Spatial and temporal analysis of drought resistance of different vegetation in the Ta-pieh Mountains based on multi-source data

Vegetation plays a crucial role in terrestrial ecosystems by comprising communities of ground-covering vegetation. It serves as a vital link between the atmosphere, hydrosphere, and soil, and also acts as an indicator of regional climatic characteristics (Ichii et al. 2010). The use of NDVI has been found to be highly effective in characterizing vegetation growth (Myneni & Williams 1994; Weber et al. 2018). Its sensitivity to meteorological threats and accurate reflection of background conditions make it efficient in capturing changes in vegetation (Philipp et al. 2021; Amin et al. 2022; Nama et al. 2022). Meteorological drought is a natural disaster caused by a prolonged lack of precipitation in a particular region (Agnew 1989). The impact of frequent droughts can be devastating to vegetated ecosystems, with dehydration causing wilting and death of vegetation in extreme cases (Jiang et al. 2022). Drought can also negatively affect vegetative productivity by limiting the amount of water available for photosynthesis (Jiang et al. 2021). Consequentially, research on how vegetation reacts to drought is one of the key topics in the field of ecological science (Ge et al. 2021; Liu et al. 2021b; Zhao et al. 2021). The impact of human activities on vegetation growth is increasing as the number of development projects rises (Sun et al. 2021; Yang et al. 2022b). Under the current conditions of global warming and frequent droughts, it is critical to investigate the trend and response characteristics of vegetation to droughts, as well as the strength of drought resistance of different vegetation and changes in drought resistance over time (Xu & Liu 2007; Zhao et al. 2022; Alvin John et al. 2023). Research contributes to ecological control at the start of a drought and gives direction for further management optimization.

More than 150 drought indices are currently used to monitor drought (Amir et al. 2023), with the Standardized Precipitation Index (SPI) and the Standardized Precipitation Evapotranspiration Index (SPEI) being the most widely used indices for regional meteorological drought assessment (Naresh Kumar et al. 2009; Cammalleri et al. 2021; Fu et al. 2022). SPI is simple to calculate and requires only precipitation data; SPEI is similar to SPI in that the difference between precipitation and potential evapotranspiration substitutes the precipitation data in SPI (Faye 2022; Sabri & Okan Mert 2023). Both indices are effective in reflecting regional drought conditions, and both are characterized by uncertainty (Laimighofer & Laaha 2022). The drought index can be used to investigate not only the occurrence of drought events and the trend of change but also the response of drought and vegetation. Droughts can limit vegetation growth, reduce vegetation cover, and possibly inflict irreversible harm to vegetation (Jiang et al. 2022; Martínez et al. 2022). In many investigations, meteorological station data are used to calculate the drought index SPI, which is then interpolated to meet regional drought monitoring requirements (Guo et al. 2019; Fu et al. 2022). The density and location of meteorological stations, as well as the interpolation technique employed, also impact the precision of the calculated results (Li et al. 2018). Meteorological grid data have more spatial coverage and resolution than station data, enabling for time series and multidimensional analysis. The limitations of the aforementioned techniques may be addressed, and the accuracy of regional drought monitoring can be greatly increased, by using continuous meteorological grid data. This method lowers the negative effect of erroneous data on research outcomes (Wei et al. 2022). Current investigation on vegetation response to drought focuses primarily on correlation and lag, as well as vegetation vulnerability in Inner Mongolia (Wei et al. 2022), the Pearl River Basin (Zhou et al. 2022), and Germany (Philipp et al. 2021). However, the previous studies mainly focused on the interannual variation characteristics of vegetation and failed to adequately reflect the seasonal differences of vegetation under climate change (Wei et al. 2022). Therefore, it is necessary to study the trend of vegetation change in different seasons for research. In this study, we explore the spatial and temporal patterns of vegetation in the Ta-pieh Mountain region and use the Mann-Kendall (M-K) significance test analysis and Theil-Sen median trend analysis to assess the trends of vegetation changes in different places (Mohamed et al. 2023; Muhammad Shehzad et al. 2023). In the context of global climate change, short-term droughts have also become more frequent. Therefore, this study evaluated and analyzed the extent of the impact of drought on vegetation growth in several seasons as well as seasonal-scale drought episodes in the study area. The stability of vegetation is usually measured by the coefficient of variation (CV) (Mann & Gupta 2022), and in this paper, the ratio of CV (RCV) in different time periods is used to reflect the changes in drought tolerance of vegetation. When examining the lag response of vegetation, sliding correlation analysis is typically used to pick the largest correlation value to define the lag period or merely to evaluate a single time series (Sun et al. 2021; Wei et al. 2022), which not only produces mistakes but also lacks spatial analysis. As a result, in this article, the phase spectrum in the cross-spectrum is utilized to more precisely characterize the lag period of vegetation, and the data are calculated raster by raster to analyze the spatial change law of the lag time of vegetation (Ebrahim et al. 2023).

