Assessing the response lag and vulnerability of terrestrial vegetation to various compound climate events in mainland China under different vegetation types

doi:10.21203/rs.3.rs-4722135/v1

Download PDF

Research Article

Assessing the response lag and vulnerability of terrestrial vegetation to various compound climate events in mainland China under different vegetation types

https://doi.org/10.21203/rs.3.rs-4722135/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

In the context of climate warming, the compound dry-hot (CDH), dry-cold (CDC), wet-hot (CWH), and wet-cold (CWC) events have become more frequent and widespread in recent decades, causing severe but disproportionate impacts on terrestrial vegetation. However, the understanding of how vegetation vulnerability responds to these compound climate events (CCEs) is still limited. Here, we developed a multivariate copula conditional probabilistic model integrating the Standardized Precipitation Index (SPI), Standardized Temperature Index (STI), and Normalized Difference Vegetation Index (NDVI) together to quantify the vegetation response to each of CDH, CDC, CWH and CWC events under diverse climates in mainland China. Results show that CDC events result in the largest probability of vegetation loss relative to other three CCEs, with the probability of NDVI below the 40% percentile being 4.8%-13.0% (0.5%-2.6%) larger than individual dry (cold) events. In contrast, CWH leads to the lowest vegetation loss probability among all CCEs, with the probability of NDVI below the 40% percentile being 5.6% ~ 6.9% (4.2% ~ 5%) less than individual wet (hot) events. The response of vegetation vulnerability to CCEs varies considerably with ecosystems and climate types. Vegetation in Loess Plateau and northwestern Xinjiang (Inner Mongolia) is highly susceptible to CDC (CDH) events, while that in northeastern and southern China (eastern coastal and southwestern regions) is more vulnerable to CWC (CWH) events. Shrubland, grassland and cropland exhibit higher vulnerability to CDC and CDH events, while deciduous (evergreen) forests are more vulnerable to CWC(CWH) events, which may be related to vegetation physiological characteristics, survival strategies, and climatic adaptations. This study enhances our understanding on the response of various vegetation types to CCEs, and provides theoretical support for the development of measures to mitigate climate hazards.

Compound climate events

copula

conditional probability

vegetation types

NDVI

mainland China

Vegetation is an indispensable component of the Earth system, which plays an important role in providing ecosystem services to the terrestrial environments, such as landscape aesthetics (Zhang et al. 2022a; Chen et al. 2023b), soil and water conservation (Broetto et al. 2017; Liu et al. 2020), climate regulation (Liu and Yin, 2013; Thom et al. 2017), carbon balance (Mekonnen et al. 2021; Yang and Pan, 2023), and water cycling (Feng et al. 2017). Extreme climate events, such as drought, flood, heat wave and cold wave, have posed significant threats to the ecosystem structure and function by impacting vegetation photosynthesis (Wu and Wang, 2022; Yu et al. 2022), respiration (Wang et al. 2018; Zheng et al. 2022), and carbon utilization processes (Wu et al. 2012; Chen et al. 2019). Under the anthropogenic climate change influences, the extreme climate events have become more frequent, widespread and intense globally (Coumou et al. 2013; Hao et al. 2013; Pendergrass et al. 2020; Chiang et al. 2022), and often manifested as the compound climate events (CCEs) formed by the agglomeration of multiple climatic events (Hao et al. 2018; Feng et al. 2020). In recent decades, the frequency of CCEs and their affected areas have increased over many regions worldwide (Hao et al. 2018; Mukherjee and Mishra, 2021), exerting more severe and disproportionate impacts on the ecosystems than individual climate events (Allen et al. 2015; Anderegg et al. 2015; Stovall et al. 2019). Consequently, it is crucial to understand the ecosystem responses to these increasingly complex compound climate events comprehensively, as it is essential to developing the effective adaptation, mitigation and resilience strategies to deal with various climate disasters.

Vegetation vulnerability refers to the propensity or predisposition of vegetation to be adversely affected (Ara Begum, 2022). Previous studies have used mainly the deterministic methods (e.g., correlation analysis and multiple linear regression method) to assess the impacts of extreme events on vegetation vulnerability (Xu et al. 2018; Wu et al. 2019a; Ding et al. 2020; Chen et al. 2023a). As temperature, precipitation and solar radiation are the main drivers of vegetation activities (Zscheischler et al. 2014; Zhang et al. 2022c; Fan et al. 2023; Wu et al. 2024), numerous studies have attempted to explore the relations between vegetation and the land surface wetness/dryness conditions by utilizing the indices based on precipitation or temperature, such as the Standardized Precipitation Index (SPI, McKee et al. 1993), Standardized Precipitation Evapotranspiration Index (SPEI, Vicente-Serrano et al. 2010), Palmer Drought Severity Index (PDSI, Palmer 1965), and Standardized Temperature Index (STI). The changes in vegetation cover are usually characterized by the remote sensing-based vegetation indices such as the Normalized Difference Vegetation Index (NDVI, Pinzon et al. 2014) and Enhanced Vegetation Index (EVI, Huete et al. 2002). The effects of climate extremes (e.g., drought or heat wave) on vegetation have been studied extensively at the regional (Xu et al. 2018; Wu et al. 2019a; Ding et al. 2020; Chen et al. 2023a) and global (Vicente-Serrano et al. 2013; Wen et al. 2019; Liu et al. 2023) scales.

CCEs, referring to the simultaneous or consecutive occurrence of multiple climate drivers and hazards (Zscheischler et al. 2018), have become more frequent under recent global warming (Hao et al. 2018; Mukherjee and Mishra, 2021) and resulted in more severe impacts than individual climate events, even when the contributing drivers are not more extreme relative to the individual events (Leonard et al. 2014; Zscheischler et al. 2018; AghaKouchak et al. 2020; Li et al. 2021). The precipitation- and temperature-related extreme events (e.g., heatwave, cold spell, flood and drought) closely related to climate change are commonly used to assess the changes in CCEs (Hao et al. 2013; Tencer et al. 2014; Wu et al. 2019b; Li et al. 2022).The simultaneous occurrence of such precipitation and temperature anomalies is typically described in terms of four compound categories, namely, the compound dry-hot (CDH), compound wet-hot (CWH), compound dry-cold (CDC), and compound wet-cold (CWC) events (Beniston, 2009; Estrella and Menzel 2012; Hao et al. 2013). Several studies have investigated the spatio-temporal distributions of these CCEs at regional scales (Beniston et al. 2009; Qian et al. 2014; Yuan et al. 2016; Wu et al. 2019b) and global scale (Hao et al. 2013; Meng et al. 2022). For example, Hao et al. (2013) conducted a 1978–2004 global analysis on the spatio-temporal variations of these four CCEs, and found that CWH and CDH events have notably increased over the high latitudes and tropical regions, while CDC and CWC events have decreased in most global regions, generally consistent with global warming. Wu et al. (2019b) explored historical changes (1961 to 2014) in CCEs in mainland China and highlighted a significant increase in their frequency associated with the anthropogenic climate warming. The frequency and areas affected by CDH and CWH events showed significant increasing trends during summer and winter seasons in most parts of China, while that by CDC and CWD showed decreasing trends for the period 1988–2014 relative to 1961–1987 (Wu et al. 2019b).

There has been a growing interest over the past decade in evaluating the ecosystem response to CCEs (Feng et al. 2019; Hao et al. 2021), with the evidence suggesting that CDH may exert substantially more negative impacts on ecosystems than individual dry or hot events (Barbosa et al. 2012; Feng et al. 2019; Hao et al. 2021; Li et al. 2021, 2022). For instance, Feng et al. (2019) investigated the probability variation of maize yield under CDH events, and found that the probability of maize yield reduction is increased from 7–31% (from 4–31%) when the extreme drought (extreme hot) condition changes to the CDH conditions. Similarly, Hao et al. (2021) quantified the global vegetation response to CDH events during growing season and found that, relative to the individual dry (hot) conditions, the probability of vegetation loss caused by CDH is increased by 7% (28%) in arid/semi-arid regions. They also found that temperate grassland is more susceptible to CDH events mainly due to stronger positive (negative) correlations between SPI (STI) and NDVI in temperate grassland than other vegetation types (Hao et al. 2021).

Although previous studies have examined the CDH impacts on vegetation growth and productivity, one important aspect often overlooked is the difference in the lagged response time to CCEs among different vegetation types. The lag effect of climate events on vegetation growth refers to the impacts of the previous climate events on the current vegetation growth (Wen et al. 2019). Most studies (Wu et al. 2015; Mulder et al. 2016; Zhao et al. 2017; Xu et al. 2018; Wen et al. 2019; Fang et al. 2019a) focused only on the associations between climatic factors with a certain fixed time lag and vegetation status. However, the current vegetation growth may show different lagged response times to the previously different climate conditions. For example, Wu et al. (2022) analyzed the multi-month time lag effects of growing season NDVI response to precipitation in the Hulunbuir region, and found that the NDVI shows a positive correlation with precipitation at the 1- and 13-month time lags, while a significant negative correlation was observed at a 9-month time lag. Furthermore, the lagged response of vegetation to climate varies considerably with the spatial patterns of underlying surfaces due to the spatial heterogeneity of ecosystems (Wu et al. 2015). The degree to which climate factors explain vegetation changes can be augmented and improved when the time-lag effect is considered (Wu et al. 2015; Zhao et al. 2017; Wen et al. 2019; Jiang et al. 2020). Therefore, an accurate assessment on the different time-lagged effects of different climate events on vegetation growth states is critical for better understanding the terrestrial ecosystem responses to CCEs.

In addition, vegetation vulnerability may be affected not only by CDH, but also by other types of CCEs (Richardson et al. 2018; Vitasse et al. 2018; Li et al. 2022). Several studies have shown that cold- and wet-related extreme events can exacerbate vegetation loss by causing the leaf frostbite, inhibiting the root respiration, shortening the growing season, and reducing the photosynthetic carbon uptake (Richardson et al. 2018; Vitasse et al. 2018; Chen et al. 2023b). Li et al. (2022) argued that CDC events impose an adverse impact on productivity at mid- to high-latitudes, surpassing the impacts of individual cold or dry events. Richardson et al. (2018) suggested that climate warming not only increases the active period of photosynthesis, but also promotes tissue de-hardening, making vegetation more susceptible to cold conditions during the pre-growth period. Although global warming continues, understanding the impact of cold-related CCEs on vegetation growth is still important, as the atmospheric circulation pattern resembling the warm Arctic-cold continents pattern results in the continued frequency of global extreme cold events (Hao et al. 2013; Li et al. 2022; Johnson et al. 2018). To the best of our knowledge, currently there is no comprehensive study focusing on assessing and comparing the vegetation loss probability among different vegetation types under various CCEs such as CDH, CWH, CDC and CWC.

