基于群落类型的寒区旱区物候特征及其对气候因子的响应——以2000-2019年柴达木盆地为例

植被物候反映了植被对环境的适应策略,物候研究对理解植被的生长机制及其对生态因子的响应有重要意义。以柴达木盆地植被为对象,基于2000—2019年NDVI数据,采用双逻辑函数法和STL时序分解算法,从区域和群落层面研究了物候的静态（空间异质性）、动态（时空变化）及其对气候的响应。结果表明：① 柴达木盆地植被生长季始期（SOS）和生长季末期（EOS）均以提前为主,分别提前了0.13天/年和0.23天/年,其中57%的植被区EOS显著提前（P<0.01）。生长季长度（LOS）缩短了0.09天/年。② SOS和EOS在空间分布和变化趋势两方面存在正相关关系,SOS较早发生的区域,EOS也较早,反之亦然,SOS提前的区域和EOS提前的区域重叠度高。③ 水热组合制约着植被物候,LOS最长的为温性草原（131天）和灌丛（128天）,最短的为高寒草甸（113天）。

Phenological characteristics of alpine arid region based on biome type and its responses to climate factors:A case study of Qaidam Basin from 2000 to 2019

Vegetation phenology reflects the adaptation of vegetation to the comprehensive environment. Phenology research is of great significance to understand the growth mechanism of vegetation in the ecosystem and its response to ecological factors, especially climate factors. Based on the MODIS NDVI data and Double Logistic Function fitting method, the vegetation phenology in the typical alpine arid region of Qaidam Basin from 2000 to 2019 was selected as the object. The characteristics of static state (spatial heterogeneity) and dynamics (temporal and spatial change) of vegetation phenology and their responses to climate change were studied at the regional and biome levels. The results are as follows. First of all, in Qaidam Basin, the start of growth season (SOS) and the end of growth season (EOS) were mainly advanced. Specifically, the start of growth season was 0.13 days/yr in advance and the end of growth season was 0.23 days/yr in advance. With regard to the end of growth season, 57% of the vegetation areas had a significant advance level (P<0.01). The length of growth season (LOS) was mainly shortened. The trend of shortening was 0.09 days/yr. Secondly, there was a positive correlation between the start of growth season and the end of growth season in spatial distribution and variation trend. In the region where the start of growth season occurred earlier, the end of growth season occurred earlier, and vice versa. The region where the variation trends of the start of growth season and the end of growth season were both in advance showed an obvious overlapping. The last but not the least, the combination of water and heat has obvious restrictive effect on vegetation phenology. No matter in distribution or variation trend of phenology, the start of growth season and the end of growth season were affected more significantly by precipitation, and the response of variation trend to precipitation was faster. Under the alpine arid condition, there was more precipitation in warm steppe and shrub, resulting in the longest growth season, 131 days and 128 days respectively. And the shortest growth season occurred in the alpine meadow (113 days).