Analysis of the Spatio-Temporal Evolution Characteristics and Influencing Factors of Extreme Climate Events in Jilin Province from 1970 to 2020

Abstract

Under global warming, the rising frequency and intensity of extreme climate events pose challenges to disaster prevention and sustainable development. Based on daily meteorological observations from 1970 to 2020 in Jilin Province, this study analyzes the spatiotemporal evolution and driving mechanisms of extreme temperature and precipitation events. Linear trend analysis and the Mann–Kendall test were employed to examine temporal trends and abrupt change years in extreme temperature and precipitation indices. Wavelet analysis was used to identify dominant periodicities and multi-scale variability. Empirical Orthogonal Function Analysis (EOF) revealed the spatial distribution characteristics of variability in extreme precipitation and temperature across Jilin Province, identifying high-incidence zones for extreme temperature and precipitation events. Additionally, Pearson correlation analysis was to investigate the correlation patterns between extreme climate indices in Jilin Province and geographical environmental factors alongside atmospheric circulation indicators. Results show that: (1) Warm-related temperature indices display significant upward trends, while cold-related indices generally decline, with abrupt changes mainly occurring in the 1980s–1990s and dominant periodicities of 3–5 years. Precipitation indices, though variable, show general increases with 3–4year cycles. (2) Spatially, most indices follow an east–high to west–low gradient. Temperature indices exhibit spatial coherence, while precipitation indices vary, especially between the northwest and central-southern regions. (3) The Arctic Oscillation (AO) exhibits a significant negative correlation with the extreme cold index, with correlation coefficients ranging from −0.31 to −0.46. It shows a positive correlation with the extreme warm index, with correlation coefficients between 0.16 and 0.18, confirming its regulatory role in cold air activity over Northeast China, particularly elevation and latitude, influence the spatial distribution of precipitation. These findings enhance understanding of extreme climate behaviors in Northeast China and inform regional risk management strategies.

1. Introduction

Global climate warming has emerged as one of the most pressing environmental challenges of the 21st century. According to the Intergovernmental Panel on Climate Change (IPCC) Sixth Assessment Report, the global mean surface temperature has increased by over 1.4 °C since pre-industrial times due to anthropogenic influences, with projections indicating that the critical thresholds of 1.5 °C or even 2 °C could be exceeded in the near future [1,2,3]. This warming trend has significantly disrupted the stability of the global climate system, leading to more frequent and intense extreme climate events that are increasingly becoming the norm rather than the exception [4,5,6,7,8]. Compared with average climatic conditions, extreme events are characterized by their pronounced anomalies and far-reaching impacts, and have become a major manifestation of global climate risk [9,10]. The rising frequency and severity of these events are posing systemic challenges to public health, agricultural productivity, water resource security, and regional ecosystems [11,12]. In recent years, extreme climate events—such as heatwaves in Europe, wildfires in Australia, and floods in Asia—have caused extensive human casualties and economic losses worldwide. Against this backdrop, in-depth research on the identification, monitoring, and underlying mechanisms of extreme climate events has gained increasing theoretical and practical significance for enhancing regional adaptation capacity and disaster risk reduction.

Since the early 21st century, Sillmann et al. [13], based on multi-model simulations, projected a substantial decrease in cold events and a sharp increase in warm extremes in the future. Empirical studies by Tavakol et al. [14], Ruml et al. [15], and Roy et al. [16] further confirmed the rising number of extreme heat days and narrowing diurnal temperature ranges in regions such as the Mississippi River Basin (USA), Serbia, and India. The IPCC also emphasized that as global temperatures approach and exceed critical thresholds, the frequency and intensity of warm extremes will markedly increase, while cold extremes will diminish [1]. Regarding precipitation, a growing body of research suggests that global warming has intensified the hydrological cycle, resulting in stronger precipitation extremes and more complex wet–dry spatial patterns. For example, Tan identified distinct regional differences in annual maximum precipitation (AMP) changes across Canada [17], while Morales et al. [18] observed increasing trends in the frequency of extreme precipitation events in Brazil’s northeastern region. Lovino et al. [19] reported a pronounced nonlinear trend in extreme precipitation in northeastern Argentina, and Jung et al. [20] noted a significant upward trend in summer precipitation extremes in South Korea, contrasted by decreasing trends in spring and winter. Continental-scale studies in Europe [21], Asia [22], South America [23], and Africa [24] consistently affirm the intensification of warm extremes and the decline of cold events, although spatial responses vary considerably due to atmospheric circulation patterns, topography, and underlying surface conditions.

