin and the eastern part of the Heilongjiang basin, with a maximum correlation coefficient of 0.21 (p < 0.01). The short-term SMD in the northwestern inland region was significantly negatively correlated with AMO (correlation coefficient of −0.19, p < 0.01). the Nino3.4 index is significantly positively correlated with the SMD in the southeast coastal region of China, with a maximum correlation coefficient of 0.23 (p < 0.01). The decrease in convective precipitation led to a stronger association between soil and meteorological drought and climatic factors. This study helps to reveal the changing patterns of SMDs and can also be used globally to identify the local development patterns of drought under climate change.

Spatiotemporal dynamic of subtropical forest carbon storage and its resistance and resilience to drought in China.

Subtropical forests are rich in vegetation and have high photosynthetic capacity. China is an important area for the distribution of subtropical forests, evergreen broadleaf forests (EBFs) and evergreen needleleaf forests (ENFs) are two typical vegetation types in subtropical China. Forest carbon storage is an important indicator for measuring the basic characteristics of forest ecosystems and is of great significance for maintaining the global carbon balance. Drought can affect forest activity and may even lead to forest death and the stability characteristics of different forest ecosystems varied after drought events. Therefore, this study used meteorological data to simulate the standardized precipitation evapotranspiration index (SPEI) and the Biome-BGC model to simulate two types of forest carbon storage to quantify the resistance and resilience of EBF and ENF to drought in the subtropical region of China. The results show that: 1) from 1952 to 2019, the interannual drought in subtropical China showed an increasing trend, with five extreme droughts recorded, of which 2011 was the most severe one; 2) the simulated average carbon storage of the EBF and ENF during 1985-2019 were 130.58 t·hm-2 and 78.49 t·hm-2, respectively. The regions with higher carbon storage of EBF were mainly concentrated in central and southeastern subtropics, where those of ENF mainly distributed in the western subtropic; 3) The median of resistance of EBF was three times higher than that of ENF, indicating the EBF have stronger resistance to extreme drought than ENF. Moreover, the resilience of two typical forest to 2011 extreme drought and the continuous drought events during 2009 - 2011 were similar. The results provided a scientific basis for the response of subtropical forests to drought, and indicating that improve stand quality or expand the plantation of EBF may enhance the resistance to drought in subtropical China, which provided certain reference for forest protection and management under the increasing frequency of drought events in the future.

Stepwise cluster ensemble downscaling for drought projection under climate change

AbstractDrought is one of the most serious natural disasters exacerbated by climate change. Changes in precipitation and temperature in the future increase the likelihood of drought in China. In this study, a stepwise cluster ensemble downscaling (SCED) model was developed to bias‐correct projections of temperature and precipitation from multiple RCM outputs, and further characterized the drought hazards. The developed SCED model was used to aggregate and correct the results of multiple regional climate models, and its performance was proved to be reliable by comparing with the observed results. The proposed SCED method has been applied for drought projections over the Fujian province, China. The results showed that the changes of precipitation and temperature in Fujian would have obvious spatial heterogeneous characteristics. The temperature in the southeast coastal areas will increase by up to 4°C and the precipitation will decrease by 3.1% in the late 21st century, while the temperature rises and precipitation increases in the southwest. Temperature in inland areas will be lower and precipitation will be less. The drought hazards were also characterized by both SPI and SPEI based on biased‐corrected projections from SCED model. According to the SPI and SPEI indices, although the number of dry months in Fujian province will not change significantly in future, the spatial and temporal heterogeneity may become more explicit. Moreover, the moderate drought (from SPI) may increase while the general drought may decrease (from both SPI and SPEI). For extreme droughts, there would not be visible changes detected by SPI, but an increasing trend characterized since the impact of temperature was included in SPEI. In addition, there would be an increasing trend on drought when increasing temperature and precipitation occurred simultaneously.

Quantitative Causality, Causality-Aided Discovery, and Causal Machine Learning

It has been said, arguably, that causality analysis should pave a promising way to interpretable deep learning and generalization. Incorporation of causality into artificial intelligence algorithms, however, is challenged with its vagueness, nonquantitativeness, computational inefficiency, etc. During the past 18 years, these challenges have been essentially resolved, with the establishment of a rigorous formalism of causality analysis initially motivated from atmospheric predictability. This not only opens a new field in the atmosphere-ocean science, namely, information flow, but also has led to scientific discoveries in other disciplines, such as quantum mechanics, neuroscience, financial economics, etc., through various applications. This note provides a brief review of the decade-long effort, including a list of major theoretical results, a sketch of the causal deep learning framework, and some representative real-world applications pertaining to this journal, such as those on the anthropogenic cause of global warming, the decadal prediction of El Niño Modoki, the forecasting of an extreme drought in China, among others.

