The Compound Heatwave and Drought Event in the Summer of 2022 and the Impacts on the Power System in Southwest China

Abstract

An unprecedented compound heatwave and drought (CHD) event occurred in the summer of 2022 in Southwest China. This extreme climate event posed significant challenges to the power system and highlights the importance of disaster risk management and adaptation to extreme climate events in the power sector. This paper assesses the complementary effects of variations in hydropower, wind, solar power generation and the power load gap in response to this CHD event. The CHD resulted in a remarkable 50% decrease in hydropower generation during the summer of 2022. Similarly, wind speeds in the southwest region slightly decreased from 2.0 m/s in mid-July to 1.7 m/s in early August. On the contrary, solar power generation doubled from mid-July to mid-August. In the summer of 2022, the increase in solar power generation could not compensate for the gap between the dramatically increased cooling demand and the reduced hydropower output. Nevertheless, it highlighted the potential synergy of power source grid load storage and hydro–wind–solar power combinations in addressing future CHD events, and the importance of early-warning for extreme climate events in the new-type power system in the future.

1. Introduction

The global mean temperature in 2022 was approximately 1.15 °C above the pre-industrial average, ranking as the sixth warmest year on record [1]. Extreme heatwaves and severe drought events were widely observed across Asia, Europe, the Americas, and Africa in 2022. The main driving force is attributed to the “triple-dip La Niña” which persisted during 2020–2023 that altered global precipitation patterns, diverting moisture away from regions like the Yangtze River Basin. In China, an unprecedented extreme heatwave took place during the summer of 2022, accompanied by a severe drought in the Yangtze River valley. According to a report by the National Climate Center in 2023, this heatwave event lasted for 79 days, with 361 stations reaching or breaking historical records. The daily maximum temperature even soared to 45 °C in Chongqing, causing significant heat stress on human health. In addition to the extreme heatwave, a severe meteorological drought also struck the Yangtze River valley simultaneously. The combination of high temperatures and reduced rainfall led to a decrease in water resources in the region. For instance, the area of Poyang Lake shrank by 65% on August 21st, marking the smallest size in the past decade. The inflow to hydropower stations was also reduced, further stressing the power balance in Southwest China amid increased electricity demand driven by the heatwave [2]. The reduction of water resources during the summer is unusual, as precipitation is typically abundant in the Yangtze River valley.

Numerous studies have examined the impacts of climate extremes on both the supply and consumption sides of power systems. On the supply side, extreme climate events affect renewable energy output, including changes in seasonal characteristics of hydropower [3], output characteristics of wind power and photovoltaic (PV) energy [4,5], and even cooling water issues in inland nuclear power plants. Extremely high temperatures are often accompanied by high-pressure climate systems, leading to decreased rainfall and increased evapotranspiration, resulting in insufficient surface runoff and lowered river levels, ultimately affecting hydropower output. On the demand side, extreme climate events influence power loads. The issue of peak load caused by extreme weather is becoming increasingly prominent, and its duration is a crucial factor affecting load characteristic changes. Based on daily household electricity consumption data in Shanghai, on warm days above 25 °C, a 1 °C increase in daily temperatures leads to a 14.5% increase in electricity consumption [6]. In the US, Deschenes and Greenstone found that climate change would increase annual residential energy consumption by a statistically significant 15% to 30% (USD 15 to 35 billion in 2006 dollars) by the end of the century [7]. Total consumption for the households considered may increase by up to 55% by the end of the century [8,9]. Similar research in Europe also indicated a demand–temperature relationship, wherein a 1 °C reduction in daily temperature typically results in a 1% increase in daily electricity demand and a 3–4% increase in gas demand [10]. The significant polarization and seasonal shifts in peak demand and consumption have important implications for the location of costly peak-generating capacity, power grid transmission infrastructure and the design of energy-efficiency policies and storage capacity [11,12]. As living standards improve and the tertiary industry develops, cooling loads, which are increasingly sensitive to temperature, grow year by year.

