Development And Application Of A LAI-based Global Vegetation Drought Index

Drought is a complex climate hazard.Under the background of global change,effectively and timely monitoring drought could mitigate the impact of droughts.Drought indices describe drought characteristics quantitatively and are becoming popular in drought related research.The healthy condition of vegetation,which could be monitored by remotely sensed Leaf Area Index(LAI),is a good indicator of vegetation drought.In this study,we developed a vegetation drought index based on a remote sensing LAI product and then applied the index globally to detect vegetation drought events,identify drought characteristics and derive features of drought propagation.Main results are listed as follows:(1)Identification of the best among of four remotely sensed LAI products.Four sets of global LAI products including GLASS,MODIS,GLOBALBNU,GLOBMAP were evaluated.Generally,four products show similar spatial and temporal distribution patterns.Referring to in situ observations,GLASS LAI has the highest accuracy,with R2 of 0.70 and RMSE of 0.96.(2)Development and validation of a LAI-based vegetation drought index(LAIpct).By applying SAD drought analysis method to GLASS LAI products,we developed a new vegetation drought index,LAIpct.The reliability of LAIpct was validated by comparing to a well-recognized vegetation drought index DSI.LAIpct shows good agreement with DSI in representing both drought area and severity,while LAIpct outperforms DSI by representing drought evolution processes more stably and smoothly.(3)Vegetation drought detection and evaluation.There were 1066 drought events detected by LAIpct in the period of 2001-2015.Most of them are flash droughts with duration less than 160 days.There were 49 drought events lasting more than one year globally,and only one drought lasting more than four years occurred in Australia.Vegetation drought occurred most frequently in the central of North America.(4)Correlation and time lag between SPI(Standardized Precipitation Index)and LAIpct.LAIpct and SPI show a significant positive correlation in the most part of the land(>99%).The lag time between SPI and LAIpct varies for different land covers,while forests have the longest lag time around four to eight months,3.75 months for shrubs,2.79 months for savannas,2.28 months for grasses,and 1.95 months for croplands.It reflects that different land covers(ecosystems)have different tolerance to the meteorological drought.For croplands,rainfed farmland has a time lag of 1.91 months,less than that of irrigated farmland(2.37 months).It means that anthropogenic disturbance could alter the drought resistance of croplands.(5)Features of drought propagation.LAIpct was compared with SPI,Palmer Drought Severity Index(PDSI),and Standardized Soil Moisture Index(SSMI),for six selected typical drought events.The propagation from meteorological drought to soil moisture drought and vegetation drought was clear identified.Results also show that for dense vegetation,the time lags are much bigger than those of sparse vegetation,which reflects the adaptability to extreme events for different vegetation types.This study demonstrated that LAIpct is an effective and stable index for monitoring vegetation drought globally and that the propagation of drought can be monitored by synthetic application of SPI(PDSI),SSMI,and LAIpct.