Atlantic Climate Variability Experiment Prospectus

Summary standing of mechanisms and models of atmo- Energetic, large-scale variability is observed spheric and ocean processes that lead to climate in the atmosphere and ocean of the Atlantic Sec- variability in the Atlantic and its global conse- tor on interannual and decadal time scales. It is quences. We must determine whether observed manifested as coherent fluctuations in tempera- and simulated variability is fundamentally ture, rainfall, surface pressure and temperature coupled, whether one component drives variabil- reaching eastward to central Europe and north- ity in the other, whether modeled variability bears ern Asia, southward to subtropical West Africa any relation to observed variability, and whether and westward to North and South America with modes of variability once identified are predict- a myriad of well documented impacts on society able. Improvement of atmospheric and oceanic and the environment. The coordinated changes circulation models, especially boundary layer in the position of mid-latitude atmospheric pres- formulations, is a central part of ACVE. sure centers and winter storm tracks over the At- The specific objectives of the Atlantic Climate lantic, as well as the strength of the trade winds, Variability Experiment are to: are commonly referred to as the North Atlantic • Describe and model coupled atmosphere- Oscillation (NAO). In the tropical Atlantic, sur- ocean interactions in the Atlantic sector, quan- face temperature variability and associated tify their influences on the regional and larger changes in the inter-tropical convergence zone scale climate system, and investigate their and the Hadley circulation occurs on interannual predictability. to decadal time scales, phenomena we collectively • Assemble a quantitative historical and real call Tropical Atlantic Variability (TAV). The NAO time data set that may be used to test, improve and TAV are reflected in marked changes in wind and initialize models of coupled Atlantic cli- stress on the ocean, as well as air-sea heat and mate variability. fresh water fluxes. These in turn induce substan- • Investigate the sensitivity of the meridional tial changes in the wind- and buoyancy-driven overturning circulation of the ocean to ocean circulations, water mass transformation and changes in surface forcing and assess the like- possibly poleward heat transport. Indeed many lihood of abrupt climate change. ocean properties show strong links to overlying Some of the processes involved in climatic atmospheric variability, suggesting that much of modulation of the North Atlantic will require the observed ocean variability is driven by the broad-scale study of entire gyres or basins; other atmosphere. Whilst it is known that the Atlantic topics can be examined in focused process ex- ocean plays a central role, through its thermoha- periments. Some investigations must extend over line circulation, in the mean climate, its role in a decade so as to observe a full cycle of the vari- modulating atmospheric variability on interannual ability of interest. But clever treatment of exist- to decadal time scales is less clear. ing historical observations will provide an impor- We propose an Atlantic Climate Variability tant long-time-scale perspective. Understanding Experiment (ACVE) to address the overarching the causes of interannual to decadal climate vari- question: In what ways does coupled atmosphere- ability demands a new type of research program. ocean dynamics in the Atlantic sector play an The current global network of sustained obser- active role in climate variability? Its primary goal vations is too thin to support detailed quantita- is to quantitatively test and improve our under- tive study of specific climate modes. Moreover,

1 typical process studies do not last long enough to of future climate fluctuations and the extent to resolve decadal-time-scale variability. which they may be predictable, and b) understand- What is needed is an integrated program of ing the potential impact of climate change due to sustained observations, process studies and mod- changing levels of greenhouse gases. eling conducted jointly by meteorologists and A substantial proportion of the climate vari- oceanographers that last a decade or longer. The ability in the Atlantic region is associated with a proposed Atlantic Climate Variability Experiment phenomenon known as the North Atlantic Oscil- needs to develop time-dependent, basin-scale lation (NAO). Although the spatial pattern of the analyses of the Atlantic ocean and atmosphere, NAO appears to be a natural mode of atmospheric based on an economically efficient observing variability, its phase may be influenced by sur- system having significantly greater coverage and face, stratospheric, or even anthropogenic pro- density than that provided by routine observations cesses. The NAO has undergone major low fre- available today. The phenomenological focus will quency variation this century. The rising NAO be the NAO and TAV modes of climate variabil- state over the past twenty years has been accom- ity. The ACVE will encompass model-data syn- panied by a strengthening of the Pacific-North thesis in which mechanisms, key processes, and America (PNA) pattern of atmospheric variabil- global teleconnections will be addressed quanti- ity, and together they have resulted in a global tatively inside an improving coupled climate pattern of cooling of the mid-latitude oceans, and model framework. A prerequisite is to develop warming over northern hemisphere land masses. an active dialogue between observers, modelers Several mechanisms have been proposed to and theoreticians. account for the observed low-frequency variabil- This document provides the scientific objec- ity in the NAO including atmospheric response tives and framework for achieving an interna- to changes in sea surface temperatures, variabil- tional Atlantic CLIVAR program. It represents a ity of deep atmospheric convection in the trop- consensus document of U.S. and European sci- ics, and the effects of changing external forces entists reflecting the findings of two U.S. work- such as solar insolation, volcanic eruptions, and shops: in Lamont (September, 1997) and Dallas anthropogenic emissions of climatically impor- (February 1998); and a joint U.S.-European meet- tant trace gases. Alternatively, the NAO may sim- ing held in Florence (May 1998). A list of scien- ply be a natural internal mode of the atmosphere. tist who attended these meetings is attached in Distinguishing between these hypothesized section 8. mechanisms is a high priority. Low frequency After a brief introduction in section 1, the sci- variation of the Atlantic Ocean is also poorly un- entific background is set out in section 2. Section derstood. SST anomalies have been observed to 3 discusses some of the societal impacts of At- propagate slowly through the upper-ocean gyre lantic-sector climate variability. The scientific circulations, formed in response to, or possibly questions are set out in section 4, and a proposed inducing changes in the NAO. The phase of the strategy for addressing them is given in section NAO also appears to be correlated to the forma- 5. Programmatic context is discussed in section tion of deep water in the Greenland, Iceland, 6, along with an outline of the next steps neces- Norwegian and Labrador Seas, which in turn in- sary to advance the planning and begin the imple- fluence the strength and character of the Atlantic mentation for this project. meridional overturning circulation (MOC). Al- though changes in the MOC can occur naturally, 1. Introduction there is also the possibility that they could be trig- The climate of the Atlantic Sector has exhib- gered by anthropogenic emissions. A breakdown ited considerable variability on a wide range of of the MOC could lead to drastic changes in glo- time scales. Improved understanding of this vari- bal climate. A program investigating Atlantic vari- ability is essential for a) assessing the likely range ability should therefore also explore those atmo-

