Fig. 6: The similarity and properties prediction of biochar.
From: The performance, pyrolysis mechanism and environmental functions of forest surface fuel biochar
a Three-dimensional principal component analysis (3d-PCA) is used to characterize similarities and dissimilarities in biochar. b The contribution of pyrolysis temperature and biomass feedstock composition (major elements, cellulose, hemicellulose, and lignin content) to explain the differences in every attribute of biochar based on correlation and optimal multiple regression models. We investigated the correlation of all attributes with these values for each biochar sample and identified the main predictors. The size of the circle represents the importance of the variable (i.e., the proportion of explained variability calculated from multiple regression models and variance decomposition analyses, and the color represents the Spearman correlation. Random forest (RF) algorithm was used to rank the importance of the attributes that determine HHV (c), EC (d), CEC (e), and Zn (Ⅱ) (f), As (V) (g), Cr (Ⅵ) (h), Cd (Ⅱ) (i), and Pb (Ⅱ) (j) adsorbed respectively; percentage increment (mean square error, %IncMSE) were used to rank attributes (* for p < 0.05, ** for p < 0.01, and *** for p < 0.001). APD average pore diameter, SA specific surface area, C&E Carboxyl & ester, PAr C–H polysubstituted aromatic C–H, MAr C–H monosubstituted aromatic C–H, Phhy phhydroxyl, Carb carboxyl, Lact lactonyl.
