Next Article in Journal
A Comprehensive Review of Industrial Workshop Oil Mist Control Technology Based on Electrostatic Collection
Next Article in Special Issue
The Spatiotemporal Evolution of Geo-Disaster Resilience in China and the Impact Mechanism of Environmental Governance
Previous Article in Journal
Forecasting Ultrafine Dust Concentrations in Seoul: A Machine Learning Approach
Previous Article in Special Issue
Assessing the Impacts of Transition and Physical Climate Risks on Industrial Metal Markets: Evidence from the Novel Multivariate Quantile-on-Quantile Regression
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Asymmetrical Time-Lagged Response of Vegetation to Drought and Extreme Precipitation Across China

1
Key Laboratory of Earth Surface Processes and Environmental Change of Tropical Islands, Haikou 571158, China
2
College of Geography and Environmental Science, Hainan Normal University, Haikou 571158, China
3
Key Laboratory of Agro-Forestry Environmental Processes and Ecological Regulation of Hainan Province, Hainan University, Haikou 570228, China
4
College of Ecology and Environment, Hainan University, Haikou 570228, China
5
Institute of Tropical Bamboo, Rattan & Flower, Sanya Research Base, International Center for Bamboo and Rattan, Sanya 572000, China
*
Authors to whom correspondence should be addressed.
Atmosphere 2025, 16(3), 240; https://doi.org/10.3390/atmos16030240
Submission received: 3 January 2025 / Revised: 8 February 2025 / Accepted: 17 February 2025 / Published: 20 February 2025
(This article belongs to the Special Issue Climate Change and Extreme Weather Disaster Risks)

Abstract

:
In this study, a study area was chosen in China to analyze the lagged response relationship between normalized difference vegetation index (NDVI) and extreme precipitation/drought from 1982 to 2015. A logistical function was applied to explain the increase in NDVI with mean annual precipitation in nine sub-regions, and the inflection point of precipitation was found to be very close to the threshold value for separating arid or humid regions. NDVI had a strong positive correlation with drought and extreme precipitation in the arid regions, while in humid regions, it presented a strong correlation with drought during 2000–2015; however, a weak correlation with drought was found before the 21st century. In this study, we quantified the time-lagged response of vegetation to drought (LTRD) and extreme precipitation (LTREP). Then, we defined four gradients ( L T R D P , L T R D T , L T R E P P , and L T R E P T ) to quantify the precipitation and temperature gradients with the lag-time response to drought or extreme precipitation, respectively. Decreasing gradients were observed for humid regions with L T R D P = −0.19 month·100 mm−1 for “wetting” and L T R D T = −0.13 month·K−1 for “warming”, while increasing gradients were found in the same regions with L T R E P P = +0.18 month·100 mm−1 for “wetting” and L T R E P T = +0.14 month·K−1 for “warming”. These results suggest that the lagging responses of vegetation to extreme precipitation and droughts exhibit opposing regional patterns across China.

1. Introduction

In recent decades, climate change has increasingly impacted human beings and ecological environments at both global and regional scales [1,2], especially against the backdrop of global warming [3]. In the context of climate change, the multi-year average climate norms and extreme climate risks have experienced complex changes across different regions worldwide [4,5]. The terrestrial vegetation ecosystem has also changed due to the long-term impacts of climate variability [6]. It is now emphasized that long-term climatic changes have undermined the stability of the global climate system, leading to a significant rise in the frequency of extreme weather events over the past two decades [7,8]. Although global warming has increased evaporation, global average precipitation has increased by 1–3% per degree of warming on average [9,10]. In many regions, including China, even a slight change in the mean state of precipitation is now established to result in a significant increase in the frequency of extreme events, leading to a complex response of vegetation to extreme climates in a warmer world [11]. Consequently, extreme events associated with precipitation anomalies, such as droughts in California [12], Southeast Australia [13], and southwest China [14], as well as floods in Virginia [15], southern China [16], and Germany [17], have gained widespread attention: These events have generated discussions due to their irreversible effects on vegetation, such as widespread crop failure [18] and vegetation degradation [19].
Changes in water and heat conditions over a period of time do not immediately affect the growth of vegetation. The time-lagged response of vegetation to precipitation and temperature differs significantly between arid and humid regions [20,21]. Asymmetrical time-lagged responses reflect the varying survival strategies of vegetation in different climatic regions [22]. For instance, in arid regions, the strong demand for water prompts a shorter time lag or even an immediate response to extreme precipitation [23,24], but a longer time lag in response to persistent droughts may occur [25,26]. This survival strategy is common among drought-tolerant plants in arid regions, resulting in substantial differences in vegetation growth between dry and wet seasons [22]. In humid regions, waterlogging-tolerant vegetation (including heliophilic and ombrophytic plants) that is adapted to environments with sufficient radiation exhibits a long time lag in the response to extreme precipitation but a short time lag in the response to droughts [27]. During rainy days, the near-surface radiation received is reduced; therefore, its combined adverse effects with continuous precipitation on vegetation growth cannot be ignored in humid regions [28,29]. Therefore, vegetation growth in humid regions is influenced by both water (precipitation) and energy (temperature and radiation).
To assess the risks of droughts and extreme precipitation, a range of drought indices (such as the Palmer Drought Severity Index, PDSI, by Palmer in 1965 [30] and the standardized precipitation index, SPI, by McKee et al. in 1993 [31]) and extreme precipitation indices (including precipitation intensity, annual maximum 1-day precipitation, and annual maximum consecutive 5-day precipitation) have been developed in recent decades. Previous studies have identified the time lag in response to climatic variables (such as precipitation and temperature) or appropriate climate indices by conducting lag correlation analyses at monthly scales [14,32]. It is crucial to understand the time lag in vegetation response to extreme droughts or precipitation in various climatic regions. Therefore, in this study, we compared the spatial differences in the time lag of response to drought and extreme precipitation at several stations or regional scales. In Section 2, we describe the datasets and methods used to identify the time lag of vegetation response at several sites in China on a monthly scale. In Section 3, we quantify the time lag of response to drought and extreme precipitation in China using drought indices and extreme precipitation climate indices through lag correlation analysis. Finally, we present our discussions and conclusions in Section 4 and Section 5.