The Ta-pieh Mountains are located in a significant region where the climate transitions between the northern and southern areas of China. These mountains are the origin of the Huai River, which is a crucial river in eastern China. Additionally, they serve as the watershed for both the Yangtze River and the Huai River. The impact of drought on the vegetative ecosystems of the Ta-pieh Mountains is extensive. Climate change is responsible for drought occurrences in the region, which are having a significant impact on plant ecology. Drought directly affects vegetation ecosystems in the Ta-pieh Mountains by impeding nutrient absorption and metabolic activity, leading to changes in vegetative shape, population dynamics, and even the extinction of vegetation species. These alterations have the potential to disrupt ecosystem diversity and integrity. Given the growing importance of drought on the ecological environment of vegetation in the Ta-pieh Mountains, it is vital to investigate and analyze the drought tolerance of various types of vegetation in the Ta-pieh Mountains. Surprisingly, there is almost a gap in current relevant research on the Ta-pieh Mountain area, where the drought tolerance of different vegetation is not clear and the mechanism of vegetation response to drought is not clear, so this study can fill the research gap in the ecological field of the Ta-pieh Mountain area, while also providing a reference for research on other similar areas in the mid-latitude. Based on this, the aim of this study was to investigate the drought tolerance and response of vegetation to drought at seasonal scales (spring, March–May, summer, June–August, fall, September–November, and winter, December–February of the following year).

In summary, the objectives of this paper are (1) to investigate the spatial and temporal changes of the NDVI for different seasons and topographies; (2) to measure seasonal trends and drought characteristics; (3) to analyze the changes of drought resistance of vegetation in the study area; and (4) to explore the effects of drought on vegetation at different seasons and to calculate the lag time.

The monthly scale rainfall datasets are from the National Tibetan Plateau Data Center (http://data.tpdc.ac.cn). These datasets were generated by the Delta spatial downscaling scheme for regional downscaling in China, and the results were verified to be credible. The temperature dataset is interpolated from weather station data using the Gaussian regression process (He et al. 2021) (https://www.zenodo.org/record). NDVI datasets from PROBA-V (https://land.copernicus.eu) have better temporal consistency than MODIS and allow for a better study of long-term changes in vegetation. The DEM altitude datasets are obtained from the Resource and Environmental Science and Data Center of the Chinese Academy of Sciences (https://www.resdc.cn), have a spatial resolution of 1 km, and are derived by resampling the most recent SRTM V4.1 data. The China Multi-period Land Use Remote Sensing Monitoring Dataset (CNLUCC) from the Resource and Environment Science Data Registration and Publication System (https://www.resdc.cn) (Xun et al. 2018). It has a higher precision when zoomed in compared to other datasets and is more suitable for long-term series studies (Suling et al. 2022).

In order to facilitate this study, 23 secondary land type classifications were regrouped and classified into 6 primary classifications based on land resources and their utilization characteristics: cropland, woodland, grassland, water, urban land, and unused.

This paper uses Matlab to extract and resample all data to a uniform spatial resolution to match all data pixels, as the spatial resolutions of the NDVI, precipitation, DEM, and land type datasets vary.

The M-K test may be applied to determine the time series trend (Yue et al. 2002). In circumstances where long-term datasets are prone to mistakes and deviations, this approach enables a more effective examination of the findings without the effects of small amounts of erroneous data. Therefore, it is widely used as a test for identifying monotonic climate trends. The trend categories are shown in Table 1.

The SPI (Naresh Kumar et al. 2009; WMO 2012) is a reliable indicator of drought and represents the probability of precipitation occurring in a region during a particular period of time. It possesses the benefits of simple calculation and remarkable stability and eliminates the effects of time and space. Moreover, it exhibits high sensitivity to drought changes and is well-suited for monitoring drought conditions (Fu et al. 2022). The SPEI is another commonly used drought index (Amir et al. 2023), and we calculated both SPI and SPEI for the study region simultaneously. The results showed a strong correlation between the two (see Supplementary Figure S1). Considering the simplicity of calculation and data availability, we chose SPI as the drought index for this study (Table 2).

In this paper, the 3-month SPI is calculated using monthly precipitation observations in the Ta-pieh Mountain region. Because the seasonal cycles of spring, summer, fall, and winter may be covered by the 3-month SPI for May, August, November, and February, respectively. To illustrate the drought characteristics of the seasons, the 3-month SPI of May, August, November, and February are employed as drought indicators.

The vegetation has some drought tolerance. Sometimes, vegetation responds to drought only after a period of time has elapsed. The correlation between NDVI and SPI was used to examine the effect of drought on vegetation (Wang et al. 2020).

The NDVI of cropland, grassland, and woodland all exhibit a distinct upward trend (Figure 3), and similar patterns exist in the Qilian Mountains (Zhang et al. 2021), in Yunnan (Sun et al. 2021) and Inner Mongolia (Wei et al. 2022). The causes of this phenomenon are inextricably linked to climate change and human activities (Chen et al. 2020; Yang et al. 2022a). China started adopting a variety of environmental programs in the early 21st century, exemplified by the return of crops to forests, and enacted relevant environmental protection laws to protect the ecological environment. In the Ta-pieh Mountains, horsetail pine (evergreen), fir (evergreen), and oak (evergreen) are the predominant tree species, which maintain the NDVI of the forest at a high level year-round (Miles & Esau 2016). NDVI was lower at lower altitudes and obviously higher in summer and autumn than in spring and winter. Most land use types are cropland, and the growing and maturing seasons of crops are in summer and autumn; wheat is dormant in winter and green in spring; rice is planted in spring and matures in autumn; and paddy fields are vacant in spring and winter, so the NDVI of cropland in spring and winter is obviously lower than that in summer and autumn (Nagai & Makino 2009; Kobata et al. 2018; Gao et al. 2020; Fu et al. 2022). The NDVI in large lakes and reservoirs in Huoqiu and Susong is significantly higher in summer, and the coverage area increases due to the combination of the summer season for the growth of aquatic plants and eutrophication of water bodies, which requires the formulation of special management policies to protect the ecosystem of water sources. When analyzing the trend of NDVI (Figure 5), we discovered that areas where NDVI is decreasing overlap significantly with cropland and urban land (Esau et al. 2016; Hwang et al. 2022), indicating that human activities have a negative effect on the growth of vegetation, and the greater the activity level, the greater the negative effect on vegetation (Hwang et al. 2022). On the other hand, in Italy, the effects of human activities like grazing reduction and afforestation have caused a decline in grassland and an increase in woodland, resulting in a definite trend of the nation becoming greener (Ebrahim et al. 2023). NDVI is lower at low altitude and low slope areas and higher at high altitude and high slope areas because cropland and urban regions occupy the low altitude and low slope sections of the study area predominantly while woodlands make up the high altitude and high slope areas; however, in winter, the NDVI diminishes above 800 m elevations and on slopes steeper than 15° (Figure 5). Steep terrain prevents effective infiltration of precipitation runoff into the soil for vegetation use. Moreover, at higher elevations (Kong et al. 2020) where temperatures are lower, vegetation's utilization of this runoff is further reduced, thereby inhibiting plant growth.

For the trend and significance test of drought in the Ta-pieh Mountain region (Figure 9), the trend of the region becoming increasingly arid and moist was insufficiently significant, and significance levels were low (P > 0.05). In the investigation of NDVI drought response (Figure 12), the proportion of positive (high) correlations for spring drought was substantially higher than for the other three seasons, demonstrating that the spring drought had the biggest effect on vegetation. Climate change is the key variables influencing plant growth (Luo 2007; Chen et al. 2020; Zhang et al. 2020). Increased temperatures and precipitation during the springtime encourage the activity of photosynthetic enzymes and hasten the mineralization and breakdown of organic materials (Luo 2007; Chen et al. 2020). Precipitation provides water for vegetation to synthesize substances necessary for life activities (organic matter, various enzymes), and if sufficient growth substances cannot be synthesized in spring, the growth of vegetation will be affected, thereby reducing crop yield or even causing extinction. However, in the spring, the intensity and frequency of droughts are high, so it is necessary to adopt measures such as irrigation in the spring to decrease the effect of drought on vegetation. In addition, the regions with a positive (high) correlation were predominantly cropland, whereas regions with a negative correlation were predominantly woodland (Wei et al. 2022). The root systems of perennial herbs and forests are strong and intricate, with fully developed organs, and they can obtain water from deeper soil layers to maintain normal growth even when drought occurs, whereas the root systems of crops on cropland have difficulty obtaining deeper soil water and maintaining normal growth (Kim et al. 2020).

In the analysis of the drought resistance of NDVI in Ta-pieh Mountain (Figure 14), the drought resistance of forest land and grassland, which are less affected by drought, has decreased, while the drought resistance of cultivated land has increased in the same period due to the scientific optimization of irrigation and fertilization methods for crops on cultivated land in recent years, which enables crops to make better use of water resources and fertilizers and therefore grow more efficiently (Cho et al. 2019; Nandan et al. 2021). The bulk of woodland and grassland are located at higher altitudes in mountainous and hilly regions that are particularly affected by climatic variables and less influenced by human activities. Therefore, the decreased drought resistance of woodland and grassland could be a result of their longer lifespans and limited adaptability to environmental changes (Lindner et al. 2010). Due to variations in vegetation type, there are areas in Figure 13 with abnormally high RCV values (RCV > 2.5) in all seasons and areas with abnormally low RCV values for the same reason.

In comparable experiments, we find that temperature lag times become shorter with increasing latitude, but there is no such pattern for precipitation, which is much more related to elevation, with higher elevations resulting in longer lag times, similar to the results in this paper (Ebrahim et al. 2023). Woodlands have the longest response time to drought, indicating that woodlands are more resistant to drought than cropland and grassland; a similar phenomenon is observed in Inner Mongolia (Wei et al. 2022). Because more root-derived carbon is conveyed from the rhizosphere to the bulk of the soil in rice soils, a broader range of rice rhizomes can make better use of soil water (Wang et al. 2018), so rice has greater overall drought tolerance than wheat.

In the present study, the spatial and temporal patterns of vegetation and drought and their responses at seasonal scales were investigated for the Ta-pieh Mountain region from 1999 to 2019 using NDVI, DEM, and precipitation data, Sen trend estimation, the M-K test, and cross-spectral analysis. NDVI increased at a higher rate in various types of vegetation, as indicated by its temporal changes. In terms of spatial pattern, NDVI exhibited a high central area and a low surrounding area, and NDVI trended higher in the majority of regions, with the exception of the northern cultivated areas in spring and winter, which exhibited large, concentrated degradation, and the degraded areas in summer and autumn, which were more dispersed. In addition, the NDVI exhibited substantial heterogeneity at various altitudes and gradients. Spring and autumn were typically wet, whereas summer and winter were typically arid. There were dry and moist areas in all seasons, but most were less significant. In the study area, drought intensity was greatest in winter and drought frequency was greatest in summer, with an overall inverse relationship between drought intensity and frequency. The analysis of the changes in drought resistance of various vegetation types during different seasons revealed that the drought resistance of cropland increased over time. Woodland demonstrated an increase in drought resistance that decreased only during the spring. In addition, grassland exhibited a decline in drought resistance during spring and summer but showed an increase during autumn and winter. Moreover, the average latency time to drought for cropland, woodland, and grassland was 1.62, 8.94, and 2.49 months, respectively, and forest land was significantly less affected by drought than cropland and grassland. The spring drought has a large influence on the majority of areas, including woodland, whereas the other three seasons of drought have vastly varying effects on the various categories of vegetation affected by it.

This research sheds light on the changing dynamics of vegetation in the Ta-pieh Mountain area, as well as its seasonal responses to drought. We address a research need in the Ta-pieh Mountain region by researching and assessing the drought tolerance of diverse flora species and its variations over time. The findings of the research have vital implications for the preservation and management of the natural environment not just in the Ta-pieh Mountain region but also in the larger context of the entire Huaihe River basin and the lower reaches of the Yangtze River. We have simultaneously filled the research gap in the studied area and provided reference for similar regions at the same latitude. The study assists us in better understanding the ecological concerns in the area and gives a foundation for building a feasible plan to maintain the ecological environment in the area.

This work was supported by the National Natural Science Foundation of China (grant number 42271084) and the Natural Science Foundation of Anhui Province (grant number 2208085US15).

H.L.: Conceptualization, Methodology, Software, Validation, Formal analysis, Writing – original draft. S.N.: Conceptualization, Methodology, Software, Supervision, Writing – review and editing, Funding acquisition. Y.Z.: Resources, Methodology, Funding acquisition, Visualization. C.W.: Methodology, Software, Investigation, Resources, Data curation. Y.C. and J.J.: Methodology, Software. X.X.: Writing – review and editing. A.R. and Y.C.: Software, Visualization.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.