摘要
植被与气候的关系非常密切,植被物候可作为气候变化的指示器。东北地区位于我国最北部,是气候变化的敏感区域,研究该区植被物候对气候变化的响应对阐明陆地生态体统碳循环具有重要意义。利用GIMMS AVHRR遥感数据集得到了东北地区阔叶林、针叶林、草原和草甸4种植被25a(1982—2006年)的物候时序变化,得出4种植被春季物候都表现出先提前后推迟的现象,秋季物候的变化则比较复杂,阔叶林和针叶林整体上呈现出秋季物候推迟的趋势,草原和草甸则表现为提前-推迟-提前的趋势。应用偏最小二乘(Partial Least Squares)回归分析了该区域植被物候与气候因子之间的关系,结果表明:春季温度与阔叶林、针叶林和草甸春季物候负相关,前一年冬季温度与草原春季物候正相关,降水与植被春季物候的关系有点复杂;4种植被秋季物候与夏季温度均呈正相关,除草原外,其余3种植被秋季物候均与夏季降水负相关。植被春季物候可能主要受温度影响,而秋季物候很可能主要受降水控制。
Climate change is a very important issue in the natural sciences,and has received much attention in various research fields. Vegetation phenology may be a good indicator of climate change at the regional or global scale,because of the close relationship between vegetation and climate. In this study,we analyzed the trend of vegetation phenology from1982 to 2006 and its driving climatic factors in northeastern China,which has experienced a rapid climate change in the past three decades partially due to its high latitude. We used a time series of 15-day averaged NDVI derived from the daily GIMMS AVHRR dataset to analyze the trend of vegetation phenology. We first used a Savitzky-Golay filter to reduce the noise in the NDVI curve to account for data contamination by random factors,then conducted a double logistic fitting to extract phonological parameters. To account for varied phenology responses to climate change among different vegetation types,we analyzed time series of those phonological parameters for the four major vegetation types in northeastern China,including broad-leaved forest,coniferous forest,steppe,and meadow. In addition,we performed a Partial Least Squares（ PLS） regression to examine the relationship between vegetation phenology and climatic variables. Results showed that spring phenology exhibited an advancing trend followed by a delay for all four vegetation types,but different vegetation types had different turning points. In contrast,the autumn phenology was somewhat complicated with inconsistent patterns across the four vegetation types. Broad-leaved forest and coniferous forest had an overall delayed trend,but the other two typesshowed a delay-advancing-delay trend. During the study period of 25 years,the spring phenology advanced 11 days for meadow,7 days for coniferous forest,5 days for broad-leaved forest,and 3 days for steppe. Autumn phenology was delayed6 days for broad-leaved forest,4 days for coniferous forest,and 1 day for meadow,while the steppe showed an advance of 8days. Partial
出处
《生态学报》
CAS
CSCD
北大核心
2016年第7期2015-2023,共9页
Acta Ecologica Sinica
基金
国家自然科学基金优秀青年科学基金项目(41222004)
中国科学院"百人计划"项目(09YBR211SS)
国家"十二五"长白山森林景观恢复项目(Y2KJB631G2)
关键词
气候变化
遥感
物候
偏最小二乘回归
climate change
remote sensing
phenology
partial least squares regression
