The nature of
Gorshkov A., Konnikova V.,
After the discovery of variability of
extragalactic radio sources it became clear that
variations on time scales longer than a year are
due to instabilities in the associated active
galactic nuclei [I.S. Shklovsky, H. Van der Laan].
Variations on shorter time scales can be either
intrinsic [S.J. Qian et al.] or due to scattering in
the turbulent interstellar medium [B.J. Rickett
et al.]. Currently, the best agreement with the
observed characteristics of the long-term
variations is achieved for shock models [for
instance, Marscher and W. K.Gear.].
Active variability studies are under way at many
observatories around the world. Many-year series
of observations at decimeter wavelengths have
been obtained for large samples of radio sources
[M. Bondi et al.]. The longest series of centimeter-
wavelength observations have been carried out at
the University of Michigan [H. D. Aller et al.]. Since
1997 the RATAN-600 radio telescope has been
used to do simultaneous observations of a sample
of ~600 sources with VLBI fluxes at 13 cm
exceeding 100 mJy [Yu.A. Kovalev et al.].
Compact radio sources have been monitored at
millimeter wavelengths for more than 20 years at
the Metsähovi Radio Observatory [E. Valtaoja, M.
Tornikoski et al.].
We have also been conducting simultaneous
observations at 0.97–21.7 GHz with the RATAN for
more than 20 years (1984–2006) . Our monitoring
differs from the data of Kovalev’s in two aspects.
First, we observe samples that are complete to a
limiting flux density; second, from 1998 the
duration of our continuous daily observations of our
samples was 60–100 days. This has enabled us to
study variations on time scales of several days
along with the longer-term variations.
As a result of more than forty-year
studies of variability became clear that
the majority of extragalactic radio
sources demonstrate variability on time
scales from tens years up to tens
variations on time scales longer than a
year (long term variability) are due to
instabilities in the associated active
variations on scales of month can be
either intrinsic or due to the scattering in
the turbulent interstellar medium;
variations on daily scales (short term
variability) most likely has the external
In this report I will talk about all
temporary scales of the variability.
Here one can see the selection
criteria of our samples:
1. S 200 mJy at 3.9 GHz (Zelenchuk Survey)
RA = 00h 24h
Declination = +04 06
|b| > 15
N = 138 objects
2. S 200 mJy at 4.85 GHz (MGB Survey)
RA = 00h 24h
Declination = +10 12.5
|b| > 15
N = 154 objects
Results in this report about long term
variability are mostly from the first sample
(Decl = +04 06)
The sub-sample of sources with the flat spectra
consists of 68 objects.
Optical identifications were done for 56 of
41 – QSO;
5 – Bl Lac;
5 – Galaxies of different types.
For all sources spectra in a range of 0.97-21.7
GHz are measured.
According to their spectra we derived all
sources into 4 classes:
1) single-component sources with no extended
component and only one compact component (in most
cases these are the GPS or HPS sources, with
maxima at the GHz frequencies);
2) two-component sources with an extended component
and a single compact component;
3) Sources with a dominant extended component and a
weak compact component (these are actually also two
component sources, but the compact component’s
influence is manifest only in the flattening of the
spectrum at high frequencies);
4) sources with complex spectra, which can have an
extended component and two or more compact
components observed simultaneously.
We divided all obtained in 20 years light curves
into 3 types, independent of object’s
classifications in the optical domain:
1)Slowly variable sources;
2)Sources with the isolated flares;
3)And sources with the chaotic variability.
Practically all single-component sources and
sources with the prevailing extended component
are the slowly variable objects.
The sources, which have 2 and 3 type of
variability, as a rule, have two-component or
Here are some examples of light
And their spectra:
СПЕКТР ИСТОЧНИКА J0509+1011 В МАКСИМУМЕ
05/2007 И МИНИМУМЕ 10/2007
Relation between the variability indices of individual
flares Vf and the variability indices V of the same
sources for the whole period of our observations.
Vf = −0.018 + 0.86V
In its majority the nature of the variability of
sources with the flare activity more or less
corresponds to model in which variability is the
result of the evolution of the shock which is
propagating along the jet.
According to this model in its development
shock passes the stages of amplification, balance
(when the energy losses and gains are comparable)
and damping. The shape of the shock spectrum
corresponds to that for a single source and
remains the same in the course of the shock
evolution. As the shock develops, the peak
frequency moves towards lower frequencies.
These are precisely the sort of spectral
variations we observe in most cases. When the
peak frequency of a flare is above the studied
range the spectrum can be approximated well
with a parabola. Sometimes the peak frequency
is high enough that the spectrum can be
described with a power-law function in our
frequency range. As the flare develops the
peak frequency moves to lower frequencies
finally entering our studied range (1-22 GHz).
