image: A typical salt-affected land view more
Credit: Image credit: Nima Shokri et al.
Researchers report using machine learning to develop a long term analysis of 40 years of topsoil salinity changes across the globe. Soil salinity varies across space and time due to natural events such as droughts or human activities like fertilization and irrigation. Excess accumulation of salts can adversely affect plants and microbial processes. Amirhossein Hassani, Adisa Azapagic, and Nima Shokri sought to understand these variations at a global level with a combination of machine learning techniques and climatic, topographic, soil, and remote sensing data. They used approximately 240,000 measurements of soil electrical conductivity and exchangeable sodium collected from 1980 to 2018 to develop predictive models of soil salinity. Their analysis showed that a soil area of 11.73 million square kilometers was salt-affected between 1980 and 2018, nearly 20% greater than the area of the United States. Globally, the likelihood of reoccurrence of salt-affected soils in the 2000-2018 period decreased compared to the 1981 to 1999 period. Brazil, Peru, and Sudan had the highest rate of annual increase in soil salinity. The results can help determine the likelihood of salt-affected soils in a particular area and inform effective remediation strategies in the face of climate uncertainty, according to the authors.
###
Article #20-13771: "Predicting long-term dynamics of soil salinity and sodicity on a global scale," by Amirhossein Hassani, Adisa Azapagic, and Nima Shokri.
MEDIA CONTACT: Nima Shokri, Hamburg University of Technology, GERMANY; tel: +49 40 42878 2870; e-mail: <nima.shokri@tuhh.de>
Journal
Proceedings of the National Academy of Sciences