2. Data and Methods

2.1. Data Processing and Controls for Consistency

To calculate climate indices, such as extremes climate indices and SPI, and to quantify the time–time in vegetation response to extreme precipitation and drought, we utilized long-term observations of daily precipitation and air temperature from 756 stations provided by the National Climate Center of the China Meteorological Administration (https://www.cma.gov.cn/ (accessed on 5 January 2021)). To ensure that the overall quality of the statistics we gathered for sites across China was at par with research requirements at the site scale, we conducted temporal quality and consistency control checks based on daily precipitation and temperature.
Firstly, we selected stations that had at least 99% of the measurements for the entire period of 1960–2015; any missing data were replaced with a reasonable value using linear extrapolation. It is important to note that when using linear extrapolation methods, we also consider the boundary conditions of the data to ensure that all values remain within a reasonable range (e.g., precipitation should not be less than zero). Secondly, we applied the RHtestV4 software package to identify homogeneity in the series. Finally, we identified a total of 510 meteorological stations with continuous measurements in 9 climatic regions across China, as illustrated in Figure 1.
To investigate and quantify vegetation growth at the site scale in China, we utilized a long-term semimonthly NDVI dataset from GIMMS NDVI3g provided by NASA from 1982 to 2015 (https://iridl.ldeo.columbia.edu/SOURCES/.NASA/.ARC/.ECOCAST/ (accessed on 5 January 2021)). The dataset of NDVI has a resolution of 0.083° × 0.083°, and we extracted the NDVI values from the grids of all meteorological stations to represent the NDVI sequences for those stations. To minimize the effects of clouds, atmosphere, and monthly phenology, we adopted the maximum-value composite method to convert the dataset from the semimonthly to monthly scale. Finally, we extracted a new monthly NDVI dataset that corresponded to the longitude and latitude of 510 meteorological stations, ensuring that the vegetation and meteorological datasets were spatially consistent. This step was crucial in ensuring accurate analysis and interpretation of the vegetation response to climate variability and extreme events at the local scale.
To gain a deeper understanding of the differences in vegetation growth across the various climatic regions, we constructed a simple biophysical model that considered changes in precipitation. We hypothesized that vegetation growth in arid regions is limited solely by precipitation. The model for arid regions was expressed as follows:
N P = r N
where N is the potential growth capacity of vegetation under the average precipitation P, and r is the parameter representing the sensitivity of vegetation to precipitation.

2.2. Assessment of Precipitation Extremes and Droughts

To gain an understanding of vegetation response to precipitation extremes, it is essential to select appropriate indices that can be used to accurately characterize such events. Luo et al. [33] indicates that vegetation is more sensitive to responses to extreme precipitation, while it is not sensitive to changes in total precipitation amounts. The Expert Team on Climate Change Detection and Indices has recently recommended a series of extreme climate indices to assess the risk of extreme events [33,34]. These indices relate to precipitation extremes, such as maximum 1-day (hereafter RX1day) and 5-day precipitation episodes (hereafter RX5day), and have been shown to effectively assess the risk of flash floods at monthly and annual scales because they can extract essential information on precipitation extremes [35]. In our study, we used both RX1day and RX5day to represent the amount of extreme precipitation on different consecutive days. This approach allowed us to accurately characterize extreme precipitation events and their impact on vegetation growth.
The standardized precipitation index (SPI) is a widely used meteorological drought index, which was first formulated by McKee et al. in 1993 [31]. It considers flexible timescales, typically of 1, 3, 6, 12, or 24 months, and can effectively capture the cumulative effect of early rainfall on ongoing drought conditions. To calculate the SPI, the gamma distribution function is selected to fit long-term precipitation records for a given accumulation period. Then, the corresponding cumulative probability distribution is transformed to a standard normal distribution. It is important to note that longer timescales are more suitable in SPIs for arid and semi-arid regions, as they consider the greater accumulation of precipitation over time, leading to a reduced probability of zero precipitation. In our study, we selected the 12-month timescale as a typical annual scale to assess annual droughts. We extracted a non-overlapping annual SPI series for a given site, using 61 years of December SPI-12 data that were collected from 1960 to 2020. Conversely, shorter timescales should be selected when considering only precipitation in the ongoing month. Therefore, we applied the 3-month timescale of SPI (SPI-3) to quantify monthly droughts, which provided a more comprehensive assessment of drought conditions at a fine temporal resolution.

2.3. Time Lags in the Response of Vegetation to Precipitation Extremes and Droughts

Cross-correlation analyses can accurately identify the length of lagged response time in climate change and vegetation response research [36]. Theoretically, the dependent variable (y) is correlated with the time series of the independent variable (x) at any time lag (l), where l with the highest correlation (the max |rl|) is the length of response latency. The equation of the cross-correlation analysis is as follows:
r l x , y = cov x , y l σ x σ y
where r is the cross-correlation coefficient; x is the independent variable series and y is the dependent variable series; l is time lag at monthly scales; σx and σy are the standard deviations, and cov(x,yl) is the covariance of x and yl that can be defined as:
cov x , y l = i = 1 n l x i x ¯ ( y i + 1 y ¯ ) n l
where xi is the independent variable (SRX1day/SRX5day) at month i, yi+l is the dependent variable at month i + l, and l is the length of time lag at the monthly scale.
In this study, RX1day/RX5day and SPI are the independent variables and indicated extreme precipitation and droughts, respectively. NDVI is the dependent variable that indicates vegetation growth status.

2.4. Eliminating Seasonal Fluctuations

In semi-arid and arid regions, the periodic fluctuation in precipitation and temperature from the wet (April to October) to the dry seasons (November to March of the following year) determines a consistent change in vegetation growth. The existence of such seasonal fluctuation (in climatic variables and vegetation growth) can lead to false results based on non-random variables in the analysis of vegetation lag response. Hence, the seasonal fluctuations in precipitation extremes (standardized RX1day/RX5day, denoted as SMRX1day and SMRX5day) and NDVI (standardized NDVI, denoted as SNDVI) should be eliminated using a standardized method for each month as follows:
X i , j = x i , j x j ¯ σ x j
where xi,j is the monthly variable (RX1day/RX5day and NDVI) and where j denotes the month (from 1 to 12, i.e., January to December) for a given year i; σ x j is the standard deviations of x in the same month j, and Xi,j is the standardized variable (μ = 0; σ = 1).