Previous studies have focused on investigating the response of vegetations to climate events by linking climate events with vegetation indices using the correlation analysis (Bao et al. 2014; Bastos, 2020; Xu et al. 2018; Ding et al. 2020; Chen et al. 2023a), and have also analyzed the direct and lagged response of vegetation to CCEs by constructing the combined stress index (Ceglar et al. 2018) or using multiple linear regression methods (Li et al. 2022). However, the relationship between CCEs and vegetation response is usually nonlinear, which makes it challenging to accurately quantify the probability of vegetation loss and its changes under different intensities of CCEs. Here, we introduce the copula function (a multivariate statistical technique), which connects the marginal distributions of two or more random variables to form their joint distribution (Nelsen, 2007; Fang et al. 2019b; Guo et al. 2023), to estimate the conditional probability of vegetation loss under various CCEs. For this purpose, a multivariate copula conditional probability (MCCP) framework is developed to quantify the loss probability of various vegetation types caused by CDH, CWH, CDC and CWC. This framework is based on a multivariate model that can be capable of addressing complex and nonlinear interactions between various compound climate events and vegetation types. These four CCEs are identified based on SPI (representing dry or wet conditions) and STI (representing hot or cold conditions) indices. The different loss levels of vegetation are represented by the percentiles of monthly NDVI data during 1982–2020. The MCCP framework is systematically evaluated during growing season (from April to September) in mainland China, encompassing the remarkable geographic diversity with a wide range of climate zones.

The main objectives of this study are to (1) explore the spatial distribution patterns of loss probability of vegetation under the conditions of CDH, CWH, CDC, and CWC; (2) evaluate the spatial discrepancies in the loss probability of vegetation caused by four CCEs and individual dry/wet (hot/cold) events; and (3) evaluate the discrepancies in the loss probability between various types of vegetation caused by CDH, CWH, CDC and CWC events. In the following, Section 2 introduces the study area, meteorological observations and vegetation data. In section 3, (a) the MCCP framework, (b) the definitions of CCEs, and (c) the methods of three-dimensional Copula model and probability of vegetation loss conditioned on the compound climate scenarios are introduced. The results and discussion are presented in Sections 4 and 5, respectively, followed the conclusions summarized in Section 6.

2.1 Study area

China has a complex and diverse topography, remarkable geographic diversity, and a wide range of climate zones (Wu et al. 2019). The eastern part of China is dominated by monsoon climate, transitioning from temperate monsoon climate in the north to subtropical and tropical monsoon climate in the south (Xu et al. 2018). In contrast, the western inland areas experience a continental climate, while the Tibetan plateau, characterized by its high altitude, falls within the plateau climate zone (Xu et al. 2018). Precipitation and temperature patterns are influenced by the monsoon climate and topography, with average annual temperature decreasing from south to north and precipitation decreasing from the southeast coast to the northwest interior (Wu et al. 2019b). This intricate climate patterns significantly affect the China's economic and social development, rendering it susceptible to the extreme climate events (Wu et al. 2019b).

2.2 Data sources and preprocessing

The 1982–2020 monthly observed precipitation and temperature data with a 0.5°× 0.5° resolution are provided by the China National Meteorological Information Center (http://data.cma.cn). The dataset is produced using the thin-plate spline interpolation method and includes monthly observations from 2472 surface meteorological stations across the country, and has been widely employed to assess and quantify extreme climate events (Wang and Chen, 2014; Wu et al. 2019b). In this study, these monthly data are used to calculate the SPI and STI, respectively.

The 1982–2020 monthly NDVI dataset is obtained from the China National Earth System Science Data Center (http://www.geodata.cn/). This dataset is produced at a spatial resolution of 5 km × 5 km based on the NOAA Climate Data Record (CDR) Advanced Very High-Resolution Radiometer (AVHRR) NDVI data by utilizing the maximum-value composition (MVC) algorithm (Xu and Yang, 2022a). To ensure the dataset quality, cross-validation was performed by comparing the NDVI dataset with the Global Inventory Monitoring and Modeling Study third-generation (GIMMS3g) NDVI and Moderate Resolution Imaging Spectrometer (MODIS) monthly NDVI datasets (Xu and Yang, 2022b). The NDVI is a derived metric calculated by the near-infrared to visible light band ratio in remote sensing data (Pinzon and Tucker, 2014; De Beurs et al. 2015). The range of NDVI vary from − 1 to 1, with a larger value (closer to 1) indicating the presence of more dense vegetation cover. A NDVI close to 0 represents the limited or no vegetation coverage, while that close to -1 indicates the cover of clouds or the found cover of water, snow, etc. In this study, the NDVI is used to identify the change of vegetation during growing season from April to September (Yuan et al. 2019). The areas with an average NDVI < 0.1 during the growing season indicating barren, snow cover or sparsely vegetated areas are excluded (Xu et al. 2018). To maintain spatial consistency, a cubic convolution method was employed to resample the 5 km × 5 km monthly NDVI data to 0.5° × 0.5° to match the resolution of the gridded precipitation and temperature data. Due to the evident long-term trend and seasonality in the vegetation sequence, it is necessary to remove these interfering signals when assessing vegetation vulnerability (Smith et al. 2023). Therefore, to achieve an approximately stationary LAI time series, we use the “detrend” function in MATLAB to remove long-term trends of LAI pixel by pixel, and then remove the seasonal trends by the mean difference method (Yao et al. 2022; Smith et al. 2023).

The Land Cover Type 1 (LC_Type1) data from the MODIS Land Cover Type product dataset (MCD12Q1), which have been widely used for vegetation classification and assessment in China (Xu et al. 2018; Wu et al. 2022a), are used to classify vegetation types (https://modis.gsfc.nasa.gov/data/dataprod/mod12.php). The majority resampling method is applied to resample the MCD12Q1 data onto the standard 0.5°× 0.5° to match the grid resolution of meteorological and NDVI data. Non-vegetated areas such as permanent water bodies, urban and built-up lands, permanent ice and snow, and barren are excluded from the map. The spatial distribution and grid percentages of 12 different types of vegetation coves across the mainland China are shown in Fig. 1, and the definitions of the different types of vegetation are shown in the supporting material Table S1. Sparse vegetation cover is mainly distributed in the arid and semi-arid regions of Northwest China. The vegetation area accounts for ~ 88.5% of the total area, with the grassland area being the largest (~ 26.54%) mainly scattered in Inner Mongolia and western China. Cropland is the second largest vegetation type distributed mainly in northeastern and northern China and some parts of central China (~ 21.35%). Woody savanna and savannah are sporadically distributed in southern and northeastern China, covering 12.3% and 10.5% of the mainland China, respectively, while the remaining 8 vegetation types account for only < 17.81% of the total area.

In this study, the MCCP framework is developed to assess the responses of vegetation vulnerability to CCEs (Fig. 2). The MCCP framework includes the following four steps:

3.1 Identifying the dry/wet and cold/hot events and CCEs

SPI (STI) is an indicator used to monitor the dry/wet and cold/hot conditions by fitting observed precipitation (temperature) data to a desired probability distribution function (McKee et al. 1993; Zscheischler et al. 2014). The main steps in calculating SPI (STI) include: a) fitting monthly precipitation or temperature data to the probability distribution function of interest, b) calculating the cumulative probability for each month based on the fitted probability distribution function, and c) converting the calculated cumulative probabilities to corresponding standard normal distribution values. In this study, the SPI and STI were calculated from 1- to 24-month scales. The threshold-based method is applied to classify SPI and STI to depict various intensities of dry/wet and cold/hot conditions (McKee et al. 1993; Zscheischler et al. 2014). The wet (hot) event is defined as the SPI (STI) ≥ 0.5, while the dry (cold) event is defined as the SPI (STI) ≤ -0.5 (Table 1).

In this study, the following four types of CCEs are considered based on that the SPI and STI are simultaneously greater or less than a certain threshold (Wu et al. 2019; Li et al. 2021; Feng et al. 2019): CDH (SPI ≤ -0.5 and STI ≥ 0.5), CDC (SPI ≤ -0.5 and STI ≤ -0.5), CWC (SPI ≥ 0.5 and STI ≤ -0.5), and CWH (SPI ≥ 0.5 and STI ≥ 0.5). Different intensity levels of SPI and STI are used to describe different intensities of CCEs based on the classifications of SPI and STI in Table 1. The frequency of CCEs is defined as the number of CCEs that occur during the study period, while the duration is defined as the consecutive months from the beginning to the end of a CCE.

Note that the two individual climate events that constitute the compound event have temporal overlap, but may occur on different time scales, because the formation of CCEs is not only influenced by the intensity of individual climate events, but also by the duration of their occurrence (Leonard et al. 2014; Raymond et al. 2020). Therefore, this study first determines the response time of NDVI to SPI (STI) for each grid point at the time scale of $\:{{T}_{spi}}_{a}$($\:{{T}_{sti}}_{a}$) (see section 3.2), then fits the marginal distributions of SPI and STI at the time scale of $\:{{T}_{spi}}_{a}$($\:{{T}_{sti}}_{a}$) and identifies the CEEs at the grid scale. For example, the lag time of NDVI response to SPI is 3 months, and hence the 3-month SPI is chosen.