In this paper, we quantitatively investigate the impact of the compound heatwave and drought (CHD) event on hydro, wind, and solar power generation, given that renewable power generation is susceptible to climate extremes. Furthermore, we develop the restricted cubic spline function to assess the impact of the CHD event on power demand and estimate the power shortage by counter-factual method, which will be useful for assessing the future impacts of climate extremes on the power load side on the power grid sector.

2. Method and Dataset

First, we used historical climate data to characterize the features of the CHD event. Second, we analyzed the impact of this CHD event on power load using the Restricted Cubic Spline (RCS) econometric regression method. Third, we illustrated its impact on power supply by utilizing datasets of hydro, wind, and solar power generation. A flow chart demonstrating the methods and data used in this study is provided in Figure 1.

To investigate the nonlinear causal relationship between the dependent variable (cooling load) and the independent variable (daily maximum temperature), researchers often use the subjective segmentation regression method, which divides the continuous variable into segments. However, this segmentation is often subjective, can lead to loss of information, and may introduce bias. In this paper, we choose to establish a Restricted Cubic Spline (RCS) function to explore the nonlinear impact of extreme high temperatures on cooling load. The spline function is essentially a piecewise polynomial, but it requires continuous and second-order derivatives at each piecewise point, and it is a linear function in the two intervals at both ends of the independent variable data range. Based on the response relationship of cooling load to extreme high temperatures, the Restricted Cubic Spline function model is set as follows:

$$L_t={\beta }_0+{\sum }_{j=1}^{k-1}{\beta }_{1j}RC{S}_j\left(T_{{\mathrm{m}\mathrm{a}\mathrm{x}}^t}\right)+{\epsilon }_t$$

(1)

in which $L_t$ is the cooling load at time t, $T_{{\mathrm{m}\mathrm{a}\mathrm{x}}^t}$ is the daily maximum temperature, ${\epsilon }_t$ is the residual item, and $RC{S}_j$ is the spline function. k represents the number of extreme high temperature segments, and j is the number of parameters in the nonlinear part that need to be estimated, excluding the constant term. Luo et al. [13] pointed out that k = 5 provided a sufficient fit for the model. In this paper, we set the commonly used percentiles (P) P25, P50, P75, and P95 as segment nodes, dividing the daily maximum temperature into 5 segments, and requiring estimation for three nonlinear parameters. The power load data, as well as the renewable energy power generation data, are sourced from the China Electricity Council and provincial power grid industry. The historical climate data are sourced from the China Meteorological Administration [14].

3. Results

3.1. The Analysis of the CHD Event in 2022 in China

To assess the meteorological drought, the daily maximum temperature (T_max) and Meteorological-Drought Composite Index (MCI) were employed. The MCI takes into account the comprehensive effects of rainfall and evaporation at different timescales, which performs better in terms of temporal–spatial continuity and accounts for more realistic durations than other drought indices [15]. Figure 2 illustrates the evolution of valley-averaged daily maximum temperature (T_max) and MCI during the boreal summer of 2022. It is clear that the Yangtze River valley experienced a hot summer, with T_max above its climatology from June to August. Particularly, the valley-averaged T_max exceeded the threshold of 35 °C for 14 consecutive days from August 7th to 23rd (Figure 2a). Figure 2b demonstrates the evolution of MCI during the summer, which reflects the cumulative effect of rainfall and evaporation. As a result, the hydropower generation in the upper reaches were severely affected. Another notable feature is that this heatwave affected the entire Yangtze River valley. As shown in Figure 2c, T_max is 4 °C higher than normal over the upper, middle, and lower reaches. This continuous heatwave, along with the reduced rainfall, led to the development of drought over the Yangtze River valley. This kind of valley-scale compound heatwave and drought (CHD) is very rare for the wet season of the Yangtze River valley.