To better identify and assess extreme climate events, scholars have continuously refined both indicator systems and analytical methodologies. The Expert Team on Climate Change Detection and Indices (ETCCDI), under the World Meteorological Organization (WMO), developed a suite of 27 extreme climate indices that are globally applicable, robust in signal-to-noise ratio, and computationally stable. These indices have been widely adopted in global and regional climate studies [25]. Methodologically, early investigations mainly relied on linear regression and the Mann–Kendall trend test. In recent years, however, researchers have integrated wavelet analysis, multi-scale decomposition, and empirical orthogonal function (EOF) analysis to construct a comprehensive analytical framework encompassing trend detection, periodicity analysis, and spatial heterogeneity assessment [26]. In terms of underlying mechanisms, a growing number of studies have begun to elucidate the links between extreme climate events and large-scale atmospheric circulation systems—particularly the Arctic Oscillation (AO), Pacific Decadal Oscillation (PDO), and North Atlantic Oscillation (NAO)—whose interannual variability has been shown to significantly affect climate extremes at mid- to high latitudes [27]. Additionally, solar activity (e.g., sunspot number, SN) has been found to exhibit coupling effects with temperature and precipitation anomalies in plateau regions [28]. Geographically, factors such as topographical complexity, elevation gradients, and latitudinal variation play important roles in modulating the spatial distribution of extreme events, especially across mountainous mid- to high-latitude areas in Asia [29].

Jilin Province, located in the central part of Northeast China, represents a typical mid- to high-latitude agricultural production region characterized by both monsoonal and continental climatic features. The province exhibits high climatic sensitivity, with ecosystems responding strongly to variations in temperature and precipitation. As the core grain-producing area of Northeast China, Jilin’s total grain output reached 38.03 billion kg in 2020, ranking among the top provinces nationwide. The frequent occurrence of extreme heat, heavy rainfall, and drought events in recent decades has posed increasing threats to food security and ecological stability. Therefore, selecting Jilin Province as the study area provides a representative case for exploring the spatiotemporal evolution of extreme climate events in mid- to high-latitude agricultural regions under global warming, and offers a valuable reference for regional climate risk assessment and adaptive management in similar climatic zones. Based on this context, the present study focuses on Jilin Province, a representative mid- to high-latitude agricultural production area in Northeast China. A framework of 21 extreme climate indices based on the ETCCDI system is constructed, and an integrated methodology—combining trend analysis, abrupt change detection, wavelet transform, and empirical orthogonal function (EOF) decomposition—is applied. The spatiotemporal evolution of extreme climate events during the period 1970–2020 is systematically evaluated. In addition, atmospheric circulation indices (e.g., AO, PDO, SOI) and geographic factors (e.g., elevation, latitude) are incorporated into a multivariate correlation analysis to identify the dominant drivers of climate extremes in the study region. Compared with existing research, this study offers two main contributions: (1) Methodologically, it integrates temporal trend, periodicity, and spatial pattern analyses to provide a holistic characterization of extreme climate dynamics; (2) Mechanistically, it builds a causal linkage framework between atmospheric circulation patterns and geographical controls at the provincial scale. These efforts not only contribute to a deeper understanding of extreme climate evolution in Northeast China, but also provide scientific support for climate risk management and adaptation strategies in major agricultural regions.

2. Materials and Methods

2.1. Study Area

Jilin Province is located at the geometric center of Northeast Asia, with geographical coordinates ranging from 121°38′ E to 131°19′ E and 40°50′ N to 46°19′ N (Figure 1). Covering a total area of approximately 187,400 square kilometers, the province accounts for about 2% of China’s national territory. Jilin features diverse landforms and pronounced topographic variation, with an overall terrain that descends from the southeast to the northwest. The province is situated in a temperate continental monsoon climate zone, characterized by distinct seasonal transitions and a high degree of temporal overlap between periods of precipitation and high temperature, reflecting a typical pattern of synchronized rainfall and heat. The annual average diurnal temperature range lies between 35 °C and 42 °C, with a frost-free period lasting approximately 100 to 160 days. Annual sunshine duration ranges from 2259 to 3016 h. The degree of climatic humidity gradually decreases from southeast to northwest, exhibiting a transitional pattern from humid and semi-humid conditions toward semi-arid conditions.

2.2. Data Sources

The meteorological observation data used in this study were primarily obtained from the National Meteorological Science Data Center of the China Meteorological Data Service Center (http://data.cma.cn/) (accessed on 8 January 2025). Definitions of the 21 selected extreme climate indices are presented in

Table 1. Selected Extreme Climate Indices.

Descriptive Name	Code	Definition	Unit
Cool Days	TX10p	Number of days when daily maximum temperature < 10th percentile	Days
Cool Nights	TN10p	Number of days when daily minimum temperature < 10th percentile	Days
Warm Days	TX90p	Number of days when daily maximum temperature > 90th percentile	Days
Warm Nights	TN90p	Number of days when daily minimum temperature > 90th percentile	Days
Frost Days	FD0	Annual count of days when TN (daily minimum temperature) < 0 °C	Days
Ice Days	ID0	Annual count of days when TX (daily maximum temperature) < 0 °C	Days
Summer Days	SU25	Annual count of days when TX (daily maximum temperature) > 25 °C	Days
Max Tmax	TXx	Monthly or annual maximum value of daily maximum temperature	°C
Min Tmin	TNn	Monthly or annual minimum value of daily minimum temperature	°C
Cold Spell Duration Indicator	CSDI	Number of days when daily maximum temperature < 10th percentile for ≥6 days	Days
Warm Spell Duration Indicator	WSDI	Number of days when daily minimum temperature > 90th percentile for ≥6 days	Days
Diurnal Temperature Range	DTR	Difference between daily maximum and minimum temperature	°C
Precipitation in very wet days	R95p	Annual total precipitation from days > 95th percentile	mm
Precipitation in extremely wet days	R99p	Annual total precipitation from days > 99th percentile	mm
Number of heavy precipitation days	R10mm	Annual count of days when precipitation ≥ 10 mm.	Days
Number of very heavy precipitation days	R20mm	Annual count of days when precipitation ≥ 20 mm.	Days
Number of Rainstorm Days	R25mm	Annual count of days when precipitation ≥ 25 mm.	Days
Simple Daily Intensity Index	SDII	Ratio of total precipitation to number of wet days (≥1 mm)	mm·d−1
Max 1-day precipitation amount	RX1day	Monthly or annual maximum 1-day precipitation.	mm
Max 5-day precipitation amount	RX5day	Monthly or annual maximum consecutive 5-day precipitation total.	mm
Annual total wet-day precipitation	PRCPTOT	Annual total precipitation from wet days (days with precipitation ≥ 1 mm).	mm
Descriptive Name	Code	Definition	Unit

