Spatial and Temporal Evolution Patterns of Droughts in China over the Past 61 Years Based on the Standardized Precipitation Evapotranspiration Index

Abstract

Based on the data of 2254 daily meteorological stations in China from 1961 to 2021, this study calculated the standardized precipitation evapotranspiration index (SPEI) of the national multi-time scale by using the FAO Penman–Monteith model to quantify the changes in dry and wet conditions. The Mann–Kendall mutation test, wavelet analysis, and other methods were used to study the spatial pattern and temporal evolution of drought. The results showed: (1) In the past 61 years, there were obvious spatial and temporal differences in drought in China, and the interannual variation in drought severity in SPEI-1, SPEI-3, and SPEI-12 gradually decreased at a rate of 0.005/10a, 0.021/10a, and 0.092/10a, respectively. (2) The time point of dry and wet mutation was 1989 according to the MK mutagenicity test. (3) Wavelet analysis showed that the drought cycle on the annual scale and the seasonal scale was consistent, and the main period was about 30 years. (4) In the past 61 years, the drought intensity of different degrees in China has shown a weakening trend, and the drought intensity reached the highest value in 61 years in 1978, at 1836.42. In 2020, the drought intensity was the lowest, at 261.55. (5) The proportion of drought stations has shown a decreasing trend. The proportion of drought-free stations has fluctuated greatly, ranging from 42.12% to 89.25%, with 2020 being the highest. This study provides a scientific basis for further research on the causes and coping strategies of drought and is of great significance for strengthening China’s drought monitoring, early warning ,and adaptation capabilities.

1. Introduction

As a result of global warming, surface evapotranspiration has increased and precipitation has become more regionalized, with the frequency, intensity, and trend of extreme climate events such as high temperatures, floods, and droughts attracting increasing attention [1,2]. Among various weather events, drought is a major environmental problem faced by humans, accounting for 70% of all meteorological disaster losses, and it thus has become the focus of risk assessments and prevention [3]. Drought is generally reviewed as a sustained and regionally extensive occurrence of appreciably below-average natural water availability, either in the form of precipitation, river runoff, or groundwater [3,4]. It has been estimated that droughts are the world’s costliest natural disaster, accounting for USD 6–8 billion annually and impacting more people than any other form of natural disaster [5,6]. According to EM-DAT statistics, there were 15 droughts worldwide in 2021, with Africa and Asia being the most affected. Unlike other weather-related disasters, drought events often involve multiple and compounding hazards, such as food shortages, pandemics, political instability, economic crises, or human/livestock/crop diseases, which increase the uncertainty of drought predictions [6,7]. The interplay between many of these factors affects exposure, vulnerability, and crisis response capabilities.

China is located in the east of Asia and the west coast of the Pacific Ocean, with high terrain in the west and low terrain in the east; the terrain is changeable, and the diversity in temperature and precipitation give it a unique monsoon climate. Affected by the monsoon climate, drought is particularly prominent in China, becoming one of the most common natural disasters in China. Since the 60s of the 20th century, drought in China has been a major threat to agricultural production due to its high frequency, long duration, and wide impact [7]. Yang et al. [8] have used data from 2190 weather stations to study the vulnerability of urban agglomerations to heatwaves in the context of global warming. Li et al. [9] have selected hail data from 2254 stations to explore hail frequency. Tang et al. [10] have found that drought events were increasing based on a study of 82 weather stations in southwest China. Huang et al. [11] have studied the temporal and spatial variations in drought in the Weihe River Basin based on data from 21 meteorological stations. The results showed that the eastern part of the basin showed an arid trend. The above studies show that the frequency of drought is on the rise in some parts of China, and the study of drought in China as a whole has become more urgent and necessary. However, there are relatively few studies on the spatial and temporal evolution patterns of drought across the country with a long span and extensive data volume, and the research on the national dry–wet cycle is also insufficient. Therefore, this study aimed to explore the spatial and temporal variability of drought in China over the past 61 years, as well as the main cycle of the national dry–wet cycle. This research has important implications for China’s climate and socio-economic development.

