中文摘要：

利用内蒙古羊草草原（Leymus chinensis）生态系统通量观测站的气象数据、野外实测和MODIS叶面积指数（Leaf area index，LAI），应用基于生态系统过程的VIP（Vegetation interface process）模型，以半小时为步长，模拟分析了羊草草原生态系统2003～2005年（分别为平水年、平水年和干旱年）蒸散及其分量的变化过程。通过与通量数据对比，VIP模型能够很好地模拟羊草草原生态系统的蒸散过程限（R^2=0．80），在峰值大小和变化趋势上，模拟值与实测值有较好的一致性。模拟结果显示：3年蒸散量分别为337、338和223mm；在降水相对充沛的2003～2004年，蒸腾量为192和171mm，而降水相对较少的2005年，蒸腾量仅为96mm；年平均蒸腾和蒸发对蒸散的贡献基本持平；生长季蒸散占全年的83％，6月开始，蒸腾大于蒸发，蒸散和蒸腾的月总值均在7、8月达到最大值，两月蒸散占全年的43％。LAI是影响蒸散的主要因素，其次是降水，而净辐射对蒸散的影响较小。在生长季，蒸发的季节变化平缓，蒸散的差异主要体现在蒸腾的差异。

英文摘要：

Aims Evapotranspiration （ET） plays an important role in arid and semiarid temperate grassland where water availability is a major limiting factor for ecosystem functions. Understanding temporal variation of ET can help explain the surface-atmosphere interaction and its ecological function in grassland ecosystems. Partitioning total ET into its components of evaporation from soil （E） and transpiration from plants （T） is important for understanding the biotic and abiotic factors that control water balance. Our objectives were to simulate the seasonal and interannual variations of ET and its components, analyze the contribution of the components to ET and analyze influencing factors. Methods We used flux data derived from eddy covariance technology over Inner Mongolia steppe （43°32′ N, 116°40′E）, measured LAI and MODIS data from 2003 to 2005 and parameterized VIP （Vegetation interface processes） model to simulate ET of the grassland. The results were validated using half-hourly latent heat fluxes （LE） and net radiation （Rn） estimated from eddy covariance measurements. Important findings VIP model can effectively simulate latent heat fluxes of the grassland （R^2=0.80）. In 2003 and 2004, precipitation （P） was near average and annual ET was 337 and 338 mm, respectively, which were greater than P. In the drier year of 2005, annual ET was 223 mm, which was higher than P. On average, E and T made relatively equivalent contributions to ET. About 83% of annual ET occurred during the growing season. E was the primary component of ET before June and was exceeded by T after that. The monthly totals of both ET and T reached maxima in July and August. Total ET during July and August accounted for 43% of the annual amount. ET was strongly correlated with LAI and moderately correlated with P. E changed little during the growing season, and the difference in ET was accounted for T.