Effect of takyr solonetzs on vegetation canopy spectral characteristics and growth prediction

Abstract: In order to quantify the effect of takyr solonetzs on canopy spectral of above vegetation (sunflower) via unifying soil background, the characteristics of canopy spectral, the relationship of spectral reflectance with SPAD value and leaf area index (LAI) were analyzed during five typical growth stages. Then, models for SPAD value and LAI prediction were established based on canopy spectral reflectance of sunflower and soil background pH value. The results showed that the tendency of SPAD value was similar to LAI of sunflower. There were increased then decreased with the advance of growth stages; they were the maximum value at budding stage and anthesis stage, respectively. Canopy spectral reflectance of sunflower in strongly alkaline soil were higher in the visible light and lower in near infrared region than reflectance of sunflower in slightly and moderately soil. Canopy spectral reflectance of sunflower at 3-leaves stage decreased by 0.10 at the wavelength of 400-520 nm and 1027-1130 nm when the soil was covered with black cloth compared to its natural conditions; meanwhile, the reflectance decreased by 0.18 at 521-1026 nm. The reflectance at other growth stages decreased in visible-light range and increased in near-infrared when the soil was covered in contrast with its natural conditions (the canopy reflectance increased by 0.059 in 720-1050 nm and its amplification was 11.29% at 7-leaves stage). The spectral reflectance of sunflower at the natural state showed a positive correlation with the SPAD value and LAI at 3-leaves stage, and the average correlation coefficients between the reflectance and SPAD value, LAI were 0.68 and 0.47, respectively. Anthesis and filling stages were same as 3-leaves stage. The average correlation coefficients between the spectral reflectance at other stages and two agronomy parameters in visible range were -0.27 and -0.52 in visible range, and the average correlation coefficients between them were 0.53 and 0.69 in near-infrared, respectively, the correlation coefficients between them were maximum values in near-infrared at budding stage. When soil was covered, the correlation coefficient was comparatively higher than its natural condition. The minimum and maximum values of determination coefficients of the models for SPAD value and LAI were predicted based on canopy spectral appeared at 3-leaves and budding stages, respectively. The effect of polynomial equations of SPAD value and LAI predicted were better than other equations at large, however, exponent and power equations of LAI predicted were the best at 3-leaves, 7-leaves and anthesis stage. The models had higher precision and capacity of predicting at all growth stage (RSPAD were 0.61-0.88，RLAI were 0.72-0.92) except 3-leaves stage. Therefore, models involved with pH value could effectively reduce the interference of soil background on the prediction of SPAD value and LAI of sunflower.