Climatological spatial scales of meteorological droughts in China and associated climate variability

This study investigates the climatological spatial scales (CSSs) of meteorological droughts in China and the linkages to climate variability. The Global Precipitation Climatology Centre monthly gridded precipitation with a spatial resolution of 25 km × 25 km for 1961–2010 is used. The standardized precipitation index at different timescales (1-, 3-, 6-, 9-, 12-, 24-, 36-, and 48-month) is applied to characterize meteorological droughts. The CSSs of meteorological droughts are calculated using a method considering spatial correlation and anisotropy. The relationships between the CSSs and main monsoons and climate teleconnections in five selected regions are quantified using dynamic spatial panel models. The five regions are South China (SC), the Yangtze River Valley (YRV), North China (NC), Northeast China (NE), and the Tibetan Plateau (TIB). The following results are obtained: (1) The CSSs of multi-timescale meteorological droughts in China are determined. The means of the CSSs at various timescales in China are 68.9, 73.8, 71.8, 69.7, 68.8, 65.9, 62.9, and 60.8 × 104 km2, respectively. The CSSs in eastern China generally show a significant decrease with increasing timescale. (2) Significant quantification relationships (R2 greater than 0.85) suggest that there is not a simple linear relationship between climate anomalies and the CSSs, but rather a complex spatiotemporal interaction including exogenous and endogenous interaction effects. The CSSs not only strongly depend on their neighbors on the same timescale but also significantly depend on the CSSs of neighbors on adjacent timescales. Three monsoon indices (Indian monsoon, East Asian monsoon, and western North Pacific monsoon) have significant impacts on the multi-timescale CSS variations in the five regions, especially for NE, YRV, SC, and TIB, while the El Niño–Southern Oscillation and Pacific Decadal Oscillation affect mainly SC and YRV. These findings could broaden our understanding of the spatiotemporal relationships of droughts and be useful for drought risk management.

Deep Reductions in Energy Use: Hobson’s Choice in Climate’s Last-Chance Saloon

In 2022, the record of extreme weather events already includes deep droughts in Sichuan province, China, and California, US; floods inundating a third of Pakistan and serious and repeated flooding in Eastern Australia; heat waves and drought in Europe; and wildfires in Europe and the western US [...]

Quantifying Flash Droughts Over China From 1980 to 2017

AbstractDrought can develop rapidly over a short period, the so‐called “flash drought”. The widely used standardized precipitation evapotranspiration index (SPEI), with traditional time scales longer than 1 month, cannot easily capture flash drought signals. Here, the SPEI with a 5‐day (pentad) time scale was proposed to investigate flash droughts in China from 1980 to 2017. New criteria for flash droughts based on the high‐resolution index were also developed. Flash droughts were stronger and longer in western and northern China. The intensity and duration of mild, moderate, and severe flash droughts increased during the study period, while severe flash droughts showed the largest increase. The high‐resolution SPEI permits us to examine the process of a drought event, that is, how a mild flash drought develops into a severe flash drought. The contribution of precipitation deficit to the duration of mild flash droughts was 50%. As the intensity of flash droughts increased, the contribution of high temperature and increased net radiation to flash drought duration increased from 50% (mild flash droughts) to 74% (severe flash droughts). When flash droughts occurred, Northwestern China and Southwestern China were dominated by unusually weak water vapor transport, while Northeastern China, Northern China, Southern China ,and Middle‐Lower Yangtze were mainly dominated by abnormal downward motion. The downward motion was generally accompanied by clear sky conditions with fewer clouds, higher air temperature, and higher surface solar radiation, which accelerated and amplified flash droughts triggered by initial precipitation deficits.

Warming intensifies severe drought over China from 1980 to 2019

AbstractClimate warming is steadily increasing drought risk in many regions of the world. However, the contribution of warming over the past few decades to drought evolution in China remains unclear. This study used the self‐calibrating Palmer drought severity index (sc‐PDSI) and a detrending experiment to quantify the contribution of warming to severe drought (sc‐PDSI < −3.0) characteristics (intensity, frequency, and duration) over China from 1980 to 2019. Differences in drought characteristics calculated from original and detrended temperature data were attributed to warming effects. The results showed that 93.8% of areas in China have experienced a significant (p < .05) warming trend from 1980 to 2019. Warming remarkably intensified severe drought characteristics across China. Without warming, the extent, frequency, and duration of severe droughts in China would reduce by 2.04%, 2.0%, and 0.82 months. In addition, there are apparent regional differences in the response of severe drought characteristics to warming. Severe drought characteristics in high elevations and high latitudes are more sensitive to warming than in other regions.

