There exists a need for simple and reliable characterization methods for biomass in several scientific areas. Not many publications report on multivariate statistical treatment of thermogravimetric data, and therefore the objectives of this study were to i) evaluate the potential for using a multivariate statistical approach for modeling degree of decomposition of thermally pretreated wood using data from conventional thermogravimetric analysis, ii) compare the predictions from the multivariate chemometric model with a gaussian curve fit...