. Based on the RClimDex 1.0 software package [30,31,32], these indices were calculated using quality-controlled daily data from 28 meteorological stations across Jilin Province for the period 1970–2020. To further explore the linkages between extreme temperature and precipitation events and large-scale atmospheric circulation systems, several representative atmospheric circulation indices were selected as external driving factors. These include the Southern Oscillation Index (SOI), North Atlantic Oscillation (NAO), Arctic Oscillation (AO), North Pacific Index (NP), and Pacific Decadal Oscillation (PDO), which were obtained from the Physical Sciences Laboratory (https://psl.noaa.gov/data/climateindices/list/) (accessed on 10 January 2025), as well as the Sunspot Number (SN), sourced from the Royal Observatory of Belgium (https://www.sidc.be/SILSO/datafiles) (accessed on 10 January 2025). Pearson correlation analysis was applied to examine the statistical relationships between these climate drivers and the extreme climate indices.

2.3. Methodology

Based on daily meteorological observations from 1970 to 2020, this study integrates Linear Trend Analysis, the Mann–Kendall mutation test, wavelet analysis, and Empirical Orthogonal Function Analysis to examine the spatiotemporal evolution characteristics of 21 ETCCDI-based extreme climate indices in Jilin Province. Furthermore, Pearson correlation analysis was applied to explore the relationships between extreme climate indices and atmospheric circulation factors (AO, SN, PDO, SOI, NP) as well as geographical variables (latitude, longitude, and elevation), providing insights into their possible influences on regional climate variability. Figure 2 presents the overall research framework and workflow of this study.

2.3.1. Linear Trend Analysis

$$Y_t=a·t+b$$

(1)

where: $Y_t$ represents the value of the climate index in a given year, $t$ denotes the time index, increasing sequentially from the initial year of 1970, $a$ is the slope coefficient indicating the trend rate, and $b$ is the intercept, representing the estimated value at the beginning of the series. A positive value of (i.e., $a$ > 0) suggests an increasing trend in the index over time, while a negative value ($a$ < 0) indicates a decreasing trend.

2.3.2. Mann–Kendall Mutation Test

The MannKendall (M-K) trend test is a widely used non-parametric statistical method in climatological trend analysis and provides a quantitative approach [33,34,35]. For $n$ independent samples, the test statistic is as follows:

$$S_k = {\sum }_{i=1}^k{r}_i k = 1, 2, 3, \dots , n$$

(2)

$$r_i=\left\{\begin{array}{c}1,x_j>x_i\\ 0,x_j>x_i\end{array}\right. (j=1,2,\dots ,i)$$

(3)

$S_k$ represents the total number of times the $i$ sample is greater than later samples $x_i>x_j (1\le i\le j)$. If the sequence $S_k$ is randomly independent, its mean and variance are:

$$E\left(S_k) = k\right(k - 1)/4$$

(4)

$$Var\left(S_k\right)=k(k-1)(2k+5)/72k=1,2,\dots ,n$$

(5)

Standardizing $S_k$:

$${UF}_k={\displaystyle \frac{(S_k-E(S_k\left)\right)}{\sqrt{Var\left(S_k\right)}}}$$

(6)

where $U{F}_1 = 0$. When $U{F}_k>0$, the sequence shows an increasing trend; otherwise, it indicates a decreasing trend. At a significance level of $\mathsf{\alpha }=0.05$, the critical value is $U_{0.05}=\pm 1.96$. If $\left|{UF}_k\right| > U_{0.05}$ exceeds this threshold, the trend change in the sequence is considered statistically significant.

The sequence $x_1,x_2,{\dots ,x}_n$ is rearranged in reverse order, and the same calculations are performed to obtain the ${UB}_k$ statistic series. When the intersection point of the ${UF}_k$ and ${UB}_k$ curves lies within the significance level boundary, that point is identified as the mutation point.

2.3.3. Empirical Orthogonal Function Analysis

EOF is a statistical method based on the analysis of the eigenstructure of a matrix, which effectively separates the spatial distribution patterns and temporal variation of meteorological variable fields. As such, it has been widely applied in meteorological and geographical research [36,37,38].