How to Manage a World Food Crisis: A Viewpoint

As this is written in mid-September 2022, the world is facing a potential food crisis worse than any since World War II. With a devastating war in Ukraine launched by Russia in late February 2022, a historic drought and heat wave in China, and an uneven monsoon in South Asia, food supplies from several of the world’s largest granaries are highly uncertain at best, and genuinely scary at worst. What should we do?

Divergent seasonal responses of carbon fluxes to extreme droughts over China

Droughts affect the interannual variability of global land carbon fluxes and are expected to exert widespread impacts on the carbon cycle given a future climate with more intense and frequent droughts. Evidence indicates that the impacts of droughts on the carbon cycle vary in different seasons, but a quantitative examination of the seasonal differences is lacking. Here we ensemble multiple data streams for greenness indices and carbon fluxes, including those from remote sensing observations, flux tower upscaling, atmospheric inversions, and dynamic global vegetation models (DGVMs), to quantify the seasonal responses of vegetation to extreme droughts in China. We find that summer droughts cause the largest negative responses of leaf area index (LAI, with median standardized anomalies of -0.40), gross primary productivity (GPP, -0.55), and net ecosystem productivity (NEP, -0.74); notably, droughts in autumn largely suppress carbon uptake. The response patterns show a high degree of heterogeneity, and we identify the factors driving these spatial variations using the extreme gradient boosting (XGBoost) machine learning approach. Climate is the dominant driver of spring and autumn GPP responses while LAI predominantly drives summer GPP loss. Looking into the biotic factors, carry-over effects (previous-season vegetation growth affecting current-season growth) contribute substantially to the alleviation of drought stress, in that previous-season greening compensate vegetation loss during droughts. Our results not only quantify the seasonal response differences in carbon fluxes and greenness, but suggest that carbon fluxes respond more sensitively to drought than greenness. Also, we show seasonal differences in the degree to which factors contribute to drought impacts, which highlight that annual-scale drought analyses may mask spring and autumn vegetation response to droughts.

Impact of ENSO Events on Droughts in China

The El Niño Southe58rn Oscillation (ENSO) is a typical oscillation affecting climate change, and its stable periodicity, long-lasting effect, and predictable characteristics have become important indicators for regional climate prediction. In this study, we analyze the Standardized Precipitation Evapotranspiration Index (SPEI), the Niño3.4 index, the Southern Oscillation Index (SOI), and the Multivariate ENSO Index (MEI). Additionally, we explore the spatial and temporal distribution of the correlation coefficients between ENSO and SPEI and the time lag between ENSO events of varying intensities and droughts. The results reveal that the use of Nino3.4, MEI, and SOI produces differences in the occurrence time, end time, and intensity of ENSO events. Nino3.4 and MEI produce similar results for identifying ENSO events, and the Nino3.4 index accurately identifies and describes ENSO events with higher reliability. In China, the drought-sensitive areas vulnerable to ENSO events include southern China, the Jiangnan region, the middle and lower reaches of the Yangtze River, and the arid and semi-arid areas of northwestern China. Droughts in these areas correlate significantly with meteorological drought, and time-series correlations between ENSO events and droughts are significantly stronger in regions close to the ocean. Drought occurrence lags ENSO events: when using the Niño3.4 index to identify ENSO, droughts lag the strongest and weakest El Niño events by 0–12 months. However, when using the MEI as a criterion for ENSO, droughts lag the strongest and weakest El Niño events by 0–7 months. The time lag between the strongest ENSO event and drought is shorter than that for the weakest ENSO event, and droughts have a wider impact. The results of this study can provide a climate-change-compatible basis for drought monitoring and prediction.