Global drought trends have gradually received extensive attention from scholars at home and abroad [12,13], and many scholars have adopted different drought indices to study droughts [14], including the Palmer Drought Severity Index (PDSI) [15,16], the Standardized Precipitation Index (SPI) [17,18], Standardized Precipitation Evapotranspiration Index (SPEI) [19,20], and the Meteorological Drought Composite Index (MCI) [21]. Among them, the PDSI is a commonly used drought index, but it is mostly applied to medium- and long-term drought changes and lacks fine temporal resolution [22]. The SPI is used in some research, but the index only involves changes in precipitation in its calculation. In recent years, many studies have shown that the increase in temperature has had a great impact on hydrometeorological conditions [23], and thus only considering precipitation is no longer sufficient to accurately characterize the changes in drought. The SPEI was first proposed in 2010 by Vicente-Serrano [24]; it takes into account precipitation, temperature, and evapotranspiration, and combines the advantages of the SPI and PDSI to calculate soil moisture balance [19,24]. As a result, the SPEI is widely used for drought monitoring. SPEI data at different time scales can reflect different drought types. One-month, three-month, and twelve-month SPEI data generally represent meteorological, hydrological or agrarian, and socio-economic droughts, respectively [7]. Stefanos Stefanidis et al. have pointed out in a recent study that shorter time scales (SPEI3 and SPEI6) were more efficient for identifying short-term droughts, while longer time scales (SPEI12 and SPEI24) were better for identifying less frequent but longer-lasting drought episodes [25]. Therefore, it makes sense to carry out multi-scale monitoring of drought. Domestic scholars have carried out a lot of research on droughts using the SPEI. Li et al. [26] have shown that the SPEI index with a time scale of 3 months and above is more applicable in most areas of China. Wu et al. [27] have also pointed out that in areas of China where the average annual precipitation exceeds 200 mm, SPEI analysis at various timescales is applicable and can characterize the drought situation more accurately.

At present, there are few studies on the spatial and temporal evolution patterns and dry–wet cycle of drought in China in the past 61 years. In this study, the more accurate FAO Penman-Monteith method was used to calculate the SPEI to explore the spatial and temporal differentiation of drought in China in the past 61 years. Secondly, MK mutagenicity test and wavelet analysis were used to analyze the distribution of drought abrupt change points and dry–wet cycles. This study can provide a scientific reference for drought mitigation and decision making for formulating corresponding ecological and environmental protection measures and coping with climate change.

2. Research Methodology and Data Sources

2.1. Calculation of SPEI

The SPEI is a drought index proposed by Vicente-Serrano et al. [24] that takes into account changes in precipitation and temperature [28,29,30]. Compared with the traditional Standardized Precipitation Index (SPI), the SPEI includes the potential evapotranspiration (PET) factor in the calculation process, which more comprehensively reflects the impact of climate change on drought. The SPEI is calculated by normalizing the difference between actual precipitation and potential evapotranspiration. For more details about the SPEI, the interested reader is referred to Danandeh Mehr et al. [31] or the Meteorological Drought Rating Criteria [32]. (The calculation of SPEI is based on Python 3.6)

$$D_i=P_i-{\mathrm{PET}}_i$$

(1)

$$\begin{array}{cc}\hfill {\mathrm{SPEI}}_i=& W_i-\frac{2.515517+0.802853W_i+0.010328{W_i}^2}{1+1.432788W_i+0.189269{W_i}^2+0.001308{W_i}^3}\hfill \\ \hfill & W_i=\sqrt{-2\ln p} \hspace{0.17em}\mathrm{for} p\le 0.5\hfill \\ \hfill & W_i=\sqrt{-2\ln \left(1-p\right)} \mathrm{for} p>0.5\hfill \end{array}$$

(2)

According to the Meteorological Drought Levels formulated by the National Meteorological Administration, the SPEI is divided into different drought levels:

In this study, 1-month, 3-month, and 12-month values of the SPEI (SPEI-1, SPEI-3, SPEI-12) were selected as drought indicators for monthly, seasonal, and annual timescales, respectively. The seasonal SPEI was calculated using the average of the input elements of the current month and the two months prior. For example, the SPEI-3 for March represented the variation in dryness and wetness in January, February, and March. In this paper, meteorological months were selected for the seasonal-scale study, with March to May as spring, June to August as summer, September to November as autumn, and December to February of the following year as winter. The drought level corresponding to a given SPEI value was determined according to the national standard of meteorological drought level [29], with droughts being defined as showing an SPEI ≤ −0.5 (mild drought and above) (

Table 1. Classifications of standardized precipitation evapotranspiration index.

