利用条件植被指数评价西藏植被对气象干旱的响应

基于2000—2014年4—10月西藏气象站遥感干旱指数 (条件植被指数，VCI) 和气象干旱指数 (标准降水指数，SPI) 之间的相关性，评估植被对气象干旱的响应特征，通过分析气候环境要素对响应特征的影响并归纳相应规则，获取西藏地区植被对气象干旱有明显响应的区域分布。结果显示:VCI与12周时间尺度的SPI具有较强相关性，说明西藏地区植被生长对降水的响应大约滞后12周；植被对气象干旱响应不敏感的原因主要包括气候极度干燥或极度湿润、土地覆盖类型为森林、年平均归一化植被指数 (NDVI) 值过小、多年NDVI变化标准差过小、有降水之外的其他水源补给等；基于对区域气候环境要素特征的分析，可以得出西藏中部偏南地区植被对气象干旱有明显响应，主要包括拉萨地区、山南地区北部、日喀则地区东部、那曲地区中部和西南部、阿里地区的东南部。

Assessing Vegetation Response to Meteorological Drought in Tibet Autonomous Region Using Vegetation Condition Index

Tibetan plateau, as the third pole, is influenced by global climate change deeply. According to the 5th IPCC assessment report, temperature on the Tibetan Plateau is rising quickly, posing serious risks to agriculture, hydrological systems and so on. Drought is becoming a main hazard to agricultural production in Tibet, and therefore it's very essential to apply effective drought monitoring techniques in agriculture management in response to climate change. Although meteorological drought indices (such as standard precipitation index, SPI) are useful in drought measurement, they often have limited spatial resolution since they rely on in situ data. Satellites-based drought indices (such as vegetation condition index, VCI) can provide drought information over large areas at a higher spatial resolution, but in a different way from station-based meteorological drought indices. It has been recognized that the existing satellite-based drought indices are more associated with agricultural drought (e.g., vegetation health, crop yield, soil moisture, etc.), and the response of vegetation to meteorological drought (precipitation deficits) varies depending on the seasonal timing, land cover type, climate, soil properties, irrigation, and other factors.Correlation coefficients between VCI and SPI at different time scales for 30 meteorological stations in Tibet during 2000-2014 are calculated. First, the time scale of SPI that is most correlated with VCI is determined. Then, climatic and environmental factors are investigated to explain the spatial variation of this correlation. With considerations of inter-correlations among environmental factors, two preconditions are recognized as favorable for a strong correlation between VCI and SPI, and regions where vegetation responds to meteorological drought obviously are identified. Results are as follows. Firstly, correlations between VCI and SPI are time scale dependent, and the lag between the occurrence of precipitation and the vegetation response is about 12 weeks in Tibet. Secondly, there are obvious spatial variations in terms of the vegetation response to meteorological drought. Insensitive vegetation response is often associated with extremely dry or wet climate, forested land cover, low annual NDVI value, low multi-annual NDVI fluctuations, and water sources other than precipitation (e.g., snowmelt, irrigation). Thirdly, according to the climatic and environmental factors, vegetation in the middle southern part of Tibet responds to meteorological drought obviously, including Lhasa region, the northern part of Shannan, the eastern part of Rikaze, the middle and southwestern part of Naqu, and the southeastern part of Ali.