A framework for determining the total salt content of soil profiles using time-series Sentinel-2 images and a random forest-temporal convolution network

• Published in: Geoderma Vol. 409; p. 115656
• Main Authors: Wang, Nan, Peng, Jie, Xue, Jie, Zhang, Xianglin, Huang, Jingyi, Biswas, Asim, He, Yong, Shi, Zhou
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.03.2022
• Subjects: China; electrical conductivity; irrigation; precision agriculture; protocols; Random Forest; salinity; salt content; satellites; Soil profile; soil profiles; soil quality; Soil salinity; soil salinization; sustainable land management; Temporal Convolution Network; temporal variation; time series analysis; Time-series images; topsoil; vegetation; China; Temporal Convolution Network; Time-series images; Soil profile; Soil salinity; Random Forest

[Display omitted] •A novel framework was developed to monitor total salt content in the soil profile to a depth of 1 m.•Field experiment and times-series remote sensing data were used to learn spatio-temporal variability.•A total of 42 parameters were evaluated the correlation with soil salinity.•Ra...