Intemational Archives of Photogrammetry and Remote Sensing. Voi. XXXII Part 7C2, UNISPACE III, Vienna, 1999 
86 
I5PR5 
UNISPACE III - ISPRS/EARSeL Workshop on 
“Remote Sensing for the Detection, Monitoring 
and Mitigation of Natural Disasters” 
2:30-5:30 pm, 22 July 1999, VIC Room B 
V ienna, Austria 
EARSeL 
serious weather conditions. 
Satellite 
Satellites overcome some of the deficiencies of aircraft. Experi 
ence shows that after a successful launch the life time of the sys 
tem and instruments is very long, to be counted in years. 
A number of features will be considered: 
(a) The operation is weather independent although some types 
of data may be dependent upon light and cloud conditions. 
(b) Satellites operate in fixed oibits so that observations may 
be repeated regularly for very long periods. This is important in 
connection with measurement of very slow surface movements, 
characteristic of pre-landslide situations. 
(c) Being in a fixed orbit with a certain repeat cycle the sensors 
may only cover an area of disaster after a certain period of days 
dependent upon the swath of the instrument in question. With the 
ERS SAR in a 35-day repeat cycle and a swath of 100 km, for 
example, an area at mid-latitudes may be observed at intervals of 
about 10 days when exploiting ascending as well as descending 
passes. Increasing the swath to 400 km as will be the case with 
ENVISAT the revisit time is shortened to two to three days al 
though with data at a coarser resolution. However, by moving the 
beam of observation over the swath the frequency of observation 
with fine-resolution data of 30 m. for instance, may be increased. 
In fact, ENVISAT ASAR may - like RADARSAT - give full 
resolution data in a swath of 100 km by this means at interv als of 
two to three days at mid-latitudes. The SPOT satellites have a 
similar possibility of changing the view angle up to 27° to both 
sides of the track. In addition to reducing the revisit time this 
features is exploited for compiling digital elevation models 
(DEM). 
(d) The ‘beam swinging’ feature requires programming of the 
satellite. Due to the orbit configuration with an orbit period of 
about 100 minutes this will invariably introduce a time delay in 
addition to the delay that may occur in the process of requests for 
acquisition. 
(e) The transfer of data from the satellite to ground is dependent 
upon dedicated receiving stations that record in real time tlie 
signals from the satellite on high-density tapes or in solid-state 
memories. Processing of the data introduces another delay. How 
ever, with present techniques these actions introduce only short 
delays of few hours. 
(f) Scenes of disaster occurring outside the coverage of the 
ground stations may only be available if recorded onboard the 
satellite and transferred to a ground station on a later orbit. This 
will cause a delay of minimum 1.5 hour but often more depend 
ing on the latitude of the station. Earth observation satellites are 
now equipped with tape recorders with sufficient capacity to 
fulfil this function and the acquisition systems are designed so 
that in most cases processed data may be transferred to the user 
within less than 24 hours. 
Thus, in order to be able to deliver timely data the satellite sys 
tem has to be designed for minimum delay in programming the 
acquisitions and subsequent data processing and dissemination. 
It might be crucial in systems with long repeat-cycles and narrow 
swaths because acquisition might otherwise be missed. 
A real-time interaction with the satellite may be desirable. Ide 
ally, the users should work directly with tire satellite, requesting 
acquisition of a specified area and receiving the processed data - 
in real time. Such interactive function may be feasible in the near 
future when space-qualified large capacity processors become 
available. Data may be of moderate spatial resolution, however, 
thus favouring the use of small (even mobile) receiving stations 
close to the end users (Marelli et al. 1997). 
VISUAL AND INFRARED SYSTEMS 
Application of sensor data in the visual and infrared bands have 
reached a state of maturity so that they are used routinely in 
many disciplines exploiting the multichannel features. Image 
analysis systems are now off-the-shelf products that enable very- 
advanced analyses. I shall therefore not dwell too much on this 
techniques except referring to few' features. 
The fine spatial resolution - 5 to 30 meter - that is achieved with 
present-day sensors in the visible bands and their multi-channel 
feature are great assets in many applications. The future will 
show commercial satellites with even finer resolution - down to 1 
meter is claimed - which might find applications in connection 
with disasters. The limited swath of these systems may be coun 
ter productive, however, for disaster management. 
Current narrow-bandwidth sensors with resolutions of 900 - 
1000 m are in orbit such as CZCS and SeaWifs, and in the future 
MERIS on ENVISAT. 
At infrared wavelengths resolutions of 80-200 m is achieved in 
some cases and 900 - 1100 m in others (NOAA AVHRR). 
It is trivial to mention that application of visual and infrared data 
is dependent on clouds, which might be a serious limitation in 
connection with floods, for instance. 
However, these types of data may be the only proposition for 
monitoring land degradation. In this context tire cloud depend 
ence is of little importance except in few areas of the Earth with 
an almost permanent cloud cover. Similarly, desertification stud-