Level	Type	SPEI
1	no drought	SPEI > −0.5
2	mild drought	−1.0 < SPEI ≤ −0.5
3	moderate drought	−1.5 < SPEI ≤ −1.0
4	severe drought	−2.0 < SPEI ≤ −1.5
5	extreme drought	SPEI ≤ −2.0

).

2.2. Mann–Kendall Nonparametric Tests

The Mann–Kendall analysis method is a non-parametric statistical test method commonly used to detect changes in temperature and precipitation series. It has the advantage of not requiring a specific arrangement of samples, has little impact on the error and loss of data, and is suitable for the analysis of various non-normally distributed time series data. The calculation is as follows:

$$S_k={\displaystyle \sum _{i=1}^k{\displaystyle \sum _j^{i=1}{\alpha }_{ij}}} \left(k = 2, 3, \dots , n\right)$$

(3)

$${\alpha }_{ij}=\{\begin{array}{c}1 X_i\succ X_j\\ 0 X_i\prec X_j\end{array} \left(1\le j\le i\right)$$

(4)

$$UF=\frac{\left[S_k-E\left(S_k\right)\right]}{\sqrt{Var\left(S_k\right)}} \left(k = 1, 2, \dots , n\right)$$

(5)

$$E\left(S_k\right)=k\left(K+1\right)/4; Var\left(S_k\right)=k\left(k-1\right)\left(2k+5\right)/72$$

(6)

$$\{\begin{array}{c}UB=-UF\\ k=n+1-k\end{array} \left(k = 1, 2, \dots , n\right)$$

(7)

UF and UB curves were used to analyze trend changes in the data while determining the onset time of mutations. When UF > 0, the sequence is in an ascending state; when UF < 0, it is in a descending state; when the curve is outside the significant horizontal range, it indicates a significant trend of strengthening or weakening. If the UB and UF curves produce an intersection point in the interval, the time corresponding to the intersection point is the beginning time of the mutation.

2.3. Wavelet Analysis

Wavelet analysis is a time-frequency analysis method that can decompose time series into the time-frequency domain to characterize the local characteristics of signals, and it is an effective tool to analyze the periodic changes in variable time series. The principle is to use a set of wavelet basis functions (parent wavelets) to describe the signal characteristics and to capture the characteristics of data changes through the oscillations of the wavelet basis functions at different scales and positions. The calculation is as follows:

$$\phi \left(t\right)=e^{ict}{e}^{-t^2/2}$$

(8)

$$W_f\left(a,b\right)={\left|a\right|}^{-1/2}\Delta t{\displaystyle \sum _{k=1}^{N}f\left(k\Delta t\right)}\overline{\phi }\left(\frac{k\Delta t-b}{a}\right)$$

(9)

where c is a constant, i is an imaginary number, W_f(a,b) is the wavelet transform coefficient, a is the scale factor, which reflects the period length of the wavelet, and b is the time factor, which reflects the translation in time.

2.4. Drought Assessment Methodologies

Drought intensity is used to characterize the severity of drought. In this paper, the SPEI ≤ −0.5 was used as the mark of the beginning of drought. The absolute value of the cumulative daily value of the SPEI in the year of drought was used to represent the drought intensity of that year, which is expressed as follows:

$$D_g={\displaystyle \sum _{i=1}^g\left|\mathrm{SPEI}\right|} \left(\mathrm{SPEI}\le -0.5\right)$$

(10)

where D_g is the cumulative drought intensity of different drought levels in the year; g is the total number of days in the year.

This paper calculates and statistically analyzes the occurrence of different degrees of drought in the SPEI over the years with the following formula:

$$P=\frac{n}{N}\times 100\%$$

(11)

where n denotes the number of stations where different degrees of drought occurred; N denotes the total number of stations in the study area.

2.5. Data Sources

The meteorological data are daily surface climate data from 1961 to 2021, and the data source is the China Meteorological Science Date Centre “https://data.cma.cn/ (accessed on 5 February 2023)”. Climate data cover precipitation (mm), maximum temperature (°C), minimum temperature (°C), average temperature (°C), wind speed (m/s), sunshine hours (h), latitude and longitude (°), and altitude (m). There are a total of 2254 meteorological stations, including national basic meteorological stations and local meteorological observation stations, and the values of stations with missing data are interpolated by neighboring years or station values.

