THE COSMOGENIC BERRYLLIUM, SOLAR ACTIVITY AND CLIMATE
/ a study in progress/

Boris Komitov, Pavel Penev and Plamen Nedev

1. Introduction

2. The main aims of this study. Data and methods

The main aims of the prasent work can be summarized in the
following few points:
1.Time series analysis of 10Be production rate ,based on Dye-3
ice probe instrumental records for the epoch 1423- 1985 AD
/fig2/ by using of T-R periodogramm analysis ( Komitov, 1986,
1997, 2001). The procedure started after cleaning the general
trend. The same procedure repeat a few times by means of
consecutively cleaning of the low-frequence / long-periodic/
oscilations and time series analysis of the residuals in each
stage. Thi final series of residuals contain only oscialtions,
which corresponding periods are less than 30 years. In the
last case the possible effects of quasy-11 and 22 years
oscilations must be easy to find and analyzed.

2.Time series analysis of Group Sunspot Numbers (Rg) series
( Hoyt and Shatten, 1998) for the epoch 1610-1995 AD plus
additional sunspot data series as the index of north-south
assymetry. The procedure is similar as for the 10Be.

3.Time series analysis of tree rings width (TRW) data
records from different places of the world, which number
is 85. The searched epoch contain the last ~ 500 years after
1610 AD. The corresponding data are taken from the International
Tree Ring Data Base (ITRDB) as an indicator for the climate
during the corresponding epoch.

4.Searching for correlative relationships between the 10Be, Rg and
TRW, icluding for phase-shifted relationships.


Fig.1 The origin of "cosmogenic" berryllium in the Earth's
atmosphere and its accumulation on the surface.


Fig2.The calendar dated 10Be concentrations in "Dye-3" ice
probe. The same are as an etalon for the corresponding berryllium
production rates in stratosphere. The starting calendar years
of Hoyt-Shatten's (Group sunspot number, Rg) and Zurich (Wolf's
number, Rz) instrumental series are pointed out by arrows and
dashed lines The 10Be maximums, corresponding of Spoerer,
Maunder and Dalton 's supercenturial minimums are signated as
"SpM","MM" and "DM" respectively.

3. Results and analysis

3.1. The 10Be series (1610-1985 AD)

The general trend in 10Be series during the epoch 1423-1985 AD
is slightly parabolic with a maximum near to 1700-1720 AD.It
correspond of the supercenturial behaviour of solar activity in
the same period , which deepest part is the Maunder's minimum
( 1645-1720 AD).
By using of procedure for removing of this trend the
residuals of 10Be variations are searched by T-R periodogramm
analysis. The results are shown on fig.3. The main cycles are
at 65, 108 and 186 years. A weak peak near 52 years is visible
too. After removing of cycles with periods >60 years the last
one stay a dominant. It is interesting to note,that quasy 52
year oscilation are found in index of assymetry of sunspot
arreas during (1874-1967), as well as in the sunspot areas in
the northern hemisphere of the Sun for the same period. This
lied to conclusion , that probably, the quasy 52-year cycle in
10Be data is caused by cyclic anisotropic effects of
GCR-absorbtion in inner parts of Solar system.
After removing of the subcenturial and supercenturial cycles
(T>30 years) the quasy 11 and 22-years oscilations are clear
shown (fig4). However they are weaker as the 14.5 and 24 years
cycles. The last fact,such as the existence of quasy 52
and 65 years cycles point, that in variations of solar
wind parameters the sunspot Schwabe-Wolf and the magnetic
Hale cycles are relatively weak.In opposite, other cycles,
such as 14 and 65 yerars oscilations, which are very weak or
not existing in sunspot activity , are obviously stronger in
solar wind phenomenas.


Fig3. The T-R spectra of 10Be series for the epoch 1610-1985 AD
after removing of the general trend.


Fig4.A T-R spectra of the 10Be series for the epoch 1610-1985 AD
after removing of the oscilations with periods T>30 years.

A phase-shifted correlation between 10Be concentrations and Rg
/Group Sunspot Numbers/ has been provided. A two extremums of
coefficient of correlation are detected. The first one is
negative ?anticorrelation/ and correspond of shifting of 6 years
/10Be be late/.The second one is positive and correspond of
shifting of 44 years /10Be be late again/.The both extremums
are agreed with Forbush-effect mechanism.They are an
independent evidence about quasy 11 and 88 years modulation
of GCR -flux by active processes on the Sun and interplanetary space
. The extrapolations of the "Dye-3" 10Be times series model for the
21st century show a genetal increaing tendency (fig.5).The last
one is in good agreement with the conclusion for the forcoming of
supercenturial solar minimum during the same period (fig.13)
( see also Komitov and Bonev ,2001)


Fig.5.A model of 10Be time series for the epoch (1610-1995 AD).
Cycles by duration T>30 years are included.

