Beruflich Dokumente
Kultur Dokumente
www.elsevier.com/locate/jastp
Received 7 March 2002; received in revised form 19 September 2002; accepted 10 December 2003
Abstract
We have used the thermodynamic model of the climate to estimate the e*ect of variations in the oceanic cloud cover on the
surface temperature of the Earth in the North Hemisphere (NH) during the period 1984 –1990. We assume that the variations
in the cloud cover are proportional to the variation of the cosmic ray :ux measured during the same period. The results indicate
that the e*ect in the temperature is slightly noticeable when we consider the surface hemispheric temperature on July 1987,
<nding an average temperature anomaly between −0:06◦ C and −0:14◦ C, along a latitudinal band between 20◦ and 40◦ . The
surface temperature averaged globally in the NH presents a decrease of ∼0:01◦ C in average, which is almost the same for
continents and oceans. However, these values are not signi<cant when compared to the overall variability of the time series
with and without forcing.
c 2003 Published by Elsevier Ltd.
the whole Earth, 0.72 at the tropics, 0.88 at mid-latitudes izontal cloud layer (the plane-parallel cloud assumption),
and 0.92 at mid-latitude oceans. an oceanic layer of 100 to 50 m in depth and a continental
Marsh and Svensmark (2000) also used the ISCCP D2 layer of negligible depth. It also includes a layer of ice and
data <nding that low altitude clouds and cosmic rays are well snow over the continents and the ocean. The basic equations
correlated (r ¿ 0:6) over ≈15:8% of the Earth’s surface. are those of hydrostatic balance, perfect gas, continuity and
When taking the global average cloud anomalies they found conservation of thermal energy applied to the atmosphere–
r = 0:63 and 0.92 for the 12-month running mean. ocean–continent system. A monthly time averaging of the
The cosmic rays are registered by neutron monitors lo- variables is used. The details of the equations can be found
cated on the surface. The lowest energy that can be detected in Adem et al. (2000).
depends on the geomagnetic latitude, 0:01 GeV in stations The TMC converges and is stable (Adem, 1962). The
located in the geomagnetic poles up to 15 GeV in the sta- principal equation for the temperature anomalies is an el-
tions near to the geomagnetic equator. The intensity of the liptic expression solved as a <nite di*erence equation by
cosmic rays :ux presents a very well-known inverse rela- the Liebmannn relaxation method described by Thompson
tionship with the sunspot cycle. This is caused by the inter- (1961). In this case, the iterations <nish when the numeric
planetary magnetic <eld which has a screening e*ect on the solution and the theoretical solution have a di*erence of
cosmic rays. Cosmic rays are the dominant source of atmo- about 0:001◦ C, which means that the model errors are about
spheric ionisation at altitudes between 0 and 35 km with a 0:001◦ C for all the points in the integration region.
maximum at ≈15 km (Dorman, 1974). Since this model is basically a thermodynamic model, the
Then, we should expect that high altitude and latitude noise level is not signi<cant, which means that the model can
clouds would be the most a*ected by cosmic rays. How- be run with di*erent initial condition and we shall always
ever, the results indicate that low altitude clouds at low lat- obtain the same solution. This is di*erent from the dynamical
itudes are more in:uenced than higher altitude and latitude models for which the noise is sometimes strong.
clouds. No satisfactory explanation is at hand, and some The integration of the model equations is carried out for
speculations point to the di*erent physical state of the cloud the North Hemisphere (NH) with the use of a grid with 1917
droplets over the polar (ice clouds) and tropical regions points, which uses a polar stereographic projection with a
(liquid clouds) and the e*ect of cosmic rays over them. constant grid distance of 408:5 km (Adem et al., 2000). The
One of the mechanisms proposed for the interaction be- integration is carried out only in the NH because the South-
tween cosmic rays and the low atmosphere involves changes ern Hemisphere has a negligible percentage of continents.
in the rate of ionisation of their compounds. This could In addition, as the human activity is mostly developed in the
modify microphysics processes such as the nucleation and NH, there are more observations.
growth of cloud particles, and therefore the formation of The TMC is suitable for obtaining hemispheric averages
the clouds themselves (higher clouds) (Tinsley and Heelis, (20 –90◦ in latitude) of the anomalies of di*erent meteoro-
1993). logical variables. The model also produces monthly, annual
If indeed the cosmic rays can in:uence the nucleation pro- and seasonal predictions. Finally, we choose the TMC be-
cesses and in turn a*ect cloud formation, we expect that this cause on the one hand it is a coupled model of the continent–
:ux should have an e*ect in the global radiation balance of atmosphere–ocean system that takes into account the
the earth–atmosphere system and therefore in the terrestrial feedback mechanisms inside this system, and on the other
surface temperature <eld. hand, it is relatively easy to manipulate.