The EOF method performs a covariance decomposition on the anomaly data matrix $X_{m \times n}$ to extract the dominant spatiotemporal modes. The covariance matrix is defined as:

$$C = {\displaystyle \frac{1}{n}}X{X}^T$$

(7)

An eigenvalue decomposition is performed on matrix $C$, yielding:

$$C = V\Lambda V^T$$

(8)

where:

(1)

$\Lambda =diag({\lambda }_1,{\lambda }_2,\dots ,{\lambda }_m)$ is the eigenvalue matrix $({\lambda }_{1 }\ge {\lambda }_2\ge \dots \ge {\lambda }_m \ge 0)$, representing the variance contribution of each mode;

(2)

$V=\left[v_1,v_2,\dots ,v_m\right]$ is the corresponding eigenvector matrix, representing the spatial modes (EOF modes);

(3)

The time coefficients (principal components) are obtained by projecting the original data: $A = V^{T}X$.

The proportion of variance explained by the $\mathrm{kth}$ mode is:

$$R_k={\lambda }_k/{\sum }_{i=1}^m{\lambda }_i\times 100\%$$

(9)

The cumulative variance explained by the first $p$ modes is:

$$G ={\Sigma }_{i = 1}^p{\lambda }_i/{\Sigma }_{i = 1}^m{\lambda }_i$$

(10)

In this study, the North test is employed to assess the significance of eigenvalue uncertainties. At the 95% confidence level, the error range of an eigenvalue $\mathit{\lambda }$ is given by:

$$e_j = {\lambda }_j\sqrt{2/n}$$

(11)

When $\Delta ={\lambda }_j-{\lambda }_{j + 1}-e_j \ge 0$ is satisfied, the modes corresponding to the two eigenvalues are considered independent, indicating a valid signal.

2.3.4. Continuous Wavelet Transform (CWT) Method

In this study, the Continuous Wavelet Transform (CWT) method is employed to perform multi-scale analysis of meteorological time series data [39,40,41]. Its mathematical expression is:

$$W(a,\tau )={\displaystyle {\int }_{-\infty }^{+\infty }x(t){\psi }^{*}(\frac{t-\tau }{a})}dt$$

(12)

where, the Morlet complex wavelet is used as the mother wavelet function $\psi \left(t\right)$. The wavelet power spectrum is defined as $P(a,\tau )={\left|W(a,\tau )\right|}^2$, which is used to reveal the localized energy variation characteristics of the signal across different time scales.

3. Results

3.1. Temporal Evolution Characteristics of Extreme Climate in Jilin Province

3.1.1. Trend Analysis of Extreme Temperature Indices

The interannual trends of extreme temperature indices in Jilin Province from 1970 to 2020 are presented in

Table 2. Temporal Variation Analysis Results of Extreme Temperature Indices.

Index	Trend Rate	Significance	Dominant Period	Mutation Year
TNn	0.39 °C/10a		3a	1980
TXx	0.31 °C/10a	*	3a	2013
DTR	−0.11 °C/10a	**	3a	1982
CSDI	−0.89 d/10a	**	7a	1977
WSDI	0.85 d/10a	**	4a	1999
FD0	−2.55 d/10a	**	4a	1988
ID0	−1.16 d/10a		5a
SU25	3.09 d/10a	**	3a	1999
TN10p	−2.70 d/10a	**	2a	1985
TN90p	−2.63 d/10a	**	4a	1988
TX10p	−1.76 d/10a	**	2a	1987
TX90p	1.75 d/10a	**	4a	1993

Note: ** and * indicate significance at the 99% and 95% confidence levels, respectively.

and Figure 3. It can be observed that both TNn and TXx exhibit upward trends, with linear trend rates of 0.39 °C/10a and 0.31 °C/10a, respectively. Among them, the increasing trend of TXx is statistically significant at the 95% confidence level (p < 0.05). DTR shows a significant decreasing trend (p < 0.01), with a trend rate of −0.11 °C/10a; although the overall trend is downward, an upward tendency is observed in the most recent decade, suggesting a widening gap between TXx and TNn. CSDI exhibits a significant decreasing trend (p < 0.01), whereas WSDI displays a clear increasing trend (p < 0.01), indicating an intensification of persistent warm events. Indices representing extreme cold events-FD0, ID0, TN10p, and TX10p-all show declining trends. In contrast, indices representing extreme warm events-TN90p, TX90p, and SU25-exhibit increasing trends and pass the 99% confidence level significance test (p < 0.01). These results collectively suggest a clear warming signal in the regional climate of Jilin Province.

The results of the mutation analysis indicate that abrupt changes in extreme temperature indices began to emerge around 1977, with the cold spell duration index (CSDI) serving as a representative case. The majority of abrupt changes occurred between 1980 and 1993, affecting several indices including TNn, DTR, FD0, TN10p, TN90p, TX10p, and TX90p. The mutation years for WSDI and SU25 were identified as 1999, while TXx showed a mutation in 2013. According to the Mann–Kendall test for the ID0 index, multiple intersections between the UF and UB curves were observed within the confidence interval, making it impossible to determine a definitive mutation point; thus, no significant abrupt change was identified for this index. Overall, the mutation years were mainly concentrated in the 1980s, characterized by a decline in cold-related indices and an increase in warm-related indices, indicating a clear trend toward a warmer climate.