High resistance of deciduous forests and high recovery rate of evergreen forests under moderate droughts in China

The resistance and recovery rate of forest ecosystems to droughts vary with the severity of the drought. Studies on the impacts of severe droughts on forest ecosystems have suggested that evergreen forests have higher resistance than deciduous forests. However, whether the resistance and recovery rate of forest ecosystems vary under moderate droughts remained largely unknown. Here, we used the Standardized Precipitation Index to identify drought characteristics in China’s forest ecosystems from 2000 to 2018. We quantified the resistance and recovery rate to moderate droughts under different drought timings based on the Enhanced Vegetation Index. We then adopted random forest regression to evaluate the relative importance of climatic variables and species richness as drivers of resistance and recovery rate to moderate droughts. We found that China’s forest ecosystems mainly experienced moderate drought events, resulting in different resistance and recovery rate compared to severe droughts. Deciduous forests had high resistance (30.39 ± 21.32) and evergreen forests had high recovery rate (1.38 ± 1.10) due to their different response strategies under moderate droughts. The resistance and recovery rate of China’s forest ecosystems varied under different drought timings. The differences between deciduous and evergreen forests’ resistance (43.88 – 78.24 %) and recovery rate (-31.04 – −52.20 %) were remarkable in spring, autumn, and winter. In evergreen and deciduous forests, climatic variables were the main influencing factors of resistance and recovery rate to moderate droughts, similar to the leading role of climatic variables under severe droughts. Global radiation (41.06 %) and temperature (30.97 %) were the dominant factors contributing to resistance, whereas the recovery rate was primarily explained by global radiation (40.30 %) and precipitation (27.33 %). Thus, the high resistance in deciduous forests and the high recovery rate in evergreen forests under moderate droughts suggest that evergreen and deciduous forests can be mixed to improve the stability of forest ecosystems under climate change.

Reconstruction and application of the temperature-vegetation-precipitation drought index in mainland China based on remote sensing datasets and a spatial distance model

In recent years, remote sensing drought monitoring indices have been gradually developed and have been widely used for global or regional drought monitoring due to their strong drought-monitoring capabilities and easy implementation advantages. However, some defects of remote sensing drought indices stand to be improved due to certain errors in the inversion of surface characteristics by remote sensing datasets. The temperature-vegetation-precipitation drought index (TVPDI) was taken as the research object, and the leaf area index (LAI), the difference between the land surface temperature (LST) and monthly average temperature, and Global Precipitation Measurement (GPM) precipitation data were selected instead of the normalized difference vegetation index (NDVI), LST and tropical rainfall measuring mission (TRMM) data to improve TVPDI. The improved remote sensing drought index was named the improved temperature-vegetation-precipitation drought index (iTVPDI). The drought-monitoring accuracy of iTVPDI was verified by the gross primary productivity (GPP), soil moisture, and crop yield per unit. The drought-monitoring ability of iTVPDI was verified with traditional drought indices, including the standardized precipitation index (SPI), standardized precipitation evapotranspiration index (SPEI), Palmer drought severity index (PDSI), temperature-vegetation drought index (TVDI), drought severity index (DSI) and crop water stress index (CWSI). The drought-monitoring accuracy of iTVPDI was verified by selecting sample areas. iTVPDI was applied to monitor drought in mainland China over the 2001–2020 period and obtained four main results. First, the correlation analyses of iTVPDI and TVPDI with GPP, crop yield per unit area, and soil moisture showed that iTVPDI had a stronger monitoring ability in Northeast, North, and Southwest China; the R2 value obtained with soil moisture was 0.62 (p < 0.05), and this value was higher than that of TVPDI. Then, the correlation analyses of iTVPDI and TVPDI with SPI, SPEI, PDSI, CWSI, DSI and TVDI showed that the correlation coefficients of iTVPDI and TVPDI with these six indicators were basically consistent, which indicated that the drought-monitoring capability of iTVPDI was consistent with that of TVPDI. In local areas such as the Qinghai-Tibet Plateau in China, the monitoring ability of iTVPDI was stronger than that of TVPDI. Third, through the sample area analysis, iTVPDI was found to moderate the NDVI-characterized vegetation factors in TVPDI in low-vegetation-cover areas affected by soil disturbances and in high-vegetation-cover areas affected by oversaturation. Finally, the results obtained from the application of iTVPDI in mainland China showed that during the warm-dry to warm-wet climate transition between 2001 and 2021, in 2010 and 2018, and in other special drought years, iTVPDI had the best response.

Assessing the Drought Variability in Northeast China over Multiple Temporal and Spatial Scales

