Solar Forcing of Regional Climate Change During the Maunder Minimum

4 Pages·2002·0.58 MB·English

by  Shindell, Drew T.

>. Mode), defined as the first EOF of 20th-
a_ Sotar Forcing of Regional century SLP (16). This pattern contains the
North Atlantic Oscillation (NAO), which
t_
Climate Change During the may be considered a different view of the
same phenomenon (17). We define the mod-
Maunder Minimum el's AO pattern as EOF i, and the AO index
as the opposite of the area-weighted change
in SLP northward of 60°. Projection of the
Drew T. ShindeU, 1Gavin A. Schmidt, 1Michael E. Mann, 2 1780-to-1680 SLP difference (Fig. 2) onto
',O David Rind, 1Anne Waple 3 the first five control EOFs shows that they
account for 37.3, 0.0, 3.4, 1.5, and 0.0% of
We examine the dimate response to solar irradiance changes between the late the variance, respectively. Thus, the regional
t_
a_ 17th century Maunder Minimum and the late 18th century. Global average climate variations primarily result from a
d temperature changes aresmall (about 0.3 to 0.4°C) inboth aclimate model and lower index state of the AO during the Maun-
z empirical reconstructions. However, regional temperature changes are quite der Minimum relative to that of a century
large. Inthe model, these occur primarily through aforced shift toward the low later (by -!.1 mbar; similarly, the SLP con-
index state of the Arctic Oscillation/North Atlantic Oscillation. This leads to trast between Iceland and the Azores decreas-
co|der temperatures over the Northern Hemisphere continents, especially in es by 1.9 mbar). This reduces onshore advec-
winter (1 to 2*(:), in agreement with historical records and proxy data for tion of warm oceanic air, yielding a 1°to 2°C
surface temperatures. cooling over large areas of North America
and Eurasia (Fig. i).
A minimum in solar irradiance, the Maunder creases with solar output. Thus, the reduced Historical data associate the European
.. Minimum, is thought to have occurred from Maunder Minimum irradiance leads to in- Maunder Minimum winter cooling with en-
,< the mid-16OOs to the early 1700s (1-3). Con- creased upper stratospheric ozone, which in hanced northeasterly advection of continental
currently, surface temperatures appear to turn causes lower stratospheric ozone to de- air, consistent with an anomalous negative
have been at or near their lowest values of the crease. Both cause negative radiative forcing NAO (18). Reconstructions of large-scale
oo
OtM past millenium in the Northern Hemisphere (14). Inclusion of ozone photochemistry surface temperature patterns in past centuries
(NH) (4-7), and European winter tempera- therefore enhances radiative forcing from (19, 20) from networks of diverse proxy in-
toO- tures were reduced by 1to 1.5°C (8). preindustriai solar variations. formation (tree rings, ice cores, corals, and
r- We used a version of the Goddard Insti- Equilibrium simulations were performed historical records) are particularly useful
r- tute for Space Studies (GISS) general circu- for spectrally discriminated irradiances in benchmarks for comparison. Correlations be-
ro-
lation model (GCM), which includes a de- 1680 and 1780 (1). An alternative reconstruc- tween these reconstructions and solar irradi-
tailed representation of the stratosphere, to tion has a 40% larger long-term solar varia- ance reconstructions (1) during the preanthro-
ll
t/1 simulate the difference between that period tion, which would give a larger climate re- pogenic interval from 1650 to 1850 provide
Ill and a century later, when solar output re- sponse (15). Initial conditions were taken an empirical estimate of the large-scale cli-
_O mained relatively high over several decades. from 1680 and 1780 in atransient simulation. mate response to solar forcing (21). The in-
The model contains a mixed-layer ocean, al- A spin-up time of about 25 years was never- stantaneous (zero-lag) frequency-indepen-
"D
lowing sea surface temperatures (SSTs) to theless required for equilibration with the dent regression pattern, primarily indicative