The purpose of the present work is to estimate the pro-
posed e*ect of the cosmic rays on the surface temperature 2.2. Modelling of the oceanic cloud cover
through the behaviour of the oceanic cloud cover for the pe-
riod 1984 –1990, when it has been found a good correlation In the TMC, the single horizontal cloud layer has the lower
between total cloud cover and cosmic rays. boundary at 3 km and the upper boundary at 4:5 km. In a
more realistic situation clouds appear in three di*erent lay-
ers, therefore in order to compensate our approximation, the
2. The model
radiative parameters in the single layer are the weighted av-
2.1. The thermodynamic model of climate erage values of the radiative parameters in each layer (Adem,
1964). In the present model, typically when the cloud layer
The model used here is the thermodynamic model of cli- fraction is about 0.5, clouds re:ect ∼25% of the solar ra-
mate (TMC) (Adem, 1962), which assumes that the energy diation, while when the cloud fraction is 1, clouds re:ect
that maintains the atmospheric circulation is the solar radi- 50% back to space. The cloud layer absorbs short-wave ra-
ation, and therefore the fundamental problem is to explain diation from the Sun proportionally to the fraction of the
quantitatively how the transformation of radiant energy in cloud cover; for example, when the cloud fraction is 0.5, the
mechanical energy is carried out. clouds absorb only 2% of solar radiation, and when the cloud
The model consists of an atmospheric layer of about fraction is 1, the clouds absorb 4%. The TMC also assumes
10 km of height which includes a uniform and single hor- that the cloud cover absorbs long-wave terrestrial radiation
J. Ram
rez et al. / Journal of Atmospheric and Solar-Terrestrial Physics 66 (2004) 1683 – 1690 1685
3. Experiments
4.2.1. Continents
The numerical results are shown in Fig. 2a and the sta-
tistical analysis of the results appear in Table 2. Observed
continental temperature anomalies are not available.
From Table 2, we notice that the statistical parameters for
the cloud forcing and no forcing cases are similar, therefore,
the e*ect of the cloud forcing over continent is not signi<-
cant. However, it is important to underline the tendency for
a cooling of the weather when the change in the cloud cover
is positive (see Fig. 2a).
4.2.2. Oceans
In Fig. 2b, we plot the pro<les of temperature anomalies.
In this case, the observed anomalies are available. In Table
3, we present a statistical analysis.
Comparing the statistical parameters in Table 3 we note
Fig. 2. Monthly anomalies of the surface global average tempera- that they are very similar, then the e*ect of the cloud forcing
ture. The dashed line corresponds to the case with no cloud forcing. is not signi<cant.
The thin solid line corresponds to the case with cloud forcing (a)
over continents; (b) over oceans, the observations over ocean are 4.2.3. The hemisphere
plotted as dotted line and squares; (c) over the NH, the observations In Fig. 2c, we have plotted temperature anomalies for the
over hemisphere are plotted as dotted line and squares. The thick whole hemisphere. In this case we can also compare our
solid line represents the variation of the cloud cover over oceans. results with observations. Table 4 shows the statistics.
J. Ram
rez et al. / Journal of Atmospheric and Solar-Terrestrial Physics 66 (2004) 1683 – 1690 1687
Fig. 3. Di*erences between the hemispheric temperature anomalies with cloud forcing and without cloud forcing during July 1987 in a chart
of isothermic lines. To obtain the correct numbers the values must be divided by 10.