Long-term drought variation provides a scientific foundation for water resource planning and drought mitigation. However, the spatiotemporal variation characteristics of drought in northeast China (NEC) are unclear. We conducted a comprehensive assessment of drought status and trends based on the Standardized Precipitation Evapotranspiration Index (SPEI) in NEC from 1990 until 2018. The findings show that: (1) the drying trend peaked in 2001, and then exhibited a mitigation tendency before drying again after 2013. The implementation of ecological restoration projects is primarily responsible for drought mitigation. (2) The areas with wetting and drying trends in the future would cover 86% and 17% of NEC, respectively. (3) There is a time lag between improved vegetation and the trend shift from dry to wet. (4) Spring and winter revealed wet trends within 71% and 84% of NEC, respectively, showing high sensitivity and resilience to drought, while 92–93% of NEC displayed dry tendencies during the summer and autumn seasons. The drought-affected area was the highest in summer and lowest in autumn. (5) The interannual drought severity was highest in May and June. (6) The highest drought impacts and trends occur within shrub and grass and sparsely vegetated land, as well as middle-temperate semiarid regions (M-semiarid). (7) The warmer the temperature zone, the more sensitive it is towards drought under the same hydrological conditions, showing a high drought-affected area. The drier the land, the higher the drought-affected area within the same temperature zone, with pronounced drought trends during the spring and summer seasons. Our findings highlight the need for the government to more explicitly develop drought mitigation strategies in accordance with NEC’s spatiotemporal drought variations and specifically the need to concentrate on droughts in M-semiarid regions occurring in summer, particularly in May and June.

High emissions could increase the future risk of maize drought in China by 60–70 %

Drought events have considerable direct and indirect economic, environmental, and social impacts, but few studies have analyzed and assessed future changes in drought disasters from a risk perspective to guide responses and adaptations thoroughly. Studying the potential climate-related impacts on future crop yield is therefore urgently needed. Intercomparison of the three Shared Socio-economic Pathway (SSP) scenarios based drought risks and yield loss of China was carried out using the climate models from the Coupled Model Intercomparison Project Phase 6 (CMIP6), and the hotspots of high drought risk regions were identified. This study found that the areas affected by severe maize drought (loss ratio larger than 0.2) accounted for 16.13 %, 20.79 %, and 18.87 % of the total national corn areas under three low, medium-to-high and high emission scenarios (SSP1-2.6, SSP3-7.0, SSP5-8.5) respectively. The northwest China maize region, the ecotone between agriculture and animal husbandry, and the western central northern China maize region have relatively high loss risk. Compared with SSP1-2.6, the yield loss rates increased with 70.73 % and 61.52 % of national corn areas for SSP3-7.0 and SSP5-8.5, respectively. There is a decrease in the areas with low-risk and a significant increase in the areas with high-risk for SSP3-7.0 and SSP5-8.5 compared to the SSP1-2.6. These results may provide theoretical support for agricultural drought risk reduction and adaptation planning to ensure food security under climate change.

Does drought increase carbon emissions? Evidence from Southwestern China

The study estimates the impact of the 2009/2010 drought in southwestern China on industrial outcomes and carbon dioxide (CO2) emissions. We focus on the outputs of the power and energy-intensive sectors and investigate the substitution of thermal power for hydropower during this extreme drought. Panel data for 93,830 firms from 2006 to 2013 were used to examine their responses to this extreme climatic event. We find that severe drought reduced hydropower production as well as the economic output of energy-intensive sectors, while it increased the power production of thermal power firms. As a result, the net CO2 emissions in the southwest increased by 6,704,364 tons, about 0.64% increase of the total regional CO2 emissions each year from 2009 to 2013. These findings suggest that climate disasters may increase carbon emissions, thereby contributing to climate change.

Spatial-temporal distribution characteristics and hazard assessment of millet drought disaster in Northern China under climate change

Due to the national nutrition crisis and the adjustment of planting structure, millet has once again become a widely planted cash crop in Northern China. However, as the climate warms in recent years, drought has occurred frequently and repeatedly in Northern China, which has seriously threatened millet production safety in this area. Therefore, it is urgent to clarify the hazard of millet drought disaster and provide reference for the formulation of millet disaster prevention and reduction measures. In this paper, combined with the water demand of millet in each growth period, Standardized Precipitation Requirement Index (SPRI) was selected to identify and quantify the drought disaster of millet. Considering the impact of different degrees of drought in different growth periods, a hazard assessment model is constructed to evaluate the hazard of millet drought disaster in Northern China under the background of climate change. The results show that: (1) SPRI can well characterize the drought of millet in Northern China. (2) In the context of climate change, millet drought in Northern China shows a trend of increasing intensity and decreasing frequency, and it is possible that millet drought will be even more severe in the future. (3) Combined with the hazard assessment model, millet drought in Northern China will still be dominated by middle hazard in the future, but the areas with high hazard will gradually increase, mainly distributed in northeast China, central and western Inner Mongolia and northern Gansu. The research results can be of great practical significance for the adjustment of agricultural production structure in the future, especially for the sustainable development of agriculture and the formulation of measures to prevent and resist agricultural disasters.