Climate change exacerbates drought risk in the southwest of China. According to a study on the annual average frequency and duration of energy droughts of hydropower in the Yangtze River Basin (YRB) where most of the hydropower plants are located in Sichuan and Chongqing, the energy drought occurs on average 10 times per year, with a duration of 54 days/year, and more energy droughts are observed during the summer (June to August) than during the winter (December to February). In comparison, the drought event in Summer 2022 was about a 1-in-52-year event over the 2007–2021 climate period, and was seven times the average severity than in the summers of 2007–2021 [16]. The global and regional climate drivers such as the El Niño (La Niña) and the Indian Ocean and Pacific Ocean summer monsoons, together with the global warming that increases evaporation, influence droughts and flooding in the southwest of China. It is projected that the risk of energy drought of hydropower in the southwest of China will keep increasing. Under the Shared Socioeconomic Pathways (SSPs) SSP2-45 and SSP5-85 scenarios, which represent business-as-usual and high emission scenarios, the return periods of an extreme drought event like that in 2022 are 10 and 6 years, and the future drought risks will increase by 80% and 88% in the Yangtze River basin, respectively [16].

3.2. Impacts of the CHD Event on the Power Load

Based on the provincial daily maximum load data of the two southwest provinces, Sichuan and Chongqing, the cooling load could be calculated by deducting the actual load data with the spring/autumn average base load [17]. The underlying rationale assumes that the cooling load is neglectable during spring and autumn, and the primary cause of the discrepancy between summer and spring/autumn averages is attributed to cooling demand. Consequently, by employing the actual daily load during summer and subtracting the corresponding spring/autumn average base load—defined as the mean daily load prevalent in spring and autumn—we can effectively estimate the cooling load. The specific steps are as follows: Initially, the typical daily load in autumn is determined by averaging the data collected between September 10 and October 15, while the typical daily load in spring is obtained by averaging the data gathered from April 15 to May 15. Subsequently, the base load calculation differentiates between working days and weekend days, taking into account the distinct patterns of energy consumption associated with each type of day. By assuming a linear growth pattern, the base load increases consistently between spring and autumn. The cooling loads are all set to be positive in this study. In instances where the actual load value on a specific day falls below the corresponding base load value, the cooling load is assigned a value of zero, as suggested by Liu et al. [17]. Figure 3 illustrates the relationship between the computed cooling power load and the daily maximum temperature.

Table 1. Estimation of the nonlinear regression model.

Variable	Sichuan	Chongqing
RCS1	1.200 *** (0.185)	0.489 *** (0.054)
RCS2	3.108 *** (0.743)	1.075 *** (0.200)
RCS3	−15.27 *** (2.813)	−7.725 *** (1.351)
Constant	−28.362 *** (5.237)	−12.787 *** (1.649)
R-squared	0.813	0.906
Bayesian crit. (BIC)	460.310	292.845
Akaike crit. (AIC)	450.223	282.758

Note: *** represents p < 0.01.

displays the spline function parameter estimation results for the nonlinear response of summer cooling load to extreme high temperatures in Sichuan Province and Chongqing. In Sichuan, the extreme high temperature percentiles P25, P50, P75, and P95 correspond to node values of 28.39 °C, 31.69 °C, 35.23 °C, and 36.90 °C, respectively. Similarly, in Chongqing, the extreme high temperature percentiles P25, P50, P75, and P95 exhibit node values of 31.13 °C, 35.92 °C, 38.53 °C, and 41.27 °C, respectively. Overall, the influence of extreme high temperatures on cooling load exhibits a significant nonlinear relationship [6]. Elevated temperatures within the RCS2 interval result in a significant increase in cooling load, with the magnitude of this increase surpassing the parameter estimation outcomes in the RCS1 interval. However, extreme high temperatures within the RCS3 interval lead to a substantial reduction in cooling load. This phenomenon can be attributed to the compound effects of CHD on both the consumption and supply sides: the impact of extreme high temperatures leading to an increased cooling load (consumption effect) is countered by the negative influence of drought on hydropower generation (supply effect). The CHD event severely affected power supply capacity, particularly in the hydropower sector, making it challenging to meet power demands. Provincial grids, such as those in Sichuan and Chongqing, implemented “power rationing” measures to ensure residents’ basic power needs were met, ultimately resulting in a decrease in cooling load.