Table 1
Statistical parameters Observed values (◦ C) Calculate values without cloud forcing (◦ C) Con<dence interval (95%)
Table 2
Continental statistical parameters
Statistic parameters Calculate values Con<dence inter- Calculate values Con<dence inter-
without cloud val (95%) with cloud forcing val (95%)
forcing (◦ C) (◦ C)
Table 3
Oceanic statistical parameters
Statistic parameters Calculate values Con<dence inter- Calculate values Con<dence inter- Observed values
without cloud val (95%) with cloud forcing val (95%)
forcing (◦ C) (◦ C)
Table 4
Hemispheric statistical parameters
Statistic parameters Calculate values Con<dence inter- Calculate values Con<dence inter- Observed values
without cloud val (95%) with cloud val (95%) (◦ C)
forcing (◦ C) forcing (◦ C)
variation of both curves (with and without cloud forcing). The fact that negative anomalies in the surface tempera-
In this case, we do not have information to validate our ture are present together with a positive change in the vari-
results with the observations. We would like to stress that ation of the cloud cover in some speci<cs regions, means
in this case as we merely isolated the region under study, that the e*ect of short-wave solar radiation income dom-
the spatial size of the integration grid is not reduced and inates slightly over the long-wave radiation in the earth–
therefore noise is not introduced. atmosphere system.
J. Ram
rez et al. / Journal of Atmospheric and Solar-Terrestrial Physics 66 (2004) 1683 – 1690 1689
Veretenenko and Pudovkin (1999), reported a decrease regions. Compared with the UV forcing, the temperature
in the entrance of solar radiation associated with an increase anomaly is ∼30 times smaller.
in cosmic ray :ux, although their study is restricted to a
latitudinal band of 60 –68◦ north. Pall-e and Butler (2000) 5. Conclusions
found that an increase of the low-level clouds at periods of
low solar activity (high levels of cosmic rays) diminished We estimate the e*ect of a cosmic ray-induced global
the surface temperature. Our results in Figs. 2–4 indicate that cloud cover over oceans on the surface temperature with the
there is a slight general tendency for cooling over the surface following outcome:
in all the cases for the period 1984 –1990, which corresponds (1) The change observed in the oceanic cloud cover,
to the minimum and ascending phase of the solar cycle. which oscillates between 3% and 4%, produces a slight
Moreover, the largest response corresponds to the latitudinal imbalance in the net :ow of radiation of the atmosphere–
band of 20 –40◦ , also in agreement with the <ndings of Pall-e ocean–continent system, implying that the surface tempera-
and Butler, of high correlations at mid-latitudes. ture averaged globally (along the period under study) in the
We have forced only the oceanic cloud cover, the ocean NH presents a decrease ∼0:01◦ C in average. This decrease
has a great thermal inertia and therefore has a small response is almost the same for continents and oceans. Nevertheless,
to radiative forcing. We can speculate that if the cloud forc- this value has the same magnitude as the control experiment,
ing is also present in the continental cloud cover then the re- where the oceanic cloud forcing is not introduced.
sponse in the surface temperature could be stronger because (2) The most intense negative anomalies in the hemi-
the continental part is more sensitive. spheric surface temperature take place during July 1987. In
Furthermore, we suggest that the obtained decrement in this period, the maximum departures of the mean value in
the temperature would be strengthened by an extra e*ect: the oceanic cloud cover occurred. There is a latitudinal band
the solar radiometers measured a decrease in the total solar from 20◦ to 40◦ where negative temperature anomalies are
irradiance of 1:3 W m−2 (0.1%) at the top of the terrestrial between 0:06◦ C and 0:14◦ C, nevertheless, these values are
atmosphere during the solar minimum of 1986, relative to not signi<cant in comparison with the overall variations of
the solar maxima of 1980 and 1990 (e.g. Lean, 1997). This the time series.
decrease in the solar irradiance induces a climatic forcing (3) In the region corresponding to Mexico, the di*erence
of about 0:22 W m−2 . This e*ect added to the e*ect of the between cloud forcing and no cloud forcing is in average
oceanic cloud forcing proposed here would produce a larger ∼0:06◦ C. This value is smaller than the overall variation of
decrement of the surface temperature of the Earth during the time series with and without cloud forcing.
epochs of solar activity minimum. (4) As the model used here has a low sensitivity factor
The space studies (GISS) general circulation model (0:35◦ C W m−2 ), the temperature anomalies found should
(GCM) predicts an NH surface temperature change of be considered as lower estimates. Stronger anomalies should
0:51◦ C for a 0.25% solar irradiance reduction (Lean and be found with higher sensitivity models.
