Geoscience Reference
In-Depth Information
plot(t,y)
title('Synthetic signal of lognormal distributed noise')
subplot(2,1,2)
plot(t,log(mwreal))
title('Results from Mann-Whitney U-test')
h e result from the Mann-Whitney test reveals that the length of the window
inl uences the amplitudes and widths of the maxima of the test parameter,
whereas the location of the transition in the means is well dei ned (Fig.
5.18 b). We next use an Ansari-Bradley test for the same three dif erent
lengths of paired sliding windows (150, 250 and 500 data points) to detect any
abrupt change in the standard deviation. We use the function
ansaribradley
introduced in Section 3.12 to perform the Ansari-Bradley test.
for j = 1:length(w)
df1 = w(j) - 1;
df2 = w(j) - 1;
for i = w(j)/2+1:length(y)-w(j)/2
[h,p] = ansaribradley(y(i-w(j)/2:i-1),y(i+1:i+w(j)/2));
abreal(j,i) = p;
end
abreal(j,1:w(j)/2) = abreal(j,w(j)/2+1) * ones(1,w(j)/2);
abreal(j,length(y)-w(j)/2+1:length(y)) = ...
abreal(j,length(y)-w(j)/2) * ones(1,w(j)/2);
end
We then display the results.
subplot(2,1,1)
plot(t,y)
title('Synthetic signal of lognormal distributed noise')
subplot(2,1,2)
plot(t,log(abreal))
title('Results from Ansari-Bradley test')
h e length of the window again clearly inl uences the amplitudes and widths
of the maxima of the test parameters, and the location of the transition in
the dispersion is again well dei ned (Fig. 5.18 c). h is method has been
successfully applied to records of terrigenous dust l ux preserved in marine
sediments of shore subtropical West Africa, the eastern Mediterranean Sea,
and the Arabian Sea, in order to detect trends, rhythms and events in the
African Plio-Pleistocene climate (Trauth et al. 2009).
5.10 Nonlinear Time-Series Analysis (by N. Marwan)
h e methods described in the previous sections detect linear relationships
in the data. However, natural processes on the earth ot en show a more
complex and chaotic behavior, and methods based on linear techniques may