The Morlet wavelet analysis results of extreme temperature indices in Jilin Province from 1970 to 2020 are summarized in Table 2, with Figure 4 illustrating the real-part contour plots of selected representative indices (CSDI, ID0, TX10p, and TX90p). Most indices exhibit clear multi-scale periodic fluctuations, primarily concentrated within short (2~7a) and medium (7~10a) periodic bands, though each index displays distinct dominant periodic structures. Specifically, CSDI shows a dominant period of approximately 7a, with a continuous high-energy band appearing in the 1980s through the early 2000s, indicating a stable medium-period oscillation. The ID0 index is characterized by a dominant 5a period, with concentrated energy around 1985, suggesting more prominent mid-frequency variations. TX10p exhibits a primary 2a cycle, with strong energy concentration across the entire time span, highlighting its pronounced interannual oscillatory behavior. In contrast, TX90p is dominated by a 4a period and displays progressively enhanced multi-scale spectral bands after 2000, indicating the emergence of more complex periodic superposition in recent decades.

3.1.2. Trend Analysis of Extreme Precipitation Indices

The interannual variation trends of extreme precipitation indices in Jilin Province from 1970 to 2020 are shown in

Table 3. Temporal Variation Analysis Results of Extreme Precipitation Indices.

Index	Trend Rate	Dominant Period
R95p	5.54 mm/10a	3a
R99p	4.15 mm/10a	3a
RX1day	1.16 mm/10a	3a
RX5day	0.71 mm/10a	3a
PRCPTOT	3.61 d/10a	3a
SDII	−0.02 d/10a	4a
R10mm	−0.11 d/10a	3a
R20mm	0.03 d/10a	3a
R25mm	0.06 d/10a	3a

and Figure 5. Among them, indices representing extreme precipitation intensity and extreme daily precipitation-such as R95p, R99p, RX1day, and RX5day-exhibit increasing trends at rates of 5.54 mm/10a, 4.15 mm/10a, 1.16 mm/10a, and 0.71 mm/10a, respectively, indicating a synchronous intensification in both the frequency and intensity of extreme heavy precipitation events in the region. The PRCPTOT index also shows an increasing trend, with a rate of 3.61 mm/10a, suggesting a general enhancement of regional moisture conditions. In contrast, SDII and R10mm exhibit decreasing trends, while R20mm and R25mm show no clear interannual trends.

Overall, the interannual variations in all extreme precipitation intensity indices are not statistically significant; therefore, mutation analysis was not conducted for precipitation indices. Wavelet power spectra of representative precipitation indices—PRCPTOT, R95p, and RX5day—are presented in Figure 6. All three indices display significant high-energy bands near the 3a period, which constitutes their dominant periodicity. Specifically, PRCPTOT shows a persistent enhancement of energy around the 3a scale from the 1980s to the early 2000s, indicating a stable short-term oscillatory pattern in annual total precipitation during this period. The R95p index exhibits strong high-frequency fluctuations around 1990 and in the most recent decade, reflecting the frequent occurrence and short-term perturbation characteristics of extreme heavy precipitation events. Although RX5day shares the same dominant periodicity as the other two indices, it displays broader spectral bands and more intense energy variation after 2000, revealing a non-stationary pattern characterized by multi-frequency superposition and phase-specific intensification.

3.2. Spatial Evolution Characteristics of Extreme Climate in Jilin Province

3.2.1. Spatial Trend Analysis of Extreme Temperature Indices

The first EOF mode of each extreme temperature index passed the significance test, and was thus selected for focused analysis to reveal the spatial distribution characteristics of extreme temperature patterns.

Figure 7 presents the spatial distribution of EOF eigenvectors of extreme temperature indices in Jilin Province from 1970 to 2020. The spatial distribution of extreme temperature indices is shown in Figure 4. The intensity-type indices (TNn, TXx, and DTR) exhibit positive loading values, indicating a consistent direction of change across Jilin Province. High TNn values occur in the southern part of the province as well as in Baicheng and northern Songyuan, showing a west-high and east-low pattern. TXx exhibits high values in Baicheng, Songyuan, and northeastern Liaoyuan, while low values appear in Yanbian Prefecture, forming a northeast-high and southwest-low distribution. DTR shows high values along the border of Yanbian, Jilin, and Baishan, and low values around Changchun and adjacent plains, presenting a north-high and south-low pattern. The duration-type indices (CSDI and WSDI) show generally consistent spatial distributions. The CSDI displays high values at the junction between Songyuan and Changchun, whereas the WSDI has high values in Tonghua, Baishan, Liaoyuan, and southern Jilin, corresponding to regions with stronger topographic and monsoonal influences. The frequency-type indices (FD0, ID0, SU25, TN10p, TN90p, TX10p, and TX90p) exhibit distinct spatial contrasts. High FD0 and ID0 values are found in western Jilin, while SU25 shows high values in eastern Baishan. TN90p and TX90p feature high-value areas in the west, whereas TX10p presents an opposite pattern with lower values in similar regions. Overall, extreme temperature indices display a general pattern of “high in the southwest and low in the northeast,” with greater variability in the western and southern areas and relatively stable conditions in the east. These spatial features reflect the combined effects of topography, latitude, and monsoonal circulation, indicating that high-latitude regions in Jilin respond more sensitively to ongoing climate warming.

