Environmental Decision-Making - New Ecoinformatics Tools in Environmental Science: Applications and Decision-making - page 244

s output depends on its
input, in some region of the input parameter space. In other words, sequential
screening allows the focusing only on the important parameters what fasters the
study process that doing the analysis over all the parameters. For example, the
problem of earlier detection of tropical cyclone beginning is solved taking into
account numerous meteorological and geophysical characteristics of the ocean/
atmosphere system delivered by the GOOS including TAO/TRITON/PIRATA/
RAMA system (Krapivin and Soldatov 2009b; Sukov et al. 2008a). It is evident that
restricted number of between these parameters is important.
The sequential decision making procedure is used to estimate the parameters of
the land surface based on the remote data received by the

Sequential method helps to understand how the model '

flying laboratory Il-18
(Fig. 2.4 ) which was used for a long time in Former Soviet Union and which was
unique remote-sensing system. The measurements were obtained by radiometers of
0.4, 0.8, 2.25, 11, 21, and 27 cm wavelength. The sets of radiobrightness tem-
peratures {T j } are registered and T Tj,min, , T j,max and T j,mean are calculated. As a result,
they are transformed to the normalized sets {T j /T j,max }or{

fl

T j,mean }.
Figure 3.10 gives examples of such transformations. The registered radiobrightness
temperatures are typical for the land covers of the eastern Caspian Sea coast. This is
a region of saline lands, dry sands, saline waters, dense bushes, and populated
landscapes. As it is seen in Fig. 3.11 , multi-channel measurements provide the data
sets, making it possible to distinguish the land surface cover. Moreover, an appli-
cation of the sequential analysis procedure guarantees the solution to this task in
real-time on-board the

ʔ T j = T j -

flying laboratory.
The radiobrightness temperature sets, made discrete by the quasi-uniform sub-
sets, are transformed to combinations of distribution characteristics which make
possible the formation of the multi-dimensional image state of the land surface. An
example of the cluster analysis application is given in Fig. 3.12 . Real brightness

fl

Fig. 3.10 Sample of the
radiobrightness temperature
registration on board the
airborne laboratory IL-18
fl flying east of the village

Kyzyk on the eastern shore of
the Caspian Sea