摘要: Visible and near infrared (Vis-NIR) reflectance spectroscopy technology, which is rapid, cost-effective, in-situ and non-destructive, is getting more and more widely used in improving the prediction and digital mapping for soil pr... 展开 Visible and near infrared (Vis-NIR) reflectance spectroscopy technology, which is rapid, cost-effective, in-situ and non-destructive, is getting more and more widely used in improving the prediction and digital mapping for soil properties. Soil available nitrogen (AN) is closely related to soil fertility and quality, assessing its content and mapping the spatial variability greatly satisfies precision agriculture. In this study, the Vis-NIR diffuse reflectance spectra collected by ASD FieldSpec Pro FR spectrometer with a performance of spectral range from 350 nm to 2500 nm, 1 nm resampling intervals, was used to model and characterize the spatial variability of available nitrogen. Firstly, the raw soil spectra was pre-processing by reducing to 400 nm - 2450 nm with transformation into apparent absorbance spectral using Log(l/R) and Savitzky-Golay smoothing. Secondly, spectral indices (normalized spectral index-NDI; difference index-DI; ratio index-RI) were convinced for seeking further relationship between AN. Afterwards, Partial Least Squares Regression (PLSR) method was employed to predict AN. The results indicated good predictions with RPD more than 1.4. Finally, the spatial variability of AN was mapped by (co)kriging method, digital mapping of the measured and predicted AN showed similar patterns and value ranges, though there are some minor differences. The resultant prediction and mapping demonstrated a promotion of assessing and mapping of soil properties by a rapid and reliable approach from lab to field in-situ. 收起