2 spheric, oceanic and ice processes which are criti- 2.1 — The North Atlantic Oscillation cal to the MOC. Earlier this century, meteorologists noticed Substantial variability is observed in the tropi- that year-to-year fluctuations in winter air tem- cal Atlantic region on interannual and decadal peratures on either side of Iceland were often out time scales. Sea surface temperature (SST) of phase. When temperatures were below normal anomalies have been observed both north and over Greenland they were above normal in south of the equator, along with regional changes Scandinavia, and vice versa. Concomitant fluc- in winds and precipitation. Following the success tuations in rainfall, and sea level pressure (SLP) of ENSO modeling in the Pacific, study of low- were also documented, reaching eastward to cen- latitude Atlantic variability has highlighted the tral Europe, southward to subtropical West Af- tropical Atlantic’s response to variable atmo- rica and westward to North America. Walker and spheric forcing, and the atmospheric response to Bliss (1932) noted a tendency for pressure to be tropical SST gradients. Key questions in both the anomalously low near Iceland in winter when it Atlantic and Pacific include the nature and con- is anomalously high near the Azores and south- sequences of tropical-subtropical interactions west Europe. This mode of climate variability has between the free atmosphere and the ocean, as come to be known as the North Atlantic Oscilla- well as inter-ocean interactions via tion (NAO), a name first used by Walker in 1924 teleconnections within the tropics or involving (see Rogers, 1990). The horizontal structure of the higher-latitude PNA and NAO patterns. the NAO is plotted in the Figure 2.1a. In con- Most important, the predictability of climate trast to the mean flow, the NAO has a pronounced fluctuations in the Atlantic region needs to be “equivalent barotropic” structure (vertical phase established. Great progress has been made in the lines, Wallace and Gutzler, 1981, Kushnir and last 20 years developing understanding of the Wallace, 1989) and increases in amplitude with evolution of ENSO in the Pacific and its influ- height in rough proportion to the strength of the ence on future weather patterns. It is timely to mean zonal wind. This result is consistent with begin comparable studies of Atlantic modes of the notion that the NAO has a structure that is climate variability. Even if it turns out that the close to that of a stationary external Rossby mode predictability of such modes is limited, improved (Held, private communication) and maybe insepa- understanding of climate variability in the Atlan- rable from true dynamics of the zonally symmet- tic region is essential to better assess the poten- ric circulation. It appears to be a “resonant struc- tial impact of climate change due to changing lev- ture” (a “neutral mode” of the climatology; els of greenhouse gases, and the possibility of Marshall and Molteni; 1993), which can be ex- rapid climate change associated with changes in cited in a number of different ways. the MOC. The vertical structure of the NAO gives us strong clues as to its likely mechanism. Eddy- 2. Modes of Atlantic Climate mean-flow interaction (i.e., dynamics internal to Variability the atmosphere) is thought to be the primary forc- To motivate the science plan described below, ing mechanism on short time-scales. Eddy-poten- we give here a brief review of what is known about tial vorticity flux forcing readily projects on to Atlantic climate variability1. the vertical structure of the NAO. The NAO index, defined as the normalized

1 Detailed discussions of climate variability phenomena in and over the Atlantic Ocean are given in a “White Paper” on the Atlantic Climate Variability (http://geoid.mit.edu/accp/avehtml.html) and elsewhere (e.g., http:// www.clivar.ucar.edu/hp.html). Some of that material may be also viewed at the ACVE web page (http:// www.ldeo.columbia.edu/~visbeck/acve/). See also (http://www.cgd.ucar.edu/cas) and the international CLIVAR documents (http://www.clivar.ucar.edu/hp.html) D1-D3.

3 a)

b)

Figure 2.1 — a) Sea Level Pressure projected on the NAO index (courtesy of J. Hurrell). b) The NAO index from 1864 to 1996 defined as the difference in normalized pressure between Lisbon and Stykkisholmur, for the winter months, December-March (Hurrell, 1995).

SLP difference between Iceland and the Azores, the seasonal variations in the NAO (Barnston and from 1864 to the present day is plotted in Figure Livezey; 1987), or the possibility that the centers 2.1b. Positive values indicate stronger-than-av- of the actual pattern may not overlap these loca- erage middle-latitude westerly winds. Such a tions. simple index is chosen because of the availabil- Paleo indicators of the NAO suggest that it ity of data - it clearly cannot accurately capture has been coherent over long periods of the 700-

4 year stable isotope record from the GISP-2 ice tic Ocean. This “Arctic Oscillation” - the AO - is core from central Greenland (White et al., 1996). an equivalent barotropic, zonally symmetric struc- Similar evidence emerges from the tree-ring data ture which couples the troposphere with the strato- (Cook et al; 1998). H. Wanner (personal commu- sphere (Perlwitz and Graf; 1995) and modulates nication, 1998), however, reconstructs the NAO the strength of the polar vortex, as in the zonal from proxy and documentary sources and hypoth- index of Rossby et al (1939). Because the energy esizes that decadal-scale cold phases during the density of the mode remains high well into the Late Maunder Minimum (1675-1715) were re- stratosphere, it has been suggested that the NAO lated to successive negative NAO events. If the could be modulated through its interaction with main forcing mechanism of the NAO were due the polar vortex and hence be sensitive to an- to its interaction with synoptic-scale eddies, then thropogenic influences in the stratosphere. Cause- one might expect its frequency spectrum to be and-effect within the stratosphere are not yet white. The spectrum of the wintertime NAO in- sorted out; one extreme view is the “Charney- dex is slightly red (power increases with period) Drazin” picture of upward propagating energy with an indication of enhanced power in some from the energetic winter troposphere, with the frequency bands (though not large enough to be stratosphere both conditioning the waveguide, and considered statistically significant; Wunsch, turning the energy into a breaking planetary wave. 1998). However, frequency domain analyses Another is that an increasingly strong polar strato- (Mann and Park, 1996; Tourre et al., 1998), sug- spheric vortex could actively excite the tropo- gests that these relatively weak spectral peaks spheric NAO. relate to basin-wide coherent phenomena and so perhaps are real. Thompson and Wallace (1998) 2.1a — Covarying patterns in the ocean have drawn attention to the fact that the NAO There are clear indications that the North may be the regional manifestation of a hemi- Atlantic Ocean varies significantly with the spheric mode of variability centered over the Arc- overlying atmosphere. Expectation of an im-

60oN

30oN

0o

30oS Correlation Wintertemperature - NAO

-0.4-0.2 0 0.2 0.4 60oS 80oW 40oW 0o 40oE 80oE

Figure 2.2 — Regression between winter SST and land surface temperatures and the NAO index (courtesy of M. Visbeck and H. Cullen)

