Investigation of greenhouse gas emissions in soil remediation activities based on the integration of machine learning models and life cycle assessment

Abstract

Addressing soil contamination and achieving carbon neutrality necessitate a comprehensive evaluation of greenhouse gas (GHG) emissions from soil remediation activities. However, the ecological values of remediated soil have long been overlooked in environmental impact assessment of soil remediation. The present study integrated machine learning (ML) models and life cycle assessment (LCA) to quantify the secondary GHG emissions resulting from remediation activities and the tertiary GHG emissions associated with carbon sequestration capacity of remediated soil. The integrated method was applied to evaluate the life cycle GHG emissions of ex-situ thermal desorption (ESTD) and ex-situ chemical oxidation (ESCO). The offsite ESTD scenario render the highest secondary GHG emissions at 83 kg CO₂-eq/m³ soil, while the onsite ESCO scenario resulted in the lowest secondary GHG emissions at 25 kg CO₂-eq/m³ soil. After considering tertiary impacts, the total GHG emissions decreased by 38 to 40 kg CO₂-eq/m³ soil, demonstrating the importance of considering soil carbon sequestration. Furthermore, the evaluation of tertiary impacts was conducted on national scale to compare the carbon sequestration capacity of remediated soil under different environmental conditions. ESTD remediated soil exhibited higher carbon sequestration capacity than ESCO remediated soil in the middle and lower reaches of the Yangtze River and northeast China, whereas ESCO remediated soil exhibited higher carbon sequestration capacity in southwest China. Overall, the present study provides a comprehensive evaluation of GHG emissions in soil remediation activities and new insights for decision making in low-carbon soil remediation.