Abstract: Based on field experiments with different rice varieties under differ ent nitrogen application levels, the quantitative relationships of rice leaf are a index (LAI) with canopy hyper-spectral parameters at different growth stages were analyzed. Rice LAI had good relationships with several hyper-spectral vege tation indices, the correlation coefficient being the highest with DI (differenc e index), followed by with RI (ratio index), and NI (normalized index), based on the spectral reflectance or the first derivative spectra. The two best spectral indices for estimating LAI were the difference index DI(854,760) (based on two spectral bands of 850 nm and 760 nm) and the difference index DI (D676,D 778) (based on two first derivative bands of 676 nm and 778 nm). In general, the hyper-spectral vegetation indices based on spectral reflectance performed better than the spectral indices based on the first derivative spectra. The test s with independent dataset suggested that the rice LAI monitoring models with di fference index DI (854,760) as the variable could give an accurate LAI estimatio n, being available for estimation of rice LAI.

Key words: bud bank, ecological adaptability, Loess Plateau, reproductive biology, soil erosion.