Figure 4 illustrates the relationship between the fitted values of cooling load and extreme high temperatures, demonstrating the nonlinear response of the cooling load to extreme high temperatures and highlighting the significant heterogeneity across different provinces. In the RCS1 and RCS2 intervals, the cooling load curves for Sichuan and Chongqing both rise in response to increasing temperatures, with steeper slopes in the RCS2 interval compared to RCS1. This suggests that the cooling load increases at a higher rate than the temperature, resulting in a U-shaped or convex curve. In the RCS3 interval, the compound impacts of CHD led to a considerable loss of power supply in Sichuan, causing a significant decrease in cooling load when temperatures exceeded 35.23 °C (Figure 4a). Meanwhile, in Chongqing, CHD resulted in stagnation and a slight decline in cooling load when temperatures surpassed 38.53 °C (Figure 4b).

To estimate the cooling demand in the intervals above 35.23 °C in Sichuan and above 38.53 °C in Chongqing, the average values of the parameter estimates for RCS1 and RCS2 were utilized to replace the parameter estimates for RCS3 (Figure 4). Subsequently, the difference between the predicted and actual values was calculated to determine the cooling load gap attributable to the combined effect of high temperature and drought. In Sichuan, the average cooling load gap amounted to 23.20 GW, equivalent to 63.6% of the spring–autumn average load. The maximum cooling load gap reached 41.98 GW, equivalent to 71% of the annual peak load. In Chongqing, the average cooling load gap was 6.55 GW, representing 45.8% of the spring–autumn average load, while the maximum cooling load gap was 10.62 GW, equivalent to 41.4% of the annual peak load.

According to statistics from the power industry, both Sichuan and Chongqing experienced record-breaking maximum loads during the summer of 2022. In Sichuan, even after the implementation of power rationing measures, the peak load still jumped to 65 GW in August due to the continuous heatwave, an increase of 25% compared to the previous year. The daily residential electricity consumption in Sichuan reached up to 473 million kWh, an increase of 263.8% when compared with annually averaged daily residential electricity consumption. The industrial and commercial daily electricity consumption increased by 50%, of which the cooling load accounted for more than 60% [2]. It is estimated that the largest power gaps in Sichuan and Chongqing reached 17 GW and 4.5 GW, respectively. During peak load periods, which coincided with the highest temperature intervals, the power gaps in Sichuan and Chongqing were approximately one-half and one-third of the annual average power load, respectively. To address this issue, large-scale controlled power consumption measures were implemented in both provinces, particularly targeting industrial power loads, as opposed to cooling loads, which primarily stem from residential and service sectors. The study results predict a theoretical cooling load based on daily maximum temperatures which is higher than real-world data. This discrepancy between predicted results and actual data can be attributed to the limitations of the industrial power load, rather than a decrease in cooling load. Consequently, it is important to note that the nonlinear relationship between cooling load and temperature found in this study may be influenced by the reduction of peak loads during extreme high temperatures (RCS3 interval), likely due to the curtailment of industrial power loads rather than a decrease in cooling load.

As a result of the CHD event, Sichuan Province had to suspend the production of industrial enterprises to guarantee the residential electricity consumption and safeguard people’s livelihood, as the drought and high temperature limited the hydropower generation capacity and the power grid transmission capacity [18]. Due to the statistics from the power sector in Sichuan Province, in the summer of 2022, Sichuan’s electricity demand surged and set a new monthly electricity sales record. In July 2022, the electricity sales were 29.1 billion kWh, a year-on-year increase of 19.8%. Among them, Sichuan’s large industrial daily electricity consumption reached 431 million kWh, a year-on-year increase of 13.1%; the average daily electricity consumption of residents was 344 million kWh, a year-on-year increase of 93.3%.

3.3. Impacts of the CHD Event on the Renewable Power Sources

In 2022, the installed capacities for hydropower, wind power, and photovoltaic (PV) power generation in Southwest China (including Sichuan Province and Chongqing Municipality) reached 110 GW, 8 GW, and 5 GW, respectively. In terms of annual electricity generation, hydropower produced 409.9 TWh, wind power generated 15.9 TWh, and PV power contributed 5.3 TWh. The installed capacities of hydropower, wind power, and PV power generation in Southwest China collectively constituted 78% of the total installed capacity in the region, with hydropower alone accounting for 70%.