Table 1

The classifications and criteria for the SPI and STI intensity levels
Range	SPI classification	STI classification
≤-2.00	Extreme drought	Extremely cold
From − 1.50 to -1.99	Severe drought	Very cold
From − 1.00 to -1.49	Moderate drought	Moderately cold
From − 0.50 to -0.99	Slight drought	Slightly cold
From 0.49 to -0.49	Near normal	Near normal
From 0.50 to 0.99	Slightly wet	Slightly hot
From 1.00 to 1.49	Moderately wet	Moderately hot
From 1.50 to 1.99	Very wet	Very hot
≥ 2.00	Extremely wet	Extremely hot

3.2 Estimating the lag time of vegetation changes in response to dry/wet and cold/hot events

The lag effect of climate events on vegetation growth refers to the impacts of the previous climate events on the current vegetation growth (Wen et al. 2019). In this study, we use the Spearman correlation analysis between monthly NDVI and SPI (or STI) from 1982 − 2020, with a lag ranging from 1 to 24 months at the grid scale, to identify the lag time of vegetation response to SPI (or STI),

$$\:\begin{array}{c}\left\{\begin{array}{c}{{R}_{spi}}_{b}^{a}=corr\left({NDVI}_{a},{SPI}_{b}^{a}\right)\\\:{{R}_{sti}}_{b}^{a}=corr\left({NDVI}_{a},{STI}_{b}^{a}\right)\end{array}\right.\:\:a=\text{4,5},\dots\:,9,\:\:\:\:b=\text{1,2},\dots\:,\text{23,24}\: \left(1\right)\end{array}$$

where a represents the a-th month of the growing season from April to September, b represents the time scale of SPI (or STI), and R represents the correlation coefficient between NDVI and SPI (or STI). For each grid, the time scale with the maximum correlation coefficient (MCC) between SPI (or STI) and NDVI is considered as the lag time of vegetation response to SPI (or STI) (Fang et al. 2019a). The response time (T) of vegetation to SPI (or STI) is defined as:

$$\:\begin{array}{c}\left\{\begin{array}{c}{{T}_{spi}}_{a}={max}\left\{abs\left({{R}_{spi}}_{b}^{a}\right)\right\}\\\:{{T}_{sti}}_{a}={max}\left\{abs\left({{R}_{spi}}_{b}^{a}\right)\right\}\end{array}\right. \left(2\right)\end{array}$$

where T_spi and T_sti represent the response times of vegetation to SPI and STI, respectively. The max {} denotes the timescale of SPI (STI) corresponding to the maximum absolute value of correlation coefficient between SPI (STI) and NDVI.

3.3 Building the three-dimensional joint distribution models of SPI, STI and NDVI

The copula function is a multivariate probability analysis method for combining the marginal distributions of multiple random variables to generate their joint distribution (Nelsen, 2007; Fang et al. 2019b). In this study, we employ the copula function to construct a multivariate joint probability model, encompassing the vegetation status (NDVI), the dry/wet conditions (SPI), and the cold/hot conditions (STI).

The Normal, Logistic, and Gumbel distributions are used to fit the 1982–2020 monthly NDVI data during the growing season (April to September) at the grid scale, and then the optimal distribution is selected based on the Kolmogorov-Smirnov (K-S) test (at the significance of 0.05) and the Akaike information criterion (AIC). The SPI and STI are fitted using the normal distribution since they were calculated after normal standardization with a mean of 0 and variance of 1 (see section 3.1).

The dependence relationship of NDVI-SPI-STI is constructed using the three-dimensional Copula function. Two elliptic family Copula functions (Gaussian and Student's t) and four Archimedes family Copula functions (Clayton, Frank, Gumbel, and Joe) are selected for constructing joint distributions. The joint distribution functions can be expressed as follows:

$$\:\begin{array}{c}{F}_{SPI,\:\:STI,NDVI}\left(spi,sti,ndvi\right)=C\left({F}_{SPI}\left(spi\right),{F}_{STI}\left(sti\right),{F}_{NDVI}\left(ndvi\right)\right)=P\left(SPI<spi,STI<sti,NDVI<ndvi\right) \left(3\right)\end{array}$$

where F_SPI (spi), F_STI (sti), and F_NDVI (ndvi) are the optimal marginal distributions of SPI, STI and NDVI sequences, respectively, and C represents the copula function. In this study, the goodness of fit of copula model is evaluated using the AIC and the Cramér-von Mises test (Schepsmeier, 2015). The Cramér-von Mises statistic measures the disparity between observed data and model, with smaller statistic values indicating a better fit between model and observed data (Schepsmeier, 2015). The significance level is set to 0.05 for the Cramér-von Mises test (Wang and Wells, 2000).

3.4 Quantifying the probability of vegetation loss caused by the four types of CCEs

In this study, the 40th, 30th, 20th, and 10th percentiles of NDVI are used to characterize different levels of vegetation stress (Fang et al. 2019; Li et al. 2021). The monthly NDVI values corresponding to the 40th, 30th, 20th, and 10th percentiles over mainland China during the growing season (April to September) from 1982 to 2020 are shown in Fig. S1. The joint probability distribution of copula (Eq. 3) and Bayesian theory are used to compute the probability of vegetation loss caused by CCEs. Four different levels of vegetation loss scenarios (monthly NDVI ≤ 40th, 30th, 20th, and 10th percentiles) under CDH (SPI ≤ -0.5 and STI ≥ 0.5), CWH (SPI ≥ 0.5 and STI ≥ 0.5), CDC (SPI ≤ -0.5 and STI ≤ -0.5), and CWC (SPI ≥ 0.5 and STI ≤ -0.5) are quantified. The probability of vegetation loss during the growing season is obtained by averaging the probability of vegetation loss from April to September. For example, the probability of NDVI ≤ 40th under the CDH condition can be expressed as:

$$\:\begin{array}{c}P\left(NDVI<ndvi\right|{spi}_{i+1}\le\:SPI\le\:{spi}_{i},\:{sti}_{j+1}\ge\:STI\ge\:{sti}_{j})=\frac{P\left({spi}_{i+1}\le\:\text{S}\text{P}\text{I}\le\:{spi}_{i},\:{sti}_{j+1}\ge\:\text{S}\text{T}\text{I}\ge\:{sti}_{j},\:NDVI<ndvi\right)}{P\left({sti}_{j+1}\ge\:\text{S}\text{T}\text{I}\ge\:{sti}_{j},{spi}_{i+1}\le\:\text{S}\text{P}\text{I}\le\:{spi}_{i}\right)}\\\:=\frac{{F}_{SPI,STI,NDVI}\left({spi}_{i},{sti}_{j+1},ndvi\right)-{F}_{SPI,STI,NDVI}\left({spi}_{i+1},{sti}_{j+1},ndvi\right)-{F}_{SPI,STI,NDVI}\left({spi}_{i},{sti}_{j},ndvi\right)+{F}_{SPI,STI,NDVI}\left({spi}_{i+1},{sti}_{j},ndvi\right)}{{F}_{STI,SPI}\left({sti}_{j+1},{spi}_{i}\right)-{F}_{STI,SPI}\left({sti}_{j+1},{spi}_{i+1}\right)-{F}_{STI,SPI}\left({sti}_{j},{spi}_{i}\right)+{F}_{STI,SPI}\left({sti}_{j},{spi}_{i+1}\right)} \left(4\right)\end{array}$$

where ndvi represents the 40th percentile of monthly NDVI during the growing season from 1982 to 2020 at each grid, F_SPI,STI,NDVI denotes the joint distribution function of SPI, STI and NDVI, F_STI,SPI denotes the joint distribution function of SPI and STI. Under the CDH condition, the range of spi is {-0.5, -1, -1.5, -2, -∞}, and the range of sti is {0.5, 1, 1.5, 2, ∞}. i= {1, 2, 3, 4}, j= {1, 2, 3, 4}, spi_i and sti_j denote the i-th and j-th values of the spi and sti sequences, respectively.

4.1 Lag response of vegetation to dry/wet and cold/hot conditions at 1-to-24 timescales

The spatial distributions of mean MCC between NDVI and SPI and STI, during the 1982–2020 growing season at 1- to 24- month scales are shown in Fig. 3a and 3b, respectively. A larger MCC indicates a more significant impact of the dry/wet or cold/hot conditions on vegetation status. Positive (negative) MCC indicates that vegetation status increases with increasing (decreasing) precipitation/temperature. A positive MCC between NDVI and SPI is observed in 59.7% of the total area (Fig. 3a), especially in the Inner Mongolia Plateau, Loess Plateau and northwestern China (MCC > 0.6), while in southwestern regions there is a weak positive correlation between NDVI and SPI (MCC < 0.2). Conversely, the negative MCC is concentrated in southeastern, northeastern and southwestern China (MCC > -0.25). A positive correlation between NDVI and STI is found for most of China (~ 80.3%), and particularly a larger MCC (> 0.4) is mainly distributed in the Loess Plateau, the central, northeastern and southeastern China (Fig. 3b). In contrast, NDVI is negatively correlated with STI in northern Inner Mongolia, some parts of northwestern and southwestern China, and eastern coastal areas.

Figures 3c and 3d show the spatial distributions of the lag time of vegetation (NDVI) response to the dry/wet and cold/hot conditions, respectively. It is evident that the lag time of vegetation response to dryness/wetness is longer in the Loess Plateau and arid areas of Northwest China than any other regions (Fig. 3c). This can be attributed to that vegetation in arid areas has undergone adaptations to the persistent water-limited environments and developed strategies to cope with adverse conditions, such as relatively short average height of plant communities (Luo et al. 2021). The lag time of 1 ~ 6 months covers 11.5% of the study area, mostly distributed in the northeastern and southeastern regions, while 58% of the study area (mainly in southern China) has a lag time of 6 ~ 12 months. A longer lag time (18 ~ 24 months) can be found in only 2.1% of the study area, mainly concentrated in the Loess Plateau and central region (Fig. 3c).

56.5% of the study areas have a lag time of 6 ~ 12 months for vegetation response to STI, which are mainly distributed in North China Plain, Northeast Plain and southeast coastal areas. A longer lag time (> 12 months) is found in northern and western China (15.2% of the study area), while a shorter lag time (1 ~ 6 months) is mainly distributed in the northeastern (Greater Khingan Mountains and Lesser Khingan Mountains) and southwestern regions of China (28.3% of the study area). The MMC and lag time of the 12 vegetation types response to SPI and STI are shown in Fig. S2. It is shown that the lag time of deciduous needleleaf forest and closed shrubland response to SPI is generally longer other vegetation types, while the lag time of evergreen broadleaf forest and closed shrubland (deciduous needleleaf forest and deciduous broadleaf forest) response to STI is generally longer (shorter) other vegetation types.

4.2 Spatial variability of frequency and duration of CCEs

Figure 4 displays the distributions of frequency and mean duration of CDC, CWH, CDH and CWC in mainland China during the 1982–2020 growing season. The frequency of CDC and CWH events (Figs. 4a and d) is slightly lower than that of CDH and CWC events (Figs. 4b and c). The higher frequency of CDC (> 13 events) and CWH (> 18 events) is observed in the Qinghai-Tibet Plateau and some parts of northeastern China relative to other regions (Figs. 4a and 4d), while the opposite spatial pattern is found in the frequency of CDH and CWC. In addition, southern China experiences more frequent CDH and CWC events (> 21 events) than northern China (< 15 events) (Figs. 4b and 4c). This phenomenon can be attributed to the thermodynamic relation between precipitation and temperature (Zhou and Liu, 2018; Wu et al. 2019b) indicating that the negative (positive) correlation between precipitation and temperature can lead to higher frequent CDH and CWC (CDC and CWH) events.

