锡林郭勒盟植被物候枯黄期对干湿变化的时间多尺度响应

基于全球监测与建模组(GIMMS)1982~2015年第三代归一化植被指数(NDVI) GIMMS NDVI 3g数据集和气象观测数据, 采用累计NDVI的Logistic曲线曲率极值法提取锡林郭勒草原植被枯黄期, 并结合不同时间尺度标准化降水蒸散指数(SPEI)分析植被枯黄期对干湿变化的响应特征.结果表明: 1982~2015年锡林郭勒草原植被平均枯黄期主要集中在265~280d, 由东北向西南推迟, 整体呈现提前趋势(-0.14d/a).年SPEI倾向率总体呈弱下降趋势(-0.10d/10a), 其中上升趋势主要分布在东北部草甸草原, 下降趋势主要分布在西南荒漠草原.近34a来, 锡林郭勒草原干旱发生频率高, 轻、中旱面积占比较大, 干旱突变始于20世纪90年代.时间多尺度SPEI影响EOS空间格局变化差异显著, 锡林郭勒盟植被EOS受月尺度SPEI-1正面影响较大, 短时间尺度干湿环境向好将延迟植被枯黄期; SPEI-3、SPEI-6、SPEI-9及SPEI-12与EOS负相关像元比例较高, 植被受干湿环境长期变化的影响提前进入枯黄期.时滞分析表明, 锡林郭勒盟SPEI-1的滞后效应主要发生在季前2个月, 长时间尺度SPEI(SPEI-3、SPEI-6和SPEI-9)累积至6月的滞后效应尤为显著, 特别在SPEI-3显著相关面积达到最大.

Response of phenological vegetation wilting period to multi-scale drying-wetting changes in Xilingol

Based on the third generation《Global Inventory Monitoring and Modelling Studies》(GIMMS) and the《Normalized Difference Vegetation Index》(NDVI) (1982~2015), this study extracted vegetation EOS (End of Growing Season) data in Xilingol grassland from 1982 to 2015, and analyzed its responses to multi-scale drying-wetting changes derived from the Standardized Precipitation Evapotranspiration Index (SPEI). The results showed that during 1982~2015, the EOS of Xilingol vegetation was 265~280 d, and delayed from northeast to southwest, with a decreasing trend of -0.14 d/a. Generally, the interannual variation in spatially averaged SPEI tended to decrease at -0.10d/10a. In addition, an increasing trend was mainly observed in the northeastern meadows, and a decreasing trend was observed in the southwestern desert grasslands. Over the past 34 years, the frequency of drought in Xilingol has increased, with the large areas experiencing the light to mid-droughts. The sudden change to frequent droughts occurred in the 1990s. There were significant spatial differences in the impacts of multiple SPEI time scales on EOS. In detail, the positive effect of monthly SPEI-1(Index for one month data) on vegetation EOS in the Xilingol was the greatest, suggesting that grassland vegetation EOS could be delayed in a favorable dry-wet environment over a short time scale. The SPEI-3, SPEI-6, SPEI-9, and SPEI-12(Index for many months data), however, had relatively large negative correlations with EOS, indicating that the long-term changes in the drying conditions could advance grassland EOS in Xilingol. The time-lag effect of SPEI-1 in the Xilingol mainly occurs two months before EOS. The long-term SPEI-3, SPEI-6, and SPEI-9 (especially SPEI-3), which are accumulated to June, have particularly considerable lag effects.