The fundamental problem of correlation between a catastrophic structural ch
ange and the noise parameters of the system was studied. Computer simulatio
ns showed the spectral parameters of noise to be correlated with instabilit
y of the Ising system. The earliest signals of the forthcoming phase transi
tion were analyzed. The Monte Carlo method was applied to simulate the stoc
hastic spin kinetics of the two-dimensional (50 x 50) Ising model in the ab
sence of an external magnetic field at various ratios between temperature a
nd the exchange integral. Transitions of two types were observed, namely, d
emagnetization when the system approached the critical temperature and a fl
uctuation transition between two metastable states of the system with oppos
itely directed total magnetic moments. The time series of total magnetizati
on were analyzed, and the power spectra, structure functions, and distribut
ion functions for various temperatures were obtained. The spectral characte
ristics were shown to be more sensitive to changes in the phase state of sy
stems and to give earlier indications of forthcoming phase changes than ave
rage statistical parameters. The parameters obtained from an analysis of po
wer spectra In S(f) and structure functions phi ((2))(tau) (time of memory
loss and spectral power at low frequencies) were close to each other.