Remote-Sensing Technologies and Data Processing Algorithms - New Ecoinformatics Tools in Environmental Science: Applications and Decision-making - page 189

From the aforesaid follows that statistical characteristics for “ spottiness ” of

brightness temperatures in microwaves can be used for detection and classi cation

of the phenomena on a surface of the ocean that was caused by a degree of sea
roughness.
Analysis of empirical histograms for spottiness of brightness temperatures in
microwaves shows that in most cases (l + , l − ) are characteristics that can be coor-
dinated with the exponential distribution while their amplitude can be coordinated
with the normal distribution. Therefore, to detect and classify phenomena on ocean
surfaces it is necessary to apply the best possible algorithms when programming/
teaching the computer to make statistical decisions about distributions.

2.10 A Device to Measure Geophysical and Hydrophysical
Parameters

According to algorithms of the multi-channel microwave monitoring discussed by
Krapivin and Shutko (2012), it is possible to promptly measure the wind speed, sea
surface roughness and water temperature as the base characteristics to calculate
other parameters of the environment. From these measurements, a mutual control of
the accuracy of correction of the estimates of the atmosphere-ocean gas exchange is
realized. Solution of the correction equations in a general case requires the use of
the criterion of discrepancy between theoretical and empirical estimates of the
parameters or correlations. Usually for this purpose, a criterion of mean square
deviation is used. The reliability and ef ciency of determination of solutions for the

system of equations relating the theoretical and empirical estimates can be increased
if their dispersive characteristics are taken as a criterion of discrepancy of these
solutions (Nitu et al. 2013a).
Let the proper emission of the sea be measured simultaneously with the help of
n radiometers, each operating at a fixed wavelength ʻ j (j =1, … , n). As a result at the

moment t i at the output of each radiometer the values Z ij (i =1, … , M) are xed, so that

Z ij = T j + ʾ ij , where T j is the brightness temperature of a sea site at the wavelength ʻ j ,

j .
Let the correlation between the brightness temperature and m parameters of the
sea be linear:

and ʾ ij is the random value (noise) with the zero mean and dispersion r

A 11 x 1 þþ A 1m x m ¼ T 1 þ n 1
A n11 x 1 þþ A nm x m ¼ T n þ n n

ð 2 : 11 Þ

where A ij are the coef cients determined in the regime of monitoring the sea site

with the known parameters {x i }. The x i parameter should be estimated for the x i

rst, the average x i

parameters so that, coincided with the accurate solution of the

¼ x i m Þ , and second, a dispersion of the xi i

system ( 2.11 ), that is, ð i ¼ 1 ; ...;