i-
t'- respond to radiative forcing, and has been increased preindustrial ozone. Results are of decadal solar-climate relationships, shows
O
O shown to capture observed wintertime solar therefore based on the past 30 years of 60- little evidence of an AO/NAO signal (Fig. 3,
O cycle-induced variations reasonably well (9). year simulations. top). This is consistent with the extremely
I]0 The GCM includes parameterizations of Modeled global annual average surface weak modeled AO/NAO response to the 1I-
L. the response of ozone to solar irradiance, temperatures were 0.34°C cooler in 1680 year solar cycle using fixed SSTs, equivalent
9 temperature, and circulation changes (al- than in 1780. This is similar to the annual to an instantaneous response (22). However,
though transport changes are noninteractive) average NH 1680-minus-1780 temperature when filtered to isolate the multidecadal/cen-
u (9), based on results for preindustrial condi- change of about -0.2 ___0.2°C from proxy tennial time scales associated with the Maun-
_O tions from our two-dimensional (2D) chem- data (6) or an energy balance model (7) der Minimum, and lagged to allow for inertia
istry model (10). These show more ozone in (model NH cooling is-0.33 C). Other proxy- in the ocean's response, the empirical regres-
i
_0 the upper and lower stratosphere, as expected based reconstructions for the NH extratropics sion (Fig. 3, bottom) shows a clear AO/NAO-
(11), primarily because of the absence of during summer indicate a -0.3 to -0.4°C type pattern of alternating cold land and
t-
i- anthropogenic halogens and a drier strato- cooling (4,5) (comparable model cooling is warm ocean temperature anomalies. Both pa-
sphere (12). Ozone's temperature sensitivity -0.3°C). leoclimate reconstructions and the GCM thus
was also increased throughout the upper Modeled surface temperature changes indicate in a remarkably consistent manner
stratosphere, more than doubling at 3 milli- show alternating warm oceans and cold con- that solar forcing affects regional scales much
bars (mbar) relative to the present (13). Solar tinents at NH mid-latitudes (Fig. 1), with more strongly than global or hemispheric
heating then speeds up chemical destruction maximum amplitude inwinter. To character- scales through forcing of the AO/NAO.
more than irradiance enhances photolytic ize regional changes, we calculated empirical Most additional paleoclimate data have
production, so that ozone concentration de- orthogonal functions (EOFs) from the NH insufficient time resolution to distinguish the
extratropical (20 °to 90°N) cold season (from Maunder Minimum but suggest a shift toward
_NASAGoddardInstitute forSpaceStudiesandCen- November to April) sea level pressure (SLP) the AO/NAO low-index state during periods
terfor ClimateSystemsResearch,ColumbiaUniver- time series. The last 40 years of each run of reduced solar forcing. Ocean sediments
sity,New York,NY 1002S, USA._Department of
were used as an 80-year control. show a decreased SST of I to 2°C in the
EnvironmentalSciences,University ofVirginia,Char-
lottesviile,VA 22902, USA._Department of Geo- The leading control EOF (Fig. 2) closely Sargasso Sea during the 16th to 19th centu-
sciences,Universityof Massachusetts,Amherst,MA matches the Arctic Oscillation pattern (AO, ries, and warmer Atlantic temperatures north
01003.USA also called the Northern Hemisphere Annular of 44°,extending to the area off Newfound-
www.sciencemag.org SCIENCE VOL 0 • MONTH 2001 1
REPORTS
........ Ir_-= I- I I 1 I:_1 I
-1.31 -.7 -.5 -.35 -.2 -.05 .05 .2 .35 .5 .7 .9
Annual Temperature change (C)
-1.75 -.7 -.5 -.35 -.2 -.05 .05 .2 .35 .5 .7 .9
Winter Temperature change (C)
Fig.1.Surfacetemperature change(in°C) from 1780 to 1680 inthe GCM.Global annua[ average (left) andNovember toApril NH extratropics (right)
areshown. Nearly allpoints arestatistically significant (not shown) becauseof theLargenumber of model years. Data for all figuresareavaitabte as
online supplemental material (40).
