Divergent Response of Leaf Coloring Seasons to Temperature Change in Northern China over the Past 50 Years

Abstract

Autumn phenology plays a critical role in terrestrial ecosystem circulations. However, the changes in autumn phenology and their correlation with temperature remain uncertain because mean temperature alone was not able to determine the changes in autumn phenology at various sites. Here, the leaf coloring season (LCS) was defined as the period when the leaves of more than half of the species had recognized changes in color. We systematically studied the changes in peak, start, end, and duration of LCS and their correlations with five temperature parameters (mean temperature, accumulated cold temperature, day temperature, night temperature, and temperature difference between day and night) in four periods. Similarly to previous findings, the start date of LCS advanced and the end of LCS delayed over the past 50 years, which consequently led to a lengthened duration of LCS in Xi'an, Harbin, Minqin, and Shenyang. In general, the rise in mean temperature, day temperature, and night temperature would delay the peak, start, and end of LCS and lengthen the duration of LCS in most cases. We also proved that the changes in LCS metrics not only could completely be explained by mean temperature but also were influenced by day temperature, night temperature, temperature difference, and even other climatic factors such as precipitation, at different sites.

Usefulness of Social Sensing Using Text Mining of Tweets for Detection of Autumn Phenology

“…Until now, the spatio-temporal variability of leaf coloring and leaf fall have been detected through the following means: visual inspection of subject trees (Klosterman and Richardson, 2017;Park et al, 2017;Klosterman et al, 2018;Xie et al, 2018b;Tao et al, 2019), fixed-point observations by time-lapse digital cameras (Liu et al, 2018;Nagai et al, 2018;Richardson et al, 2018;Xie et al, 2018a), repeated observation by drones (Klosterman and Richardson, 2017;Klosterman et al, 2018), analysis of vegetation indexes obtained by satellite observation (Garonna et al, 2014;Nagai et al, 2015;Zhao et al, 2015;Klosterman et al, 2018), and analysis of phenology information published on the internet (Nagai et al, 2020). Of these means, phenology observations by visual inspection, digital cameras, and drones excel at detailed information, but can only cover a limited number of locations.…”