Abstract: In a field plot experiment, the canopy spectral reflectance at different developmental stages of four spring wheat varieties (Dingxi24,Longchun8139, Gaoyuan602 and Dingxi38) under different planting densities was measured, and the leaf area index (LAI) corresponding to the spectra was determined. In order to estimate the LAI and to establish the best prediction model, nine spectral vegetation indices in common use were calculated, and their correlations with and predictabilities for LAI were analyzed and estimated. The results showed that the nine spectral vegetation indices had good correlations with LAI, and could better predict the LAI at four different growth stages, with the best at heading and flowering stages. The correlations between vegetation indices and LAI were more significant and the predictions were better in the whole growth period than in any growth stages. Among the vegetation indices, R₈₁₀/R₅₆₀ produced the best prediction of LAI. Therefore, by selecting vegetation index R₈₁₀/R₅₆₀(x) as independent variable, the prediction model about spring wheat LAI (y),y=0.1769x^1.5261, was established, and evaluated and tested by the experiment data of different varieties and different planting densities. The average precision (R²) and the RMSE and accuracy of the estimation were 0.9280, 0.0762 and 0.9068, respectively, indicating that this prediction model had preferable reliability and wide applicability.

Key words: CO₂ concentration, Changbai Mountains, Broad-leaved Korean pine forest, CO₂ exchange