Crude oil spectral signatures and empirical models to derive API gravity

•A spectral library with 37 types of crude oils (light to heavy) was produced.•Oil samples were spectrally characterized in the 350–15,000 nm interval.•Crude with different °APIs show particular infrared absorption features and metrics.•Multivariate predictive models based on oil spectra proved valid in estimating °API.•Oil spectra can be used to map crudes exposed at surface using remote sensing. Infrared spectroscopy is presented as a quick alternative method for establishing crude oil properties such as density (°API gravity) and content of saturates, aromatics, resins and asphaltenes (SARA). Here, thirty seven (37) samples of crude oils with different °API gravities and SARA fractions were measured in the visible, near infrared and short wave infrared (VIS – NIR – SWIR/350–2500 nm), and in the middle and longwave infrared (MWIR – LWIR/3000–15,000 nm) ranges. Main CH absorption bands and optimal intervals for the discrimination among oil types were identified in all spectral regions. Based on diffuse reflectance spectroscopy and equivalent wavelet spectra, Principal Component Analysis allowed distinguishing between three major groups of crude oils with varied densities. Partial Least Squares Regression analysis yielded sixteen predictive models tailored to estimate the °API of crude oils based on their spectral signatures in the NIR – SWIR – MWIR – LWIR intervals. These data and methods are swift and nondestructive ways to estimate the °API of crude oils, with latent applications in several chains of the petroleum industry, particularly to assess and manage the environmental impact caused by oil discharge, leakage and spill events. Additionally, considering that airborne and orbital sensor systems currently in operation can sense some of the crude oil absorption bands highlighted in this study, the oil signatures revealed here could be potentially used to map oil contamination in diverse settings and globally using remote sensing.