The novel mid-wave infrared (MWIR) retrieval method described in this paper is based on an idea derived from
retrieving accurate sea surface temperatures for the Multispectral Thermal Imager (MTI) which uses the "Same
Temperature Criterion", i.e., that retrieved skin temperatures in all bands must be the same. We apply it to simulated
MODIS datasets. In this paper we present retrieval errors for this method for a range of atmospheric conditions and skin
surface variations. The STC retrieval method is applied to actual, de-striped MODIS data, and we show that it is capable
of retrieving subtle high spatial resolution (1 km) water vapor variations over ocean surfaces. The variance of the
retrieved surface temperature is a useful criterion for finding sub-pixel area hotspots (e.g. forest fires, lava flows, gas
flares, furnaces, boilers, etc) and distinguish them from warm water bodies at night.