Dynamics and Drivers of Vegetation Phenology in Three-River Headwaters Region Based on the Google Earth Engine

• Published in: Remote sensing (Basel, Switzerland) Vol. 13; no. 13; p. 2528
• Main Authors: Wang, Jiyan, Sun, Huaizhang, Xiong, Junnan, He, Dong, Cheng, Weiming, Ye, Chongchong, Yong, Zhiwei, Huang, Xianglin
• Format: Journal Article
• Language: English
• Published: Basel MDPI AG 01.07.2021
• Subjects: Aquatic ecosystems; Carbon; Climate change; Correlation coefficient; Correlation coefficients; Datasets; Ecosystems; Environmental impact; Environmental management; Fourier analysis; Freshwater resources; Global warming; Google Earth Engine; Grasslands; Growing season; Harmonic analysis; Headwaters; Heterogeneity; Human influences; Multivariate statistical analysis; Phenology; plant phenology; Precipitation; Radiation; Remote monitoring; Remote sensing; River basins; River ecology; Rivers; Soils; Spatial heterogeneity; spatiotemporal patterns; Stations; structural equation model; Structural equation modeling; Terrestrial ecosystems; Three-River Headwaters region; Trends; Variables; Vegetation; Wetlands; China
• ISSN: 2072-4292; 2072-4292
• DOI: 10.3390/rs13132528

Phenology shifts over time are known as the canary in the mine when studying the response of terrestrial ecosystems to climate change. Plant phenology is a key factor controlling the productivity of terrestrial vegetation under climate change. Over the past several decades, the vegetation in the three-river headwaters region (TRHR) has been reported to have changed greatly owing to the warming climate and human activities. However, uncertainties related to the potential mechanism and influence of climatic and soil factors on the plant phenology of the TRHR are poorly understood. In this study, we used harmonic analysis of time series and the relative and absolute change rate on Google Earth Engine to calculate the start (SOS), end (EOS), and length (LOS) of the growing season based on MOD09A1 datasets; the results were verified by the observational data from phenological stations. Then, the spatiotemporal patterns of plant phenology for different types of terrain and basins were explored. Finally, the potential mechanism involved in the influence of climatic and soil factors on the phenology of plants in the TRHR were explored based on the structural equation model and Pearson’s correlation coefficients. The results show the remotely sensed monitoring data of SOS (R2 = 0.84, p < 0.01), EOS (R2 = 0.72, p < 0.01), and LOS (R2 = 0.86, p < 0.01) were very similar to the observational data from phenological stations. The SOS and LOS of plants possessed significant trends toward becoming advanced (Slope < 0) and extended (Slope > 0), respectively, from 2001 to 2018. The SOS was the earliest and the LOS was the longest in the Lancang River Basin, while the EOS was the latest in the Yangtze River Basin owing to the impact of climate change and soil factors. Meanwhile, the spatial patterns of SOS, EOS, and LOS have strong spatial heterogeneity at different elevations, slopes, and aspects. In addition, the results show that the drivers of plant phenology have basin-wide and stage differences. Specifically, the influence of soil factors on plant phenology in the Yangtze River Basin was greater than that of climatic factors, but climatic factors were key functional indicators of LOS in the Yellow and Lancang river basins, which directly or indirectly affect plant LOS through soil factors. This study will be helpful for understanding the relationship between the plant phenology of the alpine wetland ecosystem and climate change and improving the level of environmental management.