3. Results

3.1. Time-Varying Characteristics of SPEI

The sensitivity of the SPEI to time varies across different timescales, with smaller timescales resulting in larger fluctuations in dryness and wetness, and larger timescales resulting in smoother fluctuations in dryness and wetness; the SPEI-1 fluctuated the most frequently during 1961–2021, with more frequent changes in dryness and wetness (Figure 1). Short-term SPEI changes also revealed short-term droughts. For example, in March 1963, the monthly drought intensity reached −1.13; in December 1979, the monthly drought intensity reached −1.29; in January 1988, the monthly drought intensity reached −1.24; and in February 1999, the monthly drought intensity reached −1.37. These changes are hard to observe in seasonal and annual data. The fluctuation in the SPEI-3 was more moderate than that of the SPEI-1, and in the 1960s to 1980s, the SPEI-3 mostly fluctuated below zero. From the 1990s to the 2000s, the SPEI-3 fluctuated around zero. After the 2010s, the SPEI-3 fluctuated above zero. The SPEI-12 fluctuated with a similar magnitude as the SPEI-3 and can more intuitively reflect the trend of the SPEI changes in China over the past 61 years. The SPEI at different timescales in China all show an upward trend, indicating that China’s climate has tended toward a more humid condition over the past 61 years.

3.2. Drought Abrupt Change Test and Drought Cycle Characteristics

3.2.1. Drought Abrupt Change Test

The Mann–Kendall abrupt change test of China’s SPEI-12 from 1961 to 2021 shows that the SPEI showed a non-significant upward trend during 1982–1991, and between 1992 and 2021, there was a significant upward trend (Figure 2). The intersection points of the UF curve and the UB curve within the confidence interval are identified as abrupt change points, indicating that around 1989, the SPEI began to undergo an abrupt change.

3.2.2. Annual and Seasonal Cycle Characteristics

To further study the cycle characteristics of drought in China, Morlet continuous complex wavelet transform was chosen to analyze the SPEI-12 from 1961 to 2021. The variance distribution plot (Figure 3b) shows that the period of the strongest oscillation is 32 years, which is the first major period. In the wavelet coefficient real contour plot (Figure 3a), there are two significant changes in the wet and dry cycles during the oscillation period. Furthermore, the contour line for 2021 is not completely closed, and it is now positive, indicating that the current year is in the wet zone. This indicates that we are currently in a relatively wet period and that the next cycle will transition to negative values, indicating that climatic conditions are trending towards dryness.

The SPEI drought cycles in spring were characterized by periods of 2–4 years, 6–11 years, and 27–33 years (Figure 4a). On the 2–4-year scale, there were two dry–wet transitions in spring; on the 6–11-year scale, three more distinct dry–wet cycles occurred; and on the 27–33-year scale, there were two prominent dry–wet cycles in spring. Combining this with the wavelet variance graph, the strongest oscillation period for spring drought changes was 30 years, making it the primary cycle, with 8 years as the secondary cycle, and 3 years as the tertiary cycle (Figure 4a’). The SPEI summer drought cycles were 2–6 years and 33–39 years (Figure 4b). On the 2–6-year scale, there were six dry–wet cycles in summer; on the 33–39-year scale, there were two significant dry–wet cycles. The variance graph indicates that the strongest oscillation period for summer drought changes was 35 years (the primary cycle), with 3 years as the secondary cycle (Figure 4b’). The SPEI autumn drought cycles were 2–5 years, 10–13 years, and 29–35 years. On the 2–5-year scale, there were seven dry–wet cycles in autumn; on the 10–13-year scale, there were two more noticeable dry–wet cycles; and on the 29–35-year scale, there were two significant dry–wet cycles (Figure 4c). According to the variance graph, the strongest oscillation period for autumn drought changes was 32 years (the primary cycle), with 3 years as the secondary cycle, and 12 years as the tertiary cycle (Figure 4c’). The SPEI winter drought cycles were 4–6 years, 9–13 years, and 28–34 years (Figure 4d). On the 4–6-year scale, there were three dry–wet cycles in winter; on the 9–13-year scale, there were two dry–wet cycles; and on the 28–34-year scale, there were two significant dry–wet cycles. The variance graph shows a clear oscillation at the 32-year time scale, indicating that 32 years was the primary cycle, 11 years the secondary cycle, and 5 years the tertiary cycle (Figure 4d’).