3.2.2. Spatial Trend Analysis of Extreme Precipitation Indices

Figure 8 illustrates the spatial distribution of EOF eigenvectors of extreme precipitation indices in Jilin Province from 1970 to 2020. The intensity-type indices (R95p, R99p, RX1day, RX5day, and SDII) exhibit clear regional contrasts. R95p displays low values in the central and southern areas, decreasing from west to east and then increasing again. R99p has high values in the western and eastern parts of the province and relatively low values in the central–southern region. RX1day shows high values in the west and southeast, with lower values in the central part, whereas RX5day has high values mainly in the west. SDII presents a west-low and east-high distribution. PRCPTOT is positive overall, with high values concentrated in the central–southern region, showing a west-rising and east-declining pattern. Among the frequency-type indices, all except R20mm exhibit relatively low values. R20mm shows high values in the central–southern region, also displaying a west-rising and east-declining distribution. Overall, extreme precipitation in Jilin Province shows pronounced spatial differentiation: the southwestern and southern regions experience relatively concentrated strong precipitation events, whereas the eastern areas show weaker intensity. This pattern is closely related to topographic variation and atmospheric circulation, indicating significant regional differences in precipitation extremes—persistent dryness tends to occur in the eastern mountainous regions, while short-duration heavy rainfall is more common in the southern hilly and plain areas.

3.3. Analysis of Influencing Factors of Extreme Climate Indices in Jilin Province

3.3.1. Correlation Analysis Between Extreme Climate Indices and Atmospheric Circulation Factors

Large-scale atmospheric circulation patterns can lead to regional climate changes and subsequently influence extreme temperature events. In this study, Pearson correlation analysis was conducted between selected atmospheric circulation indices—namely SOI, SN, NAO, AO, NP, and PDO—and the extreme climate indices in Jilin Province. The results are shown in Figure 9.

Figure 9a presents the Pearson correlation heatmap between extreme temperature indices and major atmospheric circulation indices in Jilin Province. The Southern Oscillation Index (SOI) shows generally weak correlations with extreme temperature indices, displaying positive correlations only with cold indices such as FD0 and ID0, while exhibiting negative correlations with warm-related indices like WSDI, TN90p, and TX90p. This suggests that enhanced SOI phases may favor the occurrence of cold events while suppressing high-temperature extremes. The Sunspot Number (SN) displays a more complex relationship with temperature indices: it is negatively correlated with warm indices such as TX90p, TN90p, WSDI, and SU25, and positively correlated with cold indices such as DTR, CSDI, TX10p, and TN10p. Its strongest negative correlation is observed with TNn, indicating that increased solar activity may be associated with regional cooling effects. The North Atlantic Oscillation (NAO) shows positive correlations with TNn, TXx, and DTR, but negative correlations with TN90p, TX90p, and CSDI, reflecting its bidirectional influence on both warm and cold events. The Arctic Oscillation (AO) exhibits strong positive correlations with warm indices (TN90p, TX90p, SU25) and significant negative correlations with cold indices (FD0, ID0, TN10p, TX10p, CSDI), with the strongest negative correlation observed with TX10p (r = −0.60), highlighting AO as a key atmospheric driver of cold extremes. The North Pacific Index (NP) is negatively correlated with ID0 and positively correlated with TNn, but shows generally weak associations with other temperature indices. The Pacific Decadal Oscillation (PDO) has weak positive correlations with TNn and FD0, while its relationships with other indices are minimal, showing no evident regulatory effect.

Figure 9b illustrates the correlation analysis between extreme precipitation indices and atmospheric circulation factors. Similar to its relationship with temperature, SOI displays weak correlations, showing only slight positive associations with SDII, R95p, and R25mm, suggesting limited influence on precipitation intensity and heavy rainfall frequency.SN, in contrast, shows systematic negative correlations with most extreme precipitation indices, making it one of the most influential circulation factors in this context. Strong negative correlations are observed with SDII, R95p, R25mm, R99p, and RX1day, indicating that increased solar activity is often associated with suppressed regional precipitation processes. NAO shows relatively weak influence on precipitation, with slight positive correlations observed only for RX5day and R20mm. AO also exhibits generally weak correlations, with a relatively more pronounced negative correlation observed with R10mm, implying a potential role in moderating the frequency of moderate rainfall events. NP is broadly negatively correlated with precipitation indices, particularly with SDII, R20mm, and R10mm, where its influence appears more pronounced. PDO shows limited correlation with precipitation indices, displaying only a slight positive association with R10mm and no evident regulatory impact overall.

