LI Jing, WANG Ke-Ru, LI Shao-Kun, CHEN Bing, XIAO Chun-Hua, WANG Qiong, WANG Fang-Yong, LI Shuan-Ming, LAI Jun-Chen, SUN Ya-Ling, ZHANG Guo-Qiang
We conducted remote sensing dynamic monitoring of soil salinization by soil salt-spectrum inversion in cotton fields. An ASD spectrometer was used to measure the cotton soil spectra in the Communist Youth League farm of the sixth division of the Xinjiang Production and Construction Corps. The most sensitive wave band for measuring the soil salinization degree was determined by combined analysis of spectral and soil chemistry parameters, and the best salt index for monitoring soil salt content in cotton fields was determined. The results showed that soil spectral reflectance increased(0.084-1.659 g·kg-1) with increasing salinization degree, and the difference was more obvious in the near infrared region(1350-1850 nm). This band range of spectral reflectance was the most sensitive wavelength for identifying salinized soil and was significantly correlated with soil salt (r = 0.880**). By selecting sensitive bands, salt index models including SI1, BI, SI2, NDSI and SI3 were constructed to monitor soil salinization. Among the models, the SI1 and BI salt indexes were the most accurate, with a minimum RMSE of 0.151 and 0.149, and RE of 7.5% and 6.3%, respectively. We recommend these as the best models to monitor soil salinity in cotton fields.