The goal of this study was to characterize the changes in biomass with torrefaction for near infrared reflectance (NIR) and attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectroscopy for sweetgum, loblolly pine, and switchgrass. Calibration models were built for the prediction of proximate analysis after torrefaction. Two dimensional (2D) correlation spectroscopy between NIR and FTIR was found to precisely explain the depolymerization at key functional groups located within hemicellulose, cellulose, and lignin. This novel 2D technique also demonstrated the possibility of assigning key NIR wavenumbers based on mid IR spectra. Hemicellulose based wavenumbers were found to be most sensitive to torrefaction severity with complete degradation at 250–275°C. Lignin associated wavenumbers exhibited the least degradation to severity but was still detected with 2D correlation spectroscopy. Finally, calibration models for proximate analysis were performed and while both systems could be used for rapid monitoring, NIR performed better than FTIR.