Abstract:  Leaf area index（LAI） is an important parameter for describing structural characteristics of plant canopy and also an essential parameter for researching the exchange information of mass and energy associated with canopy surface. In this study，we use the VIs（Vegetation Indexes）-LAI method to inverse LAI value of the riparian forests of the lower reaches of the Tarim River. The modeling data is comprised of LAI and VIs，which were acquired by field measurements by LAI-2250 and indoor processing work of Landsat 8 OLI remote sensing image respectively. Three optimal models were built separately for the whole samples，the Tamarix forests samples and the Populus euphratica Oliv forests samples，and made precision tests by comparing with the measured LAI values which were not used in building the models previously. Then the LAI distribution map was acquired by using the whole samples VIs-LAI inversion model. The results showed as follows：（1）In the three different samples analyses，all vegetation indexes have good relevance with LAI，and among different vegetation indexes，quadratic polynomial regression model has the highest goodness of fit．（2）In the whole samples analysis which does not distinguish vegetation types，ARVI（atmospherically resistant vegetation index）has the best correlation with measured LAI，y=10.714x²+2.324x+0.157，R²=0.632. In the Tamarix forests samples analysis，NDVI（normalized difference vegetation index）has the best correlation with measured LAI，y=-0.216x²+5.744x-0.356，R²=0.689，and in the Populus euphratica Oliv forests samples analysis，ARVI has the best correlation with measured LAI，y=8.119x²+2.036x+0.095，R²=0.816. （3） All the three models have relatively high overall precision which could meet the demand of LAI inversion in this arid region. And the fitting precision of Tamarix forests samples and the Populus euphratica Oliv forests samples have improved by 4.7% and 0.2% respectively compared with the whole samples fitting precision. （4） There are three relatively high value areas in the distribution map of LAI，the Daxihaizi reservoir area，the middle part of the reach and the Taitema Lake area，due to relative high water content for vegetation. And the LAI values were mainly in a range of 0-1.5 with a mean value of 0.361. In this study，we estimated the LAI distribution of the riparian vegetation of the lower Tarim River. The results provided data support for studying riparian plant and desert ecology of the Tarim River，and also provided method support for acquiring the high spatial resolution [LAI] data in the riparian forests in the lower Tarim River.

Key words: leaf area index, vegetation index, the lower Tarim River, remote sensing inversion, Landsat 8 OLI image