As shown by the real part of each wavelet at the seasonal scale and the variance diagram in Figure 4, the oscillation period of drought change in China in four seasons was 30–32 years, consistent with the 32-year oscillation period obtained from the annual-scale wavelet analysis. From the wavelet real part contour plots Figure 4a–d, it can be seen that the year 2021 was the end of the wet period of the dry–wet cycle, and that China will enter the arid part of the dry–wet cycle, i.e., indicating a trend toward aridification.

3.3. Spatial Differentiation Characteristics of SPEI-12

3.3.1. Overall Spatial Distribution Pattern of SPEI across Different Decades

The SPEI-12 values of 2254 meteorological stations in China were interpolated by inverse distance weight space in ArcGIS 10.8.1 software, and the spatial distribution (Figure 5a–f) and average SPEI spatial distribution (Figure 5g) of each decadal drought in China over the past 61 years were obtained, as shown in Figure 5. In the 1960s, droughts were most severe in the Xinjiang region of China, followed by east-central Qinghai Province and western Gansu Province. Mild droughts were experienced in eastern and northeastern China, while Tibet, northern central China, and western Yunnan Province were wetter (Figure 5a). By the 1970s, droughts gradually expanded in Xinjiang, with droughts also occurring in northern Tibet and western Qinghai; widespread droughts in eastern China shifted northward, with most of central China being affected by drought; drought conditions increased in Heilongjiang Province in northeastern China; while the Guangdong–Guangxi region in southern China shifted from droughts to wet conditions (Figure 5b). In the 1980s, droughts gradually weakened in Xinjiang and Heilongjiang, while drought intensity increased in Tibet, Guizhou, Shandong, and northern Guangxi, with the highest decadal average drought intensity of −1.09106; meanwhile, Shaanxi, eastern Sichuan, and southern Gansu were wet, with the highest moisture intensity of 1.15548 (Figure 5c). In contrast, the spatial distribution of drought in the 1990s was quite different from that of the previous decade, with droughts converging toward the center; stronger droughts occurred in southern Gansu, Shaanxi, Shanxi, eastern Sichuan, Qinghai, and eastern Tibet, while Xinjiang exhibited a moist climate (Figure 5d). By the 21st century, stronger droughts shifted to northeastern China, with higher intensity droughts occurring over large areas in eastern Inner Mongolia, Heilongjiang, and western Jilin provinces, and the highest decadal mean drought index was −1.14854, and minor droughts also occurring in central and southern China (Figure 5e). In the last decade, 2010–2021, the spatial pattern of drought changed again, with high-intensity droughts shifting to southwestern China, with the most severe in Yunnan Province and parts of southern Tibet, where the highest decadal mean drought intensity was −1.10677; in this decade, the wettest areas were central Qinghai Province and southeastern Xinjiang, while parts of Heilongjiang and Jiangsu–Zhejiang also shifted to wetter conditions. At the same time, Heilongjiang and parts of Jiangsu and Zhejiang also changed to wet (Figure 5f).

As shown by the spatial and temporal evolutions of droughts over the 61 years, the average drought intensity was highest in northwestern Qinghai Province, while high moisture intensity occurred in the western part of Inner Mongolia Autonomous Region (Figure 5g). In terms of the changes in the spatial pattern of drought over the six decades in China, the areas experiencing high-intensity droughts generally show a pattern of initially moving from the northwest to the southwest, then from the southwest to central China, later from central China to the northeast, and finally back to the southwest.