3.3.1. Hydropower

From a river valley development perspective, the hydropower basins in Sichuan and Chongqing are primarily developed along the Dadu River, Yalong River, Jinsha River, Minjiang River, Jialing River, and other river valleys. As a major power transmission province, Sichuan’s annual electricity output can reach one-third of hydropower generation. While some rivers, such as the Liuchuan River, primarily rely on run-of-river hydropower stations on the tributaries of the Yangtze River, they lack water storage capacity and have very limited regulation ability. As a result, their power generation is highly dependent on natural water inflows and is easily affected by high temperatures and drought conditions. In total, the installed capacity of reservoir power stations with seasonal and annual adjustment capabilities in Sichuan only accounted for 36% of the total installed capacity of hydropower [2]. Thus, extreme heat and drought climate events could have substantial impact on hydropower, which in turn affects the overall power output in Southwest China.

During the period of July 1 to August 14, rainfall in the Yangtze River valley was recorded at 137.6 mm, representing a 46.2% decrease compared to the same period in the previous year (256.0 mm). This marked the lowest level in the same period of history since 1961, with the second lowest being 160.4 mm in 1972 (Figure 5a). The average number of rainy days in the valley was only 12.9 days, which is 7.1 days fewer than the average for the same period in normal years (20.0 days), making it the lowest in recorded history. Rainfall levels in most areas of the valley were more than 20% lower than those of the same period in previous years, with the central area of the upper reaches of the Yangtze River valley experiencing a 50% to 80% reduction.

Starting from mid-July, the daily power generation of hydropower began to gradually decline, in line with the decrease in incoming water. In July 2022, the per unit (PU) value of hydropower daily power generation reached approximately 0.95 (PU is calculated by dividing the daily power generation by the maximum daily power generation throughout the year). By the beginning of August, the PU value of daily power generation dropped to 0.9, and by around August 25, the daily power generation reached its lowest value of 0.7. According to the China Electricity Council (CEC), hydropower production in Sichuan decreased by half during the summer of 2022 compared to the previous year.

If we calculate the abnormal cumulative precipitation from June 1st, we can find that the cumulative precipitation began decreasing on June 25th, and the cumulative precipitation was approximately 190 mm until the end of August. As a comparison, the per unit (PU) of the hydropower generation continued to increase from 0.74 from June, reaching its peak on July 11th, and then gradually decreased during the period between July 11st and August 10th, then sharply decreased from August 10th to its lowest level of 0.71 on August 25th. From Figure 5a, we can conclude that there was significant divergence, or a negative relationship, from June to July 11th between the hydropower generation and the cumulative precipitation trend, but the relationship turns to a positive one thereafter. To account for this phenomenon, we can conclude that there is a delay between the hydropower generation management and the precipitation index. There are time lags between the prediction of climate results, delivery to the water resource management, and then the hydropower generation plan. Therefore, it is critical to develop climate predictions and climate services for the power industry.

3.3.2. Wind Power

In terms of wind power, the relationship between wind speed and wind power output was quantitatively assessed (Figure 5b). In July and August, wind speeds in the southwest region slowed down, decreasing from 2.0 m/s in mid-July to 1.7 m/s in early August. Given that wind power output is proportional to the third power of wind speed, wind power generation in Southwest China also experienced a significant decline during this period. The per unit value of daily wind power generation dropped from 0.9 in mid-July to around 0.2–0.3 in early to mid-August.

The long-term climate warming, compounded by short-term, high-pressure stagnant weather, has led to a decrease in wind power generation. On a climate time scale, the reduced pressure difference between regions due to climate warming has resulted in a long-term decline in average wind speed. On a weather time scale, high-pressure weather systems have caused stagnant weather, further reducing short-term wind speeds. In July, the average wind power generation in China (excluding South China) was approximately 50 GW, representing only 20% of the installed capacity and exacerbating power balance tensions.

3.3.3. Solar Power

In terms of photovoltaics, a quantitative evaluation was conducted on the relationship between sunshine hours and solar power generation (Figure 5c). In mid-July, sunshine duration dropped from a maximum of 12 h at the beginning of the month to a minimum of just 1 h. Since the beginning of August, sunshine duration increased, reaching about 10 h in mid-August, and then began to decrease again towards the end of August. Fluctuations in solar power generation followed a pattern similar to that of sunshine hours. The daily photovoltaic power generation was only 0.4 in mid-July and increased to 0.8 in mid-August.

