Evaluating the Changes in Soil Physicochemical Properties after Subjecting to Biochar Application using Rule-Based Machine Learning
摘要
The conversion of biomass to biochar is a promising and sustainable approach for addressing waste management issues, combating climate change, and reducing pollution, aligning with several United Nations Sustainable Development Goals (SDGs). Biochar not only sequesters carbon dioxide from the atmosphere but also improves soil chemical and physical properties, significantly impacting crop productivity. However, the reported effects of biochar on soil properties vary. Predicting these effects remains a critical gap that needs to be addressed in biochar research. Different biochar properties and application rates have led to varied responses when applied to different soil textures, ranging from negative to positive or even neutral changes in soil properties. These variations highlight the need to determine the optimal set of parameters to achieve the most desirable outcome. Five IF-THEN rules for soil pH, soil Bulk Density (BD), and soil Porosity responses, and four IF-THEN rules for soil Organic Carbon (OC) and soil Cation Exchange Capacity (CEC) responses were accepted. Their validity was further assessed using quantitative measures (using classification and specific performance metrics and 10-fold cross-validation) and qualitative assessments (using related biochar literature) to check for mechanistic plausibility. The findings of this study show that the condition attributes (feedstock type, application rate, pyrolysis temperature, and soil type) significantly influence the soil physicochemical properties. Hence, this research can guide the agricultural sector in selecting the appropriate biochar parameters to optimize soil quality, ultimately enhancing crop productivity and promoting a sustainable future for all by primarily supporting UN SDGs 2, 12, and 13.