311 
Figure 3. Correlation of reflectance spectra with clay percentage 
a. Selection of reflectance spectra 
b. Spectral correlation with clay percentage 
Figure 4. Correlation of derivative spectra with clay percentage 
a. First order derivatives of the treflectance spectra from Fig. 3 
b. Spectral correlation with clay percentage 
The precise mature of the througs and peaks in the sediment reflectance spectra shouls be further investigated. 
Ofcourse the scattering and absorption properties of water are reflected in the spectra. Furthermore there is 
some evidence of the presence of algae, according to the apparent absorption maximum around 680 nm. 
Modelling experiments and maybe new measurements could clarify this issue. 
After geometric correction of the cast data (pixel size 4x4 m) the area of the sediment test site was 
mozaiced from three adjacent tracks. After that a clay map was made on the basis of casi channels 7 and 9 
(neighbouring channels were used to estimate derivative reflectance values). Although not yet verified, this 
map seems to correspond with the view of field experts. 
3.6 Conclusions and plans for the fiiture 
The provisional result of the analysis of the data is that remote sensed optical data can be used to 
discriminate between sediment types. 
The analysis of the available data will be continued. E.g. multivariate regression, principle component 
analysis and spectral mixing techniques will be applied. 
There were too many predominantly sandy samples. The sediment grain size data set will be extended 
with samples with a more muddy character to see whether more definite conclusions can be reached. 
By virtue of its flexibility w.r.t. definition of bands and flight altitude a system like casi is very suitable 
for regular mapping of sediment types in the test area. 
The pixel size of 4x4 m as sued in this experiment is very suited for mapping of this spatially very 
dynamic area.