Extreme high temperatures have resulted in a decrease, rather than an increase, in photovoltaic (PV) output, thereby reducing solar power generation. Long-term climate change has contributed to a reduction in direct radiation and sunshine hours. Additionally, extreme high temperatures can diminish the output power of PV modules and even cause damage to components, impacting the overall performance of PV power generation. For instance, in 2022, the installed PV capacity in Sichuan increased by 5% compared to the same period in 2021; however, PV power generation in July experienced a year-on-year decrease of 6%.

3.3.4. Complementarity of the Hydro–Wind–PV Power

Based on the per unit (PU) value of hydro, wind, and PV power (in Figure 5d), we can observe that there is complementarity between hydro, PV, and wind power output [19]. For instance, hydropower was decreasing while wind and PV power were increasing during the period spanning July 11th–25th. Additionally, there is complementarity between wind and PV, where wind is low while PV is high during the period spanning July 1st–10th and in August. It should be noted that the installed capacities of wind and PV are still relatively small compared to hydropower in the southwest, but they are growing rapidly and can play a much more significant complementary role in the future.

It is estimated that the installed capacities of hydro, wind, and PV power in Sichuan province will increase from 80.2 GW, 4.5 GW, and 3 GW in 2020 to 132 GW, 30 GW, and 30 GW in 2050, and the installed capacities of hydro, wind, and PV power in Chongqing will increase from 7.8 GW, 0.8 GW, and 0.7 GW in 2020 to 10.2 GW, 9.5 GW, and 7.5 GW in 2050 [20]. Wind and PV capacities will increase by 7.4 and 10.2 times in Sichuan and Chongqing over the next three decades, while hydropower will only double.

In terms of energy balance, PV and wind power are negatively correlated. On days when PV output is larger, wind power output is smaller. PV and wind power form a more complementary relationship. As wind power and PV installations expand in the future, PV and wind power can help alleviate the electricity shortages caused by hydropower deficits. Therefore, the configuration of the installed hydro–wind–PV combination can be further developed to better stabilize total output fluctuations in future power supply.

4. Improving the Climate Predictability and Adaptation Capacity of Power System

4.1. Adaptation Measures in the Power System

From the power system perspective, it is important to improve the disaster risk management of such CHD events and enhance the climate resilience of the new-type power system. Given the double-sided impact of the CHD event on renewable power generation and electricity consumption in summer 2022, the Sichuan and Chongqing provinces successfully addressed the power shortage by implementing various measures. These measures included cross-regional power support through power grid reversal transmission, deployment of power generation vehicles, implementation of orderly power load management during peak periods, and restrictions on industrial power consumption to ensure household power supply. Furthermore, Sichuan announced that it will build facilities allowing for a 7.7 GW gas-fired capacity, as well as an 8.5 GW power transmission capacity, which will focus on linking major hydropower dams to the Sichuan provincial grid and more regional high-voltage lines interconnecting Sichuan to its neighbors’ grids in order to fight against future droughts [21].

To build a more climate-resilient power system for the future, three key adaptation measures should be prioritized. Firstly, proactive planning for multi-energy integration is essential, with a focus on promoting the coordinated development of fossil and renewable power sources. Leveraging the complementary advantages of wind, solar, and hydropower can enhance emergency supply capacity. Secondly, proactive integration of power sources, grids, loads, and storage is crucial. Flexible energy resources like power storage (such as the pumped hydropower storage) and flexible power loads such as vehicle-to-grid (V2G) and virtual power plants (VPP) can play a vital role during periods of power shortage. Thirdly, enhancing the application of meteorological information across various time and space scales is important to improve prediction accuracy.

4.2. Early Warning of Climate Service

Encouraging cross-departmental cooperation and fostering innovation in both meteorology and the power sectors are key steps to bolstering the climate resilience of the power system. From the climate service perspective, early warning of extreme climate events is crucial for adapting to climate change. During COP27, the United Nations introduced “Early Warning for All (EW4ALL)”, which calls for a $3.1 billion investment from 2023 to 2027 to establish a global early warning system aimed at enhancing preparedness for climate-related disasters. It is critical to improve the prediction and early-warning of such CHP events at the sub-seasonal to seasonal (S2S) time scale in the future.

