Comparative role of charcoal, biochar, hydrochar and modified biochar on bioavailability of heavy metal(loid)s and machine learning regression analysis in alkaline polluted soil

Abstract

Pot experiment was performed aimed to assess the comparative role of charcoal, biochar, hydrochar and thiourea-vegetable modified biochar at 1 and 2 % doses, and <1 mm particle size on the bioavailability of Cd, Pb, As, Ni, Cu and Zn, and enhance NPK, and mustard growth in a slightly alkaline polluted soil. Furthermore, machine learning method was used to examine the systematic evaluation of the impact of feature selection based on Pearson's correlation on the performance of the linear regression model. The results revealed that maximum fresh and dry biomass of mustard was observed by 26.38 and 38.18 % with hydrochar 1 %, whereas lemon biochar at 2 % reduced fresh and dry biomass up to 34.0 and 53.0 % than control. The immobilization of Cd and Pb was observed by 83.70 and 71.15 % with thiourea-vegetable modified biochar at 2 %, As 71.62 % with hydrochar 2 %, Ni 80.84 % with thiourea-vegetable modified biochar 2 %, Cu 66.32 % with and Zn 36.30 % with thiourea-vegetable modified biochar at 2 % than control. However, the maximum mobilization of Cu in soil was observed by 30.3 % with lemon biochar 2 %, similarly for Zn 37.36 % with hydrochar 2 % as compared with other treatments. The phyto-availability of Cd, Pb, As and Cu in the mustard shoot and root biomass was reduced except Ni and Zn in soil than control. It was observed that using the machine learning regression analysis approach, variability in treatments effectiveness is evident across different feature correlation thresholds. This study clearly shows that the beneficial role of studied amendments on mustard growth and reduced bioavailability of heavy metal(loid)s and enhance primary macronutrients in alkaline polluted soil. It is suggested that future studies may be conducted on combined application of studies amendments on plant growth, immobilization of heavy metal(loid)s in multi-metal polluted soil under different field conditions.

Keywords: Heavy metal(loid)s; Immobilizing material; Machine learning; Mustard growth; Remediation.