The mean duration of CDC and CWH events is generally shorter than CDH and CWC events in the northeast China plain, southern and eastern China (Figs. 4e-h), while the longer mean duration of CDH and CWC is found in the North China Plain and Inner Mongolia (1.8 ~ 2.6 months, Figs. 4f and 5g). The high frequency of CDC and CWH is usually accompanied by a longer duration in some regions of the Qinghai-Tibet Plateau and northeastern China (2 ~ 2.5 months, Figs. 4e and 4h). Conversely, the low frequency of CDH and CWC events is usually accompanied by a shorter mean duration in the Qinghai-Tibet Plateau and northeast China (1 ~ 1.6 months, Figs. 4f and 4g).

4.3 Probability of vegetation loss caused by CDC and CDH events

Based on the optimal marginal distribution of NDVI identified at the grid scale (Fig. S3 and Table S2), the three-dimensional joint distribution of NDVI-SPI-STI is constructed based on the optimal copula functions selected for each grid based on the AIC and the Cramér-von Mises test (Fig. S4 and Table S3). The probability distributions of vegetation loss (NDVI ≤ 40th, 30th, 20th, and 10th) caused by individual drought, wet, hot and cold events during the growing season are displayed in Figs. S5-S8, respectively. The probability distributions of vegetation loss (NDVI ≤ 40th, 30th, 20th and 10th) caused by different CDC intensity levels (as defined in Table 1) during growing season are shown in Fig. 5 (dry & extremely cold), Fig. S9 (dry & slightly cold), Fig. S10 (dry & moderately clod), and Fig. S11 (dry & very cold), respectively. The conditional probabilities of vegetation loss caused by different CDH intensity classes are shown in Fig. 6 (dry & extremely hot), Fig. S12 (dry & slightly hot), Fig. S13 (dry & moderately hot), and Fig. S14 (dry & very hot).

The CDC results in a larger probability of vegetation loss (Figs. 5 and S9-S11) than the individual dry or cold events in most regions, especially in the Loess Plateau (Figs. S5 and S8). For example, the average probability of NDVI below the 40th percentile caused by the CDC events is 3.1%~6.3% (2.0%~3.3%) larger than that caused by individual dry (cold) events in the mainland China, demonstrating a stronger impact of the cold events on vegetation than dry events. As the intensity of CDC increases, the probability of vegetation loss increases in most areas, particularly in the Yellow River basin and northwestern Xinjiang, while it decreases in southwestern China and eastern coastal regions (Figs. 5 and Figs. S9-S11). The spatial patterns of the probability of NDVI ≤ 30th, 20th, and 10th percentiles under the CDC conditions are similar to that of the 40th percentile (Figs. S9- S11).

The probability of vegetation loss due to CDH in the Inner Mongolia is significantly higher than that in other regions (Figs. 6 and Figs. S12- S14). Compared to the hot (dry) events, CDH events result in a larger (smaller) vegetation loss probability in most regions, and the probability increases (decreases) with the dry (hot) intensity increases. However, in the Inner Mongolia, the probability of vegetation loss due to CDH is relatively higher than that caused by both dry and hot conditions (Figs. 6, Figs. S5, S7, and S12- S14) due to the significant positive (negative) correlation between NDVI and SPI (STI) in this region (Figs. 3a and 3b), which suggests that the drier or hotter conditions accelerate vegetation loss over this region. The average probability of NDVI below the 40th percentile due to CDH is 0.9%~5.5% (3.5%~10.2%) less (greater) than that due to dry (hot) events in mainland China except for Inner Mongolia (Figs. 6, S12- S14), indicating that the dry condition can lead to larger vegetation loss probability in these regions than the hot conditions.

Figure 7 shows the loss probabilities of different vegetation types caused by CDC and CDH. As shown, CDC events result in a higher probability of almost all vegetation types than the dry events, indicating that vegetations are more susceptible to loss when the dry events encounter cold events. Conversely, under the CDH the hot conditions show positive effects on all types of vegetation growth (except for open shrubland), reducing the vegetation vulnerability caused by dry events (Fig. 7). Overall, the loss probability caused by CDC (CDH) increases (decreases) with the increased intensity of cold (hot) events for all vegetation types. Interestingly, the loss probability of some vegetations (e.g., evergreen needleleaf forests, evergreen broadleaf forests and deciduous needleleaf forests) due to CDC tends to decrease with the increasing dry intensity. This phenomenon could be attributed to that these vegetations mostly grow in the wet environments where the occurrence of drought evaporates excess soil water and alleviates the soil water oversaturation problem, making it more conducive to vegetation growth (Nicolai-Shaw et al. 2017; Ding et al. 2020; Pascoa et al. 2020).

We further explore the differences in the vegetation loss probability (≤ 40th quantile) of 12 vegetation types under various CDC (Fig. S15) and CDH conditions (Fig. S16). Under CDC conditions, the deciduous broadleaf forests show the highest loss probability (46.2%~59.4%), followed by grasslands (48.4%~ 58.9%) and permanent wetland (47.2%~56.8) (Fig. S15). In contrast, the lowest loss probability is found in evergreen broadleaf forests (40.0%~46.3%), evergreen needleleaf forests (39.9% ~ 46.3%), and deciduous needleleaf forests (37.4% ~55.0%). Under CDH conditions, the closed shrublands exhibit the highest loss probability (45.5% ~58.5%), followed by grasslands (40.5% ~54.0%) and croplands (39.5% ~48.4%), while the low loss probability is detected in the deciduous needleleaf forests (25.1% ~32.7%) and deciduous broadleaf forests (28.1% ~36.6%) (Fig. S16). These findings suggest that the deciduous broadleaf forests, grasslands and permanent wetland (closed shrublands and grasslands) are more vulnerable to CDC (CDH) events than the evergreen forests (deciduous forest ecosystems).

4.4 Probability of vegetation loss caused by CWC and CWH events

The spatial distributions of probability of vegetation loss (NDVI ≤ 40th, 30th, 20th, and 10th percentiles) during growing season under various CWC conditions are shown in Fig. 8 (wet & extremely cold), Fig. S17 (wet & slightly cold), Fig. S18 (wet & moderately clod), and Fig. S19(wet & very cold). The probability of vegetation loss due to CWC generally decreases with increasing intensity of wet conditions (Fig. 8), and increases with the increasing intensity of cold conditions (Fig. 8 and Figs. S17-S19). Spatially, CWC leads to a larger (smaller) probability of vegetation loss mainly in northeastern and southern China (Yellow River Basin, Inner Mongolia, and northern Xinjiang) (Fig. 8 and Figs. S17-S19). Moreover, the average probability of NDVI below the 40th percentile due to CWC is 4.8%-13.0% larger than that due to the wet conditions in mainland China, and 0.5%-2.6% smaller than that due to cold conditions (Fig. 8, Figs. S6, S8 and S17-S19), indicating that vegetation is more affected by the cold than the wet events. However, in some regions such as Inner Mongolia and northern Xinjiang (northeastern, southwestern, and southeastern coastal regions), the vegetation loss probability caused by CWC is significantly lower (higher) than the wet (cold) events.

The spatial distributions of vegetation loss probability under four types of CWH events are displayed in Fig. 9 (wet & extremely hot), Fig. S20 (wet & slightly hot), Fig. S21 (wet & moderately hot), and Fig. S22 (wet & very hot), respectively. The vegetation loss probability is slightly higher in the northeastern, southwestern, and eastern coast regions than in other regions (Fig. 9 and Figs. S20-S22). Compared to the wet (hot) events, CWH leads to a lower vegetation loss probability in most regions, and the probability increases (decreases) with the increasing intensity of wet (hot) events (Fig. 9, Figs. S6, S7, and S20-S22). In particular, the average probability of NDVI below the 40th percentile caused by CWH events is 5.6% ~ 6.9% (4.2% ~ 5%) less than that caused by the wet (hot) events, suggesting that the hot events encountered wet events can reduce the vegetation loss probability in most areas, and that the impacts of hot events on vegetation are greater than the wet events.

Figure 10 compares the loss probability of different vegetation types under various CWC and CWH conditions. Almost all CWC (CWH) events increase (decrease) the loss probability compared to the wet events, and the loss probability increases (decreases) with the increasing cold (hot) intensity (Fig. 10). However, the loss probability due to CWC and CWH in open shrublands, grasslands and cropland tends to decrease with increasing wetness intensity, suggesting that the humid conditions contribute to the growth of these vegetation types.

Under CWC conditions, deciduous needleleaf forests show the highest loss probability (52.3%~ 66.6%), followed by deciduous broadleaved forests (47.0%~ 64.1%) and woody savanna (47.1%~ 57.3%) (Figs. S23), while the lower loss probability is found in the closed shrublands (32.6%~ 39.4%), croplands (40.2%~ 49.2%), and grasslands (36.9%~ 48.5%) (Figs. S23). Under CWH conditions, the higher vulnerability is observed in the deciduous needleleaf forests (34.1%~ 50.7%), evergreen needleleaf forests (39.1%~ 45.6%) and evergreen broadleaf forests (40.2%~ 47.6%) than in the deciduous broadleaf forests (28.3%~ 40.0%) and grasslands (25.4%~ 30.9%) (Fig. S24). These results suggest that the deciduous needleleaf and deciduous broadleaf forests (evergreen forests and deciduous needleleaf forests) are more vulnerable to the CWC (CWH) events than the closed shrubland and cropland (deciduous broadleaf forests and grasslands).

5.1 Difference in vegetation loss probability between compound and individual climate events

The results indicate that most vegetation types are rather sensitive to CDC events, and the average probability of NDVI below the 40% percentile is 46.9%~54.9% (Fig. 5 and Figs. S9-S11), followed by CWC (42.3%~51.1%, Fig. 8 and Figs. S17-S19), CDH (38.3%~47.7%, Fig. 6 and Figs. S12- S14) and CWH (30.6%~34.1%, Fig. 9and Figs. S20-S22) events. CDH events are generally considered as exerting a substantial adverse influence on vegetation growth globally (Feng et al. 2020; Hao et al. 2018, 2021). However, our findings suggest that vegetation is more vulnerable to the CDC and CWC events than CDH events in mainland China.

