MICHAEL MATSON NOAA National Earth Satellite Service Washington, DC 20233 JEFF DOZIER Department of Geography University of California Santa Barbara, CA 93106 Identification of Subresolution High Temperature Sources Using a Thermal IR Sensor Steel mills, and gas flares from oil fields, were identified by using the 3.8-pm and 11-pm sensors on board the NOAA-6 satellite. INTRODUCTION T HE THIRD-GENERATION, polar orbiting, TIROS-N series of environmental satellites provides two thermal infrared (IR) channels for twice-daily monitoring of the Earth's surface (see Table 1). Part of the four-channel Advanced Very High Res- olution Radiometer (AVHRR), the 3.8 pm thermal IR channel, was designed to supplement the 11-pm thermal IR channel in the computation of sea sur- face temperature by providing corrections for at- mospheric water vapor and cloud contamination. The best spatial resolution for the two thermal studies, initial examination of the 3.8-pm data over land areas revealed differences between them and the 11-pm data. Specifically, as shown by the computer-enhanced NOAA-6 satellite image of the midwestem United States in Figure 1, the 3.8-pm data accentuate high temperature areas (Table 2). These areas are not as apparent in the 11-pm data, as is readily revealed by a brightness temperature plot across a high temperature area (Figure 1, number 5) in Cleveland, Ohio (see Figure 2). At this high radiation source the 3.8-pm brightness temperature is 34.5 K higher than the 11-pm bright- ABSTRACT: Simultaneous use of the 3.8-pm and 11-pm thermal infrared chan- nels on board the 3rd-generation NOAA-6 environmental satellite provides the capability to detect subresolution scale high-temperature sources, and to esti- mate both the temperature and size of such sources. Examples resented include gas flares from oil fields in the Middle East and steel mills in the midwestern United States. channels is 1.1 km at the satellite subpoint, and the minimum noise-equivalent differential tem- perature (NEAT) is 0.12 K for 300-K scene (Schwalb, 1979). Calibration is provided on board for con- verting the satellite-measured thermal energy emitted by the Earth and its atmosphere into brightness temperatures. Unless these tempera- tures are corrected for land-surface emissivity, as well as for absorption and re-emission in the inter- vening atmosphere, they cannot be compared strictly with Earth surface thermodynamic tem- peratures. Although designed for sea surface temperature PHOTOCRAMMETRIC ENGINEERING AND REMOTE SENSING, Vol. 47, No. 9, September 1981, pp. 1311-1318. ness temperature. A typical brightness tempera- ture difference between the two channels, 1 to 1.5 K, is shown by the rest of the plot and is caused primarily by less water vapor attenuation in the 3.8-pm channel (Weinreb and Hill, 1980). The purposes of this paper are to (1) present the radiance-temperature relations that explain the different responses of the two thermal-IRchannels to a common target; (2) demonstrate how the dif- ferent responses of the two channels can be used to determine both the brightness temperature of the source causing the two-channel difference and the area of the pixel (picture element) covered by 0099-1 112/81/4709-1311$02.25/0 @ 1981 American Society of Photogrammetry