5 print of the changing atmosphere on the ocean, spreading of SST anomalies along the path of the and perhaps a coupling between climate change Gulf Stream and North Atlantic Current is of par- signals in the two fluids, follows from several ticular interest (Sutton and Allen, 1997). Here related points: winter SST anomalies born in the western sub- First, changes in the NAO, are reflected in tropical gyre are found to propagate eastward with marked changes in surface wind stress, air-sea a transit time of a decade or so. Subsurface ocean heat exchange, and the evaporation-precipitation observations provide an even clearer depiction balance (Cayan, 1992a) and sea-ice. Heat flux of long-term climate variability because the ef- anomalies directly drive month-to-month SST fect of the annual cycle and month-to-month vari- tendencies in the North Atlantic (Cayan, 1992b). ability in the atmospheric circulation decays rap- Figure 2.2 shows the regression of SST on the idly with depth. A large freshening event (dubbed NAO index during winter. It reveals a tri-polar the Great Salinity Anomaly, GSA) occurred in pattern - a cold subpolar region, warmth in middle the sub-polar gyre in the early 1970s (Dickson et latitudes and a cold region between the equator al., 1988). Observations suggest it may have been and 30°N. This is also the leading pattern of SST linked to coherent changes in the overlying at- variability during winter. Its emergence is con- mospheric circulation and an extreme manifesta- sistent with the spatial form of the anomalous tion of the quasi-decadal fluctuations (Reverdin surface fluxes associated with the NAO pattern, et al., 1997; Belkin et al; 1998). The GSA event as pointed out by Cayan (1992a, b). It appears followed the deep reaching changes in tempera- that the strength of the correlation increases when ture and salinity in the North Atlantic, from the the NAO index leads SST indicating that SST is late 1950s to the early 1970s (Levitus, 1989). responding to atmospheric forcing on monthly These changes coincide with a multi-decadal fluc- time scales (Frankignoul 1985; Battisti et al 1995; tuation of SST and atmospheric conditions Delworth; 1997). Changes in sea-ice cover in both (Kushnir, 1994; Hansen and Bezdek, 1996). Vari- the Labrador and Greenland Seas are well corre- ability in the temperature of the upper subtropi- lated with the NAO, with a particularly strong cal ocean has also been linked with fluctuations atmospheric feedback. in the NAO index (Molinari et al., 1997). On Second, the NAO has its largest amplitude in decadal time scales, the intensity of convection winter. The interior ocean has a selective memory tends to see saw between the Labrador Sea and for winter conditions, both in the subduction of the Greenland-Iceland Seas (Dickson et al., 1996) winter conditions in to the thermocline (Stommel, orchestrated by fluctuations in the NAO. The link 1979; Marshall et al, 1993; Williams, et al., 1995), between propagating SST anomalies and the prop- and in the seasonal sequestering of winter condi- erties of water (McCartney et al., 1996) mixed at tions from below the seasonal thermocline. Third, the end of winter, shows that the end-product of the ocean acts as an integrator of the high-fre- the transformation process, Labrador Sea Water quency forcing acting on it. If an SST anomaly is (LSW), underwent extended warming and cool- ‘tied’ to a deep thermal anomaly, which is re-ex- ing trends consistent with the phase of the NAO. posed each winter (Alexander and Deser, 1995), then the SST anomaly can re-emerge year after 2.2 — Tropical Atlantic Variability year and so might have a greater ability to alter, The dominant low-frequency climate phe- through anomalous fluxes, the overlying atmo- nomenon in the tropical Atlantic is the covarying sphere. Sea surface temperature observations fluctuation of tropical SST and trade winds display a myriad of long-term (interannual and (Nobre and Shukla, 1996, and Figure 2.3a). The decadal) responses. The pattern of SST variabil- time series of the joint SST and wind pattern (Fig- ity is indicative of the ability of the ocean to store ure 2.3b) clearly contains long-term components heat locally and move it around (Deser and and a noticeable imprint of decadal variability. Blackmon, 1993; Hansen and Bezdek, 1996). The These fluctuations exhibit a pattern of basin-wide

6 SST anomalies straddling the mean position of than the Pacific, and do not appear to be consis- the ITCZ, with concomitant changes in trade wind tently sustained (Zebiak, 1993). However, they intensity. The Northern Hemisphere trades, and do exert some influence over the adjacent land the associated cross equatorial flow, seem to de- masses. pend sensitively on SST. Very little wind vari- The Atlantic sector clearly responds to the ability is found south of the equator. Weaker than Pacific El Niño although the relative importance normal trades are associated with positive local of this remote forcing in tropical Atlantic vari- SST anomalies, and vice versa (Nobre and Shukla, ability is still unresolved. (Hastenrath et al., 1987; 1996; Enfield and Mayer, 1997). The northern Hameed et al., 1993; Nobre and Shukla, 1996; trades are also affected by SST changes south of Enfield and Mayer, 1997). The ENSO influence the equator, such that colder than normal SST in is transferred through the atmosphere, via changes the South Atlantic correspond to weaker than in both the Walker circulation and Northern Hemi- normal trades in the North Atlantic and vice versa. sphere extratropical teleconnections (i.e., the Pa- Early statistical analysis of SST data revealed a dipolar structure (particularly when referenced to meteorological anomalies, Moura and Shukla, 1981) and led investigators to dub the pattern the “tropical Atlantic SST dipole.” However, SST variability north and south of the equator are not correlated (Houghton and Tourre, 1992; Enfield and Mayer, 1997, Mehta, 1998), suggesting that it is the change in the cross-ITCZ SST contrast to which the trades respond. While it is becoming apparent that SST variability on both sides of the Atlantic ITCZ is not well described as a tempo- rally coherent seesaw fluctuation, it does seem that climate variability in the tropical Atlantic is sensitive to the cross equatorial SST differences (Hastenrath and Heller, 1977; Servain, 1991; Ward and Folland, 1991; Folland et al., 1991). The build-up of the North Atlantic SST anomaly lags behind the local change in the wind circula- tion by a month or two (Nobre and Shukla, 1996), much like the lag between the atmosphere and ocean in mid-latitudes (Frankignoul, 1985). How- ever, as the SST anomaly builds up, the Northern Hemisphere trades adjust, so that on seasonal time Figure 2.3 — The dominant pattern of tropical scales the atmosphere feels the SST anomaly Atlantic variability revealed by a joint EOF analysis buildup and responds to it (Nobre and Shukla, of surface wind stress and SST monthly anomalies, September 1963 to August 1987. Spatial structures 1996). A mode of tropical Atlantic atmosphere- of the first EOF mode and the associated time ocean variability has been identified that is akin coefficients are illustrated in (a) and (b), respec- to ENSO in the Pacific (Covey and Hastenrath, tively. Contours represent SST loading values, 1978; Philander, 1986; Zebiak, 1993) and in- interval of 0.1°C. Vectors depict wind-stress volves variations in SST along the equator. These loadings. The unit vector on the lower-left corner represents an easterly wind-stress anomaly of 1.0 perturbations in equatorial winds and SST are not dyn cm-2. The time coefficients have been normal- as dominant as their Pacific counterpart, appar- ized by their own standard deviations (Nobre and ently because the Atlantic basin is much narrower Shukla, 1996).