3.3.2. Relationship Between AO and Extreme Climate Indices

It is evident that atmospheric circulation exerts a certain degree of influence on the extreme climate indices in Jilin Province. Among the selected indices, the Arctic Oscillation (AO) shows relatively strong correlations and clear variability over time. Therefore, AO is selected for further analysis of its time-frequency relationships with representative extreme climate indices: TN10p, TN90p, R95p, and RX1day. The results are presented in Figure 10.

As shown in the figure, AO and TN10p exhibit significant resonance in a high-energy region from 1996 to 2002, with a dominant period of 3~4.5 a. The phase arrows point to the lower right, indicating a negative correlation, with TN10p lagging behind AO (Figure 10a). Additionally, a weaker resonance region appears around 1980, with a period of approximately 6~8 a, also showing a negative relationship. The resonance between AO and TN90p is relatively weak, with intermittent high-energy zones observed from the late 1990s to around 2005. The dominant periods mostly fall within the 3~6 a range, and the phase arrows generally point to the upper right, indicating a positive correlation, with TN90p lagging behind AO (Figure 10b). AO and R95p show two notable periods of significant resonance: one around 1980, with a dominant period of 6~8 a, and another from 2002 to 2010, with a 2~3 a dominant cycle. In the former, phase arrows point to the lower left, suggesting a negative correlation with R95p lagging AO; in the latter, the arrows point to the upper right, indicating a positive correlation with R95p also lagging AO (Figure 10c). For AO and RX1day, a strong resonance band is observed from 1996 to 2004, with a dominant period of 2.5~4 a. The phase arrows point to the upper right, suggesting a positive correlation with RX1day lagging behind AO (Figure 10d). In addition, a weaker resonance region with a 6–8year period is found in the mid-1980s.

3.3.3. Correlation Analysis Between Extreme Climate Indices and Geographic Factors

Figure 11a presents the Pearson correlation heatmap between extreme temperature indices and geographic factors, including latitude, longitude, and elevation. Among the intensity indices, TNn and TXx are positively correlated with latitude, while TNn is significantly negatively correlated with elevation (p < 0.01), indicating that extreme low temperatures are more likely to occur in high-altitude areas. TXx is significantly correlated with all three geographic factors. DTR shows significant correlations with both latitude and elevation, reflecting notable spatial variability. For duration indices, CSDI is significantly positively correlated with elevation, whereas WSDI is primarily and significantly correlated with latitude. Among the frequency indices, ID0 and SU25 show positive correlations with latitude, and SU25 is significantly negatively correlated with elevation. TX90p is significantly positively correlated with elevation but negatively correlated with latitude. These results suggest that elevation is the most critical geographic factor influencing the spatial distribution of extreme temperature events, followed by latitude. Cold events are more likely to occur in high-elevation areas, while low-elevation and low-latitude regions exhibit a more pronounced warming trend.

Figure 11b shows the correlation between extreme precipitation indices and geographic factors. Except for SDII, all extreme precipitation indices are significantly negatively correlated with latitude, with PRCPTOT and R10mm exhibiting the strongest correlations (as high as −0.83). Longitude is positively correlated with PRCPTOT and R10mm, and negatively correlated with SDII and RX1day, with SDII showing a statistically significant negative correlation. Elevation is positively correlated with PRCPTOT and R10mm, but significantly negatively correlated with SDII, suggesting that precipitation amounts may increase with altitude, while precipitation intensity tends to decrease. It is evident that topography and latitude jointly regulate precipitation variability in Jilin Province. The gradual transition from the western plains to the southeastern mountainous regions enhances orographic precipitation.

4. Discussion

Based on daily meteorological data from 1970 to 2020, this study systematically revealed the spatiotemporal evolution characteristics of 21 extreme climate indices in Jilin Province. The results indicate a significant upward trend in extreme warm events, a marked decline in extreme cold events, and an intensification of extreme precipitation intensity, whereas changes in precipitation frequency were not significant. Such characteristics are consistent with the IPCC Fifth Assessment Report, which highlighted enhanced climate variability in mid-to-high latitudes under global warming [42], and also correspond to the findings of Alexander et al. [43], who reported an increase in warm nights and a reduction in cold nights over more than 70% of global land areas. Compared with national-scale studies [44,45,46], the warming amplitude in Jilin Province is more pronounced, suggesting a higher sensitivity to global temperature rise.

In terms of temperature, warm event indices such as TX90p, SU25, and WSDI show persistently significant upward trends, indicating a notable increase in thermal resources across the region, with more frequent and prolonged occurrences of warm nights and heat events. In contrast, cold event indices including FD0, TN10p, and TX10p exhibit declining trends, suggesting a reduction in the number of extreme low-temperature nights and a convergence in temperature variability. This indicates a weakening in the frequency and intensity of cold surges under a warming climate background. This finding agrees with Gan et al. [47], who demonstrated intensified heavy precipitation across mainland China during 1961–2022 but with little change in precipitation days. Similar conclusions were reported by Ren et al. [48] and Ren et al. [49], who revealed that precipitation extremes in Northeast China show strong spatial variability and are influenced by multiple circulation factors and terrain conditions. The spatial distribution pattern observed in this study, characterized by higher precipitation intensity in the central and southern regions and lower values in the eastern part of the province, is consistent with Chen et al. [50], confirming the regional characteristics of precipitation change.