3. Results

3.1. Spatial Relationship Between Precipitation and Vegetation

As illustrated in Figure 2, the water demand of vegetation varies across different climatic regions. Therefore, it is necessary to classify various climatic regions into arid or humid regions for accurate analysis. In our study, we divided nine sub-regions across China into four climatic categories: arid (annual P < 200 mm), semi-arid (200 mm < annual P < 400 mm), semi-humid (400 mm < annual P < 800 mm), and humid (annual P > 800 mm) regions. To further sub-classify these regions, we used an annual precipitation threshold of 400 mm. Regions with annual precipitation of <400 mm were classified as arid, including semi-arid regions. Conversely, regions with annual precipitation of >400 mm were classified as non-arid regions [37]. This classification allowed us to accurately distinguish among different climatic regions and evaluate vegetation response to precipitation variability and extremes in various settings.
To confirm that precipitation is the key factor limiting vegetation growth in arid regions, we analyzed NDVI and annual precipitation across the nine sub-regions. The NDVI in Northwest China, which was the driest region among all our regions of interest, was only 0.32, while the NDVI in South China, the region with the most abundant precipitation, was as high as 0.76.
Overall, these findings indicate that vegetation growth is better in regions with higher precipitation, suggesting that it is a crucial factor in determining vegetation growth in arid regions. These findings also highlight the importance of understanding precipitation variability and extremes in arid regions for effective management of vegetation resources.
In humid regions, the negative impact of reduced near-surface radiation on vegetation cannot be ignored, especially when the number of days with precipitation increases. Therefore, we introduced the negative impact of increased precipitation on vegetation using the logistical function. The model could be expressed as follows:
N P = r N 1 N K
where K is the maximum potential growth capacity of vegetation, and the integral form of the model could finally be expressed as:
N = K 1 + e a b P
where a and b are the two parameters of the model.
We plotted the NDVI against the mean precipitation in each sub-region and fitted a logistical curve that is shown in Figure 2b (determination coefficient, r2: 0.89 N D V I = 0.78 1 + e 1.962 0.008 P   ). Interestingly, the inflection point of the vegetation response to precipitation (P = 416 mm) was very close to the amount of precipitation that distinguished between arid and humid regions (400 mm). This finding suggests that the relationship between precipitation and vegetation growth is nonlinear and has a threshold. Below this threshold, vegetation growth is limited, whereas, above this threshold, the growth of vegetation increases until it reaches a saturation point. This information is valuable to understand vegetation response to precipitation variability and extremes in different climatic regions and can inform effective management and conservation strategies.

3.2. The Response of Vegetation to Extreme Precipitation and Droughts Across China

To demonstrate the response of vegetation to extreme precipitation and droughts in China, we prepared a plot of the standardized time series of RX5day and NDVI (mean ± standard) for arid regions (annual precipitation < 400 mm), including 123 sites across China. We selected the sum of monthly RX1day/RX5day for each year to take into account the extremes of each month.
As shown in Figure 3a, the NDVI gradually increased before 1990 and further increased after 2010, while only some change was noted from 1990 to 2010. Although the trends of RX5day were inconsistent with NDVI, the higher/lower NDVIs were always accompanied by higher/lower SMRX5day. For example, the highest SMRX5day (+2.5) and SNDVI (+2.8) appeared in the same year, i.e., 2012. A significant positive correlation (r = 0.54, p < 0.05) between these two parameters suggested that vegetation has a relatively more sensitive response (against to humid regions) to extreme precipitation in arid regions. Similarly, a significant correlation (r = 0.47, p < 0.05) between SNDVI and SPI confirmed that vegetation is sensitive to drought in these regions (Figure 3b).
We prepared the same set of graphs for sub-humid and humid regions (annual precipitation > 400 mm). A significant correlation between standardized RX1day (instead of RX5day) and standardized NDVI (r = 0.43, p < 0.05) was found during the same period (Figure 3c). This observation suggests that plants in humid regions do not have an urgent need for surplus water supply, such as from heavy rainfall over multiple days. Thus, we have a reason to believe that vegetation in humid regions may have lesser demand for water from extreme precipitation than in arid regions.
When droughts occur in humid regions, vegetation is not sensitive to the decrease in water supply due to more precipitation at earlier times. Nevertheless, a relatively weak response of vegetation to drought (r = 0.31, p = 0.07) could still be captured on an annual scale from 1982 to 2015 (Figure 3d). Consistent with the change in vegetation in arid regions, NDVI in humid regions increased rapidly from 2000 to 2012 and then decreased rapidly after 2013. A significant correlation was found to exist between NDVI and SPI in humid regions after 2000 (r = 0.63, p < 0.05). This significant correlation could be related to the increase in NDVI. As the conditions for vegetation growth improved, its demand for water increased due to the increase in transpiration rates. With global warming, the risk of extreme climate events occurring in China has increased [4,10]. Additionally, it has been observed that the growth of vegetation in arid regions is more susceptible to the impacts of extreme climate events due to water availability constraints (Figure 3b,d).