The probability of vegetation loss caused by CDC events is significantly larger than that caused by cold or dry events in most regions (Fig. 5, Figs. S5, S8, S9-S11), aligning with the previous global-scale studies (Li et al. 2022) which pointed out that the negative impact of CDC events on vegetation growth in mid- to high-latitude regions (> 23.5°N) is greater than that of cold and dry events. Under CWC conditions, the vegetation loss probability increases (decreases) significantly relative to wet (cold) conditions in most regions except for the central region of Inner Mongolia (northeast, southwest and southeast coastal regions) (Fig. 8, Figs. S6, S8, S17- S19). This suggests that the wet events weaken the adverse impact of cold events. This can be attributed to positive correlation between SPI and NDVI in these regions, where the wet conditions could favor vegetation growth (Fig. 3a), reducing the probability of vegetation loss. In addition, under CWC conditions, vegetation in most regions is more affected by cold events than wet events (Fig. 8, Figs. S6, S8, S17-S19) which leads to the fact that even when moisture is sufficient, the extreme cold condition can still have a significant negative impact on vegetation growth.

Compared with the dry (hot) events, the probability of vegetation loss due to CDH experiences a reduction (intensification) in most regions except for Inner Mongolia, and it also increases (decreases) as the intensity of dry (hot) conditions increases (Fig. 6, Figs. S5, S7, S12- S14), likely due to that the anomalous warmth during growing season might have facilitated vegetation growth, partially offsetting the adverse impacts of drought conditions (Nitzbon et al. 2020; Wang et al. 2020; Lian et al. 2021; Shao et al. 2021; Li et al. 2022). However, the barren and arid soils of the Inner Mongolian Plateau have the limited water-holding capacity, and the rapid depletion of soil moisture under CDH conditions further exacerbates vegetation loss (Bastos et al. 2020; Lian et al. 2020; Bevacqua et al. 2021; Li et al. 2022). Compared to the other three types of CCEs (CDC, CWC and CDH), CWH events are closer to the suitable conditions for vegetation growth, since the probability of vegetation loss caused by CWH events is significantly lower than that caused by the wet (or hot) events (Fig. 9, Figs. S6, S7, S20-S22).

5.2 Differences in vulnerability of different vegetation types under four CCEs

Our results show significant regional differences in vegetation vulnerability to CCEs (Figs. 5–6, 8–9), which may be influenced by a combination of factors such as vegetation type, climatic, and soil conditions (Hao et al. 2021; Fang et al. 2019a; Xu et al. 2018). Due to the superimposed effect of drought and low temperature, the growth and survival ability of most vegetation are restricted in arid and semi-arid regions (Richardson et al. 2018; Vitasse et al. 2018), and hence the vegetation in arid and semi-arid regions (e.g., the Loess Plateau and northwestern Xinjiang) is more susceptible to CDC events (Figs. 7 and S9-S11). In contrast, the vegetation in Inner Mongolia is predominantly composed of grasslands and savannas that are more reliant on the shallow soil water sources. High temperatures and drought exacerbate evapotranspiration and cause insufficient water supply for the growth of grasslands and savannas (Bao et al. 2014; Bastos et al. 2020; Ding et al. 2020; Hao et al. 2021; Li et al. 2022) and hence more susceptible to CDH events (Figs. 6 and S12-S14). Compared with the arid and semi-arid regions (e.g., Loess Plateau, Inner Mongolia, and northern Xinjiang), the probability of vegetation loss caused by CWC events is relatively higher in the humid regions (such as northeastern and southern China) (Fig. 8 and Figs. S17-S19). One possible explanation is that CWC events cause the over-humid soil in humid regions, posing excessive moisture challenges for certain vegetation, and thus showing adverse impact on vegetation growth, while in arid and semi-arid regions, wet events help to alleviate drought stress thereby promoting vegetation growth (Bao et al. 2014; Na et al. 2018; Pei et al. 2021; Li et al. 2022; He et al. 2023). A higher probability of vegetation loss due to CWH is found in only the eastern coastal and southwestern regions (Fig. 9 and Figs. S20-S22), mainly due to the stronger negative correlation between NDVI and SPI/STI over these regions (Figs. 3a and 3b) where the increased temperature and precipitation limit vegetation growth.

The differences in vegetation vulnerability are not only dependent on geographic location, but also closely related to vegetation types (Ding et al. 2020; Chen et al. 2023a; Chen et al. 2023b). Our results indicate that the closed shrubland, grassland, and cropland exhibit higher vulnerability to CDC and CDH events with the loss probability increasing with drought intensity, while the forests (except for deciduous broadleaf forests) show higher resistance to these events (Figs. 7 and S15-S16). This is consistent with the previous findings in Europe (Chen et al. 2023a) and global regions (Hao et al. 2021), suggesting that forest vegetation typically have extensive and deep root systems that can capture water from deeper soil layers, while shrublands, grasslands and cropland rely more on the near-surface soil moisture storage, rendering them more sensitive to moisture fluctuations.

There are also significant differences in the loss probability between forest vegetations. The deciduous broadleaf forests show the highest loss probability under CDC conditions (Fig. S15), likely related to their survival strategies and climate adaptability. When facing the external moisture and temperature stresses, the deciduous broadleaf forests may reduce leaf area or shed leaves to minimize evapotranspiration in order to protect themselves under drought impacts (Munné-Bosch and Alegre, 2004; Breda et al. 2006; Schuldt et al. 2020). The mixed forests show the lower (higher) vulnerability than the deciduous broadleaf forests (coniferous and evergreen broadleaf forests) under CDC conditions (Fig. S15), mainly due to that the mixed forest exhibits different vegetation compensatory strategies (broadleaf vs. coniferous) to cope with CCEs (Migliavacca et al. 2021; Pardos et al. 2021). The evergreen coniferous and evergreen broadleaf forests are distributed in humid and warm regions (Fig. 1a), where excessive wet conditions may lead to root asphyxiation, root rot, and other problems, restricting growth and even causing mortality (Carnicer et al. 2013; Pei et al. 2021; Li et al. 2022; He et al. 2023). Therefore, the evergreen coniferous and evergreen broadleaf forests are more vulnerable to CWH events than other vegetation types with the loss probability increasing with increasing wetting intensity (Fig. 10 and Figs. S24).

Under CDH conditions, grasslands demonstrate more pronounced vulnerability compared to wooded savannas and savannas (Fig. S16), meaning that the mixed vegetations of grassland and forest are more resistant to CDH events than the single grasslands. This can be attributed to the fact that the grassland-forest mixture has a richer vegetation structure than grasslands, and the deeper root systems and diversified vegetation composition in forests help to better utilize soil moisture thereby alleviating moisture stress during CDH events (Geng et al. 2019). A similar pattern can also be observed in croplands where the vegetation is more sensitive to CDH events than that in the cropland/natural vegetation mosaic (Fig. S16).

5.3 Limitations of this study

There are several limitations in this study, therefore caution should be exercised in interpreting the vegetation loss probability under the compound climatic conditions. First, some previous studies argued that NDVI cannot accurately reflect the actual physiological state for the areas with dense vegetation cover (e.g., forests) (Morton et al. 2014; Xie et al. 2019) due to the issues such as the red channel saturation in tropical forested areas (Morton et al. 2014), and it may introduce bias in assessing the probability of vegetation degradation, likely causing underestimation/overestimation of vegetation loss probability. In addition, our study focused on the co-occurrence of NDVI with CCEs during growing season without considering the effects of underlying surface factors (such as soil properties and soil moisture) and human activities (e.g., urbanization, agricultural expansion, irrigation, afforestation and overgrazing). Zhang et al. (2022b) found a bi-directional causality between soil moisture and vegetation productivity identified over 66% of the vegetated land areas. Fang et al. (2019) suggested that human activities such as irrigation and afforestation in the Loess Plateau region have alleviated the limitation of soil moisture on vegetation growth, increased (decreased) soil water storage capacity (soil erosion), and effectively reduced the vulnerability of vegetation to extreme drought events. Therefore, ignoring the impact of irrigation and afforestation may underestimate the vegetation loss probability caused by CCEs. Future research should consider utilizing the finer resolution vegetation and meteorological datasets and integrating multiple sources of data (e.g. geographic information system (GIS) data, human activity statistics, and records of natural disasters), and it is expected that the incorporation of these non-climate factors and underlying surface factors into the analysis framework can substantially contribute to a more comprehensive depiction of the real state of the environment.

This study developed a multivariate copula conditional probability (MCCP) framework for quantifying the loss probability of various vegetation types caused by four types of CCEs (CDH, CWH, CDC, CWC) with the aim to explore the vegetation response to these CCEs during growing season in mainland China. The CDH, CWH, CDC and CWC were identified based on the SPI (representing the dry or wet conditions) and STI (representing the hot or cold conditions) indices. The differences in the loss probability among 12 vegetation types caused by four types of CCEs were investigated. The following main conclusions can be drawn from this study:

(1) Vegetation is more vulnerable to CDC events than other three types of CCEs, with the average probability of NDVI below the 40th, 30th, 20th, and 10th percentiles estimated to be 46.9%-54.9%, 36.7%-47.4%, 25.7%-38.6% and 13.7%-26.9%, respectively. In contrast, CWH events lead to the lowest average probability of vegetation loss, with the average probability of 30.6%-34.1%, 24.1%-27.0%, 17%-19.6%, and 8.6%-11.7% for NDVI below the 40th, 30th, 20th, and 10th percentiles.

(2) The average probability of vegetation loss caused by CDC events is larger than that caused by dry (cold) events in the mainland China. Under CWC events, the average probability of vegetation loss is larger (smaller) than that under wet (cold) conditions, and the loss probability due to CWC decreases (increases) with the increasing intensity of cold (wet) events. Compared to hot (dry) events, CDH events result in a larger (smaller) probability of vegetation loss, and the probability increases (decreases) with the increasing intensity of dry (hot) events. Conversely, CWH events lead to a lower probability of vegetation loss than that caused by both the wet and hot events, and the loss probability increases (decreases) with the increasing intensity of wet (hot) conditions.

(3) Spatially, vegetation in the Loess Plateau and northwestern Xinjiang is highly susceptible to CDC events due to the superimposed effect of drought and low temperature, but the loss probability caused by CDC is slightly lower than that due to drought events in these regions. Vegetation (grasslands and savannas) in the Inner Mongolia region is more susceptible to CDH events. In contrast, vegetation in northeastern and southern China (eastern coastal and southwestern regions) is more vulnerable to CWC (CWH) events, while that in the Loess Plateau, Inner Mongolia, and northern Xinjiang is less affected by both CDH and CDC events.

(4) Shrubland, deciduous broadleaf forests, grassland, and cropland vegetation demonstrate the higher vulnerability to CDC and CDH events, with the loss probability increasing with drought intensity. Grasslands (cropland) exhibits more pronounced vulnerability than the woody savannas and savannas (cropland/natural vegetation mosaic) under CDH conditions. Deciduous (evergreen) forests face a high risk of vegetation decline under CWC (CWH) conditions, and the loss probability increases with the increasing wetness intensity. Furthermore, significant differences in the probability of loss are found between different forest types. Under CDC conditions, the mixed forests show less (more) vulnerability than the deciduous broadleaf forests (coniferous and evergreen broadleaf forests), while the evergreen coniferous and evergreen broadleaf forest are more vulnerable to CWH events than the deciduous and mixed forests.