(27, 28), and zonal wind and planetary wave
propagation changes over recent decades are
well reproduced in the model (22).
Reduced irradiance during the Maunder
Minimum causes a shift toward the low-index
AO/NAO state via this same mechanism.
During December to February, the surface
cools by 0.4 ° to 0.5°C because of reduced
incoming radiation and the upper stratospher-
ic ozone increase. Cooling in the tropical and
subtropical upper troposphere is even more
pronounced (-0.8°C) because of cloud feed-
backs, including an -0.5% decrease in high
cloud cover induced by ozone through sur-
face effects. A similar response was seen in
simulations with a finer resolution version of
the GISS GCM (14). This cooling signifi-
I I..... : I I I I_: TI I
cantly reduces the latitudinal temperature
-7.0-2.5-1.9-1.4-.8 -.3 .3 .8 1.4 1.9 2.5 3 -2.2-1.6-1.2-.9 -.5 -.2 .2 .5 .9 1.2 1.6 2.3
gradient in the tropopause region, decreasing
Pressure (mb) Pl_mSure (rob)
the zonal wind there at _40°N. Planetary
Fig.2. The leadingEOF ofNH extratropicaL November-to-April SLPover the past40 years of the waves coming up from the surface at mid-
1680 and 1780 simulations (left), and the SLPchange from 1780 to 1680 (right), both in mbar latitudes, which are especially abundant dur-
(mb). The SLPdifference isfiltered inEOFspacebyshowing the projection onto thefirst 20 EOFs, ing winter, are then deflected toward the
which contain 70% of the total difference. This removes some ofthe high-frequency noiseinthe
equator less than before (equatorward Elias-
model induced byoverly large energies at high wavenumber_ Hatched areas indicate statistical
significance at the 90% Level. sen-Palm flux is reduced by 0.41 m2/s 2, 12°
to 35°N, 300 to 100 mbar average), instead
land (23), which isconsistent with a reduced processes. This results in (iii) an increased propagating up into the stratosphere (in-
NAO but not with uniform basin-wide cool- latitudinal temperature gradient at around 100 creased vertical flux of 6.3 × l0 -4 m2/s 2, 35°
ing. Other evidence suggests colder North to 200 mbar, because these pressures are in to 60°N, 100 to 5 mbar average) (29). This
Atlantic temperatures during those centuries, the stratosphere at higher latitudes, and so do increases the wave-driven stratospheric resid-
however, including evidence for increased not feel the surface warming (26). The tem- ual circulation, which warms the polar lower
sea ice around Iceland (24), a region of min- perature gradient leads to (iv) enhanced lower stratosphere (up to I°C), providing a positive
imal change in the GCM. stratospheric westerly winds, which (v) re- feedback by further weakening the latitudinal
Our previous studies have demonstrated fract upward, propagating tropospheric plan- temperature gradient. The wave propagation
how external forcings can excite the AO/ etary waves equatorward. This causes (vi) changes imply a reduction in northward an-
NAO in the GISS GCM (22, 25). Briefly, the increased angular momentum transport to gular momentum transport, hence a slowing
mechanism works as follows, using a shift high latitudes and enhanced tropospheric down of the middle- and high-latitude west-
erlies and a shift toward the low AO/NAO
toward the high-index AO/NAO as an exam- westerlies, and the associated temperature
ple: (i) tropical and subtropical SSTs warm, and pressure changes corresponding to a high index. Because the oceans are relatively
leading to (ii) a warmer tropical and subtrop- AO/NAO index. Observations support a warm during the winter owing to their large
ical upper troposphere via moist convective planetary wave modulation of the AO/NAO heat capacity, the diminished flow creates a
2 • HONTH 2001 VOL 0 SCIENCE www.sciencemag.org
REPORTS
8O ilar experiments (32).