3.3. Time-Lagged Response of Vegetation to Droughts and Extreme Precipitation

Vegetation survival strategies differ widely across various climatic regions, such as from arid to humid, resulting in a large gap in the time-lagged response of vegetation to drought (LTRD) and extreme precipitation (LTREP). To better understand the delayed response of vegetation to extreme climates in different climatic regions, we conducted further investigations of LTRD and LTREP at monthly scales, respectively. Our findings show that LTRD varies spatially, as shown in Figure 4a. Overall, the time lag in the response of vegetation in southern China is shorter than that in northern China. Stations with short-term LTRD of 1–2 months are mainly distributed in southern humid regions, including Southeast China, Southwest China, Central China, South China, and a small part of arid regions with NDVI less than 0.2. In contrast, longer LTRD of more than 3 months was mainly recorded for Northeast China.
The spatial pattern of LTREP is inverse to that of LTRD, and larger spatial differences for LTREP were found between arid and humid regions in this study (Figure 4b). Stations with short LTREP of 1–2 months are mainly distributed all across China, except for Southeast China, South China, and the southern part of Southwest China, which have larger LTREP of >3 months.
Of interest in this study are the inter-regional differences in the time-lagged response of vegetation to drought and extreme precipitation. To this end, we defined a series of gradients ((1) LTRD with increasing annual precipitation, L T R D P ; (2) LTRD with increasing annual air temperature, L T R D T ; (3) LTREP with increasing annual precipitation, L T R E P P ; and (4) LTREP with increasing annual air temperature, L T R E P T ) to assess the changes in LTRD and LTRED from semi-humid (temperate) to humid (tropics) regions, respectively.
The marked gradients of LTRD with precipitation ( L T R D P ) or temperature ( L T R D T ) could be evaluated among various sub-regions. It was noted that the time-lagged response of vegetation to drought tends to be shorter in wet and warm regions (Figure 5a,b). For all the humid sub-regions, we further calculated the descending gradients of L T R D P = −0.18 month·100 mm−1 with “wetting” (r2 = 0.67, p < 0.05) and L T R D T = −0.13 month·K−1 with “warming” (r2 = 0.66, p < 0.05) using a simple linear regression model. The similar regional patterns of LTRD and LTREP suggested that both the LTRD and LTREP can reach 5-month-long time lags in the semi-humid region of Northeast China and 2-month-long time lags in the humid region of Southeast China.
Interestingly, the time-lagged response of vegetation to extreme precipitation presents an opposing regional pattern (Table 1). Specifically, the LTREP tends to be larger in wet and warm regions. The same method was applied to calculate the gradient to assess the change in LTREP with precipitation and temperature. The absolute value of the gradient of the time-lagged response of vegetation to drought was almost equal, but with the opposite sign (Table 1); the increasing gradient of L T R E P P = +0.18 month·100 mm−1 with “wetting” (r2 = 0.88, p < 0.01) and L T R E P P = +0.14 month·K−1 with “warming” (r2 = 0.92, p < 0.001), indicating that sub-regions with larger LTRD values were coupled with shorter LTREP (Figure 5c,d).

4. Discussion

Vegetation growth varies significantly among different climatic regions in China. The logistical function, introduced by Lieth (1973) [38] as the “Miami model”, evaluates the net primary productivity of terrestrial vegetation by considering the limitation of climatic variables. Luo et al. (2004) [39] verified that the leaf area index (LAI), an important vegetation index, can be fitted by the logistical curve. Numerous studies have confirmed that the spatial distribution of NDVI, which is the most widely used vegetation index, is mainly controlled by precipitation in many regions. In this study, we investigated whether NDVI exhibits threshold-like logistical patterns associated with mean annual precipitation (MAP) at regional scales (Figure 1). We found that NDVI showed significant logistical relationships with precipitation. Li et al. (2019) [40] reported that changes in precipitation have a bidirectional effect on vegetation growth in the regions in China where MAP fluctuates around 400 mm. Of great interest here is the threshold of MAP = 400 mm that divides sub-regions into arid or humid regions and is very close to the inflection point, where NDVI increases with precipitation (MAP = 416 mm).
The responses of vegetation to extreme climate events, such as droughts, floods, and heat waves, are becoming increasingly complex. Previous studies have noted that these responses can be decomposed into two types (Figure 6). The first type is the sensitivity of vegetation to climate change, which tends to focus on the consistency of vegetation and climate variables at annual scales. The NDVI-drought/NDVI-extreme precipitation correlation is stronger in arid and semi-arid regions (MAP < 400 mm) than in humid and semi-humid regions (MAP ≥ 400 mm). This is consistent with previous studies, which have confirmed that precipitation is the dominant factor driving vegetation growth in arid regions, while non-precipitation factors, such as radiation and temperature, are of significance in humid and semi-humid regions [22,41]. Overall, there is a negative correlation between drought/extreme precipitation and vegetation in most arid regions and some humid regions. The “Grain for Green Project” was initiated in 1999 to reduce soil erosion and improve local ecological conditions; it has contributed to improved vegetation in vulnerable areas, such as the Loess Plateau in the north of Shaanxi Province [42], upper and middle Yangtze River over Sichuan Province [43], and the typical arid/semi-arid rain-fed agricultural areas in Gansu Province [44]. Improved vegetation is bound to increase transpiration, and thus, the demand for water. In this study, we found a significantly increased occurrence of fluctuations in vegetation and drought/extreme precipitation (determination coefficient, r2) in the 21st century (shown in Figure 3d), which can be explained by a wide range of vegetation restoration efforts across China.
The other type of response of vegetation to climate is its lagged response to extreme events at monthly/weekly/daily scales, which is indicative of appropriate survival strategies of various vegetation types or climate regions. Therefore, we further studied the time lag in responses of vegetation (de-seasonalized NDVI) to drought (SPI) and extreme precipitation (de-seasonalized RX1day and RX5day) across China. The time lag in the responses of vegetation to drought/extreme precipitation differs among the nine sub-regions studied here. These time lags in responses were analyzed based on the maximum correlation coefficient between monthly de-seasonalized NDVI and SPI (or de-seasonalized RX1day and RX5day) using cross-correlation analysis. Inherent seasonal variations have been noted in vegetation and climate variables in the monthly scale studies. In this study, we did not consider the changes in the seasonal cycles of vegetation due to the earth’s revolution around the sun. The standardized method was used to remove seasonal cycles for a given long-term monthly NDVI, RX1day, or RX5day. The longest time-lagged response of vegetation to drought in China was ~5 months, which occurred in a sub-humid climate region (MAP = 600 mm). Vicente-Serrano et al. (2013) reported that semi-arid and sub-humid biomes respond to drought at longer timescales due to the lack of rapid response of arid biomes to such climatic conditions [22]. On the contrary, the shortest time lag in the response of vegetation to extreme precipitation, ~1 month, was detected in the same sub-humid regions in China. In addition, we quantified the variation in the time of lagged response to drought/extreme precipitation with temperature and precipitation, which tended to be shorter/longer in wetter and warmer regions, respectively.
Ding et al. (2020) compared the lag times of global vegetation responses to climate change and found that the lag times for vegetation responses vary from 1 month to 3 months across different regions on a global scale [45]. Wu et al. (2015) found that in the arid regions of the Americas, vegetation exhibits a shorter lag time in response to extreme precipitation, which is consistent with the findings of this study regarding China [46]. Ivits et al. (2016) quantified the lag time of vegetation response to climate change in Europe [47], noting that the lag time for drought response during the growing season ranges from 1 to 3 months. In arid regions, vegetation exhibits a longer lag time in response to drought, which is also consistent with our findings. In summary, the lag response time of vegetation to drought events in arid regions is longer compared to extreme precipitation events, while the results in humid regions are the opposite. Therefore, we revealed the differences in vegetation response patterns to extreme events between arid and humid regions under climate change (see Figure 6). In arid regions, the strong demand for water prompts a shorter lag time response to extreme precipitation, but a longer lag in response to persistent droughts (as shown in Conceptual Model 1 in Figure 6a). Moreover, the response of vegetation to the climate in humid regions follows two opposing patterns. The response of ombrophytic/shade-enduring plants to precipitation is similar to that of vegetation in arid regions (as illustrated by the green solid line in Figure 6b). In contrast, heliophiles with adequate water supply are limited by insufficient solar radiation (as indicated by the green dotted line in Figure 6b).