7 cific North American Pattern, PNA). ENSO is masses, move north within the S. Atlantic sub- thought to particularly effect the western flank of tropical circulation and enter the tropical regime the Northern Hemisphere subtropics, during the largely at the western boundary. There, signifi- winter and spring following the onset of ENSO. cant water mass transformation occurs within the It appears that ENSO effects further east in the energetic equatorial circulations, with much Gulf of Guinea is rather small (Enfield and Mayer, warmer waters emerging northward. The warm- 1997). water link between tropical and northern subtropi- Finally, there exists a link between tropical cal circulations is complex, involving both climate variability and the NAO (see Figure 2.2). strongly varying Ekman transport in the interior Variability of the NAO can change the intensity and highly time-dependent western boundary of the trades in the eastern subtropical North At- current flows. Upon entering the subtropical gyre, lantic and thence SST. Moreover, the NAO itself the northward thermocline water flow is incor- may be sensitive to subtropical SST anomalies porated into the Florida Current/Gulf Stream sys- (S. Jewson, personal communication, 1998, show tem and related interior circulation. The role of that there is a significant model response with and the Atlantic in a world of increasing CO2 emis- without the tripole SST pattern of Figure 2.2, and sions has received increased attention of late. argues that it is the subtropical SST anomalies About 60 percent of the global oceanic CO2 up- which are the most important). It has even been take may take place in the Atlantic sector, a con- suggested (Tourre et al; 1998, Rajagopalan, per- sequence of its intense meridional overturning sonal communication; Robertson et al, personal circulation. However, some model projections communication) that there is a link between the suggest that in only a few decades, Atlantic cli- NAO and tropical SST in the South Atlantic. It mate might radically shift into a different equi- remains to be seen whether the induced changes librium with much reduced MOC. Evidence for in tropical SST can feed back on the extratropi- rapid climate shifts in the past has been found in cal circulation through mechanisms akin to those several paleo records and is attributed to changes acting in the Pacific. in the strength of the MOC. A weaker MOC will result in a reduced poleward oceanic heat trans- 2.3 — Thermohaline circulation and abrupt port and might dramatically reduce oceanic CO2 climate change uptake and more importantly, rapidly cool Eu- The mean circulation of the ocean plays a key rope and the northeastern American continent. role in meridionally transporting water proper- One might rightfully question the accuracy of ties such as heat and freshwater, carbon and nu- such projections. However, the dramatic impact trients. In concert with meridional atmospheric on the western world of such a scenario, even if fluxes, ocean transports balance the earth’s glo- highly unlikely, justifies a concerted effort to bal heat and hydrologic budgets. At 25°N, the monitor the evolving climate in the Atlantic sec- Atlantic’s subtropical gyre circulation carries tor and improve fundamental understanding of some 1.2 PW of heat northward: approximately the coupled climate system. 60 percent of the net poleward ocean flux and 30 percent of the total flux by ocean and atmosphere at this latitude (Hall and Bryden, 1982). This 3. Impacts of Atlantic Climate poleward heat flux is of course intimately associ- Variability ated with the watermass transformations that take place as thermocline waters move north and are 3.1 — The North Atlantic Oscillation ultimately converted by air-sea interaction into • Temperature and global warming cold North Atlantic Deep Water (NADW). Ther- The NAO exerts a dominant influence on mocline waters of the South Atlantic, represent- the wintertime temperatures of the Northern ing some mix of Indian Ocean and Pacific water Hemisphere. Surface air temperature and SST

8 in wide regions across the North Atlantic ba- spheric mean wintertime temperature. Hurrell sin, in eastern North America, the Arctic, (1996) demonstrated that both NAO and Eurasia and the Mediterranean, are signifi- ENSO are linearly related to this climate trend cantly correlated with NAO variability. with NAO dominating a larger land area than Changes in temperature over land (and related ENSO. Thompson and Wallace (1998) argue changes in rainfall and storminess, see below) that an index of the AO is more strongly cor- are of serious consequence to a wide range of related to the temperature trend over Eurasia human activities. than the NAO. Perlwitz and Graf (1995) sug- Changes of more than 1°C in surface tem- gest that the trend in the NAO and the associ- perature can be associated with a one stan- ated trend in wintertime temperature over dard deviation change in the NAO index oc- Eurasia (and other details of the trend pattern) cur over the northwest Atlantic and extend are a manifestation of the anthropogenic from northern Europe across much of Eurasia greenhouse effect. They hypothesize that (Hurrell, 1995; Hurrell and van Loon, 1997, stratospheric increases in CO2 result in en- see Figure 3.1). Temperature changes over hanced radiative cooling of the polar strato- northern Africa, the Middle East and the sphere in winter, leading to a strengthening southeast U.S. are also notable. Of particular of the polar vortex. They then invoke strato- interest is the contribution of the NAO vari- sphere-troposphere interaction to explain the ability to the recent trend in Global/hemi- recent positive trend in the NAO index

Figure 3.1 — Observed surface temperature change associated with a one standard deviation of the NAO index. The regression coefficient was computed for the winters of 1935-1996 (from Hurrell and van Loon, 1997). Contours are in 0.1°C. Dark (light) shading indicate positive (negative) changes.

9 through enhancement of tropospheric station- the past several winters have been among the ary waves. lowest recorded this century, causing eco- • Precipitation and Storms nomic hardship to those industries dependent Changes in mean circulation patterns over on winter snowfall (Beniston, 1997). Severe the North Atlantic are accompanied by pro- drought also prevailed throughout Spain and nounced shifts in the storm tracks and asso- Portugal affecting olive harvests. In contrast, ciated synoptic eddy activity. These effect the increases in wintertime precipitation over transport and convergence of atmospheric Scandinavia was related to recent positive moisture and can be directly tied to changes mass balances in the maritime glaciers of in regional wintertime precipitation (Figure southwest Norway, one of the few regions of 3.2, see also van Loon and Rodgers, 1978; the globe where glaciers have not been re- Lamb and Peppler, 1987; Hurrell, 1995). Drier treating. than normal conditions occur during high Storminess in the midlatitudes is associ- NAO index winters over much of central and ated with the path of synoptic disturbances southern Europe, the northern Mediterranean and their frontal systems. Indeed, this is the countries, and west North Africa. At the same source of the above- indicated changes in time, wetter than normal conditions occur midlatitude rainfall patterns. It is well known from Iceland through Scandinavia. Over that changes in low-frequency patterns are North America the effect of NAO on precipi- associated with changes in storm activity and tation is not strong. However an out of phase shifts in the storm tracks (Lau, 1988; Rogers, relationship between the NAO index and 1990). The NAO is clearly linked to system- snowfall over New England was found. The atic changes in storm tracks over the North- NAO-related changes in precipitation patterns ern Hemisphere from North America to have directly affected different European Eurasia and the Mediterranean (Rogers, economies, in particular because of the long 1997). stretch of consecutive winters in which the There is a connection between the season- NAO continued to intensify. Over the Alps, ally-averaged NAO and low frequency vari- for instance, snow depth and duration over ability within a season (e.g., blocking) over