3.3.2. Trends in SPEI in Seven Geographic Regions

The drought intensity of sites in northeast China fluctuated greatly, mostly concentrated between −0.75 and 0.25, and the mean value of drought intensity of all sites was close to zero (Figure 6). The distribution of drought intensity of the sites in north China exhibited a wider range, with the highest moisture intensity at 1.5 or more, the highest drought intensity close to −1.5, and the drought intensity of most of the sites being concentrated between −0.25 and 0.125. The mean value of drought intensity of all stations was less than zero. The intensity of stations in east China was mostly concentrated between −0.125 and 0.375, the mean value of drought intensity of all stations exceeded zero, and the trend of drought intensity change across 61 years was clearly apparent; the increment of drought intensity was 0.139 per decade. The intensity of fluctuation at stations in south China was more clearly apparent and was within the range of −0.75–0.75, with the mean of drought intensity of all stations exceeding zero. The drought intensity of stations in central and northwestern China were evenly distributed between −0.75 and 0.75, and the annual change in drought intensity in central China was relatively smooth, with the trend clearly apparent. The drought intensity of sites in northwest China fluctuated slightly, but the increase in the drought index was also clearly apparent, with a trend of 0.124 per decade. The drought intensity of sites in southwest China exhibited the smallest range, and fluctuations were relatively smooth before the end of the 1990s, but during the 21st century, fluctuations in drought intensity were more significant, showing a downward trend in the 2000s and a rapid upward trend in the 2010s. Overall, severe drought areas included the northeast, central China, northwest, and southwest regions, while the central China, northwest, and southwest regions exhibited periods of higher moisture intensity. Over time, these three regions exhibited pronounced changes in wetness and dryness. In addition, the trend across the seven major regions indicates that the aridity index of all major regions increased, i.e., a shift to increased wetness. Therefore, in the short term, the trend of drought in China was weakening, while the trend of moistness strengthened.

3.3.3. Change in Average Annual SPEI in 31 Provinces

The occurrence of drought in 31 Chinese provinces and municipalities in the past 61 years was analyzed (Figure 7). For SPEI values exceeding −0.5, they were classified according to Liu et al. [30], with −0.5 < SPEI < 0.5 classified as normal, 0.5 ≤ SPEI < 1 as mildly wet, 1 ≤ SPEI < 1.5 as moderately wet, 1.5 ≤ SPEI < 2 as severely wet, and SPEI ≥ 2 as extremely wet. From the 1960s to the 1980s, most of the 31 provinces and municipalities were at a high level of drought, and the average SPEI values of Guangdong, Guangxi, and Jiangxi in 1963 were −1.58, −1.62, and −1.99, respectively, indicative of severe droughts. The year 1966 saw a number of provinces in central China with drought levels exceeding severe droughts, and the SPEI of Henan in 1978 was −2.17, which indicates extreme drought. In 1978, some provinces in central and eastern China were in extreme drought, with average drought indices of −2.16, −2.34, and −2.06 in Anhui, Jiangsu, and Shanghai, respectively. In 2012 and onward, the occurrence of severe and extreme droughts decreased significantly, occurring only in individual years, including in the Tibet Autonomous Region in 2015, Hubei Province and Yunnan Province in 2019, and Guangdong Province in 2021. In terms of the drought years in each province, 24 of 31 provinces had an average drought index indicative of mild droughts and above for 15 years of the 61 years, 27 provinces had an average drought index indicative of moderate droughts or above for more than 4 years, and 23 provinces had an average drought index indicative of severe droughts or above for at least one year. However, only four provinces had an average drought index indicative of extreme droughts or above for at least one year in the past 61 years: Henan in 1966, and Anhui, Jiangsu, and Shanghai in 1978. This indicates that the extent of drought in China was serious in 1966 and 1978. Overall, in recent years, because of the increase in precipitation, the number of wet provinces has gradually increased, the number of dry provinces has gradually decreased, the number of provinces with extreme and severe drought has been decreasing, and there has been a tendency for China to become wetter.

3.4. Drought Intensity and Station Ratio in Each Year

3.4.1. Drought Intensity

The calculated daily SPEI values were used to calculate the cumulative drought intensity at each drought level in each of the past 61 years (Figure 8). The intensity of mild droughts in China has essentially been stable throughout the past 61 years and has been maintained at 200 or less, while the intensity of moderate droughts has fluctuated greatly in the past 61 years, reaching a maximum cumulative intensity of 663.78 in 1986. The intensity of severe drought has fluctuated considerably in the past 61 years, and was at a high level in the periods of 1963–1972, 1978–1981, and 1997–2001. The intensity of extreme droughts was the highest in 1965, reaching 682.66; since the 1980s, the intensity of extreme droughts has been gradually decreasing, observed at below 100 in most years (Figure 8). This indicates that the intensity of extreme droughts has been gradually decreasing in the past 61 years in China. Analyses of the temporal distribution of drought intensity in China over the 61 years have demonstrated that in the 1960s and 1970s, drought intensity was generally high; the drought intensity in 1978 reached the highest value in 61 years (1836.42). During the 1980s and 1990s, drought intensity declined, but during the 2000s onwards, drought intensity gradually increased, attaining a maximum of 1320.04 in 2011. Between 2012 and 2021, annual drought intensity gradually decreased, with a minimum of 261.55 in 2020. Drought intensity exceeded 1800 in two years, 1966 and 1978. Over 52 years experienced a drought intensity exceeding 400, which shows that China has often been affected by droughts, although the cumulative value of drought intensity in the past decade has decreased. However, it remains important to implement drought prevention measures and protect farmlands to reduce the impact and extent of drought.