5. Conclusions

This study emerged to compensate for the lack of a systematic understanding of the different survival strategies of vegetation during drought or extreme precipitation in various regions with differing climates. To investigate these differences, we selected sites in China, including 510 meteorological stations, and assessed the response of vegetation to drought (SPI) or extreme precipitation (RX1day and RX5day) using GIMMS-NDVI from 1982 to 2015 at annual scales. On average, NDVI showed a significant positive correlation with drought (r = 0.54, p < 0.01) and extreme precipitation (r = 0.47, p < 0.001) in arid regions. By contrast, NDVI had a significant correlation with drought (r = 0.63, p < 0.05) during 2000–2015 in humid regions, but a weak correlation with drought (r = 0.31, p = 0.07) was observed before the 21st century. To explain the increase in NDVI with the variation in mean state precipitation in the nine sub-regions, we applied a logistical function. Using this function, we found that NDVI depends on a logistical curve based on the mean state of precipitation in each sub-region (r = 0.94), and the inflection point of precipitation (MAP = 416 mm) is very close to the threshold value (MAP = 400 mm) that divides the nine sub-regions into arid and humid regions in this study. To quantify the sensitivity of vegetation to drought and extreme precipitation at annual scales, we analyzed the correlation coefficient between NDVI and SPI (RX5day or RX1day) in arid (including semi-arid) and humid (including semi-humid) regions from 1982 to 2015. We found a strong relationship between NDVI and extreme precipitation in these regions, but only a weak relationship was observed between NDVI and drought in humid regions before the 21st century.
We systematically quantified the time-lagged response of vegetation to extremes at monthly scales; four gradients ( L T R D P , L T R D T , L T R E P P , and L T R E P T ) were defined to quantify the precipitation and temperature gradients with the time-lagged response to drought and extreme precipitation, respectively. In humid regions, decreasing gradients were found, which were L T R D P = −0.19 month·100 mm−1 for “wetting” and L T R D T = −0.13 month·K−1 for “warming”. On the other hand, increasing gradients were found in the same regions with L T R E P P , = +0.18 month·100 mm−1 for “wetting” and L T R E P T = +0.14 month·K−1 for “warming”. These results suggest that the response of vegetation to extreme precipitation is slow in humid regions, but its response to drought is faster in the same regions.

Author Contributions

Conceptualization: W.L., Y.C. and J.Z.; Funding acquisition: W.L. and J.Z.; Methodology: Y.C., W.L. and H.Y.; Resources: J.Z. and H.Y.; Software: W.L. and Y.C.; Validation: W.L.; Writing—original draft: J.Z. and W.L.; Writing—review and editing: H.Y. and W.L. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by Sanya Science and Technology Special Fund (2022KJCX04) and the Hainan Provincial Natural Science Foundation of China (No. 420RC601 and 320QN253).

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The precipitation and temperature datasets provided by the National Climate Center of the China Meteorological Administration (https://www.cma.gov.cn/ (accessed on 5 January 2021)). The GIMMS-NDVI3g dataset is from NASA (https://iridl.ldeo.columbia.edu/SOURCES/.NASA/.ARC/.ECOCAST/ (accessed on 5 January 2021)).

Conflicts of Interest

The authors declare that there are no conflicts of interest regarding the publication of this paper.