The ECHAM3 GCM exhibits a similar
spatial pattern of response to solar forcing,
6O
with warming over the NH continents and
cooling over the North Atlantic and North
4O Pacific (33). In that model, however, the sur-
face temperature and AO sensitivity were
much less than in our simulation. This is
probably due to the lack of interactive ozone
and the poorly resolved stratosphere, two im-
portant positive feedbacks in our model.
The importance of SST changes to initiate
-.9 -.7 -.S -.35 -.2 -.05 .05 .2 .35 .5 .7 the AO/NAO response is consistent with our
Temperature change (C) fixed-SST solar cycle experiments (9, 22),
whose response did not project strongly onto
the AO/NAO, and with the weak empirical
instantaneous climate sensitivity (Fig. 3, top).
Present-day observations and modeling also
support links between long-term tropical SST
changes and the NAO (34,35).
Surface temperatures in an energy bal-
ance model (36) driven by volcanic forcing
gave a poor correlation with proxies during
the 17th and 18th centuries. Correlations
between calculations of solar-induced cli-
mate change and temperature proxies were
quite good, however, although the impacts
-.0 -.7 -.s --.36 -2 -.05 .05 .2 .35 .s .7 .0 were very small. Inclusion of ozone chem-
ical feedback would amplify the response,
Temperature change (C)
potentially reproducing observed ampli-
Fig. 3.Regression duringthe period from 1650 to 1850 between reconstructed solar irradiance (1) tudes as well as variability.
and annual average surface temperatures (19). The top panel shows the instantaneous regression
The GISS model results and empirical
over alltime scales, whereas the bottom panel shows the regression when filtered to onLyinclude
contributions from time scales longer than 40 years, with a20-year lag(resuLtsare similar fortime reconstructions both suggest that solar-forced
lagsof 10to 30years). Thecorrelation isgivenin °C per-0.32 W/m z(the solar forcing usedinthe regional climate changes during the Maunder
modeL) for comparison with the modeLedtemperature changes shown in Fig.1.Gray areasindicate Minimum appeared predominantly as a shift
no data; hatched areas indicate statistical significance atthe 90% Level. toward the low AO/NAO index. Although
global average temperature changes were
cold-land/warm-ocean surface pattern (Fig. and climate feedbacks are included, the cool- small, modeled regional cooling over the
1). ing drops to --0.19°C. With current atmo- continents during winter was up to five times
To quantify the importance of ozone spheric composition, ozone enhancement greater. Changes in ocean circulation were
changes, an identical pair of 1680 and 1780 from increased irradiance outweighs temper- not considered in this model. However, given
experiments was run, leaving out the ozone ature-induced changes, leading to more mid- the sensitivity of the North Atlantic to AO/
response. These simulations show areduction die and upper stratospheric ozone (9), which NAO forcing (37), oceanic changes may well
in the AO index of only 0.5 mbar, and the dampens the net radiative forcing. Ozone's have been triggered as a response to the
SLP change projection onto the AO was reversal from apositive (preindustrial) to a atmospheric changes (38). Such oceanic
15.1%. Thus, there was an analogous but negative feedback supports results showing changes would themselves further modify the
weaker response without interactive ozone. that solar forcing has been a relatively minor pattern of SST in the North Atlantic (39) and,
The ozone changes further reduce the tropo- contributor to late 20th-century surface to a lesser extent, the downstream air temper-
pause radiative forcing by -0.02 W/m 2 (in- warming (7, 19, 31). However, ozone feed- ature anomalies in Europe.
