An improved genetic algorithm for determining modified water-retention model for biochar-amended soil

摘要

Biochar has been globally recognized as a soil amendment to ameliorate the degraded soil structures. We investigated the different biochar percentages contributed to the changes in soil water retention, soil infiltration, and water-holding capacity of one dimensional scale. Besides, infiltration models were compared, and an improved genetic algorithm (GA) combined with multi-objective optimization and elitist strategy was proposed to upgrade the modified van-Genuchten (VG) model. Results indicated that observed cumulative infiltration displayed reductions by 14.06%, 46.62%, and 71.78% for the soil mixed with 5%, 10%, and 15% biochar, respectively, relative to the pure soil. The Kostiakov model was more effective than the Philip model in predicting cumulative infiltration. Furthermore, the constructed modified VG model based on the inversed hydraulic parameters was capable of predicting soil moisture at suction less than 2070 kPa (i.e., 1.38 times wilting point) but caused an underestimation beyond it. This research has the potential to replace the soil water retention curve (SWRC) measurement by one-dimensional infiltration experiment with parameters inversed from the improved GA combined with a modified VG model. It is time-saving and efficient during the SWRC study.