The observed maximum variation amplitude
normalized so that ΔSmax(νmax) = 1
0.1 1 n/nmax
Now I’ll talk a little bit about QSOs properties of
our sample: 41 sources (from 68).
4 have single-component spectra with
Zmean = 1.7 ± 0.3);
13 two component spectra (Zmean = 1.6 ± 0.2);
3 spectra with a dominant extended component
(Zmean = 2.0 ± 0.4);
and 21 complex spectra (Zmean = 1.6 ± 0.2).
The deviations of the mean redshifts of the QSOs
with different spectral classes from the value for
all the quasars in our sample (Zmean = 1.64 ± 0.12) are
not statistically significant.
We were able to determine timescales for
flares detected in 26 quasars. The apparent
linear sizes lie within 0.18–2.3 pc, while their
true sizes lie within 0.06–1.0 pc (E. Valtaoja).
The flare Tb values exceed the Compton
limit in 14 quasars. With one exception the
linear sizes of the emitting regions in these
objects are within 0.29 pc.
Among objects with smaller angular sizes
none has Tb below the Compton limit.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 R,пк
The variability indices for the objects with
Tb > 1012 K are higher than those for the objects
with Tb < 1012 K.
Relations between the variability indices
and Z (Tb < 1012 K):
0 0.5 1 1.5 2 2.5 3 z
But if we do correction to rest frame there is
no any dependence
Some conclusions from the long-term
1) For 40 sources it was possible to determine
the visible characteristic time of flare which
allowed for the objects with known red shifts
to determine the visible angular dimensions and
2) In 14 sources Tb > 1012 K. For these sources
the average value of the Doppler-factor is
3) In all QSOs with the linear sizes less than
0.29 pc Tb exceeds Compton limit.
The reason for so sharp boundary is not clear.
4) There exists a small number of radio
sources when the shape of spectrum
remains constant with variations in the flux
density. We suppose that for multi-
component sources such variability is due to
changes in the angle between the jet and
the line of sight. In the case of flat-
spectrum sources (a 0) we probably
observe a single inhomogeneous compact
The characteristics of the compact components of
QSO’s with brightness temperatures within the
Compton limit are independent of redshift. This
follows from the lack of Z-dependence for the
the linear sizes of the emitting regions;
the both variability indices (V and Vf );
the peak frequencies of compact components.
We suppose that these can be interpreted as
evidence for a lack of cosmological evolution of
QSOs at least up to Z = 3.
(Astronomy Reports, 2008, Vol. 52, No. 4, pp. 278–298)
we named as
1998 at the
cyclic with the
time of 12-16
Since 1998 we observed 112 sources with the
flat spectra of two different samples. “Weekly”
variability is detected in 38 objects. The main
Variability usually is manifested at two or
three frequencies both at low and at high
Characteristic times are from 6 to 40 days;
In the same source characteristic time can
change with the frequency and in the different
epochs of observations;
In the strongest sources the variation is
correlated at the adjacent frequencies.
Light curve: 2123+0535
The autocorrelation and cross-correlation functions
There is correlation with long-term
variability: most frequently the “weekly”
variability is observed near the maximum
of the long-term variability.
The spectrum of variable component
corresponds to the spectrum of uniform
0527+0331 1998 0527+0331 1999
But in general spectrum of the “weekly” variable
component not as simple as in 0527+0331.
For example - 0409+1217: BL Lac, Z=1.020
The most likely that the falling spectra of
“weekly” variations is the strongest
manifestations of scintillations (“flicker”)
discovered by Heeschen in 1984 (variations of
flux density of the flat spectra radio sources at
cm-wavelengths on the scales of several days).
The variability index is just several
percentages but they are observed practically
in all sources with the flat spectra. Variations
have average characteristic time ~4 days.
For the definition
of temporal scale
D = f (t ) f (t )]
- temporal lag
In this case we suggest that we observe
“flicker” as well as intrinsic variability
2123+0535 1999 1000
2123+0535 2004 2123+0535 2006
1 10 1 10
0527+0331: observations at 32 m dish of
the Zelenchukskaya Observatory of
Institute of Applied Astronomy; = 3.5 cm.
05 jan 2005
15 16 17 18 19 20 21 22 23 0 1
1819+3845: observations at 32 m dish of
the Zelenchukskaya Observatory of
Institute of Applied Astronomy;
February 5, 2005; = 3.5 cm.
0 0.5 1 1.5 2 2.5 3 3.5
Thank you for attention!