Climate extremes become increasingly fierce in China

Main text

Heatwaves wreaked havoc across the Northern Hemisphere in summer 2022 and resulted in at least 15,000 deaths in Europe. Eastern China also experienced an unprecedentedly hot and dry summer. The maximum 2-m air temperature (Tmax) at around 300 national meteorological stations broke the historical record, and high temperature warnings sounded for 41 consecutive days. As a consequence, a devastating chain of disasters took shape. With Tmax >40°C, more than 270 million people suffered from heat stress, and there were hundreds of casualties reported due to thermoplegia. The long-lasting heatwaves and precipitation deficit subsequently led to severe drought in the Yangtze River basin. Decreased runoff reduced the hydropower generation by half in Sichuan Province, while the super heatwave led to a remarkable increase in electricity consumption and further aggravated the power shortage. Meanwhile, the drought induced a 10-day wildfire in Chongqing over August 17 to 26, 2022, that burned forests and emitted large quantities of pollutants and carbon into the air. Furthermore, Poyang Lake (the largest freshwater lake in China) lost more than 75% of its water in the summer, and the out-of-season drought continued until November 9, when the water level was only 6.67 m, which broke the historical minimum. The severe heatwave and drought in summer 2022 significantly highlight the importance and necessity to better understand climate extremes in China.

Actually, dozens of climate extremes have converged in the last 3 years (from summer 2020 to summer 2022) in China (Figure 1), challenging scientists and decision-makers alike. Examples include but are not limited to the following events: (1) the super meiyu that occurred over the Yangtze-Huaihe River basin in summer 2020 with an uncharacteristically early onset and late withdrawal, causing numerous casualties and huge economic losses; (2) no typhoon activity in the western North Pacific in July 2020, breaking the record since the beginning of the satellite era; (3) the severe drought in South China that persisted from October 2020 to May 2021 and threatened the water supply for residents in the Guangdong-Hong Kong-Macao Greater Bay Area; (4) the most sharp transition from the cold early winter to the warm late winter of 2020/2021 that sent an obviously different sign compared with the winter mean air temperature in the east of China; (5) the super sandstorm that swept across North China, the likes of which have been absent for almost 10 years¹; (6) the torrential rain that hit Zhengzhou over July 19 to 21, 2021, and directly led to 398 deaths and a total financial loss of 120 billion RMB²; and (7) the record-breaking dragon-boat-rain and heatwaves in early summer 2022 that influenced South China and North China, respectively.

Figure 1.

Climate extremes that severely impacted East China from summer 2020 to summer 2022

The colored shading is the influenced areas of each extreme, and the gray shading indicates cumulated influences of climate extremes in the 3 years.

It is clear that climate extremes and resultant hazards have been intensifying and occurring more frequently in China. In response to climate change, climate extremes exhibit new features such as being long lasting, record breaking, impact compounding, and clustering. Although the changing climate has been described as globally and chronically hotter, and “dry gets drier, wet gets wetter,”³ significant interannual variation is noteworthy. The patterns of extremes could be opposite as seen in the last 3 years, for example, the super meiyu in 2020 versus the ultra-weak meiyu in 2022, and the cold-to-warm transition in winter 2020/2021 versus the warm-to-cold transition in winter 2021/2022. Such ultra-strong climate system variabilities indicate an extremely challenging prediction issue, which has been highlighted as the “grand science challenge” by the Intergovernmental Panel on Climate Change (IPCC) and the World Climate Research Programme (WCRP).

A large number of studies consistently documented that human activity has increased the probability of hot extremes.³ Without human influences on the climate system, some recently observed hot extremes would have been unlikely to occur. The confidence of changes in heavy precipitation and drought attributed to anthropogenic influences is relatively lower at regional and local scales. Although most studies agree on the conclusion that anthropogenic influences increase the probability of heavy precipitation at the continental scale in East Asia, Zhou et al.⁴ pointed out that anthropogenic forcing reduced the probability of a super meiyu in 2020 by 46%. Overall, attribution studies focus on the extent to which anthropogenic forcing influences climate extremes, yet far less attention has been paid to understand the underlying physical mechanisms. The reliability of attribution heavily depends on the ability of climate models in simulating the underlying physical processes responsible for the observed extremes. Correct attribution could support the projection, adaptation, and mitigation of climate change. However, it is unable to support subseasonal-to-seasonal prediction of the climate system, let alone extreme events.

Air-sea-land interactions significantly modulate interannual-decadal variations in extremes and naturally provide a source of predictability because of the climate “memory.” The joint effects of the three oceans and the Arctic are believed to have driven the occurrence of the 2020 super meiyu. The subseasonal reversal of the Arctic air-ice system, the La Niña condition, and the warmer sea surface temperature (SST) in the North Atlantic concurrently reached record-breaking intensity in autumn/winter 2020. Physically, a cold-to-warm extreme transition was triggered in the east of China in winter 2020/2021, and severe sandstorms then reoccurred in spring 2021 in North China.¹ The preceding shallower Eurasian snow cover and warmer North Atlantic SST are found to strengthen moisture convergence in South China, resulting in the unprecedented dragon-boat rainfall in 2022. The large-scale western Pacific subtropical high and the Iran subtropical high merged and resulted in severe and broadscale heatwaves in July and August 2022. A specific climate extreme does not result from one inducement but is triggered jointly by several high-strength climate factors.

Further, a specific climate driver does not always correspond to the occurrence of the same climate extreme. For example, the La Niña in autumn/winter 2020 contributed to the persistent drought in South China; however, the same La Niña that continued until 2021 was associated with heavier-than-normal rainfall in South China. Thus, the prediction of extremes in general is even more challenging. Although attempts to predict climate extremes have been boosted in recent years, the level of prediction skill remains low. One possible reason for this is that the key processes involved in the air-sea-land interactions are not well or accurately depicted in both numerical and statistical models. Another reason is that the features and climate effects of El Niño/Southern Oscillation, record-breaking SST anomalies, and Arctic Sea ice (just to name a few) have also changed in the context of global warming. Therefore, instead of unilateral studies, comprehensive studies of how human influences and the internal variability of climate systems as well as their interactions contribute to climate extremes must be put on the agenda.

Climate extremes are the starting gun of a risk and disaster chain. Compound events and low-likelihood, high-impact events mostly lead to higher risks to population health and food and energy security, but their indistinctive statistical regularities mean low predictability. Thus, usable predictions of climate extremes and subsequent risks remain a grand global challenge and an urgent task in China. One of the best approaches for extreme climate predictions is probably through a statistical-dynamical hybrid scheme. Particularly, the thriving development of artificial intelligence (AI) for climate prediction brings a brand new promotion and now steps into the explainable stage.⁵ Probabilistic forecasts estimate the probability for extreme events and can better support level-to-level disaster prevention and reduction. An area that requires particular attention is the rapid assessment of risk and disaster chains, both before and after the occurrence of extreme events. In addition, rapid attribution and mechanism analysis systems would be helpful to address the question of to what extent one specific climate extreme is influenced by human activity and internal variability. The answer to this question is a constant demand from both decision-makers and the media.

Published Online: February 21, 2023