祁连山不同植被类型的物候变化及其对气候的响应

基于1982—2006年GIMMS NDVI和2000—2014年MODIS NDVI遥感数据,利用double logistic拟合方法提取了1982—2014年祁连山区不同植被的生长季始期、生长季末期和生长季长度3个重要的物候参数,分析了不同植被物候期的时间变化趋势、空间分异特征及对气候因子的响应。结果表明:(1)祁连山区不同植被的生长季始期和生长季末期随年际变化表现出波动提前或推迟,其中沼泽植被的变化波动最大;草甸植被、灌丛植被、阔叶林植被和栽培植被生长季长度出现延长趋势;(2)祁连山区植被生长季始期集中在5月初,其中阔叶林植被生长季开始最早,荒漠植被生长季开始最晚,植被生长季末期集中在9月,栽培植被生长季结束较早,荒漠植被、沼泽植被生长季结束较晚,植被生长季长度集中在110—140 d,其中阔叶林植被、针叶林植被生长季长度较长,而荒漠植被、高山植被生长季长度较短;(3)祁连山植被物候期变化趋势的空间分布表明植被生长季始期、生长季末期主要表现为提前不明显和推迟不明显,生长季长度主要表现为缩短不明显和延长不明显;(4)物候要素与气候要素相关性表明前期温度的积累有利于植被的开始生长,但当年3月的降水量对植被生长季始期同样有重要作用,不同植被生长季末期与8月、9月温度相关性较大,而与10月、11月降水的相关性较大。

Phenological variation in different vegetation types and their response to climate change in the Qilian Mountains, China, 1982-2014

Vegetation phenology is used in the study of climate change because of stable distribution, and ease of observation and interpretation. In this study, a 15-day time series of averaged normalized difference vegetation indices (NDVI) derived from the daily Global Inventory Modeling and Mapping Studies (GIMMS) NDVI dataset and 16-day averaged NDVI values derived from the daily Moderate Resolution Imaging Spectroradiometer (MODIS) NDVI dataset were used to analyze trends in vegetation phenology. Firstly, using the ENVI tool to cut remote sensing data, NDVI time series data for the study area from 1982 to 2014 were obtained. Secondly, using a Savitzky-Golay filter, noise contamination caused by random factors was reduced by producing a smooth NDVI curve. Thirdly, using double logistic fitting, three important phenological parameters, including the start of growing season (SOS), the end of growing season (EOS), and the length of growing season (LOS) were extracted for different vegetation types. Temporal change trends, as well as its spatial distribution characteristics, of the nine major vegetation types in the Qilian Mountains were analyzed. These types included meadow, steppe, desert vegetation, shrub vegetation, alpine vegetation, coniferous forest, broad-leaved forest, cultivated vegetation, and swamp vegetation. In addition, the effects of climatic factors on phenology were analyzed by correlation analysis. The results showed that:(1) The annual variation of SOS and EOS of different vegetation types in the Qilian Mountains fluctuated in its advance or or delay; the maximum variation was observed in the swamp vegetation. The LOS of meadow, shrub, coniferous forest, and cultivated vegetation was longer, but the LOS of desert vegetation shortened. (2) The SOS of vegetation was primarily in May; the growing season of the broad-leaved forests began the earliest and desert vegetation began the latest. The EOS of vegetation occurred primarily in September; the growing season of cultivated vegetation ended earlier, whereas the growing season of the desert and swamp vegetation ended later. The LOS of vegetation was from 110 days to 140 days, among which the LOS of broad-leaved and coniferous forests was longer, whereas the LOS of desert and alpine vegetation was shorter. (3) The spatial distribution of variation trends for vegetation phenology indicated that SOS and LOS were advanced or delayed but not substantially, and the LOS was primarily shortened or prolonged, but not substantially. (4) The correlation of phenological metrics and climatic factors indicated that the accumulation of early stage temperature was beneficial to the growth of vegetation, but the amount of precipitation in March was also important to the SOS of vegetation. The EOS of different vegetation types was related with temperature in August and September, and correlated with precipitation in October and November, but the correlation was not significant. At different altitudes, the phenological parameters were different. In particular, the LOS appeared to shorten with increasing elevation, which is consistent with the spatial distribution of phenological parameters. These results explain the relationship between climate change and phenological phases. Climate change has an obvious effect on vegetation phenological parameters, and the phenological period can explicitly indicate the climate change.