REFLECTIVE SPECTROSCOPY OF GASOLINES FOR THE DEVELOPMENT OF A HYPERSPECTRAL LIBRARY AND DEVELOPMENT OF AGE ESTIMATION TECHNIQUES FOR SPILL INVESTIGATION

Hyperspectral remote sensing (HRS) has high potential for identifying petroleum product spills in the environment with a high degree of accuracy. Identifying these product spills quickly and in a cost effective way is an important first step towards addressing remediation in a rapid manner as HRS can define the area of a spill and monitor the increase and changes in size and shape over time. In order to utilize HRS a detailed library must be built containing reflective spectra of relevant materials. Reflectance spectra of two grades of octane (87 and 93) and three brands of gasoline (Certified, Meijer, and Sunoco) as well as a sample of diesel and kerosene were investigated on examples of two sand substrates. Combinations of 5 ml fuel and 50 g of substrate were measured over a period of 2 weeks of evaporation under a low flow hood using an ASD spectroradiometer at conditions of 75 (+/- 5) % relative humidity and 72 (+/-3) degrees Fahrenheit. Our findings indicate that there was no major variation among the gasolines, including 87 and 93 octane ratings. Furthermore gasoline could be identified on the substrates, supporting previous work. Absorption features at ~1701 nm for gasoline and ~1725 nm for diesel are the most pronounced, and these were used to create age estimation curves. Age estimation time windows are very different for gasoline, being only 3 hours, whereas for heavier fuels, at least 1 to 2 weeks must pass before features differentiate. Substrate seems to play an important role in establishing the aging curve. The finer the grain size, the more the fuel evaporates. The detection limit for gasolines under these lab conditions was 30% of the original gasoline mass. These results support progress in developing comprehensive HRS fuel spill detection protocols.