References

  1. Zhang, Y.; Gentine, P.; Luo, X.; Lian, X.; Liu, Y.; Zhou, S.; Michalak, A.M.; Sun, W.; Fisher, J.B.; Piao, S.; et al. Increasing Sensitivity of Dryland Vegetation Greenness to Precipitation Due to Rising Atmospheric CO2. Nat. Commun. 2022, 13, 4875. [Google Scholar] [CrossRef]
  2. Jiao, W.; Wang, L.; Smith, W.K.; Chang, Q.; Wang, H.; D’Odorico, P. Observed Increasing Water Constraint on Vegetation Growth over the Last Three Decades. Nat. Commun. 2021, 12, 3777. [Google Scholar] [CrossRef]
  3. Vicente-Serrano, S.M.; Domínguez-Castro, F.; Murphy, C.; Peña-Angulo, D.; Tomas-Burguera, M.; Noguera, I.; López-Moreno, J.I.; Juez, C.; Grainger, S.; Eklundh, L.; et al. Increased Vegetation in Mountainous Headwaters Amplifies Water Stress During Dry Periods. Geophys. Res. Lett. 2021, 48, e2021GL094672. [Google Scholar] [CrossRef]
  4. Trenberth, K.E.; Fasullo, J.T.; Shepherd, T.G. Attribution of Climate Extreme Events. Nat. Clim. Change 2015, 5, 725–730. [Google Scholar] [CrossRef]
  5. Thornton, P.K.; Ericksen, P.J.; Herrero, M.; Challinor, A.J. Climate Variability and Vulnerability to Climate Change: A Review. Glob. Change Biol. 2014, 20, 3313–3328. [Google Scholar] [CrossRef] [PubMed]
  6. Rahmstorf, S.; Coumou, D. Increase of Extreme Events in a Warming World. Proc. Natl. Acad. Sci. USA 2011, 108, 17905–17909. [Google Scholar] [CrossRef] [PubMed]
  7. Meehl, G.A.; Zwiers, F.; Evans, J.; Knutson, T.; Mearns, L.; Whetton, P. Trends in Extreme Weather and Climate Events: Issues Related to Modeling Extremes in Projections of Future Climate Change. Bull. Am. Meteorol. Soc. 2000, 81, 427–436. [Google Scholar] [CrossRef]
  8. Zscheischler, J.; Lehner, F. Attributing Compound Events to Anthropogenic Climate Change. Bull. Am. Meteorol. Soc. 2022, 103, E936–E953. [Google Scholar] [CrossRef]
  9. Rajah, K.; O’Leary, T.; Turner, A.; Petrakis, G.; Leonard, M.; Westra, S. Changes to the Temporal Distribution of Daily Precipitation. Geophys. Res. Lett. 2014, 41, 8887–8894. [Google Scholar] [CrossRef]
  10. Zhang, W.; Zhou, T. Increasing Impacts from Extreme Precipitation on Population over China with Global Warming. Sci. Bull. 2020, 65, 243–252. [Google Scholar] [CrossRef]
  11. Zhang, J.; Sun, F.; Lai, W.; Lim, W.H.; Liu, W.; Wang, T.; Wang, P. Attributing Changes in Future Extreme Droughts Based on PDSI in China. J. Hydrol. 2019, 573, 607–615. [Google Scholar] [CrossRef]
  12. Jasechko, S.; Perrone, D. California’s Central Valley Groundwater Wells Run Dry During Recent Drought. Earths Future 2020, 8, e2019EF001339. [Google Scholar] [CrossRef]
  13. King, A.D.; Pitman, A.J.; Henley, B.J.; Ukkola, A.M.; Brown, J.R. The Role of Climate Variability in Australian Drought. Nat. Clim. Change 2020, 10, 177–179. [Google Scholar] [CrossRef]
  14. Jiang, P.; Wang, Y.; Yang, Y.; Gu, X.; Huang, Y.; Liu, L.; Liu, L. Spatiotemporal Features and Time-Lagged Effects of Drought on Terrestrial Ecosystem in Southwest China. Forests 2023, 14, 781. [Google Scholar] [CrossRef]
  15. Praharaj, S.; Chen, T.D.; Zahura, F.T.; Behl, M.; Goodall, J.L. Estimating Impacts of Recurring Flooding on Roadway Networks: A Norfolk, Virginia Case Study. Nat. Hazards 2021, 107, 2363–2387. [Google Scholar] [CrossRef]
  16. Yang, W.; Yang, H.; Yang, D. Classifying Floods by Quantifying Driver Contributions in the Eastern Monsoon Region of China. J. Hydrol. 2020, 585, 124767. [Google Scholar] [CrossRef]
  17. Schmidt, L.; Heße, F.; Attinger, S.; Kumar, R. Challenges in Applying Machine Learning Models for Hydrological Inference: A Case Study for Flooding Events Across Germany. Water Resour. Res. 2020, 56, e2019WR025924. [Google Scholar] [CrossRef]
  18. Anderson, W.B.; Seager, R.; Baethgen, W.; Cane, M.; You, L. Synchronous Crop Failures and Climate-Forced Production Variability. Sci. Adv. 2019, 5, eaaw1976. [Google Scholar] [CrossRef]
  19. Avetisyan, D.; Borisova, D.; Velizarova, E. Integrated Evaluation of Vegetation Drought Stress through Satellite Remote Sensing. Forests 2021, 12, 974. [Google Scholar] [CrossRef]
  20. Zhao, J.; Huang, S.; Huang, Q.; Wang, H.; Leng, G.; Fang, W. Time-Lagged Response of Vegetation Dynamics to Climatic and Teleconnection Factors. Catena 2020, 189, 104474. [Google Scholar] [CrossRef]
  21. Wang, M.; An, Z.; Wang, S. The Time Lag Effect Improves Prediction of the Effects of Climate Change on Vegetation Growth in Southwest China. Remote Sens. 2022, 14, 5580. [Google Scholar] [CrossRef]
  22. Vicente-Serrano, S.M.; Gouveia, C.; Camarero, J.J.; Beguería, S.; Trigo, R.; López-Moreno, J.I.; Azorín-Molina, C.; Pasho, E.; Lorenzo-Lacruz, J.; Revuelto, J.; et al. Response of Vegetation to Drought Time-Scales across Global Land Biomes. Proc. Natl. Acad. Sci. USA 2013, 110, 52–57. [Google Scholar] [CrossRef] [PubMed]
  23. Jiang, H.; Xu, X. Impact of Extreme Climates on Vegetation from Multiple Scales and Perspectives in the Agro-Pastural Transitional Zone of Northern China in the Past Three Decades. J. Clean. Prod. 2022, 372, 133459. [Google Scholar] [CrossRef]
  24. Shi, X.; Chen, F.; Shi, M.; Ding, H.; Li, Y. Construction and Application of Optimized Comprehensive Drought Index Based on Lag Time: A Case Study in the Middle Reaches of Yellow River Basin, China. Sci. Total Environ. 2023, 857, 159692. [Google Scholar] [CrossRef] [PubMed]
  25. Cui, L.; Shi, J. Temporal and Spatial Response of Vegetation NDVI to Temperature and Precipitation in Eastern China. J. Geogr. Sci. 2010, 20, 163–176. [Google Scholar] [CrossRef]
  26. Guo, L.; Sun, F.; Liu, W.; Zhang, Y.; Wang, H.; Cui, H.; Wang, H.; Zhang, J.; Du, B. Response of Ecosystem Water Use Efficiency to Drought over China during 1982–2015: Spatiotemporal Variability and Resilience. Forests 2019, 10, 598. [Google Scholar] [CrossRef]