Figure 3.2 — Precipitation anomalies associated with the NAO; E-P is plotted, computed as a residual of the atmospheric moisture budget using ECMWF global analyses, for high NAO index minus low index winters - see Hurrell, 1995.

10 the Atlantic (see Nakamura, 1996). There is namical processes (Post et al., 1997). Influ- weaker intra-seasonal variability near ences of the NAO are evident among five Greenland and the Labrador Sea and stron- species of ungulates in populations in ger intra-seasonal variability over Europe, Greenland, Canada, the U.S., Great Britain, when the NAO index is positive (and visa- Norway and Finland. versa). The ocean integrates the effects of storms 3.2 — Tropical Atlantic Variability in the form of surface waves, to exhibit a • Rainfall and droughts marked response to long lasting shifts in the Although sea surface temperature anoma- storm climate, as is the case with variability lies in the tropical Atlantic are weaker than related to the NAO. The recent rising trend in those associated with the Pacific El Niño, they the intensity of the NAO has had a profound can cause disastrous climate hazards over the effect on the surface wave climate of the North Americas and Africa. In the tropical Atlantic, Atlantic. The trend in measured wave heights rainfall is governed by the annual migration has been supported by many anecdotal reports of the subtropical high pressure cells on both about increasing wave heights encountered by sides of the equator and the changes in their mariners and North Sea oil rig operators. Re- strength, as well as the north-south swings of cent studies by Bacon and Carter (1993) and the Intertropical Convergence Zone. Many of Kushnir et al. (1997) have tied, in a more the continental regions bordering the tropical quantitative manner, the increase in wave Atlantic experience a sharp seasonal contrast heights to the increase in wintertime stormi- in rainfall where half the year is wet and the ness and mean wind speeds in the North At- other half dry. Some of these regions are lantic during the last 30 years or so. blessed with abundant rainfall, while others • Fisheries and ecosystems are semiarid. It is in these semiarid regions Changes in the NAO have been associ- that a delay in seasonal rainfall or a signifi- ated with a wide range of effects on the ma- cant drop in its total amount can bring seri- rine ecosystem too. This includes changes in ous hardship to their inhabitants. Such is the the production of zooplankton and the distri- situation in sub Saharan Africa (the Sahel, bution of fish (e.g., Fromentin and Planque 15°W-15°E,15-20°N) with its boreal summer 1996; Friedland et al., 1993, Drinkwater, rainfall, and the northeast corner of Brazil 1994). The NAO has been linked to blooms (Nordeste, 35 45°W, 2-10°S) where it rains of the “brown tide” unicellular algae in coastal in the boreal winter. Many meteorologists waters of Eastern Long Island, which devas- have for quite some time been concerned with tate the local commercial scallop fishery the dynamics of climate anomalies in these (LaRoche et al., 1997). two regions. Hastenrath and collaborators Evidence has also recently emerged that, (Hastenrath and Heller, 1977; Hastenrath, in Scandinavia, the NAO influences tempo- 1978; Hastenrath et al., 1984; Hastenrath and ral and spatial variability in timing of plant Greischar, 1993, Moura and Shukla, 1981; growth: the length of the plant growth season Ward and Folland, 1991; Nobre and Shukla, varied by 20 days between extremes of the 1996) found that tropical Atlantic SST distri- NAO index (Post et al., 1997). Through ef- bution and the associated anomalies in sea fects on vegetation and climatic conditions, level pressure and wind are leading factors in the NAO was furthermore observed to influ- determining the anomalies in seasonal ence several aspects of life history and ecol- Nordeste rainfall. These same climatic fac- ogy of terrestrial, large mammalian herbi- tors were also found to be strongly linked with vores, including: phenotypic variation, fecun- rainfall anomalies over the Sahel and west- dity, demographic trends, and population dy- ern tropical Africa (Lamb, 1978; Hastenrath,

11 1984; Folland et al., 1986; Folland et al., Markham and McLain, 1977; Moura and 1991). Shukla, 1981). Dry Nordeste years tend to Rainfall anomalies in the Nordeste dis- occur when SSTs north of the equator are play a wide range of time scales with clear warmer than normal and SSTs south of the decadal components (Figure 3.3a). Such long equator are colder than Normal. Rainfall in time scales are also present in the Sahel rain- subtropical west Africa is more complex but fall time series (not shown). The correlation also displays considerable dependence on a between seasonal rainfall in the Nordeste and similar out of phase variation of SST in the tropical Atlantic rainfall indicates an out of tropical Atlantic (Lamb, 1978; Lough, 1986; phase relationship between SST anomalies in Lamb and Peppler, 1987). It rains more when the ocean regions underlying the Northern and Northern Hemisphere subtropical SST are Southern Hemisphere trade wind region, re- warm, and Southern Hemisphere SST are spectively (Hastenrath and Heller, 1977; cold. Recent attempts to understand the dy-

a) rainfall index (Nordeste) 2

es 1 li 0

anoma -1

-2 1840 50 60 70 80 90 1900 10 20 30 40 1950 60 70 80 90 Year b) rainfall, SLP, and surface vector wind 30N

20

10

EQ

10

20

30S

c) SST and surface vector wind

80W 60W 40W 20W 0 20E 5 m/s

Figure 3.3 — a) Seasonal rainfall anomalies over the Nordeste region of Brazil. b) Correlation of Nordeste rainfall index with rainfall anomaly (shaded), wind vector and sea surface pressure (contours). c) Correlation of Nordeste rainfall index with SST (shaded) and wind vector (courtesy of Mitchell).

