Using the HyMap airborne hyperspectral sensor for offshore seepage mapping

Abstract

The HyMap airborne spectrometer is an airborne remote sensing instrument collecting data in 126 spectral channels from the visible (VIS) to the shortwave infrared (SWIR) wavelength regions. Applications are wide spanned ranging from mineral exploration to environmental monitoring. Offshore seepage mapping has always been a potential application, but not many open file reports can be found in freely accessible literature though oil & gas exploration companies are known to use the technology. As part of a 2003 airborne survey campaign in the USA, HyVista Corporation collected airborne HyMap data over one of the most active natural oil seeps in the world offshore Santa Barbara, California. These seeps are just off Coal Oil Point and spread over an area of 10's of sqkm. UCSB has been studying this seepage for many years and slick distribution and flow rates are monitored regularly. A production platform called `Holly' is also located in the seep area and can be clearly identified on the imagery. Airborne hyperspectral seep mapping offers a new possibility of characterising seeps in an exploration area of interest. It offers the advantage of allowing spectral discrimination of seep components which may not be separable with conventional techniques like radar. HyMap's capability to sense in the SWIR offers the additional possibility of properly separating an effect called `sunglint' from any data over open water, since water will absorb radiation almost 100 % at these wavelengths. Due to the same effect it is also possible to discriminate seepage with a strong surface expression from others, which is not possible with sensors having no SWIR spectral region. The Santa Barbara oils seeps were surveyed with 4 parallel HyMap lines covering the major seepage and migration path of the slicks at the time. Different spectral components in the open water as well as the seep area are identified and mapped. Seamless data products are produced using advanced mapping techniques and utilising the hyperspectral nature of the data. Results of these different proc+F13essing steps will be presented.