Competing interests:

The authors declare no conflicts of interest.

Author Contribution

C.W. Conceptualization; T.Y. Data curation; C.W. and T.Y. Formal analysis; C.W. B.X.H. and Y.J. Funding acquisi-tion; C.W. and T.Y. Methodology; C.W., B.X.H. and Y.J.Project administration; T.Y. Visualization; T.Y. Writing - original draft; C.W. and P.J.Y. Writing - review & editing; All authors have read and agreed to the published version of the manuscript.

Acknowledgments

This work was supported by funding from the National Natural Science Foundation of China (Grant No. 52279016, 51909106), the Guangdong Basic and Applied Basic Research Foundation (Grant No. 2023A1515011760), the Natural Science Foundation of Shandong Province, China (Grant No. ZR2023QD090), the project of Jinan Science and Technology Bureau (2021GXRC070) and State Key Laboratory of Simulation and Regulation of Water Cycle in River Basin, China Institute of Water Resources and Hydropower Research (IWHR-SKL-KF202318).

Data Availability

The 1982–2020 monthly observed precipitation and temperature data are available for download from http://data.cma.cn. The 1982–2020 monthly NDVI dataset is obtained from http://www.geodata.cn/. The Land Cover Type 1 (LC_Type1) data is obtained from https://modis.gsfc.nasa.gov/data/dataprod/mod12.php.