  27. Liu, K.; Harrison, M.T.; Archontoulis, S.V.; Huth, N.; Yang, R.; Liu, D.L.; Yan, H.; Meinke, H.; Huber, I.; Feng, P.; et al. Climate Change Shifts Forward Flowering and Reduces Crop Waterlogging Stress. Environ. Res. Lett. 2021, 16, 094017. [Google Scholar] [CrossRef]
  28. Liang, S.; Xu, D.; Luo, D.; Xiao, A.; Yuan, X. Study on the Impact of Land Use and Climate Change on the Spatiotemporal Evolution of Vegetation Cover in Chongqing, China. Atmosphere 2025, 16, 25. [Google Scholar] [CrossRef]
  29. Roebroek, C.T.J.; Melsen, L.A.; Hoek Van Dijke, A.J.; Fan, Y.; Teuling, A.J. Global Distribution of Hydrologic Controls on Forest Growth. Hydrol. Earth Syst. Sci. 2020, 24, 4625–4639. [Google Scholar] [CrossRef]
  30. Palmer, W.C. Meteorological Drought; U.S. Department of Commerce, Weather Bureau: Washington, DC, USA, 1965; Volume 30. [Google Scholar]
  31. McKee, T.B.; Doesken, N.J.; Kleist, J. The Relationship of Drought Frequency and Duration to Time Scales. In Proceedings of the 8th Conference on Applied Climatology, Anaheimet, CA, USA, 17–22 January 1993; Volume 17, pp. 179–183. [Google Scholar]
  32. Wen, Y.; Liu, X.; Yang, J.; Lin, K.; Du, G. NDVI Indicated Inter-Seasonal Non-Uniform Time-Lag Responses of Terrestrial Vegetation Growth to Daily Maximum and Minimum Temperature. Glob. Planet. Change 2019, 177, 27–38. [Google Scholar] [CrossRef]
  33. Luo, M.; Sa, C.; Meng, F.; Duan, Y.; Liu, T.; Bao, Y. Assessing Extreme Climatic Changes on a Monthly Scale and Their Implications for Vegetation in Central Asia. J. Clean. Prod. 2020, 271, 122396. [Google Scholar] [CrossRef]
  34. Ying, H.; Zhang, H.; Zhao, J.; Shan, Y.; Zhang, Z.; Guo, X.; Rihan, W.; Deng, G. Effects of Spring and Summer Extreme Climate Events on the Autumn Phenology of Different Vegetation Types of Inner Mongolia, China, from 1982 to 2015. Ecol. Indic. 2020, 111, 105974. [Google Scholar] [CrossRef]
  35. Xu, H.; Chen, H.; Wang, H. Detectable Human Influence on Changes in Precipitation Extremes Across China. Earths Future 2022, 10, e2021EF002409. [Google Scholar] [CrossRef]
  36. Guo, M.; Yue, W.; Wang, T.; Zheng, N.; Wu, L. Assessing the Use of Standardized Groundwater Index for Quantifying Groundwater Drought over the Conterminous US. J. Hydrol. 2021, 598, 126227. [Google Scholar] [CrossRef]
  37. Bai, X.; Fan, Z.; Yue, T. Dynamic pattern-effect relationships between precipitation and vegetation in the semi-arid and semi-humid area of China. Catena 2023, 232, 107425. [Google Scholar] [CrossRef]
  38. Lieth, H. Primary Production: Terrestrial Ecosystems. Hum. Ecol. 1973, 1, 303–332. [Google Scholar] [CrossRef]
  39. Luo, T.; Pan, Y.; Ouyang, H.; Shi, P.; Luo, J.; Yu, Z.; Lu, Q. Leaf Area Index and Net Primary Productivity along Subtropical to Alpine Gradients in the Tibetan Plateau. Glob. Ecol. Biogeogr. 2004, 13, 345–358. [Google Scholar] [CrossRef]
  40. Li, Y.; Xie, Z.; Qin, Y.; Zheng, Z. Estimating Relations of Vegetation, Climate Change, and Human Activity: A Case Study in the 400 Mm Annual Precipitation Fluctuation Zone, China. Remote Sens. 2019, 11, 1159. [Google Scholar] [CrossRef]
  41. Sun, W.; Wang, Y.; Fu, Y.H.; Xue, B.; Wang, G.; Yu, J.; Zuo, D.; Xu, Z. Spatial Heterogeneity of Changes in Vegetation Growth and Their Driving Forces Based on Satellite Observations of the Yarlung Zangbo River Basin in the Tibetan Plateau. J. Hydrol. 2019, 574, 324–332. [Google Scholar] [CrossRef]
  42. Li, G.; Sun, S.; Han, J.; Yan, J.; Liu, W.; Wei, Y.; Lu, N.; Sun, Y. Impacts of Chinese Grain for Green Program and Climate Change on Vegetation in the Loess Plateau During 1982–2015. Sci. Total Environ. 2019, 660, 177–187. [Google Scholar] [CrossRef]
  43. Peng, W.; Kuang, T.; Tao, S. Quantifying Influences of Natural Factors on Vegetation NDVI Changes Based on Geographical Detector in Sichuan, Western China. J. Clean. Prod. 2019, 233, 353–367. [Google Scholar] [CrossRef]
  44. Liu, S.; Huang, S.; Xie, Y.; Wang, H.; Huang, Q.; Leng, G.; Li, P.; Wang, L. Spatial-Temporal Changes in Vegetation Cover in a Typical Semi-Humid and Semi-Arid Region in China: Changing Patterns, Causes and Implications. Ecol. Indic. 2019, 98, 462–475. [Google Scholar] [CrossRef]
  45. Ding, Y.; Li, Z.; Peng, S. Global analysis of time-lag and -accumulation effects of climate on vegetation growth. Int. J. Appl. Earth Obs. Geoinform. 2020, 92, 102179. [Google Scholar] [CrossRef]
  46. Wu, D.; Zhao, X.; Liang, S.; Zhou, T.; Huang, K.; Tang, B.; Zhao, W. Time-lag effects of global vegetation responses to climate change. Glob. Change Biol. 2015, 21, 3520–3531. [Google Scholar] [CrossRef]
  47. Ivits, E.; Horion, S.; Erhard, M.; Fensholt, R. Assessing European ecosystem stability to drought in the vegetation growing season. Glob. Ecol. Biogeogr. 2016, 25, 1131–1143. [Google Scholar] [CrossRef]
Figure 1. Study area covering nine large sub-regions: 1, Northeast China (NEC); 2, Inner Mongolia (IM); 3, Northwest China (NWC); 4, North China (NC); 5, Central China (CC); 6, Southeast China (SEC); 7, South China (SC); 8, Southwest China (SWC); and 9, the Tibetan Plateau (TP) with various normalized difference vegetation index (NDVI) ranging from 0.1 to 0.9. In this study area, 510 out of 756 meteorological stations (gray dots) with continuous instrumental meteorological records were selected.