12 namics of this asymmetric SST distribution 4. The scientific challenges are described in Carton et al., 1996, Nobre Given the societal, economic and ecological and Shukla, 1996 and Chang et al., 1997. impact of interannual to decadal time scale cli- • Hurricanes mate variability, it is of paramount importance to Interannual variability of the seasonal fre- improve our dynamical understanding of climate quency of Atlantic hurricanes is related to, fluctuations in and over the Atlantic Ocean, to amongst other factors, the sign and amplitude study the predictability of these fluctuations, and of the SST anomaly in the subtropics (Gray, to quantify their impact on regional weather pat- 1990; Shaeffer, 1996). Moreover, Gray (1990) terns and short time-scale climate variability. Sev- also suggests that tropical cyclone intensity eral potentially important mechanisms of basin- is correlated with the multi-decadal fluctua- scale atmosphere-ocean interaction and coupling tion of North Atlantic SST (Kushnir, 1994). have been proposed. These physical processes In fact, Gray anticipates a return to a stormier need to be explored and dynamical hypotheses tropical climate, similar to that of the 1950s, advanced and tested within the Atlantic Climate when the North Atlantic finally recovers from Variability Experiment. In particular, data analy- its recent cold phase. Such an increase would sis and the collection of new observations must threaten the densely populated coastal regions be guided by theoretical and modeling work that in the Gulf of Mexico and the North Ameri- explore mechanisms underlying the coupled sys- can Atlantic seaboard which have developed tem. markedly during the relatively benign 1970s We will first review the basic hypothesis for and 1980s. coupled Atlantic Climate Variability and then identify a set of related key issues that need to be 3.3 — Abrupt Climate Change addressed during ACVE. We then focus on the The possibility of a major change in the oce- component parts of the atmosphere-ocean system anic meridional overturning circulation under to reveal their inherent nature in isolation, and greenhouse forcing has been proposed by finally proceed to a consideration of the coupled Manabe, Stouffer, Broecker, and Rahmstorff, in behavior both regionally and on large scale. separate works based on a variety of coupled GCMs, and ocean-only GCMs. “Abrupt climate 4.1 — Hypotheses for Coupled Atlantic Cli- change” has received a particular burst of atten- mate Variability tion in the literature and the popular press through On interannual to decadal time scales there the suggestion that the increasing freshwater are preferred modes of variability of the coupled transports and warmer atmospheric temperatures atmosphere-ocean system. The basic hypothesis in a greenhouse world will blanket the high-lati- which needs to be explored during ACVE is: The tude oceans with a buoyant layer, and cause in- preferred modes of climate variability are signifi- stability or even collapse of the MOC. Model for- cantly influenced by ocean-atmosphere interac- mulation of the air-sea fluxes of heat and mois- tions local to the Atlantic and distinct from the ture, the sites of deep convection in simulations, modes of either the ocean and atmosphere in iso- and many other dynamical aspects of the models lation. For this hypothesis to hold, SST anoma- are in question. Yet one cannot ignore the possi- lies must exert a significant influence on atmo- bility that increased fresh-water input and atmo- spheric flows. In particular, the atmospheric re- spheric high-latitude temperature could suppress sponse to SST anomalies must significantly in- the MOC. The impact, particularly on Eurasian fluence the flux of heat, moisture and momen- climate is potentially severe. The fairly low prob- tum at the ocean surface. ability is thus multiplied by a large potential im- In contrast, the null hypothesis is: The atmo- pact, and thus demands attention. sphere has internal modes of variability with a somewhat red spectrum. The ocean integrates this

13 essentially stochastic forcing, responding more rily driven by the surface fluxes associated with robustly to the lower frequencies and thereby red- internal variability of the atmosphere that reflects dening the response spectrum. Interdecadal vari- the spatial pattern of the NAO. The NAO not only ability then is either a result of decade to decade provides a dominant forcing to middle/high lati- changes in atmospheric boundary conditions due tude SST variability and meridional overturning to anthropogenic, solar, or volcanic effects, or, cell (MOC) in the Atlantic ocean, but also may more likely, represents the low frequency tail of play a significant role in exciting variability in a red spectrum. the tropics (Figure 4.1). In contrast to heating in The major challenge of ACVE is to determine higher latitudes, convective heating in the trop- the validity of these two hypotheses by means of ics extends much deeper into the atmosphere, observations, numerical simulations and theoreti- making the tropical atmosphere more responsive cal studies. Because the source of climate vari- to SST forcing. Thus, active ocean-atmosphere ability can be internal (natural) and external due coupling is more likely to occur in the lower lati- to anthropogenic forcing, attempts to test the hy- tudes. potheses must take these factors into the consid- A coupled mechanism has been proposed for eration. the tropical Atlantic whereby the cross-equato- The strength of active ocean-atmosphere cou- rial SST gradient plays a crucial role for the posi- pling is likely to vary spatially. In middle/high tion of the ITCZ. This in turn will influence the latitudes, the diabatic heating is confined within strength of the trade winds and feedback to tropi- the lower boundary of the atmosphere, making if cal SST anomalies (Figure 4.2). Moreover, the difficult to thermally force the NAO directly at variability of the ITCZ may have an important the ocean surface. Therefore, weak air-sea feed- remote impact on the NAO, possibly through re- backs are anticipated. The ocean, then, is prima- arranging the Hadley circulation (Figure 4.1). It

Figure 4.1 — Mechanisms of Atlantic Variability (courtesy of J. Marshall).

14 Figure 4.2 — Mechanisms of Tropical Atlantic Variability (courtesy of P. Chang).

is, therefore, plausible that the NAO and TAV are 4.2 — Modes of Variability interrelated. Exploring interrelationships between the NAO and TAV will be one of the major focus THE UNCOUPLED SYSTEM of ACVE. Additionally, TAV is subject to remote influ- The Atmosphere ence of the ENSO related variability in the tropi- It is well known that atmospheric GCMs, with cal Pacific (Figure 4.2). The remote influence may prescribed SST, display NAO-like fluctuations. take place either through an anomalous walker Although no single theoretical explanation for the circulation due to an eastward shift of the atmo- NAO is widely accepted, it seems that its funda- spheric convection, or through North Hemispheric mental dynamics arise from atmospheric pro- atmosphere teleconnections (i.e., PNA). How cesses. There is probably no single cause; rather these two processes are related is currently not the NAO appears to be a preferred mode of the clear. atmosphere that can be excited in a number of different ways, its position and spatial structure