stantaneous, but with adjusted stratospheric backs remain important to the dynamical re- These results provide evidence that rela-
temperatures) for 1680 conditions (compared sponse associated with decadal and shorter tively small solar forcing may play a signif-
to the direct solar forcing of-0.32 W/m2). term solar variability, which is initiated by icant role in century-scale NH winter climate
Cloud feedbacks caused by ozone magnify its upper stratospheric temperature changes, change. This suggests that colder winter tem-
influence somewhat, so that the global annual which are proportional to solar heating and peratures over the NH continents during por-
average surface temperature change increases hence to ozone (9). tions of the 15th through the 17th centuries
from -0.28°C to -0.34°C, including ozone The AO/NAO response is dependent on (sometimes called the Little Ice Age) and
chemistry. The largest impacts resulted from the overall forced change in tropical and sub- warmer temperatures during the 12th through
the enhancement of the tropopause region tropical SSTs, translation of these changes 14th centuries (the putative Medieval Warm
latitudinal temperature gradient through high into upper tropospheric temperature changes Period) may have been influenced by long-
cloud changes and subsequent wave-driven by hydrological feedbacks, and the resulting term solar variations.
feedbacks. wind and planetary wave refraction changes.
In contrast to the preindustrial simula- These processes vary between models and in
References and Notes
tions, 20th-century simulations (30) show particular are sensitive to the inclusion of a
1.J.Lean,J.Beer,R.Bradley,Geophys.Res.Lett.22,
that solar irradiance changes alone warm the realistic stratosphere (22, 25), emphasizing 319s 099s).
surface by -0.25°C, but when ozone changes the need for intermodel comparisons of sim- 2. E.Bardetal.,TellusB52,985 (2000).
www.sciencemag.org SCIENCE VOL 0 • MONTH 2001 3
REPORTS
3. M. Stuiver, T. F. Braziunas, Radiocarbon 35. 137 the rate-limiting reactions for the chlorine, nitrogen, Thompson, J.R.Holton. Proc. Natl. Acad. 5d. U.S.A.
(1993). and hydrogen cycles isweakly negative, whereas for 97, 1412 (2000).
4. K.Briffa et al., Nature 393, 350 (1998). oxygen itisstrongly positive. The relative importance 29. The flux changes are roughly -1/3 those seen inour
5. P. Jones et al., Holocene 8, 455 (1998). of the oxygen cycle was greater during the preindus- greenhouse gas forcing simulations (22, _>5),which is
6. M. E.Mann, R.S.Bradley, M. K.Hughes, Geophys. Res. trial period, leading to large increases inoverall tem- consistent with the AO response ratio.
Lett. 26, 759 (1999). perature sensitivity in the upper stratosphere. 30. Temperature change was calculated as the differ-
7. T. J.Cmwley, Science 289, 270 (2000). 14. J.Hansen, M. Sato, R.Ruedy, ]. Geophys. Res. 102, ence between simulations, using 1900 and 1990s
B. C. Pfister, in Climate Since A.D. 1500, R.S. Bradley, 6831 (1997). irradiance, both with 1990s atmospheric composi-
P. D.Jones, Eds.(Routledge, London, 1995), pp. 118- 15. D. V. Hoyt, K. H. Schatten, J. Geophys. Res. 98, tion. However, the influence of irradiance chan_es
142. 18,895 (1993). cannot be fully separated from that of the chang-
9. D. T. Shindel[, D. Rind, N. Baiachandran, J.Lean, P. 16. D.W. J.Thompson, J.M Wallace, Geophys. Re_. Lett. ing atmospheric composition on ozone photo-
Lonergan, Science 284, 305 (1999). 25. 1297 (1998). chemistry or the nonlinear long-term dynamical
10. D. T.Shinde[I, D. Rind, P. Lonergan, J.Clim. 11, 895 17. J.M.Wallace, Quart.J. tboooo_ HeteroL 5oc. 126, 791 and chemical responses.
(1998). (2000). 31. G. C.Hegerl et al., Clim. Dyn. 13, 613 (1997).
11. D.J. Wuebbles, C.-F. Wei, K.O. Patten, Geophys. Res. 18. V. C.Slonosky, P. D.jones, T.D.Davies, Int. J. Clirna- 32. N. Gillett et al.,J. Geophys. Res., in press.
