基于TVDI的横断山区干旱时空演变特征及影响因子研究

以横断山区为研究区，通过对比2001—2019年基于 MODIS获取增强型植被指数（EVI）和归一化差分植被指数（NDVI）反演的温度植被干旱指数（TVDI_N、TVDI_E）及潜在蒸散量（PET）和实际蒸散量（ET）反演的作物缺水指数（CWSI），与土壤含水量进行相关性分析，选择适用于横断山区干旱监测指标，采用Mann-Kendall 检验等统计方法研究干旱的时空变化特征，并分析干旱在不同土地类型、海拔和气象要素影响下的空间演变特征．研究结果表明:1）基于EVI反演的TVDI_E与土壤含水量相关性最高，更适合于监测横断山区的干旱情况．2）TVDI_E监测结果表明横断山区近19 年来干旱变化情况整体呈下降趋势，空间分布呈现南高北低的变化趋势，严重干旱主要集中分布在攀枝花市附近及湿热地带；北部则集中于三江流域（澜沧江、怒江、金沙江）附近及红原草原地区；干旱程度最严重的阶段是夏季，由春季逐渐向夏季过渡期阶段干旱面积明显增加．3）随海拔的增加，耕地分布多集中于海拔<3 000 m的地区，林地分布海拔为1 000~5 000 m的地区，草地主要生长于海拔>5 000 m的地区；19年来耕地、林地和草地中TVDI_E总体呈减弱趋势，但处在高原过渡带及干热河谷周围的植被的干旱呈增加趋势．在生长季缺水期时，南部地区的植被、高原过渡带林地和北部高原的草地受干旱影响严重．4）TVDI_E与日照时间的正相关性最高，与相对湿度的负相关性最高．气象因子对春初和秋末的TVDI_E复合作用最强，大部分区域呈现显著正相关． 

Spatial-temporal characteristics and possible impacts of drought in Hengduan Mountain: a TVDI study

Drought in Hengduan Mountain from 2001 to 2019 was studied．Comparison was made between two temperature vegetation dryness index (TVDI), calculated from enhanced vegetation index (EVI) and normalized difference vegetation index (NDVI), and crop water stress index (CWSI), calculated from potential evapotranspiration (PET) and actual evapotranspiration (ET)．Theil-Sen trend analysis and Mann-Kendall test were used to study spatial distribution, temporal variation and space-time evolution of drought．Meteorological data were correlated．EVI-based TVDI was found more suited for drought research．Ts-EVI characteristic space constructed from MODIS EVI and land surface temperature (LST) products showed an annual drought decrease in the study period, average annual drought was found higher in the south and lower in the north．Severe drought occurred mainly in Panzhihua, in hot and humid areas．In the north, concentration was noted in vicinity of the three rivers (Lancang River, Nujiang River and Jinsha River) and Hongyuan Grassland．The most severe drought occurred in summer, and drought increased significantly during transition from spring to summer．Significant correlation was found between drought and sunshine duration, but weak or no correlation between drought, precipitation and humidity respectively．Meteorological conditions could be used to predict disaster to facilitate disaster prevention and mitigation．