Background • Natural methane seeps found in the Santa Barbara Channel ▫ Coil Oil Point • Mostly dissolves in the ocean or broken down by microbial activity • A small fraction is able to escape to atmosphere Atmospheric Methane • Sources of methane total roughly 540 Tg yr-1 ▫ Geological sources are often ignored • Terrestrial and marine seepage can contribute 35-45 Tg yr-1 (about 7%) • CH4 is one of the 4 major atmospheric absorbers ▫ measurable absorption features in the VIS–NIR spectrum • Detecting methane is important for understanding global budget Objectives • Useful to develop a method of CH4 detection using remote sensing ▫ high resolution ▫ can compare with field measurements • Remote sensing strengthens greenhouse gas studies by allowing measurement at multiple scales • Goal: to detect a methane gas “plume” with MASTER spectral imagery ▫ Mapping methods ▫ Methane index Measurement • Remote data collected using MODIS/ASTER airborne simulator (MASTER)
• 16 air samples were collected throughout SBC
• CH4 over the open ocean - 1.8 ppm • CH4 over a marine seep - 36.6 ppm • High signal seen at 34°24’7N, 119°52’59W
• Radiance and irradiance spectra also collected in
the field using the HydroRad 3 spectrometer Field data as endmembers • Spectra collected for 4 different kelp beds, open ocean, grass, pool, wet sand, oil slick and methane seep • Endmember – a “pure” spectrum of a material or target area with unique spectral signature • Data was used to develop a spectral library
• Resampling - Used to match the response of an
instrument with wavelengths of an image input file Detecting Methane: Mapping • Spectral Angle Mapper (SAM) - uses angles to match pixels to the field reference spectra Detecting Methane: Mapping • Spectral Information Divergence (SID) - uses divergence measure to match pixels to spectra ▫ Smaller divergence means greater similarity between pixels Methane from MODTRAN • Computer simulations run to model methane radiance at sensors. • CH4 held at 1.8 ppm and then at 36 ppm for the first kilometer • Spectral range of 705-33000 cm-1 (0.3 to 14 µm)
• Radiance values from MODTRAN were used to
create reference spectra for high and low atmospheric methane Detecting Methane: Mapping • SID map created with MODTRAN methane endmembers included Detecting Methane: Mapping • Matched Filtering (MF) – provides the abundances per pixel of each defined endmember ▫ Linear mixing model Detecting Methane: Developing an Index • Difference between two simulated methane models Regions of Interest • Two ROIs selected from the MASTER image: ▫ Area where air samples were taken ▫ Region without methane seep • Radiance and reflectance statistics extracted for each region Methane Index • Methane index developed taking spectral features into account • high feature – bands 43 to 47 • low feature – bands 12, 13, 23 to 25, 27
• Radiance to radiance comparison
• Radiance to reflectance comparison Conclusion – Epic Fail • Noise ▫ Instrument noise ▫ Atmospheric noise ▫ std dev of .45 W/m/sr/µm • Higher resolution images over area needed • MASTER band 26 not working ▫ Overlaps with an absorption region of 3.31 µm • Interference from chemical reactions occurring • Weak signal References • Barnhouse Jr, WD. Methane plume detection using passive hyper- spectral remote sensing: A Thesis. Graduate College of Bowling Green State University. • Cavanaugh, KC; Siegel, DA; Kinlan, BP; Reed, DC. Scaling giant kelp field measurements to regional scales using satellite observations. Marine Ecology Progress Series 403 (2010) 13-27. • Dickey, T., M. Lewis, and G. Chang (2006), Optical oceanography: Recent advances and future directions using global remote sensing and in situ observations, Rev. Geophys., 44. • Leifer, I; Roberts, D; Margolis, J; Kinnaman, F. In situ sensing of methane emissions from natural marine hydrocarbon seeps: A potential remote sensing technology. Earth and Planetary Science Letters 245 (2006) 509–522. Questions?