天山北坡前山带降水分布型对荒漠植被的影响——基于逐日降水数据和NDVI分析

降水分布型对荒漠植被的影响规律的揭示是理解全球气候变化下干旱区地表植被覆盖变化过程的基础。基于新疆天山北坡前山带1999-2014年生长季MODIS 1B逐日遥感资料和7个气象站点降水观测数据，运用GIS软件提取出气象站点周围荒漠植被的归一化植被指数（Normalized Difference Vegetation Index，NDVI）值，利用相关分析、决策树判识分析等方法，系统分析天山北坡前山带生长季荒漠植被NDVI在不同量级降水和降水型下的时空变化及其持续性等特征，并探究两者的相关程度。研究表明：生长季NDVI在不同时间、空间尺度上对降水存在一定规律的响应特征。从时间尺度来看，NDVI对降水的响应延迟现象随时间越来越不明显，两者相关性先增后减，在生长季中期相关性最强；空间尺度上，各区域差异较大，NDVI对降水的响应快慢与研究区各地的降水量大小明显相关。不同地区的年内NDVI变化趋势也不相同，18天以上的干期长度影响显著。无论研究区域中任一地区年均降水量大小，都一致表现出在生长季期间降水较多的时期，植被覆盖普遍出现为基本不变或轻微改善的趋势，相反在降水较少的时期，植被覆盖普遍出现轻微退化或基本不变（趋势斜率为负的极小值）的趋势。未来可以通过计算体现某地区植被覆盖变化趋势的值，来判断这一时期该地区大气的干湿程度。另外不同量级降水和降水分布型对NDVI变化影响明显，在生长季早、中期，降水量在6.1-12.0 mm时对荒漠植被的生长最为有利，NDVI增长幅度在0.15以上，并且对生长季早、中期植被生长最佳的降水分布型分别为B型降水（指连续两天出现降水且第一天的降水量大于第二天的降水量）和A型降水（指连续两天出现降水，第一天的降水量小于第二天的降水量）。揭示NDVI对降水的响应规律，可为干旱区、半干旱区合理生产灌溉、节约用水、生态恢复与重建等提供理论依据。

Influence of precipitation distribution on desert vegetation of Northern Piedmont, Tianshan Mountains: analysis based on daily NDVI and precipitation data

Revealing the influence characteristics of precipitation distribution on desert vegetation is helpful for understanding the process of vegetation cover change in arid area under the background of global climate change. Spatio-temporal pattern of Normalized Difference Vegetation Index (NDVI) and its possible relations to precipitation in the northern piedmonts of the Tianshan Mountains were analyzed based on MODIS 1B data and daily precipitation data in 1999-2014 from 7 meteorological stations. The results show that there are regular characteristics in the NDVI responding to precipitation at different time and spatial scales in the growing season. The lags of NDVI responding to precipitation become less and less obvious over time. The correlation between daily precipitation and NDVI increases firstly, then decreases, and reaches the maximum value in the medium term of the growing season. Lag of NDVI is clearly related to the precipitation of the region. The annual NDVI of each region is also different. When dry length is over 18 days, the significant effect on NDVI can be found. Overall, whether the annual precipitation in the study region is large or small, the vegetation all brings out essentially unchanged or slight improvement tendency during periods of heavy precipitation in growing season. Conversely, it all brings out degradation tendency or essentially unchanged (the slope of the trend is a negative relative minimum) during periods of little precipitation. Therefore, it can be estimated by the exploration of vegetation changes how dry and wet of atmosphere. Furthermore, precipitation and distribution type have obvious influence on NDVI. Precipitation of 6.1-12.0 mm is the most favorable condition for the growth of desert vegetation, and it can increase NDVI by more than 0.15. B type (referring to two consecutive days of precipitation, and precipitation on the first day more than next day) and A type (referring to two consecutive days of precipitation, and precipitation on the first day less than next day) respectively are the best distribution types of rainfall in the early and middle of the growing season. Exploration of the response pattern of NDVI to precipitation can not only offer a scientific basis for rational irrigation and water conservation, but it can also provide important theoretical guidance for restoration of damaged ecosystem in northern piedmonts of the Tianshan Mountains.