3.4.2. Drought Station Percentage

The percentage of stations with mild droughts was approximately 15% in most years, with small fluctuations, and the percentage of stations with moderate droughts was at least 10% in most years, with a clear downward trend in the last decade (Figure 9). The percentage of stations with severe drought fluctuated slightly, with a maximum of 14.08% (1978), while in other years, the percentage of stations with severe drought was mostly below 6%. The percentage of stations with extreme drought was below 2% in most years, except in 1965 when it reached 12.74%; a gradual decline in the last decade was apparent. The percentage of drought-free stations fluctuated greatly in the 61 years, ranging from 42.12% to 89.25%, with the highest value in 2020; this percentage has been steadily increasing in the past decade. This implies that the number of drought-free stations will be increasing, and the number of drought-affected areas in China will be gradually reduced in the near future. Overall, a period of low drought occurrence was seen from the 1980s to the mid-1990s and from 2010 to 2021, while a period of high drought occurrence was seen in the 1960s.

4. Discussion

Over the past 61 years, the SPEI in China has exhibited an increasing trend, indicating an overall tendency towards a more humid climate. Wang et al. [33] have analyzed two indices (SPI and SPEI_PM) in China from 1961 to 2011, and they have concluded that there was a general wetting phenomenon in northwest China and the central to northeastern part of the Qinghai–Tibetan Plateau, but there was an increase in drying in the central and southwestern part of the country. Li et al. [34] have analyzed changes in national drought using the SC-PDSI, and they concluded that there was no obvious trend of increasing aridity nationally between 1961 and 2015, but that there were differences in local areas. In the context of global warming, there is no evidence that the severity of drought in China has increased nationwide, while the SPI and SPEI_PM in both the extreme arid and arid regions have shown a clear wetting trend in the last 52 years [35,36,37]. Continued warming will further increase the water vapor content of the atmosphere, which in turn will lead to non-uniform changes in precipitation intensity and distribution across the country [38,39,40]. Therefore, despite the overall trend towards a more humid climate in China, regional variations in drought and moisture conditions still exist due to the uneven distribution of precipitation [40,41,42,43].

The spatial location of drought-prone areas in China has varied over the past 61 years. In the 1960s and 1970s, drought-prone areas were mainly located in the northwest region; in the 1980s, drought-prone areas were mainly located in the southwest; in the 1990s, drought-prone areas were shifted to central China and the southeast part of northwest China; from the 2010s to the present, drought-prone areas shifted to the southern part of the southwest region and places like Yunnan. Wang et al. [44] have explored the spatial and temporal evolution characteristics of China’s decadal droughts from 1960 to 2011 based on the SPI index, and they found that China’s drought-prone areas shifted from northwest China to north China and to the east part of southwest China; then to central China and the southeast part of northwest China; and then to northeast China, the east part of northwest China, and the east part of southwest China. Liao et al. [21] have analyzed the spatial and temporal distributions of drought in China from 1961 to 2015 based on the MCI and demonstrated that north China, the Yellow River and Huaihuai River, eastern northwest China, western northeast China, most of southwest China, and Inner Mongolia are drought-prone areas in China. Our results are consistent with the conclusions obtained from earlier studies. Atmospheric circulation anomalies are the main cause of precipitation anomalies and wet and dry changes [45,46]. Monsoon and circulation have a direct influence on the geographical distribution and amount of precipitation in China. Some studies show that the high latitude and strong westerly circulation in Asia at the end of the 20th century [47] caused a strong descent of high-altitude air currents, resulting in adiabatic warming of the air, leading to positive feedback with the strengthening of continental warm high pressure, which is the main cause of drought and low rainfall in north China. The wet trend in southeastern China at the end of the 20th century has been attributed by some to the anomalous anticyclone in southern Japan that transported warm and humid air from the tropical Pacific Ocean to southern China [48].

