Abstract: Leaf area index (LAI) of a semiarid Artemisia ordosica community at Yanchi, Ningxia was measured with both direct (allometric equations) and indirect (Li-2000 canopy analyzer) methods throughout the growing season of 2012. Our objectives were to compare the direct and indirect measurements, to explore how actual leaf area index (LAI_a) values obtained from allometric equations can be used to calibrate effective leaf area index (LAI_e) measured by Li-2000, and to model seasonal LAI_a variations using climatic variables. Results showed that bivariate functions of height and canopy area generally predicted individual total leaf area better than did the functions of canopy area alone, and the fitting achieved highest coefficient of determination (R²) in July and August. Both LAI_e and LAI_a showed humpshaped seasonal patterns. A midseason stable period started from early July and ended in midSeptember. LAI_e was always greater than LAI_a over the growing season, with the mean values of LAI_e and LAI_a in the stable period being 1.15 and 0.31 m²·m^-2, respectively. LAI_e and LAI_a showed a twophase linear relationship over the growing season, separated by late August. LAI_e and LAI_a showed a weak relationship with rainfall or soil water content  in the growing season. LAI_a varied with accumulated temperature above 5 °C (∑D), first increasing with ∑D, and then being stable during the midseason. LAI_a showed a negative exponential relationship with the ratio of photosynthetically active radiation reflectance (ρ_p) to shortwave radiation reflectance (ρ_s) (ρ_p/ρ_s). Our results contribute to the understanding of LAI patterns in semiarid A. ordosica communities, could aid in the development of efficient and accurate methods for measuring LAI, and provide ground-based data for calibrating remotely-sensed information and for parameterizing processbased ecosystem carbon cycle models.

Key words: improved TOPSIS method, early warning, Zhangjiajie, grey GM (1,1) model., tourism ecological security