Remote Sensing of SST Around the Outfall of a Power Plant from
LANDSAT and NOAA Satellites
Ruo-Shan Tseng
Departemtn of Marine Resources
National Sun Yat-sen University
Kaohsiung, Taiwan 804
Tel: 886-7-5252000 ext 5033 Fax: 886-7-5255033
E-mail: rstseng @mail.nsysu. edu tw
Abstract
Satellite data from Landsat Thematic Mapper (TM) and NOAA Advanced Very High Resolution Radiomether (AVHRR) were used to derive sea surface temperature (SST) in coastal waters of Hsinta that receive cooling effluent from a power plant. Ground truth (GT) temperature measured simultaneously from a ship as Landsat passed were applied in this paper to improve the atmospheric correction process for monitoring SST. Local radiosonde measurements, used in Lowtran7 adjustments for atmospheric effects, produced corrected ocean surface radiances and atmospheric transmittances. Secondly, a scheme combining NOAA-AVHRR and Landsat-TM data was used to derive SST. The advantage of this scheme is that no atmospheric correction process is required. Both methods showed good agreement between the GT and satellite-derived SST for thermal plume of Hsinta power plant.
Introduction
The capability of the TM flown on Landsat 4 and 5 to estimate the SST for the purpose of monitoring thermal plumes produced by power plants has been demonstrated by Gibbons and Wukelic (1989) and Liu and Kuo (1994). Both studies used the thermal band 6) data combining with some sorts of atmospheric processes which require local radiosonde measurements used in LOWTRAN 6 or 7 adjustments. Although the spatial resolution of the thermal band is only 120m, it can provide some insight into the distribution of SST in coastal waters that receive cooling effluent from a power plant. On the other hand, multi-channel sea surface temperature (MSST) techniques have been commonly used for the AVHRR onboard NOAA series satellites for over twenty years, and the root mean square differences of about 0.6°C were found between satellite and in situ data from drifting buoys (Strong and McClain, 1984). In spite of this high accuracy of SST measurements by the AVHRR, its spatial resolution of 1.1 km is too rough for thermal plume analysis in a relatively small area.
An attempt was made in this paper to incorporate both TM and AVHRR data to monitor power plant thermal discharges. The main idea was to make use of the high-resolution advantage or TM data and the high-accuracy advantage of AVHRR data. A simple scheme was then developed to derive the SST without going through the atmospheric correction processes.
Ground Truth and Satellite Data
The study area for this research is at Hsinta power plant which is located in the southwestern Taiwan coast (Fig. 1). The cooling water discharges via an open channel into shallow water of a coastal embayment south of Hsinta harbor that connects to deeper water in the open ocean. In order to validate and compare with the satellite results, one field trip was made to measure the surface-layer water temperature by a ship when the Landsat overpassed this study area at 09:40 local time on 27 June 1996. The temperature of the upper 1m of the sea surface was measured with a conductivity-temperature-depth recorder (CTD), which has an error of ±0.01°C. Ship's loacation was determined by a DGPS, which has an error of within 3 m. The sampling points of SST were centered at the plant outfall and spreaded radially out to a distance of about 2 km. Radiosonde data acquired at the nearby Tungkang readiosonde station were obtained for the overflight data. The vertical measurements of atmospheric pressure,temperature, and dew point, from sea level to 24 km, were used as input to LOWTRAN to correct atmosphere scattering and absorption. The level 10, digital data of TM acquired by the National Central University were obtained for the data of June 27, 1996. Digital AVHRR data of NOAA-12 and NOAA-14 satellites were aquired by our own HRPT receiving station situated at the National sun yat-sen University. A software of MAPIX-OCEAN was used to read the digital data and calculate the SST according to the MCSST equations.
Fig. 1 The site of Hsinta Power Plant