In this paper, it is concluded that the abrupt change point of the drought time-series China over the last 61 years is 1989, and that the main cycle of annual and seasonal dry and wet changes has a 32-year period. Liang et al. [49] have studied the characteristics of dry and wet changes in the Tibetan Plateau based on the 1980–2014 daily SPEI, and they observed that the abrupt change in droughts in the Tibetan Plateau occurred in 1990 and that the drought intensity decline was more pronounced after 1990, showing a more obvious trend of drought mitigation. Wu et al. [50] have studied drought characteristics in Xinjiang across multiple temporal- and spatial-scales from 1961 to 2020 by using the MCI, and they demonstrated that 1987 was the key abrupt change year between wet and dry conditions in Xinjiang, which is consistent with the present study. Li et al. [51], using wavelet analysis to study the drought trend of the Haihe River Basin from 1961 to 2010, have found that within the large-scale cycle of 25–40 years, the relative humidity index anomalies experienced three quasi-oscillations of dry and wet alternation, and the contour line was not closed in 2010, indicating a period of relative wetness thereafter. While the study time of this paper is 1961–2021, in the past 10 years, the dry and wet changes in China match the prediction of Li et al. [51], indicative of the reliability of the large-scale cycle range of dry and wet changes in China that was obtained through wavelet analysis.

In recent years, the overall drought intensity in China has shown a fluctuating downward trend, and the proportion of stations with mild drought has increased year by year. Gao et al. [52] have analyzed the spatiotemporal evolution characteristics of drought in China over the past 30 years based on the MCI, and they noted that drought intensity has gradually decreased in the past 30 years and that the years of strongest annual drought intensity in some of the northern river basins mainly appeared at the end of the 20th century and beginning of the 21st century. Wang et al. [53], based on SPI analysis, have observed a downward trend in both the annual-scale drought station percentage and drought intensity in the Hexi Corridor region from 1960 to 2016, and also that drought intensity in the study area was mainly dominated by mild and moderate droughts. The results on drought intensity and drought station percentage in this paper are consistent with these earlier studies [54]. With global warming, the water cycle accelerates, and the increase in precipitation exceeds the increase in evaporation, leading to a shift in climate towards warm and humid conditions and resulting in a weakening of drought intensity [55,56,57,58].

In addition, in terms of choosing a spatial interpolation method, this study selects a more suitable inverse distance weight interpolation method with reference to the conclusions of Rahman et al. [59] and Chen et al. [60], comparing the applicability of different spatial interpolation. However, due to the uneven distribution of meteorological stations, the distribution of meteorological stations in southwest China is relatively scattered, and the research accuracy is lacking, which leads to the limitations in regional drought research. In view of this, in future drought research work, we can explore the use of satellite remote sensing technology and climate models to assist in supplementing and verifying the data of meteorological stations in southwest China, so as to make up for the limitations of insufficient data.

5. Conclusions

(1)

Both drought frequency and drought intensity in China have fluctuated between 1961 to 2021, with a decreasing trend in the SPEI at different timescales between the 1960s and 1970s and a more pronounced upward trend in the SPEI at different timescales since the 2010s. Overall, the SPEI-1, SPEI-3, and SPEI-12 demonstrated that drought intensity gradually decreased at rates of 0.005 per decade, 0.021 per decade, and 0.092 per decade, respectively.

(2)

In the past 61 years, there has been a shift in areas with high drought intensity in China: northwest China → central southwest China → central China → northeast China → southwest China; each province generally experienced a downward trend in drought intensity over the 61 years. The number of provinces experiencing drought has been significantly reduced, with most provinces showing a wetting trend.

(3)

From 1961 to 2021, abrupt changes from dry to wet conditions in China occurred in 1989. The duration of the annual dry and wet cycle is 29–35 years, and the duration of seasonal dry and wet cycle is 30–32 years; the oscillatory cycles of dry and wet changes essentially have the same time scale.

(4)

Over the past 61 years, the drought intensity in China has shown fluctuations but has generally exhibited a decreasing trend, with the most significant decline observed in extreme drought events. The percentage of stations experiencing some level of drought has also shown a downward trend, although the percentage of stations experiencing mild drought has remained relatively stable. The variations in the percentage of stations experiencing severe and extreme drought are noticeable. In recent 61 years, the drought severity in China has gradually weakened, indicating an overall trend towards a more humid climate.