The predictability of this CHD event is evaluated using the China Meteorological Administration’s (CMA) new-generation dynamical prediction system, CMA-CPSv3 [22]. While the prediction of such events at seasonal time scales remains challenging, current climate models such as the CMA-CPSv3 are capable of back-casting the general trend of below-normal rainfall and above-normal temperature when initialized in early March in 2022. At the subseasonal time scale, predicting heatwave changes is possible. CMA-CPSv3 successfully captured the heatwave development in the Yangtze River valley from the 1st pentad of July to the 5th pentad of August and the heatwave decay at the end of August in 2022. Nevertheless, it still remains unable to predict the rainfall at the subseasonal timescale. Consequently, improving the subseasonal prediction skill for compound climate extremes is urgently needed to build the resilience of the power system. Advancements in climate modeling and prediction would enable better preparedness and response to extreme climate events, ultimately helping to minimize the impacts on power systems and maintain reliable electricity supply during CHD events.

5. Conclusions and Discussion

There are three key findings and policy implications from this study. Firstly, the compound heatwave and drought (CHD) event could cause simultaneous decreasing hydro, wind and solar power generation, pushing up the power load, and exacerbate the power shortage. Renewable power generation, including wind, hydropower, and solar power, is sensitive to climate variability. Fluctuations in power generation during extreme climate events can place significant stress on the power system. In this study, we found that hydropower and solar power generation exhibit a notable complementary relationship. As hydropower generation experienced a remarkable decrease of 50% in Southwest China due to severe drought, solar power generation increased by two times from mid-July to mid-August, driven by enhanced solar irradiation. However, the increase in solar power generation was insufficient to compensate for the decrease in hydropower generation. This was primarily due to two factors. On the one hand, the installed capacity of hydropower in Southwest China is significantly larger than that of solar power, limiting the ability of solar power to fully offset the decline in hydropower generation. On the other hand, the concurrent compound heat and drought (CHD) event led to a substantial increase in cooling load, further exacerbating the strain on the power system. However, as the amount of installed solar power generation facilities increases, the future role of the compensation effect could become more important.

Secondly, the disaster risk management for the extreme climate events are critical for improving the climate resilience of the new-type power system in China. With the development of ultra-high voltage (UHV) technology for regional electricity transmission [20], the new-type power system will be characterized by a high proportion of renewable energy, and a high proportion of power transmission from west to east [23], together with summer and winter double power load peaks [24]; therefore, the compensation effects could arise from the fluctuations of renewable power in remote regions [25,26]. This would enable better load balancing and more effective integration of renewable energy resources across larger geographical areas, enhancing the overall resilience and stability of power systems [27].

Lastly, the early warning of the extreme climate event is essential for enhancing the adaptation ability for the power sector and other climate-sensitive sectors. The heatwave change predictions of the current climate model show promise at the subseasonal timescale. With useful early warning prediction, decision makers can formulate policies to increase water storage during the water storage period based on effective climate prediction information to cope with CHD events. The cross-departmental cooperation and innovation in both the meteorology and power sectors are important to enhance the climate resilience of the power system.

This study has caveats, and here we suggest directions for further study. This study primarily focuses on the power system fluctuations in Southwest China, while neglecting the remote effects. For instance, hydropower generation in Southwest China not only supplies local electricity consumption but also contributes to meeting a large share of electricity demand in East China (e.g., Shanghai, Jiangsu, Zhejiang, etc.). The CHD event in 2022 also led to an increase in electricity demand in East China, which in turn placed additional stress on the power system in Southwest China due to the fact that the latter is the main power-exporting region. This aspect has not been investigated in the current study. Moreover, the analysis in this study only considered the complementary effect among hydro, solar, and wind power generations at a local scale within Southwest China. Future research could consider these remote effects and inter-regional compensation mechanisms to provide a more comprehensive understanding of the power system dynamics in China, and across the globe, during extreme climate events [28]. This would help inform strategies for optimizing the development and deployment of renewable energy resources to ensure a stable and secure power supply under climate change.