10 Remote Sensing of Surface Water
233
peratures in degrees Kelvin and are proportional to the product of the surface temperature and the emissivity of the medium. Using the differences between vertically
and horizontally polarised brightness temperatures (~T) minimizes the effects of
atmospheric interference. For 37 Ghz data from the Nimbus-7 satellite scanning multichannel microwave radiometer (SMMR), heavily vegetated land shows the lowest ~T
« 4 K) grading up to about 30 K as vegetation cover decreases. Calm surface water
is easily distinguishable with a ~ T of about 60 K.
Sippel et al. (1994) used daytime 37 Ghz SMMR brightness temperatures at 6-day
intervals for 1979-1985 to estimate seasonal changes in inundation area for a 34,500
Ian 2 area along the Amazon River near Manaus. The resolution ofSMMR at the 37
Ghz frequency is about 25 Ian and Sippel et al. (1994) used a linear mixing model
incorporating the 4 K and 60 K endpoints to determine inundation fractions of mixed
pixels. Hamilton et al. (1996) extended the fractional analysis of SMMR data to study
seasonal and inter-annual inundation patterns in the Pantanal, an area which varies
between 11 and 110 thousand square kilometres located mainly in Brazil. By developing a regression between flooded area and river stage Hamilton et al. (1996) were
able to reconstruct a 95-year flood record using recorded historic river levels
It is also possible to estimate sediment concentration and total sediment transport
during major flood events. Mertes (1994) used Landsat TM data to measure sediment
concentration during the 1977 and 1989 flood events on the Amazon River. Field data
were combined with depth averaged velocities from a 2D hydrodynamic model and
the Landsat distribution of suspended sediment concentration to give vertical deposition rates and mass balances of flood deposits.
10.8 Conclusion
Studies have suggested that the use of remotely sensed data in hydrology and water
resources could yield benefit/cost ratios in the order of 100: I from savings in flood
damage and from improved planning. However, the current use of remotely sensed
data in operational monitoring of surface waters is limited. This is partly because
visible data from satellite are limited to relatively cloud-free and daylight conditions.
Even for relatively simple tasks such as estimating the areal extent of water bodies
such as large lakes, clouds obscure the analysis. Microwave satellites offer the
potential of (almost) all-weather application and the synergistic application of combined electro-optical remote sensing for surface water estimates needs to be examined
more closely.
Efforts should also be made by the data suppliers to educate potential customers in
the advantages of remotely sensed data. Certainly, the real strength of satellite remote
sensing applications is in alleviating some of the hydrometric data collection and
management problems facing many of the developing nations. Developed nations with
large expanses of low populated or remote areas can also benefit greatly from some
of the observational techniques presented. In both cases, the lack of appropriate
education and training of operational agencies and consultants in remote sensing
techniques limits the thinking to traditional techniques.
233
peratures in degrees Kelvin and are proportional to the product of the surface temperature and the emissivity of the medium. Using the differences between vertically
and horizontally polarised brightness temperatures (~T) minimizes the effects of
atmospheric interference. For 37 Ghz data from the Nimbus-7 satellite scanning multichannel microwave radiometer (SMMR), heavily vegetated land shows the lowest ~T
« 4 K) grading up to about 30 K as vegetation cover decreases. Calm surface water
is easily distinguishable with a ~ T of about 60 K.
Sippel et al. (1994) used daytime 37 Ghz SMMR brightness temperatures at 6-day
intervals for 1979-1985 to estimate seasonal changes in inundation area for a 34,500
Ian 2 area along the Amazon River near Manaus. The resolution ofSMMR at the 37
Ghz frequency is about 25 Ian and Sippel et al. (1994) used a linear mixing model
incorporating the 4 K and 60 K endpoints to determine inundation fractions of mixed
pixels. Hamilton et al. (1996) extended the fractional analysis of SMMR data to study
seasonal and inter-annual inundation patterns in the Pantanal, an area which varies
between 11 and 110 thousand square kilometres located mainly in Brazil. By developing a regression between flooded area and river stage Hamilton et al. (1996) were
able to reconstruct a 95-year flood record using recorded historic river levels
It is also possible to estimate sediment concentration and total sediment transport
during major flood events. Mertes (1994) used Landsat TM data to measure sediment
concentration during the 1977 and 1989 flood events on the Amazon River. Field data
were combined with depth averaged velocities from a 2D hydrodynamic model and
the Landsat distribution of suspended sediment concentration to give vertical deposition rates and mass balances of flood deposits.
10.8 Conclusion
Studies have suggested that the use of remotely sensed data in hydrology and water
resources could yield benefit/cost ratios in the order of 100: I from savings in flood
damage and from improved planning. However, the current use of remotely sensed
data in operational monitoring of surface waters is limited. This is partly because
visible data from satellite are limited to relatively cloud-free and daylight conditions.
Even for relatively simple tasks such as estimating the areal extent of water bodies
such as large lakes, clouds obscure the analysis. Microwave satellites offer the
potential of (almost) all-weather application and the synergistic application of combined electro-optical remote sensing for surface water estimates needs to be examined
more closely.
Efforts should also be made by the data suppliers to educate potential customers in
the advantages of remotely sensed data. Certainly, the real strength of satellite remote
sensing applications is in alleviating some of the hydrometric data collection and
management problems facing many of the developing nations. Developed nations with
large expanses of low populated or remote areas can also benefit greatly from some
of the observational techniques presented. In both cases, the lack of appropriate
education and training of operational agencies and consultants in remote sensing
techniques limits the thinking to traditional techniques.