15 linked to the basic characteristics of the large- eling experiments yield perplexing results which scale atmospheric circulation. However, it is im- seem to depend on model properties, model inte- portant to determine whether atmospheric pro- gration methodology, and the geographical and cesses alone can lead to long-term, decadal cli- seasonal distribution of the SST anomalies. Re- mate variability in the Atlantic Basin with ob- solving this issue is crucial to the understanding served amplitudes, phases and scales. Distin- of decadal variability in the North Atlantic (and guishing between coupled and uncoupled modes elsewhere) and to making progress in our studies of variability is observationally difficult. Research of the coupled interaction. The atmospheric re- specifically designed to explore the dynamics of sponse to the evolving pattern of low-latitude SST NAO-type variability is required. These investi- in different basins of the oceans remains an im- gations should address the following questions portant research topic, in particular the atmo- regarding various aspects of planetary wave dy- spheric linkages between the tropical Atlantic and namics governing tropospheric quasi-stationary other ocean basins. Key subjects include the at- anomalies over the Atlantic: mospheric response to dipole-like SST perturba- • What is the relative importance of topographic tions. Atmospheric studies should also be de- and thermal forcing of the extra-tropical plan- signed to understand the remote influences of the etary wave field over the Atlantic? Pacific ENSO on the tropical Atlantic variability. • Does the Atlantic-sector atmosphere display Questions that need to be answered regard- sensitivity to atmospheric deep convective ing midlatitude ocean-atmosphere coupling are: activity within the basin, particularly in the • What is the sensitivity of the atmosphere to tropics? the location and amplitude of SST anomalies • What is the role of land surface processes and and subsurface thermal anomalies? topography on both sides of the basin in ex- • What are the spatial patterns of the citing and regulating NAO variability and its atmosphere’s response to mid-latitude SST variants? anomalies. What are the atmospheric dynami- • What is the relevance of instability mecha- cal mechanisms responsible for setting it up nisms (barotropic/baroclinic) and synoptic- these patterns? Is the atmospheric response scale excitation and feedback on Atlantic ex- linear, or do the patterns depend on the polar- tra-tropical low-frequency atmospheric vari- ity of the SST anomaly? ability? • How does the response of the atmosphere • How do the properties of horizontal and ver- evolve in time? Does the response depend on tical propagation (stratospheric connections), the season, and on the time scale of the oce- as well as non-linear atmospheric interactions, anic thermal anomalies? affect NAO and other teleconnections? • Is the observed near-surface circulation pat- Could they lead to multi-year variability as tern over the tropical Atlantic Ocean repro- an internal atmospheric phenomenon? ducible by forcing an AGCM with a dipole- The response of the atmosphere to sea-sur- like SST perturbation? If so, what are the un- face temperature anomalies has been the subject derlying dynamics determining the circula- of many observational and modeling studies go- tion pattern? How does the modified circula- ing back to the very first atmospheric GCMs. A tion pattern affect the surface heat flux? broad consensus between observations, theory, • How important are continental heat sources and models exists with regard to the response of to the tropical Atlantic circulation? the atmosphere to tropical SST anomalies, in par- • What are the atmospheric connections be- ticular over the Pacific Basin. However, there is tween tropical and extra-tropical climate vari- no clear theoretical understanding of, or consen- ability in the Atlantic? What influence do sus about, the response of the atmosphere to SST tropical Atlantic SST anomalies have on ex- anomalies in mid-latitudes. Indeed, GCM mod- tra-tropical atmospheric variability?

16 • Through what physical mechanisms does the ity. It is important to resolve the spatial and tem- Pacific Ocean exert influence on the tropical poral characteristics of such variability, their pa- Atlantic? rameter sensitivity, and model dependence. Com- • Why does the ENSO influence appear to be parisons with observations are crucial in deter- stronger in the northern tropical Atlantic than mining the degree of realism in the model simu- in the southern tropical Atlantic? lations. As indicated above, processes that lead to SST variability should be a primary focus, but The Ocean the model-observation comparisons should en- It has been shown through theory and mod- compass other measures of the dynamical sys- els that because of its large thermal heat capac- tem to ensure a complete description of the vari- ity, the ocean can respond locally to high-fre- ability. quency stochastic atmospheric forcing on sea- Studies are required that enhance understand- sonal and inter-annual time-scales (Hasselmann, ing of oceanic processes that influence SST in 1976; Frankignoul, 1985). Decadal variability in the tropics, including the role of gyre exchange the ocean surely involves non-local processes as processes. Specific questions that must be an- well, such as the vertical and horizontal exchange swered include: of water masses. It is thus important to explore • What is the role of ocean dynamics in on- and the following issues: off-equatorial SST variability? • What are the preferred patterns of oceanic • What is the relative importance of horizontal variability and their characteristic time scales? heat transport and vertical mixing/subduction What dynamical mechanisms govern them? processes? • How do anomalies propagate through the • How significant are cross-equatorial and ocean? What is the oceanic teleconnectivity? tropical-subtropical gyre exchanges? What is • What changes in horizontal gyre transports the relative importance of western boundary and meridional overturning intensity result current transports and interior Ekman trans- from random and sustained surface forcing, ports? possibly orchestrated by slowly propagating • Do oceanic waves play an important role in internal Rossby waves or advection of strati- creating or modulating SST fluctuations in the fication anomalies? tropical and mid-latitude Atlantic Ocean? • Is the Atlantic much affected by conditions imposed at its northern and southern bound- THE COUPLED SYSTEM aries with the Arctic, and Southern Oceans respectively. In studying the combined ocean-atmosphere sys- • What are the relationships between sub-sur- tem we need to identify what modes of variabil- face ocean thermal anomalies and SST? ity owe their existence to active coupling between In most modeling studies of ocean variabil- the two systems and could not otherwise be mani- ity, a simplified surface interaction is assumed fested. This is clearly a question that can only be where the heat and moisture exchange at the sur- addressed through a combination of modeling and face is either prescribed or derived from an at- observational studies: modeling/theory to offer mosphere that adjusts to the changes in SST in and test plausible physical mechanisms, obser- prescribed ways. Such studies of the forced re- vations to examine evidence for their authentic- sponse must be complemented by investigations ity. In the tropical Atlantic system, coupled model of internal (free) modes of ocean variability. This experiments must also be performed to distinguish avenue is particularly relevant to the hypothesis the relative importance of external forcing and that decadal variability originates in the ocean. local air-sea processes to variability there. Exter- Studies of both forced and free variability must nal forcing must include the deterministic forc- be carried out with models of varying complex- ing associated with climate anomalies (e.g.,

