Algorithms for Detecting Sub-Pixel Elevated Temperature Features for the NASA Surface Biology and Geology (SBG) Designated Observable

One of the top priorities of the Surface Biology and Geology (SBG) Earth Observing System is the detection and retrieval of elevated temperature features (ETF) usually found in the vicinity of active fires and volcanic activity. We test the ability of currently proposed midwave (MIR: 3-5 mu m) and thermal infrared (TIR: 8-12 mu m) bands to detect ETF within the 400-1200 K range. Specifically, our investigation aims to compare and contrast the use of the 4 and 4.8 mu m MIR bands. We use land surface temperature data obtained by the airborne Hyperspectral Thermal Emission Spectrometer (HyTES) instrument over active fire and lava flows to model at-sensor SBG radiances in the 3-12 mu m range. This is achieved using the Temperature Emissivity Uncertainty Simulator (TEUSim) with the designated/proposed SBG MIR and TIR band characteristics. For ETF detection, we applied the Normalized Thermal Index (NTI) and Enhanced Thermal Index (ETI) to determine a suitable threshold for a wide range of ETF sizes and temperatures. We find that combining an NTI threshold of -0.7 followed by an ETI threshold of 0.02 accurately identifies ETFs at a 97% rate. Sensor noise up to 0.5 K has negligible effects on ETF detection in the 400-1200 K range. The currently proposed SBG MIR and TIR bands are sufficient to detect unsaturated ETFs caused by wildfire and volcanic activities at a similar to 3 day revisit and subpixel ETF area of similar to 9 m(2) (at 500 K) that is unattainable by current satellite TIR instruments. Plain Language Summary The upcoming NASA Earth Observing System focusing on surface biology and geology (SBG) science and applications aims to improve the detection sensitivity of elevated surface temperature (ETF) anomalies such as wildfires and lava flows at high spatial resolution (60 m). In this study, we test whether the proposed SBG midwave and thermal infrared bands can detect ETFs in the 400-1200 K range. We use surface temperature data collected over active wildfires and lava flows from airborne instrument and model the at-sensor SBG radiances in the 3-12 mu m range. We used a combination of two thermal indexes (Normalized and Enhanced) previously developed for other infrared satellite data to determine a suitable threshold for a range of ETF sizes and temperatures. We find that the proposed SBG MIR and TIR bands can detect ETFs with a high accuracy rate of 97%, even at subpixel areas up to 9 m 2 that current satellite TIR instruments cannot achieve. Additionally, sensor noise has negligible effects on ETF detection. These findings could help improve the detection and monitoring of wildfires and volcanic activities and they may also be relevant for other thermal anomaly investigations, including those from anthropogenic sources.