Symposium on Remote Sensing for Resources Development and Environmental Management / Enschede / August 1986
755
Satellite data in aquatic area research: Some ideas
for future studies
Jouko T.Raitala
Jet Propulsion Laboratory, Caltech, Pasadena, Calif., USA (on leave from Dept, of Astronomy, University of Oulu, Finland)
ABSTRACT: Computer-aided digital remote sensing techniques were used to evaluate the usefulness of Landsat MSS
data in aquatic area studies. These investigations unravelled some of the Landsat data potentials in monitoring
factors critical to limnology, aquatic botany, geomorphology and engineering: 1) Besides depth (in relatively
shallow Finnish lakes) the MSS data may also include useful information about Secchi disc values, humus content
in water (colour, iron) and productivity (nutrients, chlorophyll). 2) Aquatic vegetation classification is
possible only where vegetation units are big enough in respect to the 0.5 hectares ground resolution. Different
life-forms (helophytes, nympheids, elodeids, bryophytes etc.) are mapped quite easily, but when using the super
vised classification procedure even some minor nuances can be traced. 3) Multitemporal satellite imagery has
been used to evaluate alterations within the littoral areas of some Finnish water reservoirs between successive
periods of high water and also along the shallow coastal sea of the Gulf of Bothnia.
The limitations of the MSS data in exact small-scale interpretations are, however, so apparent that the most
critical support of relevant water chemical and field data must not be underestimated. The use of MSS data with
poor ground resolution and with too broad and too few spectral channels represents only a small range of the
possibilities which will be gained by using the more advanced TM and SPOT data. The MSS data may nevertheless
be important when trying to trace the changes in aquatic and other environments during the last decade.
1. INTRODUCTION
Different lakes and coastal sea areas are amongst
the few most essential environments in Finland
forming a natural surrounding for almost all activi
ties of men. This means that aquatic areas are the
object of various and often conflicting efforts of
utilization. Research, investigation and planning
of water areas for recreational and residential use
and as a source of livelihood and raw water or as a
dumping place for waste water all require accurate
information, some of which can in principle be
obtained and prepared by using Landsat or other
satellite data and the computer-aided analysis of
such data.
The Landsat MSS data includes information of
aquatic areas down to the Secchi disc depth. The
green and yellow wavelengths of channel 4 indicate
variations within a few uppermost metres in brownish
Finnish lakes and water features even down to tens
of metres within certain open ocean areas. The red
radiation of channel 5 usually comes from within the
uppermost one metre water layer, and the near-in
frared channels 6 and 7 include information from
within the uppermost centimetres or millimetres, re
spectively. In addition the uppermost part of the
water column is best represented within all recorded
channels.
The present study describes some attempts to apply
aquatic remote sensing to the preparation of
parametric map-like presentations, quantitative
evaluations and time-related investigations within
various water areas in Finland.
2. WATER QUALITY
Laborious and expensive efforts are called for when
attempts are made to inspect water quality and its
changes over large Finnish lake areas by means of
field work. Satellite images and their computer-
aided analysis may help to locate the most critical
areas where changes in water quality have taken
place. Continuous satellite monitoring of water
areas -- in terms of data availability -- can
provide information for example of movements of dif
ferent water bodies, changes in production rate,
overgrowth, pollution and purification.
However, more basic studies and progress are still
needed before remote sensing will be able to answer
all questions unambiguously. The wavelength
channels of the Landsat MSS sensor are too wide and
too few to allow distinct parametric map-like
presentations of only one of the aquatic variables
to be performed. Indications of different phenomena
are intermingled, resulting in difficulties and mis
interpretations in monitoring. Naturally brownish
Finnish lake waters are easily confused with waste
water pollution and even a sparse nymphaeid vegeta
tion may prevent recognition of water quality. We
must, then, be extremely cautious when trying to
generalize results obtained from one water area to
another.
In applying the remote sensing technique to
studies of aquatic areas the most critical support
of the relevant water chemical data acquired by the
Water District Offices of Finland makes it possible
to find the main differences between various water
areas in advance and to limit studies to within the
same type lakes. Satellite data is thus not used to
find major differences but rather to indicate minor
variations within basically similar water units.
Even certain small nuances and changes in water
quality can be traced and followed once a critical
choice has first been made of what factors can be
monitored and when and where (Lindell 1980, Raitala
et al. 1984c). Channel 7 for example displays very
sensitively changes in the quality of the uppermost
surface water and in vegetation above, on and just
below the water surface. Near-infrared radiation is
almost totally absorbed by even a thin water layer,
while the existence of green vegetation, turbidity,
algae production and surface-reaching shallows in
crease the quantity of reflected radiation.
Statistical correlation between certain water
chemical values and Landsat record derivatives from
within the Kuusamo and Kemijarvi areas in north-