Figure 1. Study area covering nine large sub-regions: 1, Northeast China (NEC); 2, Inner Mongolia (IM); 3, Northwest China (NWC); 4, North China (NC); 5, Central China (CC); 6, Southeast China (SEC); 7, South China (SC); 8, Southwest China (SWC); and 9, the Tibetan Plateau (TP) with various normalized difference vegetation index (NDVI) ranging from 0.1 to 0.9. In this study area, 510 out of 756 meteorological stations (gray dots) with continuous instrumental meteorological records were selected.
Atmosphere 16 00240 g001
Figure 2. Explaining the spatial pattern of vegetation with various annual precipitation among nine sub-regions using the logistical function: (a) spatial difference in precipitation and NDVI; (b) the scatter plots of annual mean precipitation and NDVI fitted by the logical curve ( N D V I = 0.78 1 + e 1.962 0.008 P   , r2 = 0.88). In (b), the blue solid circles indicate the scatter points representing the relationship between annual precipitation and NDVI.
Figure 2. Explaining the spatial pattern of vegetation with various annual precipitation among nine sub-regions using the logistical function: (a) spatial difference in precipitation and NDVI; (b) the scatter plots of annual mean precipitation and NDVI fitted by the logical curve ( N D V I = 0.78 1 + e 1.962 0.008 P   , r2 = 0.88). In (b), the blue solid circles indicate the scatter points representing the relationship between annual precipitation and NDVI.
Atmosphere 16 00240 g002
Figure 3. The relationship between standardized NDVI and climate indices from 1982 to 2015 in arid (a,b) and humid (c,d) regions. The shaded range in the four plots was estimated using σ / n .
Figure 3. The relationship between standardized NDVI and climate indices from 1982 to 2015 in arid (a,b) and humid (c,d) regions. The shaded range in the four plots was estimated using σ / n .
Atmosphere 16 00240 g003
Figure 4. Spatial distribution of the time-lagged response of vegetation to (a) drought and (b) extreme precipitation across China from 1982 to 2015.
Figure 4. Spatial distribution of the time-lagged response of vegetation to (a) drought and (b) extreme precipitation across China from 1982 to 2015.
Atmosphere 16 00240 g004
Figure 5. The time-lagged responses of vegetation to drought (a,b) and extreme precipitation across the nine sub-regions studied (c,d). In this study, the gradients for different types of lagged responses could be quantified by linear fitting. Error bars represent ± σ / n . The coefficients of determination and their significance levels are also shown in all four graphs. The results of the gradient changes in the lagged response are quantified by the slope of the linear (solid line) in each subplot.
Figure 5. The time-lagged responses of vegetation to drought (a,b) and extreme precipitation across the nine sub-regions studied (c,d). In this study, the gradients for different types of lagged responses could be quantified by linear fitting. Error bars represent ± σ / n . The coefficients of determination and their significance levels are also shown in all four graphs. The results of the gradient changes in the lagged response are quantified by the slope of the linear (solid line) in each subplot.
Atmosphere 16 00240 g005
Figure 6. Conceptual models explaining the different responses of vegetation to drought and extreme precipitation. (a) Typical patterns in arid regions with a longer time lag in response to drought, but shorter time lag in response to extreme precipitation; (b) Typical patterns in humid regions with a shorter time lag in response to drought, but longer time lag in response to extreme precipitation. In the figures, the black dotted line represents the multi-year average state of vegetation or climate.
Figure 6. Conceptual models explaining the different responses of vegetation to drought and extreme precipitation. (a) Typical patterns in arid regions with a longer time lag in response to drought, but shorter time lag in response to extreme precipitation; (b) Typical patterns in humid regions with a shorter time lag in response to drought, but longer time lag in response to extreme precipitation. In the figures, the black dotted line represents the multi-year average state of vegetation or climate.
Atmosphere 16 00240 g006
Table 1. Summary of the gradients in time-lagged response of vegetation to increase in drought and extreme precipitation with increasing temperature using linear fitting among the various regions (annual P > 400 mm or annual T > 5 °C).
Table 1. Summary of the gradients in time-lagged response of vegetation to increase in drought and extreme precipitation with increasing temperature using linear fitting among the various regions (annual P > 400 mm or annual T > 5 °C).
L T R D P (Month·100 mm−1) L T R D T (Month·K−1) L T R E P P (Month·100 mm−1) L T R E P T (Month·K−1)
Gradient−0.19−0.13+0.18+0.14
r20.670.660.880.92
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Lai, W.; Chen, Y.; Zhang, J.; Yang, H. Asymmetrical Time-Lagged Response of Vegetation to Drought and Extreme Precipitation Across China. Atmosphere 2025, 16, 240. https://doi.org/10.3390/atmos16030240

AMA Style

Lai W, Chen Y, Zhang J, Yang H. Asymmetrical Time-Lagged Response of Vegetation to Drought and Extreme Precipitation Across China. Atmosphere. 2025; 16(3):240. https://doi.org/10.3390/atmos16030240

Chicago/Turabian Style

Lai, Wenli, Yongxiang Chen, Jie Zhang, and Huai Yang. 2025. "Asymmetrical Time-Lagged Response of Vegetation to Drought and Extreme Precipitation Across China" Atmosphere 16, no. 3: 240. https://doi.org/10.3390/atmos16030240

APA Style

Lai, W., Chen, Y., Zhang, J., & Yang, H. (2025). Asymmetrical Time-Lagged Response of Vegetation to Drought and Extreme Precipitation Across China. Atmosphere, 16(3), 240. https://doi.org/10.3390/atmos16030240

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop