Vegetation phenology is the research of events in the plant life cycle, and how these events adapt to climate changes (Zhang et al., 2018a; Nilsson et al., 2019), which makes it the most intuitive and sensitive biological indicator of climate change. Climate change such as modified precipitation, increasing temperature, and increasing frequency of extreme weather events will change the vegetation phenology (He et al., 2020), which have extensive effects on plant community structure, plant distribution, energy cycle, vegetation ecosystem, and primary production of vegetation (Gonsamo et al., 2013; Schwieder et al., 2016; Meng et al., 2019; Pabon-Moreno et al., 2020; Shen et al., 2020). Altered phenological timing, furthermore, may have negative effects on species interactions in the food chain such as migration, survival, reproduction, and occupying feeding habits (Burgess et al., 2018; Misra et al., 2021). Therefore, a better understanding of how vegetation phenology to climate change can improve the accuracy of phenological simulation models and provide a scientific reference for future management of vegetation ecosystems.
The plant growth such as leaf unfolding, flowering, fruiting, and leaf senescence is recorded to reflect the phenological characteristics of vegetation based on ground monitoring sites of phenology. However, this method has some limitations in spatiotemporal scale and biome-scale (Zhang et al., 2021a). Therefore, the remote sensing technique for phenological monitoring, which is used to simulate various logistics regression models through long-term series remote sensing data and to extract vegetation phenology (i.e., the start of the growing season (SOS), length of the growing season (LOS), and end of the growing season (EOS)) from suitable models, has been widely applied to explore the spatial-temporal variations of vegetation phenology at the regional and global scales (Steven P. Norman 2017; Araghi et al., 2019; Tong et al., 2019; Bornez et al., 2020; Geng et al., 2020; Xu et al., 2020). However, there is a wide range of mixed pixels in satellite remote sensing data, the spatial resolution of remote sensing data will affect the accuracy of vegetation phenological feature inversion, especially in the extreme arid areas where the vegetation distribution is more sparse and the coverage is lower. In addition, previous researches on vegetation phenology have focused on the plant phenology change trends (Delbart et al., 2015; Suepa et al., 2016; Lim et al., 2020), phenology changes of different vegetation types (Yuan et al., 2020), temperature (Ma et al., 2020; Zou et al., 2020; Xu et al., 2021), rainfall (Meier et al., 2015; Huang et al., 2020; Li et al., 2020), aspect (An et al., 2018), altitude (Zou et al., 2020), solar radiation (Jin et al., 2017), and other single climatic factors affecting vegetation phenology. Therefore, there is still a need to demonstrate the relationship of the preseason, interannual, and multi-climatic factors with phenology, and whether what kind of interaction exists between the SOS and EOS requires further study.
Arid and semi-arid ecosystems are important parts of the Earth's climate system, covering 41% of the Earth's land surface (Schimel, 2010; Yuan et al., 2014; Yu et al., 2021). These regions are characterized by harsh environments and little rainfall, especially in highly arid areas, where the average annual rainfall is often < 60-100mm (Noy-Meir, 1973). The Tarim River Basin is located in the hinterland of Eurasia, an extremely arid area. This region is highly vulnerable and sensitive to climate change due to the particular features of comprising mountains and a mosaic of inner-mountain basins (Zhou et al., 2020). Populus euphratica (P. euphratica), an ancient, precious, endangered species, is a constructive tree in extremely arid areas (Zhou et al., 2020). P. euphratica is a vital part of the extreme drought ecosystem in this region due to its advantageous characteristics such as cold resistance, heat resistance, drought resistance, wind-sand resistance, and salt-alkali resistance (Li et al., 2019a). In the upper Tarim River basin, since there are less affected by implementing the Ecological Water Conveyance Project than the lower Tarim River, the results that the response of P. euphratica phenology to climate change is more believable. However, research in this field mainly focuses on the relationship between vegetation and groundwater (Gou and Miller, 2014; Zhou et al., 2020), changes in rainfall and temperature under climate change and global warming, and its impact on water resources (Xu et al., 2010; Chen et al., 2015; Lang et al., 2016; Keram et al., 2019; Wu et al., 2020), and the impact of water on physical and chemical indicators of vegetation growth (Zhou et al., 2010; Eusemann et al., 2013; Yuan et al., 2014; Yu et al., 2021). To date, few if any studies have revealed the dynamics of P. euphratica phenology and its response to climate change in arid areas.
The novelty of this study is that a modified method was proposed to explore P. euphratica phenology and its response to climate change using 15-year Global LAnd Surface Satellite (GLASS) leaf area index (LAI) time-series data (2002–2016) in the upper Tarim River basin. A gridded spatial sampling strategy, which can reduce the accuracy error caused by many mixed pixels in satellite remote sensing data, was used to extract the LAI information of P. euphratica from the GLASS LAI time-series data, and the phenological information (i.e., SOS, LOS, EOS) was obtained with the dynamic threshold method from the reconstructed growth time series curve by using the Savitzky-Golay filtering method. The grey relational analysis (GRA) and canonical correlation analysis (CCA) methods were utilized to explore the relationship between phenology and multi-climatic factors. Specifically, The main objectives of this study were to ① quantitatively reveal the spatio-temporal characteristics and change trends of P. euphratica phenology, ② identify the key meteorological factors and key periods that affect the phenology of P. euphratica, and ③ confirm whether the beginning and the end of the P. euphratica growing season influence each other. The results provide insights into quantitative assessment of spatio-temporal variations of P. euphratica phenology under climate change and provide a scientific reference for future management of the desert ecosystem.