中国农业气象 ›› 2018, Vol. 39 ›› Issue (09): 559-566.doi: 10.3969/j.issn.1000-6362.2018.09.001

• 论文 • 上一篇    下一篇

基于去趋势分析的中国温带旱柳开花始期对气候变暖的响应

周孝煌,郑彦佳,徐琳,张煦庭,魏培   

  1. 中国农业大学资源与环境学院,北京 100193
  • 出版日期:2018-09-20 发布日期:2018-09-13
  • 作者简介:周孝煌(1993-),硕士生,从事中国北方植物物候对气候变化的响应研究。E-mail:274685287@qq.com
  • 基金资助:
    国家自然科学青年科学基金(41401048)

Detrending Time Series Improves the Response of Salix matsudana’s Flowering to Climate in China’s Temperate Zone

ZHOU Xiao-huang, ZHENG Yan-jia, XU Lin, ZHANG Xu-ting, WEI Pei   

  1. College of Resource and Environment Sciences, China Agricultural University, Beijing 100193, China
  • Online:2018-09-20 Published:2018-09-13

摘要: 为了更准确地揭示植物在物候水平上对气温变化的响应与适应机制,在去气温和物候时间序列趋势的基础上,建立日尺度的气温?物候回归模型,分析中国温带地区40个站点1986-2011年旱柳开花始期对气温变化的响应与适应机理。结果表明:中国温带地区77.5%的站点旱柳开花始期呈提前趋势,相应的敏感时段春季气温则呈上升趋势(0.8±0.5℃·10a-1)。其中,75%的站点物候期或气温时间序列存在显著变化趋势。进行去趋势处理后,17.5%的站点相关性发生变化,去趋势前其相关性较低,去趋势后,表现为显著负相关;模拟的均方根误差(RMSE)从去趋势前的5.5d降至去趋势后的4.9d,说明去趋势方法能提高气温-物候响应关系诊断的精确度。基于去时间序列线性趋势的气温-物候回归模型,得到92.5%的站点旱柳开花期与敏感时段春季气温呈显著负相关关系,即随着气温的升高,旱柳开花期呈提前趋势,其线性响应速率为-2.5±2.1d·℃-1。

关键词: 气候变化, 旱柳, 开花始期, 时间序列, 去趋势分析

Abstract: In order to reveal the response and plasticity mechanism of the plant to climate change more accurately, we established the daily mean air temperature-based phenology model on the basis of detrending inter annual variation in climate and phenology and analyzed Salix matsudana’s first flowering response to climate change and plasticity mechanism at 40 stations in the temperate zone of China from 1986 to 2011.The results showed that 77.5% of first flowering at the stations in the temperate zone of China advanced, and the climate in the corresponding sensitive period became warmer (0.8±0.5℃·10y-1). 75% of sites phenology or temperature time series had a significant trend at 40 stations. Comparing with detrending and actual values, 17.5% of relationships changed. Before detrending, the correlation was not significant. However, the results showed significant negative correlations after detrending. The root-mean-square error of the simulation ( RMSE) decreased from 5.5 days to 4.9 days, indicating that the detrending method can improve the accuracy of the temperature response to phenology. Therefore, based on the temperature-phenology regression model with detrending time series, 92.5% of stations had a significant negative correlation between the first floweing and spring temperature at the sensitive period, and the linear response rate was -2.5±2.1 d·℃-1.

Key words: Climate change, Salix matsudana, First flowering, Time series, Detrending