Regarding precipitation, intensity-related indices such as RX1day, R95p, and SDII show an upward trend to varying degrees, while frequency-related indices such as CDD and R10 exhibit no significant trends. This finding agrees with Gan et al. [47], who demonstrated intensified heavy precipitation across mainland China during 1961–2022 but with little change in precipitation days. Similar conclusions were reported by Wang et al. [44] and Xin et al. [46], who highlighted large interannual fluctuations and significant spatial differences in precipitation extremes, suggesting that topography and atmospheric circulation jointly contribute to precipitation instability in Jilin Province. The spatial distribution pattern observed here, characterized by higher precipitation intensity in the central-southern region and lower values in the east, is consistent with Han et al. [51] and He [52], indicating that large-scale circulation and the AO jointly modulate precipitation variability in Northeast China.

The variability of extreme climate events is driven by multiple factors, among which large-scale atmospheric circulation plays a crucial role. This study revealed that the AO is negatively correlated with cold indices but positively correlated with warm indices (TX90p, WSDI), indicating that AO modulates cold air activity and thus influences temperature extremes. This agrees with Tong et al. [53], who demonstrated the influence of AO on cold air trajectories and temperature anomalies in Inner Mongolia. Furthermore, the interaction of circulation systems such as NAO and PDO with the East Asian monsoon has also been identified as a key driver of extreme climate variability in Northeast China [50,54]. When AO or PDO phases are positive, Jilin Province tends to experience more warm extremes and fewer cold events, indicating the role of multi-scale circulation resonance in shaping regional anomalies. Geographical factors are equally important. The southeastern mountainous region of Jilin, affected by monsoonal uplift, shows stronger precipitation extremes, while the northwestern plains exhibit lower precipitation intensity, revealing clear spatial heterogeneity. This pattern aligns with El Kenawy et al. [55], who emphasized the sensitivity of temperature changes in mountainous regions. The response differences across elevation and latitude suggest that topographic conditions exert significant modulation on regional extreme climate behavior [56].

It should be noted that this study has some limitations. The sparse distribution of meteorological stations may limit the ability to capture localized extremes, and human activities (e.g., urbanization and land-use change) have not been incorporated into the analysis framework. Future research should integrate reanalysis datasets and high-resolution climate models to quantify the coupling effects between circulation and topography and expand multi-scale risk assessments of extreme events.

This study systematically analyzed the spatiotemporal evolution and driving mechanisms of extreme climate events in Jilin Province, enriching empirical evidence of extreme climate research in mid–high latitudes. The findings reveal that both atmospheric circulation and topographic conditions jointly influence temperature and precipitation extremes in Jilin Province, although the formation of extreme events involves more complex multi-factor interactions. Besides circulation and terrain, human activity intensity, land-use patterns, and urbanization may also play crucial roles at local scales. Future studies should integrate multi-source data and coupled models with higher spatial resolution and longer time series to better quantify the combined effects of different drivers, thereby providing a more scientific basis for regional climate risk assessment and adaptive management.

5. Conclusions

From 1970 to 2020, extreme climate events in Jilin Province exhibited a significant warming–cooling contrast, characterized by rising temperatures and enhanced precipitation extremity. Among the extreme temperature indices, warm events intensified markedly while cold events weakened. TXx, SU25, WSDI, TN90p, and TX90p increased at rates of 0.31 °C/10a, 3.09 d/10a, 0.85 d/10a, −2.63 d/10a, and 1.75 d/10a, respectively. In contrast, FD0, TN10p, and TX10p decreased significantly at −2.55 d/10a, −2.70 d/10a, and −1.76 d/10a, with abrupt changes concentrated during the 1980s–1990s. Most indices displayed dominant interannual oscillations with cycles of 2–7 years. Extreme precipitation indices showed no uniform trend overall, but intensity-related indices strengthened: R95p, R99p, and PRCPTOT increased at rates of 5.54 mm/10a, 4.15 mm/10a, and 3.61 mm/10a, respectively, with dominant periodicities of approximately 3 years, indicating intensified precipitation yet stable frequency. EOF analysis revealed that extreme temperature exhibited spatially coherent variations across the province, while precipitation intensity was higher in the western and central–southern regions and lower in the east, reflecting clear spatial heterogeneity. Correlation analysis showed that AO was significantly negatively correlated with cold event indices such as FD0, ID0, and TN10p, with correlation coefficients ranging from −0.31 to −0.46, and positively correlated with warm indices such as TX90p and TN90p, with coefficients between 0.16 and 0.18, suggesting its role in modulating cold air activity over Northeast China. SN showed negative correlations with most precipitation indices, with coefficients from −0.20 to−0.40, indicating a suppressive effect on precipitation intensity. Geographically, extreme temperature increased with latitude but decreased with longitude and elevation, whereas precipitation intensity increased with longitude and topographic uplift while decreasing with latitude. Overall, during the past five decades, Jilin Province has undergone pronounced warming and intensified precipitation extremes, consistent with the general pattern of climate change observed in mid–high latitudes under global warming.