Allen CD, Breshears DD, McDowell NG (2015) On underestimation of global vulnerability to tree mortality and forest die-off from hotter drought in the Anthropocene. Ecosphere 6(8):129. 10.1890/es15-00203.1
Anderegg WRL, Hicke JA, Fisher RA, Allen CD, Aukema J, Bentz B, Hood S, Lichstein JW, Macalady AK, McDowell N, Pan YD, Raffa K, Sala A, Shaw JD, Stephenson NL, Tague C, Zeppel M (2015) Tree mortality from drought, insects, and their interactions in a changing climate. New Phytologist 208(3): 674–683. DOI:101111/nph13477
AghaKouchak A, Chiang F, Huning LS, Love CA, Mallakpour I, Mazdiyasni O, Moftakhari H, Papalexiou SM, Ragno E, Sadegh M (2020) Climate Extremes and Compound Hazards in a Warming World Annual Review of Earth and Planetary. Sciences 48: 519–548. DOI101146/annurev-earth-071719-055228
Ara Begum R, Lempert R, Ali E, Benjaminsen TA, Bernauer T, Cramer W, Cui X, Mach K, Nagy G, Stenseth NC, Sukumar R, Wester P (2022) Point of Departure and Key Concepts. In: Pörtner -O, Roberts DC, Tignor M, Poloczanska ES, Mintenbeck K, Alegría A, Craig M, Langsdorf S, Löschke S, Möller V, Okem A, Rama B (eds) Climate Change 2022: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change [H. Cambridge University Press, Cambridge, UK and New York, NY, USA, pp 121–196. doi:10.1017/9781009325844.003.
Bao G, Qin ZH, Bao YH, Zhou Y, Li WJ, Sanjjav A (2014) NDVI-Based Long-Term Vegetation Dynamics and Its Response to Climatic Change in the Mongolian Plateau. Remote Sensing 6(9): 8337–8358. DOI:103390/rs6098337
Barbosa J, Rambal S, Soares AM, Mouillot F, Nogueira JMP, Martins GA (2012) PLANT PHYSIOLOGICAL ECOLOGY AND THE GLOBAL CHANGES. Cienc Agrotecnol 36(3):253–269 DOI:101590/s1413-70542012000300001
Bastos A, Ciais P, Friedlingstein P, Sitch S, Pongratz J, Fan L, Wigneron JP, Weber U, Reichstein M, Fu Z, Anthoni P, Arneth A, Haverd V, Jain AK, Joetzjer E, Knauer J, Lienert S, Loughran T, McGuire PC, Tian H, Viovy N, Zaehle S (2020) Direct and seasonal legacy effects of the 2018 heat wave and drought on European ecosystem productivity. Science Advances 6(24): eaba2724. DOI:101126/sciadvaba2724
Beniston M (2009) Trends in joint quantiles of temperature and precipitation in Europe since 1901 and projected for 2100. Geophys Res Lett 36:L07707 01029/2008GL037119
Bevacqua E, De Michele C, Manning C, Couasnon A, Ribeiro AFS, Ramos AM, Vignotto E, Bastos A, Blesic S, Durante F, Hillier J, Oliveira SC, Pinto JG, Ragno E, Rivoire P, Saunders K, van der Wiel K, Wu WY, Zhang TY, Zscheischler J (2021) Guidelines for Studying Diverse Types of Compound Weather and Climate Events. Earths Future 9: e2021EF002340. DOI:101029/2021ef002340
Breda N, Huc R, Granier A, Dreyer E (2006) Temperate forest trees and stands under severe drought: a review of ecophysiological responses, adaptation processes and long-term consequences. Ann For Sci 63:625–644
Broetto T, Tornquist CG, de Campos BHC, Schneider JC (2017) Relationships between Agriculture, Riparian Vegetation, and Surface Water Quality in Watersheds. Revista Brasileira De Ciencia Do Solo 41: e0160248. DOI:101590/18069657rbcs20160286
Carnicer J, Brbeta A, Sperlich D, Coll M, Penuelas J (2013) Contrasting trait syndromes in angiosperms and conifers are associated with different responses of tree growth to temperature on a large scale. Frontiers in Plant Science 4: 409. DOI:103389/fpls201300409
Chen Q, Timmermans J, Wen W, van Bodegom PM (2023a) Ecosystems threatened by intensified drought with divergent vulnerability. Remote Sensing of Environment 289: 113512. DOI:101016/jrse2023113512
Chen SL, Huang YF, Wang GQ (2019) Response of vegetation carbon uptake to snow-induced phenological and physiological changes across temperate China. Science of the Total Environment 692: 188–200. DOI:101016/jscitotenv201907222
Chen ZR, Huang YL, Shen YP, Fu WC, Yao X, Huang JK, Lan YX, Zhu ZP, Dong JY (2023b) How Vegetation Colorization Design Affects Urban Forest Aesthetic Preference and Visual Attention: An Eye-Tracking Study. Forests 14(7): 1491. DOI:103390/f14071491
Chiang F, Greve P, Mazdiyasni O, Wada Y, AghaKouchak A (2022) Intensified Likelihood of Concurrent Warm and Dry Months Attributed to Anthropogenic Climate Change. Water Resour Res 58(6):411 DOI:101029/2021wr030411
Coumou D, Robinson A, Rahmstorf S (2013) Global increase in record-breaking monthly-mean temperatures. Climatic Change 118(3–4): 771–782. DOI:101007/s10584-012-0668-1
De Beurs KM, Henebry GM, Owsley BC, Sokolik I (2015) Using multiple remote sensing perspectives to identify and attribute land surface dynamics in Central Asia 2001–2013. Remote Sensing of Environment 170: 48–61. DOI:101016/jrse201508018
Ding YB, Xu JT, Wang XW, Peng XB, Cai HJ (2020) Spatial and temporal effects of drought on Chinese vegetation under different coverage levels. Science of the Total Environment 716: 137166. DOI:101016/jscitotenv2020137166
Estrella N, Menzel A (2012) Recent and future climate extremes arising from changes to the bivariate distribution of temperature and precipitation in Bavaria Germany. Int J Climatol 33: 1687–95. DOI: 101002/joc3542
Fan FF, Xiao CW, Feng ZM, Yang YZ (2023) Impact of human and climate factors on vegetation changes in mainland southeast asia and yunnan province of China. Journal of Cleaner Production 415: 137690. DOI:101016/jjclepro2023137690
Fang W, Huang SZ, Huang GH, Huang Q, Wang H, Wang L, Zhang Y, Li P, Ma L (2019a) Copulas-based risk analysis for inter-seasonal combinations of wet and dry conditions under a changing climate. International Journal of Climatology 39(4): 2005–2021. DOI:101002/joc5929
Fang W, Huang SZ, Huang Q, Huang GH, Wang H, Leng GY, Wang L, Guo Y (2019b) Probabilistic assessment of remote sensing-based terrestrial vegetation vulnerability to drought stress of the Loess Plateau in China. Remote Sensing of Environment 232: 111290. DOI:101016/jrse2019111290
Feng HH, Zou B, Luo JH (2017) Coverage-dependent amplifiers of vegetation change on global water cycle dynamics. Journal of Hydrology 550: 220–229. DOI:101016/jjhydrol201704056
Feng SF, Hao ZC, Zhang X, Hao FH (2019) Probabilistic evaluation of the impact of compound dry-hot events on global maize yields. Science of the Total Environment 689: 1228–1234. DOI:101016/jscitotenv201906373
Feng SF, Wu XY, Hao ZC, Hao Y, Zhang X, Hao FH (2020) A database for characteristics and variations of global compound dry and hot events. Weather and Climate Extremes 30: 100299. DOI:101016/jwace2020100299
Geng SB, Shi PL, Song MH, Zong N, Zu JX, Zhu WR (2019) Diversity of vegetation composition enhances ecosystem stability along elevational gradients in the Taihang Mountains. China Ecological Indicators 104: 594–603. DOI:101016/jecolind201905038
Grossiord C (2020) Having the right neighbors: how tree species diversity modulates drought impacts on forests. New Phytologist 228(1): 42–49. DOI:101111/nph15667
GuoWW, Huang SZ, Huang Q, Leng GY, Mu ZX, Han ZM, Wei XT, She DX, Wang HY, Wang ZX, Peng J (2023) Drought trigger thresholds for different levels of vegetation loss in China and their dynamics. Agricultural and Forest Meteorology, 331:109349. Doi:101016/jagrformet2023109349
Hao Y, Hao ZC, Fu YS, Feng SF, Zhang X, Wu XY, Hao FH (2021) Probabilistic assessments of the impacts of compound dry and hot events on global vegetation during growing seasons. Environmental Research Letters 16(7): 074055. DOI:101088/1748-9326/ac1015
Hao ZC, AghaKouchak A, Phillips TJ (2013) Changes in concurrent monthly precipitation and temperature extremes. Environmental Research Letters 8(3): 034014. DOI:101088/1748-9326/8/3/034014
Hao ZC, Hao FH, Singh VP, Zhang X (2018) Changes in the severity of compound drought and hot extremes over global land areas. Environmental Research Letters 13(12): 124022. DOI:101088/1748-9326/aaee96
He L, Guo JB, Yang WB, Jiang QN, Chen L, Tang KX (2023) Multifaceted responses of vegetation to average and extreme climate change over global drylands. Science of the Total Environment 858: 159942. DOI:101016/jscitotenv2022159942
Huete A, Didan K, Miura T, Rodriguez EP, Gao X, Ferreira LG (2002) Overview of the radiometric and biophysical performance of the MODIS vegetation indices. Remote Sens Environ 83:195–213 DOI: 101016/S0034-4257(02) 00096 – 2
Jiang P, Ding W, Yuan Y, Ye W (2020) Diverse response of vegetation growth to multi-time-scale drought under different soil textures in China's pastoral areas. Journal of Environmental Management 274: 110992. DOI: 101016/jjenvman2020110992
Johnson NC, Xie SP, Kosaka Y et al (2018) Increasing occurrence of cold and warm extremes during the recent global warming slowdown. Nat Commun 9:1724 Doi: 101038/s41467-018-04040-y
Leonard M, Westra S, Phatak A, Lambert M, van den Hurk B, McInnes K, Risbey J, Schuster S, Jakob D, Stafford-Smith M (2014) A compound event framework for understanding extreme impacts. WILEY INTERDISCIPLINARY REVIEWS-CLIMATE CHANGE 5(1): 113–128. DOI:101002/wcc252
Li HW, Li YP, Huang GH, Sun J (2021) Quantifying effects of compound dry-hot extremes on vegetation in Xinjiang (China) using a vine-copula conditional probability model. Agricultural and Forest Meteorology 311: 108658. DOI:101016/jagrformet2021108658
Li J, Bevacqua E, Chen C, Wang ZL, Chen XH, Myneni RB, Wu XS, Xu CY, Zhang ZX, Zscheischler J (2022) Regional asymmetry in the response of global vegetation growth to springtime compound climate events. Communications Earth & Environment 3: 123. DOI:101038/s43247-022-00455-0
Lian X, Piao SL, Chen AP, Wang K, Li XY, Buermann W, Huntingford C, Penuelas J, Xu H, Myneni RB (2021) Seasonal biological carryover dominates northern vegetation growth. Nature Communications 12(1): 983. DOI:101038/s41467-021-21223-2
Lian X, Piao SL, Li LZX, Li Y, Huntingford C, Ciais P, Cescatti A, Janssens IA, Penuelas J, Buermann W, Chen AP, Li XY, Myneni RB, Wang XH, Wang YL, Yang YT, Zeng ZZ, Zhang YQ, McVicar TR (2020) Summer soil drying exacerbated by earlier spring greening of northern vegetation. Science Advances 6(1): eaax0255. DOI:101126/sciadvaax0255
Liu HY, Yin Y (2013) Response of forest distribution to past climate change: An insight into future predictions. Chin Sci Bull 58(35):4426–4436 DOI:101007/s11434-013-6032-7
Liu YF, Liu Y, Shi ZH, Lopez-Vicente M, Wu GL (2020) Effectiveness of re-vegetated forest and grassland on soil erosion control in the semi-arid. Loess Plateau Catena 195: 104787. DOI:101016/jcatena2020104787
Liu ZB, Zhu JY, Xia JY, Huang L (2023) Declining resistance of vegetation productivity to droughts across global biomes. Agricultural and Forest Meteorology 340: 109602. DOI101016/jagrformet2023109602
Luo W, Griffin-Nolan RJ, Ma W et al (2021) Plant traits and soil fertility mediate productivity losses under extreme drought in C3 grasslands. Ecology 102(10): e03465. DOI: 101002/ecy3465
Maurel C, Nacry P (2020) Root architecture and hydraulics converge for acclimation to changing water availability. Nat Plants 6(7):744–749 DOI:101038/s41477-020-0684-5
McKee TB, Doesken NJ, Kleist J (1993) In: The relationship of drought frequency and duration to time scales American Meteorological Society, Boston, MA, pp
Mekonnen ZA, Riley WJ, Berner LT, Bouskill NJ, Torn MS, Iwahana G, Breen AL, Myers-Smith IH, Criado MG, Liu YL, Euskirchen ES, Goetz SJ, Mack MC, Grant RF (2021) Arctic tundra shrubification: a review of mechanisms and impacts on ecosystem carbon balance. Environmental Research Letters 16(5): 053001. DOI:101088/1748-9326/abf28b
Meng Y, Hao ZC, Feng SF, Zhang X, Hao FH (2022) Increase in compound dry-warm and wet-warm events under global in CMIP6 models. Glob Planet Change 210:103773
Migliavacca M, Musavi T, Mahecha MD et al (2021) The three major axes of terrestrial ecosystem function. Nature 598:468–472 DOI:101038/s41586-021-03939-9
Morton DC, Nagol J, Carabajal CC, Rosette J, Palace M, Cook BD, Vermote EF, Harding DJ, North PRJ (2014) Amazon forests maintain consistent canopy structure and greenness during the dry season. Nature 506:221–224 DOI:101038/nature13006
Mulder CP, Iles DT, Rockwell RF (2016) Increased variance in temperature and lag effects alter phenological responses to rapid warming in a subarctic plant community. Global Change Biology 23(2): 801–814. DOI: 101111/gcb13386
Mukherjee S, Mishra AK (2021) Increase in Compound Drought and Heatwaves in a Warming World. Geophys Res Lett 48: e2020GL090617. DOI:101029/2020gl090617
Munné-Bosch S, Alegre L (2004) Die and let live: leaf senescence contributes to plant survival under drought stress. Funct Plant Biol 31(3):203–216 DOI:101071/fp03236
Na L, Na RS, Zhang JQ, Tong SQ, Shan Y, Ying H, Li XQ, Bao YL (2018) Vegetation Dynamics and Diverse Responses to Extreme Climate Events in Different Vegetation Types of Inner Mongolia. Atmosphere 9(10): 394. DOI:103390/atmos9100394
Nelsen RB (2007) An Introduction to Copulas. Springer Science & Business Media
Nicolai-Shaw N, Zscheischler J, Hirschi M, Gudmundsson L, Seneviratne SI (2017) A drought event composite analysis using satellite remote-sensing based soil moisture. Remote Sensing of Environment 203: 216–225. DOI:101016/jrse201706014
Nitzbon J, Westermann S, Langer M, Martin LCP, Strauss J, Laboor S, Boike J (2020) Fast response of cold ice-rich permafrost in northeast Siberia to a warming climate. Nat Commun 11:2201 DOI:101038/s41467-020-15725-8
Palmer WC (1965) Meteorological Drought. US Weather Bur Res Pap: No 45
Pardos M, del Rio M, Pretzsch H, Jactel H, Bielak K, Bravo F, Brazaitis G, Defossez E, Engel M, Godvod K, Jacobs K, Jansone L, Jansons A, Morin X, Nothdurft A, Oreti L, Ponette Q, Pach M, Riofrio J, Ruiz-Peinado R, Tomao A, Uhl E, Calama R (2021) The greater resilience of mixed forests to drought mainly depends on their composition: Analysis along a climate gradient across. Europe Forest Ecology and Management 481: 118687. DOI:101016/jforeco2020118687
Pascoa P, Gouveia CM, Russo AC, Bojariu R, Vicente-Serrano SM, Trigo RM (2020) Drought Impacts on Vegetation in Southeastern Europe. Remote Sensing 12(13): 2156. DOI:103390/rs12132156
Pei FS, Zhou Y, Xia Y (2021) Assessing the Impacts of Extreme Precipitation Change on Vegetation. Activity Agriculture-Basel 11(6):487 DOI:103390/agriculture11060487
Pendergrass AG, Meehl GA, Pulwarty R, Hobbins M, Hoell A, AghaKouchak A, Bonfils CJW, Gallant AJE, Hoerling M, Hoffmann D, Kaatz L, Lehner F, Llewellyn D, Mote P, Neale RB, Overpeck JT, Sheffield A, Stahl K, Svoboda M, Wheeler MC, Wood AW, Woodhouse CA (2020) Flash droughts present a new challenge for subseasonal-to-seasonal prediction. Nature Climate Change 10(3): 191–199. DOI:101038/s41558-020-0709-0
Pinzon JE, Tucker CJ (2014) A Non-Stationary 1981–2012 AVHRR NDVI3g Time Series. Remote Sens 6(8):6929–6960 DOI:103390/rs6086929
Qian X, Miao QL, Zhai PM, Chen Y (2014) Cold–wet spells in mainland China during 1951–2011. Nat Hazards 74(2):931–946 DOI: 101007/s11069-014-1227-z
Richardson AD, Hufkens K, Milliman T, Aubrecht DM, Furze ME, Seyednasrollah B, Krassovski MB, Latimer JM, Nettles WR, Heiderman RR, Warren JM, Hanson PJ (2018) Ecosystem warming extends vegetation activity but heightens vulnerability to cold temperatures. Nature 560(7718): 368–371. DOI:101038/s41586-018-0399-1
Raymond C, Horton MR, Zscheischler J et al (2020) Understanding and managing connected extreme events. Nat Clim Change 10:611–621
Schepsmeier U (2015) Efficient information based goodness-of-fit tests for vine copula models with fixed margins: A comprehensive review. J Multivar Anal 138:34–52 DOI:101016/jjmva201501001
Schuldt B, Buras A, Arend M, Vitasse Y, Beierkuhnlein C, Damm A, Gharun M, Grams TEE, Hauck M, Hajek P, Hartmann H, Hiltbrunner E, Hoch G, Holloway-Phillips M, Korner C, Larysch E, Lubbe T, Nelson DB, Rammig A, Rigling A, Rose L, Ruehr NK, Schumann K, Weiser F, Werner C, Wohlgemuth T, Zang CS, Kahmen A (2020) A first assessment of the impact of the extreme 2018 summer drought on Central European forests. Basic and Applied Ecology 45: 86–103. DOI:101016/jbaae202004003
Shao H, Zhang YD, Gu FX, Shi CM, Miao N, Liu SR (2021) Impacts of climate extremes on ecosystem metrics in southwest China. Science of the Total Environment 776: 10. DOI:101016/jscitotenv2021145979
Smith T, Boers N (2023) Reliability of vegetation resilience estimates depends on biomass density. Nat Ecol Evol 7:1799–1808 Doi:101038/s41559-023-02194-7
Stovall AEL, Shugart H, Yang X (2019) Tree height explains mortality risk during an intense drought. Nature Communications 10: 145979. DOI:101038/s41467-019-12380-6
Tencer B, Weaver A, Zwiers F (2014) Joint Occurrence of Daily Temperature and Precipitation Extreme Events over Canada. Journal of Applied Meteorology and Climatology 53(9): 2148–2162. DOI:101175/jamc-d-13-03611
Thom D, Rammer W, Seidl R (2017) The impact of future forest dynamics on climate: interactive effects of changing vegetation and disturbance regimes. Ecological Monographs 87(4): 665–684. DOI:101002/ecm1272
Vicente-Serrano SM, Beguería S, L´opez-Moreno JI (2010) A multiscalar drought index sensitive to global warming: the standardized precipitation evapotranspiration index. J Clim 23:1696–1718 DOI:101175/2009JCLI29091
Vicente-Serrano SM, Gouveia C, Camarero JJ, Beguería S, Trigo R, López-Moreno JI, Azorín-Molina C, Pasho E, Lorenzo-Lacruz J, Revuelto J (2013) Response of vegetation to drought time-scales across global land biomes. Proc Natl Acad Sci 110:52–57
Vitasse Y, Schneider L, Rixen C, Christen D, Rebetez M (2018) Increase in the risk of exposure of forest and fruit trees to spring frosts at higher elevations in Switzerland over the last four decades. Agricultural and Forest Meteorology 248: 60–69. DOI:101016/jagrformet201709005
Wang L, Chen W (2014) A CMIP5 multimodel projection of future temperature, precipitation, and climatological drought in China. Int J Climatol 34(6):2059–2078 DOI:101002/joc3822
Wang SH, Zhang YG, Ju WM, Porcar-Castell A, Ye SS, Zhang ZY, Brummer C, Urbaniak M, Mammarella I, Juszczak R, Boersma KF (2020) Warmer spring alleviated the impacts of 2018 European summer heatwave and drought on vegetation photosynthesis. Agricultural and Forest Meteorology 295: 108195. DOI:101016/jagrformet2020108195
Wang XH, Ciais P, Wang YL, Zhu D (2018) Divergent response of seasonally dry tropical vegetation to climatic variations in dry and wet seasons. Global Change Biology 24(10): 4709–4717. DOI:101111/gcb14335
Wen Y, Liu X, Xin Q, Wu J, Xu X, Pei F et al (2019) Cumulative effects of climatic factors on terrestrial vegetation growth. J Geophys Research: Biogeosciences 124:789–806 DOI:101029/ 2018JG004751
Wu X, Zhang R, Bento VA, Leng S, Qi J, Zeng J, Wang Q (2022a) The Effect of Drought on Vegetation Gross Primary Productivity under Different Vegetation Types across China from 2001 to 2020. Remote Sens 14: 4658. https://doiorg/103390/rs14184658
Wu T, Bai H, Feng F, Lin Q (2022b) Multi-month time‐lag effects of regional vegetation responses to precipitation in arid and semi‐arid grassland: A case study of Hulunbuir, Inner Mongolia. Natural Resource Modeling 35: e12342. https://doiorg/101111/nrm12342
Wu CH, Zhong LL, Yeh Pat J-F, Gong ZJ, Lv WH, Chen B, Zhou J, Li JY, Wang SS (2023) An evaluation framework for quantifying vegetation loss and recovery in response to meteorological drought based on SPEI and NDVI. Science of the Total Environment 906: 167632. DOI: 101016/jscitotenv2023167632
Wu CY, Gonsamo A, Chen JM, Kurz WA, Price DT, Lafleur PM, Jassal RS, Dragoni D, Bohrer G, Gough CM, Verma SB, Suyker AE, Munger JW (2012) Interannual and spatial impacts of phenological transitions, growing season length, and spring and autumn temperatures on carbon sequestration: A North America flux data synthesis. Global and Planetary Change 92–93: 179–190. DOI:101016/jgloplacha201205021
Wu DH, Zhao X, Liang SL, Zhou T, Huang KC, Tang BJ, Zhao WQ (2015) Time-lag effects of global vegetation responses to climate change. Global Change Biology 21(9): 3520–3531. DOI:101111/gcb12945
Wu XC, Guo WC, Liu HY, Li XY, Peng CH, Allen CD, Zhang CC, Wang P, Pei T, Ma YJ, Tian YH, Song ZL, Zhu WQ, Wang Y, Li ZS, Chen DL (2019a) Exposures to temperature beyond threshold disproportionately reduce vegetation growth in the northern hemisphere. National Science Review 6(4): 786–795. DOI:101093/nsr/nwy158
Wu XY, Hao ZC, Hao FH, Zhang X (2019b) Variations of compound precipitation and temperature extremes in China during 1961–2014. Science of the Total Environment 663: 731–737. DOI:101016/jscitotenv201901366
Xie XM, He B, Guo LL, Miao CY, Zhang YF (2019) Detecting hotspots of interactions between vegetation greenness and terrestrial water storage using satellite observations. Remote Sensing of Environment 231: 111259. DOI:101016/jrse2019111259
Xu HJ, Wang XP, Zhao CY, Yang XM (2018) Diverse responses of vegetation growth to meteorological drought across climate zones and land biomes in northern China from 1981 to 2014. Agricultural and Forest Meteorology 262: 1–13. DOI:101016/jagrformet201806027
Xu Y, Yang YP (2022a) A 5 km resolution dataset of monthly NDVI product of China (1982–2020). National Earth System Science Data Centre centre. DOI: 1012041/geodata239118756960240ver1db (In chinese)
Xu Y, Yang YP (2022b) A 5 km resolution dataset of monthly NDVI product of China (1982–2020). China Sci Data 7(1):1–9 DOI: 1011922/11-6035csd20210041zh (In chinese)
Yang WL, Pan JH (2023) The role of vegetation carbon sequestration in offsetting energy carbon emissions in the Yangtze River Basin, China. Environment Development and Sustainability 26. DOI:101007/s10668-023-03572-8
Yao Y, Fu BJ, Liu YX et al (2020) Evaluation of ecosystem resilience to drought based on drought intensity and recovery time. Agric For Meteorol 314:108809
Yu HL, Zhang J, Kong XC, Du GG, Meng BP, Li M, Yi SH (2022) The consequences of urbanization on vegetation photosynthesis in the Yangtze River Delta, China. Frontiers in Forests and Global Change 5: 996197. DOI:103389/ffgc2022996197
Yuan WP, Cai WW, Chen Y, Liu SG, Dong WJ, Zhang HC et al (2016) Severe summer heatwave and drought strongly reduced carbon uptake in Southern China. SciRep 6:18813
Yuan X, Wang LY, Wu PL, Ji P, Sheffield J, Zhang M (2019) Anthropogenic shift towards higher risk of flash drought over China. Nature Communications 10: 4661. DOI:101038/s41467-019-12692-7
Zhang N, Zheng XR, Wang X (2022a) Assessment of Aesthetic Quality of Urban Landscapes by Integrating Objective and Subjective Factors: A Case Study for Riparian Landscapes. Frontiers in Ecology and Evolution 9: 735905. DOI:103389/fevo2021735905
Zhang WX, Wei F, Horion S, Fensholt R, Forkel M, Brandt M (2022b) Global quantification of the bidirectional dependency between soil moisture and vegetation productivity. Agricultural and Forest Meteorology 313: 1087350. DOI: 101016/jagrformet2021108735
Zhang XH, Zhang BP, Yao YH, Wang J, Yu FQ, Liu JJ, Li JY (2022c) Dynamics and climatic drivers of evergreen vegetation in the Qinling-Daba Mountains of China. Ecological Indicators 136: 108625. DOI:101016/jecolind2022108625
Zhao WQ, Zhao X, Zhou T, Wu DH, Tang BJ, Wei H (2017) Climatic factors driving vegetation declines in the 2005 and 2010. Amazon droughts Plos One 12(4): e0175379. DOI:101371/journalpone0175379
Zheng JS, Xi XY, Jia GS (2022) Effects of Shifting Spring Phenology on Growing Season Carbon Uptake in High Latitudes. Journal of Geophysical Research-Biogeosciences 127(12): 1–15. DOI:101029/2022jg006900
Zhou P, Liu ZY (2018) Likelihood of concurrent climate extremes and variations over China. Environmental Research Letters 13(9): 094023. DOI:101088/1748-9326/aade9e
Zscheischler J, Michalak AM, Schwalm C et al (2014) Impact of large-scale climate extremes on biospheric carbon fluxes: An intercomparison based on MsTMIP data. Glob Biogeochem Cycles 28(6):585–600 DOI:101002/2014gb004826
Zscheischler J, Westra S, van den Hurk BJJM, Seneviratne SI, Ward PJ, Pitman A et al (2018) Future climate risk from compound events. Nature climate change 8(6): 469–477. DOI:101038/s41558-018-0156-3

