An examination of enhanced atmospheric methane detection methods for predicting performance of a novel multiband uncooled radiometer imager

Abstract

To evaluate the potential for a new uncooled infrared radiometer imager to detect enhanced atmospheric levels of methane, three different analysis methods were examined. A single-pixel brightness temperature to noise-equivalent delta temperature (NEdT) comparison study performed using data simulated from MODTRAN6 revealed that a single thermal band centered on the 7.68 µm methane feature leads to a detectable brightness temperature difference exceeding the sensor noise level for a plume of about 17 ppm at ambient atmospheric temperature compared to an ambient plume with no enhanced methane present. Application of a normalized differential methane index method, a novel approach for methane detection, demonstrated how a simple two-band method can be utilized to detect a plume of methane that is 10 ppm above ambient atmospheric concentration and −10 K from ambient atmospheric temperature with an 80 % hit rate and 17 % false alarm rate. This method was capable of detecting methane with similar levels of success as the third method, a proven multichannel method, matched filter. The matched-filter approach was performed with six spectral channels. Results from these examinations suggest that given a high enough concentration and temperature contrast, a multispectral system with a single band allocated to a methane absorption feature can detect enhanced levels of methane.