3.2. The Maunder's minimum in 10Be data

The problem about the reality of the 11-years Schwabe-Wolf cycles
during Maunder minimum is discussed by many authors. On the base
of analysis mainly of 14C data some authors conclude, that this
cycle not exist at the end of 17th century. However according
other authors (Beer et al ,1998) the Schwabe-Wolf cycle exist
during the same period.
In the present work we will made a short commentar of our results
and conclusions in this course. The part of 10Be series, which
cover the Maubnder's minimum (1640-1720 AD) is investigated by
using of T-R periodogramm analyses for detecting of cycles. The
local extremums of 10Be concentrations has been compared by calendar
years of 11-years solar cycle sunspot maximums according Schove
(1983) and local maximums of Rg data series between 1640 and 170 AD.
The results are summarized on fig 6 and 7.
As can be seen on fig.6 the 14 and 22-years oscilations are
dominated in the high frecuences region and the "classical" 11-years
Schwabe -Wolf cycle is not detectable. However from other hand the
local minimums of 10Be are in good correspondance to the calendar
years of Schwabe-Wolf cycles maximums in Schove's series, as well
as to the local maximums of observed Rg iodex except the sunspot
maximum near 1685 AD (fig.7). The last one correspond rather to a
maximum of 10Be concentrations , as to a minimum , which requires
the "Forbush-effect" mechanism.
This facts may be interpreted as an evidence that during Maunder's
minimum the 11-years oscilations of solar wind parameters has been
usually under the sensivity threshold for its detection by "Forbush
-effect" and 10Be production. However there are spoors of this cycle
in sunspot activity and the local peaks of observed Group Sunspot
Number are aproximately in the same calendar years, which are
pointed in Schove's series as a moments of Schwabe-Wolf cycle
maximums. The last one is longer during Maunder's minimum and it
is rather 13-14 than 11 years. An extremally low is the Schwabe
Wolf cycle, which maximum is near to 1685 AD.


Fig.6 The "Dye-3" 10Be concentrations during the Maunder's
minimum (1640 -1720 AD).


Fig.7 The moments of the Schwabe-Wolf's cycles maximums
according Schove(1985,1983) are pointed out by gray circles;
by black and white squares- the local Rg maximums.

By combining of all this facts for 10Be and sunspot data we
conclude that the most possible calendar years of Schwabe-Wolf
cycle maximums during the Maunder's minimum are 1648, 1660,
1675, 1685, 1694, 1705 and 1718 AD. The resident time of
~ 1-1.5 years for the 10Be data in this estimation is taken
into account.

3.3. The dominant cycles in the tree rings width
time series (after 1610 AD)

85 time series of tree rings width data from different
regions of Earth has been investigated for existing of
cycles The relative frequences for detection of
different cycles is shown on fig. 8.The main maximums of
frequences are in intervals of periods, which centers are
at 22.5, 32.5 42.5, 52.5, 62.5, 72.5 and 105 years .They
corresponded of cycles in solar activity and 10Be data.
The maximum near 32.5 year is 1/2 of the observed in 10Be
data 65-years cycle . It is equal to 3 Schwabe-Wolf sunspot
solar cycles. May be the quasy 30-35 years cycle is caused
by generation of two dendroclimatical optimums during the
65-years cycle -the first over the increasing and the second-
over the decreasing phase of the last one
. All observed cycles can be associated and explained by solar
modulation over the climate.


Fig.8 The relative frequences of detectable cycles in 85 tree
ring width time series.

---

R E F E R E N C E S

Beer J., Tobias S. and Weiss N., 1998, Solar Physics, v181,
pp237-249

Gleissberg W.,1944,Terr. Magn.Atm.Electr.,v49

Hoyt D. and Shatten K.,1998, Solar Physics,v181,p491

Komitov B.,1986, Solnechnie dannie,No5,pp73-78/in Russian/

Komitov B., 1997, Bulgarian Geophys.J.,v23,No1-2,pp69-78

Komitov B.,2001, in Statistical methods in Bulgarian
Biochronology an Biometeorology, ed. by N.Madjirova,
Macros ,Plovdiv /in Bulgarian/

Komitov B. and Bonev B., 2001, Astrophys J.Let.,v554,
LL119-122

Komitov B. and Kaftan V.,2003, Geomagnetism and Aeronomy,
/in press/

Sharma M.,,2002,Earth and Planet Sci.Let.,v.199,pp459-472

Schove D.J.,,1955,Journ.Geophys.Res.,v60,pp127-146,

Schove D.J.,,1983, Solar Cycles