城市化影响下北京市植被物候变化特征

Spatiotemporal Dynamics of Vegetation Phenology in Beijing

摘要: 植被物候是环境变化的敏感指示器，植被物候的时空演化及其对城市化的响应研究有助于优化城市空间与绿地格局。基于MOD13Q1 EVI数据，本文采用S-G滤波进行时间序列重建，并借助动态阈值法提取物候参数，解析2002~2021年北京市物候的时空变化，同时在城–郊–村分区尺度上评估其对城市化强度的响应。结果表明：20年来全市植被物候呈现SOS提前(0.72 d∙a⁻¹)、EOS推迟(0.55 d∙a⁻¹)、GSL延长(1.26 d∙a⁻¹)的趋势；城市SOS最早(97.58 d)、EOS最晚(308.06 d)、GSL最长(210.02 d)。物候年际变化幅度呈梯度特征：城区最小(0.30 d∙a⁻¹)、郊区次之(1.07 d∙a⁻¹)、农村(1.41 d∙a⁻¹)更大。不透水面(ISP)增加普遍导致SOS提前、EOS推迟与GSL延长；在城市，郊区，农村三个分区中，ΔISP每增加10%，SOS分别提前0.96 d、0.97 d、0.56 d；EOS分别推迟0.28 d、0.31 d、0.26 d；GSL分别延长1.24 d、1.30 d、0.82 d，物候对城市化的响应具有明显的分区特征；将城郊村分区与不透水面占比结合，能够观察到物候对城市化敏感性的差异，从而提升物候对城市化响应的准确性。

Abstract: Vegetation phenology is a sensitive indicator of environmental change; elucidating its spatiotemporal evolution and responses to urbanization can inform urban spatial planning and green-space configuration. Using MOD13Q1 EVI, we reconstructed time series with a Savitzky-Golay filter and extracted phenological metrics via a dynamic-threshold method to characterize phenological changes in Beijing from 2002~2021, and assessed responses to urbanization intensity across urban-suburban-rural partitions. Citywide phenology exhibited earlier start of season (SOS; 0.72 d∙a⁻¹), later end of season (EOS; 0.55 d∙a⁻¹), and longer growing-season length (GSL; 1.26 d∙a⁻¹). In urban areas, SOS occurred earliest (DOY 97.58), EOS latest (DOY 308.06), and GSL was longest (210.02d). Interannual variability showed a clear gradient—smallest in urban zones (0.30 d∙a⁻¹), intermediate in suburban zones (1.07 d∙a⁻¹), and largest in rural zones (1.41 d∙a⁻¹). Increasing impervious surface percentage (ISP) generally advanced SOS, delayed EOS, and lengthened GSL. For each 10% increment in ΔISP, SOS advanced by 0.96 d, 0.97 d, and 0.56 d; EOS was delayed by 0.28 d, 0.31 d, and 0.26 d; and GSL extended by 1.24 d, 1.30 d, and 0.82 d in urban, suburban, and rural areas, respectively. These results demonstrate marked partition-dependent sensitivities of phenology to urbanization; combining the urban-suburban-rural framework with ISP effectively reveals heterogeneous sensitivities and improves attribution of phenological responses to urbanization.