Rind, 1998). Furthermore, Rind et al. (1999) showed that (5) We conclude that the modulation of the oceanic cloud
an estimated solar forcing increase of ∼0:25% accounts for cover does not a*ect signi<cantly the surface temperature
a 0:45◦ C temperature increase since 1600 and an increase of the NH.
of about 0:2◦ C over the past 100 years.
Finally, we must point out that we have worked with a low References
sensitivity model; stronger temperature anomalies should be
expected if more sensitive models are used. The present Adem, J., 1962. On the theory of the general circulation of the
atmosphere. Tellus 14, 102–115.
model gives a lower estimate of the forcing.
Adem, J., 1964. On the physical basis for the numerical prediction
of monthly and seasonal temperatures in the troposphere–ocean–
continent systm. Monthly Weather Review 92, 91–104.
4.2.4. Comparisons with other solar forcings Adem, J., Mendoza, V.M., Ruiz, A., Villanueva, E.E., Garduño, R.,
Cosmic rays are not the only proposed mechanism for 2000. Recent numerical experiments on three-months extended
solar activity in:uencing the climate. For instance, Soon and seasonal weather prediction with a thermodynamic model.
et al. (2000) compared the equilibrium climate responses of Atm-osfera 13, 53–83.
a quasi-dynamical energy balance model to radiative forcing Dorman, L.I., 1974. Cosmic Rays. Variations and Space
by equivalent changes in CO2 , solar total irradiance and solar Exploration. North-Holland, New York.
ultraviolet (UV). They found that when the solar constant Garduño, R., Adem, J., 1994. Initial radiative perturbations and
their responses in the Adem thermodinamic model. World
is increased by 0.5%, which causes a 1 W m−2 increase in
Resource Review 6, 343–349.
the net down-going solar radiation at 250 mb, the change Gierens, K., Ponater, M., 1999. Comment on “variation of cosmic
in global surface temperature is about 1:2◦ C. In the case ray :ux and global cloud coverage—a missing link in solar
of solar ultraviolet, a similar forcing produces a change of climate relationship” by H. Svensmark and H. Friis-Christensen
∼3◦ C in the global surface temperature. Our low sensitivity (1997). Journal of Atmospheric and Solar-Terrestrial Physics
model gives at most changes of −0:14◦ C and only in some 61, 795–797.
1690 J. Ram
rez et al. / Journal of Atmospheric and Solar-Terrestrial Physics 66 (2004) 1683 – 1690
Lean, J., 1997. The Sun’s radiation and its relevance for Earth. Svensmark, H., Christensen, E.F., 1997. Variation of Cosmic
Annual Review of Astronomy and Astrophysics 35, 33–67. Ray Flux and Global cloud Coverage—A missing Link
Lean, J., Rind, D., 1998. Climate forcing by changing solar in Solar-Climate Relationships. Journal of Atmospheric and
radiation. Journal of Climate 11, 3069–3093. Solar-Terrestrial Physics 59, 1225.
Marsh, N.D., Svensmark, H., 2000. Low cloud properties Thompson, P.D., 1961. Numerical Weather Analysis and
in:uenced by cosmic rays. Physical Review Letters 85, 5004. Prediction. Macmillan, New York.
Pall-e, E., Butler, C.J., 2000. The in:uence of cosmic rays Tinsley, B., Heelis, R.A., 1993. Correlations of atmospheric
on terrestrial clouds and global warming. Astronomy and dynamics with solar activity: evidence for a connection via the
Geophysics 41, 418–422. solar wind, atmospheric electricity, and cloud micro-physics.
Rind, D., Lean, J., Healy, R., 1999. Simulate time-dependent Journal of Geophysical Research 98, 10375.
climate response to solar radiative forcing since 1600. Journal Veretenenko, S.V., Pudovkin, M.L., 1999. Variation of solar
of Geophysical Research 104, 1973–1990. radiation input to the lower atmosphere associated with
Soon, W., Posmienter, E., Baliunas, S., 2000. Climate hyper- di*erent heliogeophysical factors. Journal of Atmospheric and
sensitivity to solar forcing? Annales Geophysicae 18, 583–588. Solar-Terrestrial Physics 61, 521–529.