749 
Symposium on Remote Sensing for Resources Development and Environmental Management / Enschede / August 1986 
Sea surface temperature studies in Norwegian coastal areas 
using AVHRR- and TM thermal infrared data 
J.P. Pedersen 
University of Tromse, Norway 
ABSTRACT: This work presents an algorithm for deriving sea surface temperatures from infrared satellite data. 
The algorithm is based upon physical solution of the equation of radiative transfer. The theory for calcu 
lating the atmospheric transmittance and -radiance is briefly discussed. Calculated transmittances are com 
pared to values reported by others. Results from applications of the algorithm on NOAA/AVHRR-data are pre 
sented. At last, sea surface temperature data derived from the Landsat/Thematic Mapper are presented. Due to 
lack of TM calibration data, the temperatures are derived from comparisons of digital values and in-situ 
measured temperatures at knownlocations in the data set. 
1 INTRODUCTION 
The thermal infrared data from the NOAA-series of 
satellites have been applied by different Norwegian 
institutes for studying sea surface temperatures and 
currents for many years. In recent years there has 
been a growing interest in the development of the 
natural coastal-zone resources in Norway. Today a lot 
of sea-farms are in operation all over the country, 
and a lot more are planned for the future. In selec 
ting the most suitable location for a sea farm, it is 
often necessary to know about the annual variations 
of the currents and the temperatures in the actual 
area. One way of collecting this information is by 
using infrared data from satellites. 
The NOAA-series of polar orbiter, sunsynchronous 
satellites, from which data are read out at Troms0 
Telemetry Station, offers the opportunity to study 
surface phenomena in the Arctic regions with a high 
frequency of repetivity. The thermal infrared data 
from the NOAA-satellites are frequently used by dif 
ferent Norwegian institutes for studying currents 
and the sea surface temperatures (SST). The spatial 
resolution of the NOAA/AVHRR-dta of 1 km limits the 
applications to open ocean areas. 
However, the new generation of satellites, represen 
ted by the Landsat/TM offers the opportunity to study 
surface phenomena at an increased spatial resolution. 
In studying currents and SST's, the 120 meter reso 
lution of the thermal TM-channel is more adapable for 
coastal-zone applications, as compared to the NOAA/ 
AVHRR. 
2 SATELLITE INSTRUMENTATION 
The primary surface observing sensor onboard the NOAA 
series of satellites is the AVHRR (Advanced Very High 
Resolution Radiometer). The AVHRR is a five channel 
radiometer observing at visible, near-infrared and 
thermal infrared wavelengths. The observing channels 
are 0.58-0.68 urn, 0.7-1.1 urn, 3.55-3.93 um, 10.3- 
11.3 um, and 11.5-12.5 um. The AVHRR scans the earth 
surface within +/- 55.4 degrees from nadir, represen 
ting a surface swath width of approximately 2500 km. 
The spatial resolution of the AVHRR channels is lxl 
km (at nadir). The satellite altitude is approximate 
ly 830 km, the orbital period approximately 102 minu 
tes, which represents 14.1 orbits pr. day (Schwalb, 
1978). Technical data for the NOAA-satellites are 
listed in table 1. 
The TM onboard the Landsat satellites represents 
the next generation of earth observing sensors. TM is 
a seven channel radiometer observing at visible, near- 
infrared, and thermal infrared. The observing wave 
lengths are 0.45-0.52 um, 0.53-0.60 um, 0.63-0.69 um, 
0.76-0.90 um, 1.55-1.75 um, 2.08-2.35 um, and 10.4- 
12.5 um. The TM scan angle is +/- 7.7 deg.,represen 
ting a swath width of 185 km at earth surface. The 
spatial resolution is 30x30 meter, except for the 
thermal channel which has a resolution of 120x120 
meter. Satellite altitude is app. 705 km, the orbi 
tal period 99 minutes. The TM repeat cycle is 16 days. 
As for the NOAA satellite, the Landsat is a punsyn- 
chronous, polar orbiter (NASA, 1984). Technical data 
for TM are listed in table 1. 
For the purpose of this application, the thermal 
channel of theAVHRR and the TM have to be considered. 
The AVHRR channel 4 (eventually channel 5) and the TM 
channel 6 cover the same part of the electromagnetic 
spectrum, although the spectral width of the TM chan 
nel is twice the width of the AVHRR channel. Dtle to 
the spectral coincidence of the TM and the AVHRR 
channels, they are applicable for comparable studies 
of SST's. The predicted absolute accuracies are 0.5 K 
and < 0.12 K for TM and AVHRR respectively. 
The most significant difference between the two 
channels is thespatial resolution, 120 meter for TM 
compared to 1 km for AVHRR. The spatial resolution 
limits the AVHRR applications primarily to medium 
scale phenomena studies in open oceans. For studies of 
small scale phenomena often observed in the coastal- 
zone areas and in the fjords, the TM thermal channel 
is the most suitable one. 
Table 1. Technical data of NOAA/AVHRR and Landsat/TM. 
AVHRR 
TM 
Spectral 
chi 0.58-0.68 
chi 0.45-0.52 
bands (um) 
ch2 0.7 - 1.1 
ch3 3.55-3.93 
ch4 10.3-11.3 
ch5 11.5-12.5 
ch2 0.53-0.60 
ch3 0.63-0.69 
ch4 0.76-0.90 
ch5 1.55-1.75 
ch7 2.08-2.35 
ch6 10.4-12.5 
Spatial 
app. 1 x 1 km 
30 x 30 m 
resolution 
(at nadir) 
ch6 120 x 120 m 
Swath width 
+/- 55.4 deg. 
app. 2500 km 
+/- 7.7 deg. 
185 km 
Orbit 
sunsynchronous 
, polar 
Data 
availab. 
From Troms0 
Telemetry Station 
From Earthnet/ 
Kiruna