The SBG Level 3 TIR Surface Minerology (SM) Product

The proposed SBG thermal infrared (TIR) Level-3 (L3) Surface Mineralogy (SM) product relies upon the L2 TIR surface emissivity data as input together with a spectral library of common Earth surface minerals. These spectra are acquired in emission in the laboratory at 2 cm-1 and resampled to the SBG six-band resolution. The SM algorithm uses the principal of linear spectral mixing in TIR region where the larger absorption coefficients typical of most rock-forming minerals limit photon transmission and scattering within the mineral grains. The emitted sample/pixel spectrum, therefore, has features in linear proportion to the areal abundance of those minerals in that sample/pixel. SM will be applied to a limited subset of the TIR land surface data using a seasonally adjusted global emissivity mask. Only those data with an average emissivity of < 0.95 (~ 30% of the Earth’s land surface) will be mapped using the SM algorithm. This threshold excludes surfaces with significant vegetation, snow/ice, and water cover. The emission spectra from all pixels that meet this threshold test is modeled using the pre-determined spectral library as input  resulting in a best-fit suite of mineral endmember images plus their corresponding wavelength-specific residual error images. The blackbody endmember image accounts for the spectral depth of the library endmembers; however, the values are normalized out of each pixel’s mineral endmember total. In addition, a root-mean-squared (RMS) error image (to assess the overall goodness-of-fit) and a weight percent silica (WPS) image are produced.