No competing interests reported.

SupplementaryMaterials.docx

Download PDF

Reviewers agreed at journal
11 Sep, 2024
Reviews received at journal
09 Aug, 2024
Reviewers agreed at journal
31 Jul, 2024
Reviewers invited by journal
31 Jul, 2024
Editor assigned by journal
12 Jul, 2024
Submission checks completed at journal
11 Jul, 2024
First submitted to journal
11 Jul, 2024

You are reading this latest preprint version

Assessing the response lag and vulnerability of terrestrial vegetation to various compound climate events in mainland China under different vegetation types

Status:

Version 1

Abstract

Figures

1. Introduction

2. Study area and data

2.1 Study area

2.2 Data sources and preprocessing

3. Methods

3.1 Identifying the dry/wet and cold/hot events and CCEs

3.2 Estimating the lag time of vegetation changes in response to dry/wet and cold/hot events

3.3 Building the three-dimensional joint distribution models of SPI, STI and NDVI

3.4 Quantifying the probability of vegetation loss caused by the four types of CCEs

4. Results

4.1 Lag response of vegetation to dry/wet and cold/hot conditions at 1-to-24 timescales

4.2 Spatial variability of frequency and duration of CCEs

4.3 Probability of vegetation loss caused by CDC and CDH events

4.4 Probability of vegetation loss caused by CWC and CWH events

5. Discussion

5.1 Difference in vegetation loss probability between compound and individual climate events

5.2 Differences in vulnerability of different vegetation types under four CCEs

5.3 Limitations of this study

6. Conclusions

Declarations

Author Contribution

Acknowledgments

Data Availability

References

Additional Declarations

Supplementary Files

Status:

Version 1