Lett. ZS, 523 (1998). tol. Zl, 285 (2001). 33. U. Cubasch et aL, Clim. Dyn. 13, 757 (1997).
12. Because 2D photochemical models have difficulty 19. M. E.Mann, R.S. Bradley, M. K. Hughes, Nature 392, 34. B. Rajagopalan, Y. Kushnir, Y. M. Tourre, Geophys.
accounting for the dynamically driven stratospher- 779 (1998). Res. Lett. 75, 3967 (1998).
ic hydrological cycle, we prescribed stratospheric 20. Because the pattern reconstructions aresmoothed by 35. M. P.Hoer[ing, J.W. Hurrell, T. Xu, Science Z9Z, 90
water vapor changes due to methane oxidation by retaining only atruncated eigenvector basis set, the (Z001).
conserving xxxxxxx (2 x CH4 + H20), which sat- amplitude of regional temperature variations is gen- 36. M. Free, A. Robock, J. Geophys. Res. 104, 19,057
ellite data show is roughly constant. Long-term erally underestimated. This isparticularly true before (1999).
methane trends were taken from ice core data, 1730, when five or fewer eigenvectors are retained. 37. T.L Delworth, K.W. Dixon, J. C6m. 13, 3721 (2000).
which show approximately 60% less methane dur- 21. A. M. Waple, M. E.Mann, R.S.Bradley, C6m. Dyn., in 38. W. S.Broecker, Proc. Natl. Acad. Sd U.S.A. 97, 1339
ingthe preindustrial period. Observations from the press. (2000).
HALOE instrument, version 18data, obtained from 22. D. T. Shindell, G.A. Schmidt, R.L Miller, D. Rind, J. 39. M. Visbeck, H. CuUen, G. Krahmann, N. Naik, Geo-
the NASA Langley data center, were used to de- Geophys. Res. 106, 7193 (2001). phys. Res. Lett. 25, 4521 (1998).
termine the fraction of methane oxidized through- 23. L D. Keigwin, I_ S.Pickart, Science 286, 520 (1999). 40. Supplemental Web material is available on Science
out the stratosphere. For example, at 1 mbar in 24. P. Bergth6rsson, J&kul! 19, 94 (1969). Online at www.sciencemag.org/cgi/contentJfull/
mid-latitudes, -1.6 parts per million by volume 25. D.T. Shindell, R.L Miller, G. A.Schmidt, L Pandolfo, VOL/ISSUE/PAGE/DCl.
(ppmv) of water was removed relative to the Nature 399, 452 (1999). 41. Stratospheric climate modeling at GlSS is supported
present based on 80% oxidation of the 1.0-ppmv 26. For forcing by greenhouse gases, the enhanced lati- by NASA's Atmospheric Chemistry Modeling and
methane change. This assumes that the preindus- tudinal gradient isfurther strengthened bytheir cool- Analysis Program. GA.S. and D.T.S. acknowledge the
trial methane oxidation rate in the stratosphere ing effect in the stratosphere. Ozone depletion and support of NSF grant ATM-00-02267. M.E.M ac-
was aslarge asthat at present. This isfundamental volcanic forcing enhance the AO by an abbreviated knowledges support from the National Oceanic and
to the negative ozone response, which differs from mechanism, directly altering temperatures near the Atmospheric Administration- and NSF-supported
that seen in aprevious study (11). tropopause as instep (iii). Earth Systems History program. We thank J.Lean for
13. Ozone's temperature sensitivity isgoverned by cat- 27. Y. Ohhashi, K.Yamazaki, J. Meteorol. Soc. Jpn. T'I, providing the solar irradiance reconstruction.
alytic cycles involving chlorine, nitrogen, hydrogen, 495 (1999).
and oxygen radicals. The temperature dependence of 28. D.LHartmann, J.M.Wallace, V. Limpasuvan, D.W. J. 13 July 2001; accepted 31 October 2001
Artic[e is 585 picas
4 • MONTH 2001 VOL0 SCIENCE www.sciencemag.org