17 ENSO) teleconnected into the Atlantic basin, and evidence exists that SST and surface salinity the stochastic forcing associated with dynamical (SSS) anomalies propagate cyclonically around instabilities internal to either the atmosphere or the gyre and may be important in modulating the ocean. The following questions must be ad- intensity of convection and the sea-saw between dressed: convective activity in the Labrador and GIN Seas. • What are the coupled mechanisms that can The following questions need to be addressed: lead to variability that exhibits characteris- • What are the feedbacks and coupling mecha- tics of the observed variability in the North nisms that maintain these anomalies? Atlantic? • How is the formation of persistent sub-polar • Is the coupling inherently regional or is it re- SLP-SST anomalies linked to local and re- mote and/or basin-scale? If the coupling is mote (e.g., Arctic, subtropics) anomalies in regional, how does it depend on local proper- surface properties (such as ice, heat fluxes and ties? precipitation) and subsurface properties (such • What is the role of remote atmospheric and as mixed layer depth and intermediate/ deep ocean behavior in shaping the spatial and tem- water formation rates)? poral characteristics of the variability? • Do significant upwelling branches of the • What is the “minimum physics” needed to MOC exist which are local to, rather than re- capture realistic coupled behavior? To what mote from, the sinking branch, and how do degree is this physics adequate to describe and they effect SST and SSS? predict such phenomena? • What is the role of the annual cycle in quasi- • How do these simple mechanisms manifest decadal variability? themselves in more complex and more real- • What is the role of sea ice? istic models? Can particular coupled mecha- • How do the exchanges between the Arctic nisms be identified in these models? How do Ocean and the North Atlantic affect North they interact or compete with one another? Atlantic climate variability? • What new observation and analysis methods are required to test coupled variability found (b) Subtropical/Subpolar Gyre interactions in simple or complex models? Prominent SST anomalies tend to form along • Is the dipole-like SST variability better de- the North Atlantic Current at the boundary be- scribed as a self-sustained oscillator due to tween the subtropical and subpolar gyres. They unstable air-sea interactions, or as a stable are of particular interest because they have been dynamical system driven by stochastic pro- linked to changes in climate over Western Eu- cesses? rope. As reviewed above, it has been suggested • What is the relationship between the Atlantic that quasi-decadal anomalies in this region are dipole, NAO and equatorial modes? the result of an inherently coupled phenomenon. • How do these patterns of climate variability Coupled model studies indicate that this region interact with the seasonal cycle? is a sensitive indicator of, and may be a driver of, multi-decadal variability of the thermohaline cir- 4.3 — Regional aspects culation. Moreover, coupled models also suggest (a) Subpolar gyre and Northern Marginal Seas that on decadal time scales, upper-ocean gyre Important interactions occur between cold- dynamics play an important role. In this scenario, fresh and warm-salty water masses in the Labra- the life cycle of SST anomalies is determined by dor and the Greenland-Iceland-Norwegian Seas, the interplay between the local response of the where winter-season convection forms interme- ocean to variations in surface heat flux and the diate and deep water. Large, quasi-decadal and advection of heat by the gyre circulation. The lat- multi-decadal SST anomalies, coherent with SLP ter counteracts the former, but lagged in time. The anomalies, are observed in these regions. Strong atmosphere, reacting to the change in SST, is the

18 bridge connecting the local and non-local re- mospheric circulation. This process was likewise sponse of the ocean. In studying this mechanism, identified as a central element for the Pacific the following questions should be addressed: Basin Extended Climate Study. • What are the essential dynamics of the ocean Two possible corollaries of are worth noting. and atmosphere required to capture the inter- First, because there is a net poleward transport of action? Are the mechanisms that have been upper-ocean water associated with the Atlantic’s proposed robust? Can they be reproduced in thermohaline circulation, nearly all of the ther- models of various types? mocline water that feeds the equatorial undercur- • What is the role of the seasonal cycle in the rent comes most directly from the southern hemi- interaction? sphere, a conclusion that is supported both by • How is the interaction affected by variability observations and models. Thus, it is more likely in the other parts of the Atlantic, and how does that variability of the southern STC leads to sur- it affect variability elsewhere? face anomalies in the equatorial cold tongue. Sec- ond, because this process leads to a change in the (c) Subtropical Gyre equatorial upwelling, it initially causes an SST In the northeastern part of the subtropical anomaly pattern that is essentially symmetric gyres much of the upper ocean thermocline ven- about the equator. It is not clear whether subse- tilation takes place. Variability of the sea-surface quent ocean-atmosphere interactions can gener- properties in those subduction zones can subduct ate asymmetric, dipole-like anomaly patterns and beneath the surface layer. Now shielded from the cause a shift in the position of the ITCZ. The most direct air-sea interaction they can propagate important questions are: largely undisturbed within the gyre circulation • Can subtropical thermocline water mass throughout the subtropics/tropics unless exposed anomalies indeed transit to the tropics and to surface entrainment zones in upwelling and the influence surface conditions in the equatorial western boundary current regions. Specific issues upwelling zones? are: • Do feedback mechanisms exist between the • Can subducted water mass anomalies survive equatorial surface ocean, atmosphere and many years subsurface to finally reemerge and the STCs to sustain decadal time scale modify sea surface conditions? oscillations? • How important are changes in the circulation relative to changes in the surface fluxes in the (e) Tropical Ocean generation of to subducted temperature Ocean-atmosphere interactions are gener- anomalies? ally expected to be important in the tropics because 1) the atmosphere is sensitive to (d) Tropical/Subtropical Gyre Interaction changes in the SSTs and 2) the adjustment time Decadal TAV may be driven remotely from scales of the tropical oceans are relatively short. the subtropics through variation of the shallow The high correlation between the tropical meridional Subtropical Cells (STCs) that carry Atlantic SSTs and rainfall variability in the subtropical thermocline water into the tropics. neighboring continents reflects the atmosphere- One hypothesized process involves zonal-wind ocean coupling in the region. Modeling studies anomalies about the tropical-subtropical bound- further suggest that local air-sea interactions in ary latitudes: the winds that set the strength of the tropical Atlantic can lead to a decadal TAV the STCs. STC modulation eventually alters the and an ENSO-like interannual mode. For the amount of cold water that flows into the eastern decadal TAV mode, the dominant air-sea feed- equatorial Atlantic. This in turn may modify the back appears to be between interhemispheric extent/intensity of the equatorial cold tongue, the SST gradient and wind-induced latent heat flux air-sea heat flux and ultimately, the tropical at- in the tropical North and South Atlantic, al-

19 though two other related feedback mechanisms The philosophy of the Atlantic Climate Vari- may also contribute to air-sea interactions in the ability Experiment is to take a basin-wide per- tropical Atlantic: spective in the study of coupled ocean-atmo- i) weakening of the trades in the warm SST spheric variability in the Atlantic-sector on region, which decreases entrainment into interannual to decadal time scales. Our proposed the surface mixed layer thereby causing research program has been jointly developed be- even warmer SST; and tween U.S. and European scientists. Its primary ii) weakening of the trades along the North goal is to understand the mechanisms responsible African coast, which reduces coastal up- for Atlantic Climate Variability and quantitatively welling thereby warming SST in the north- test, and thereby improve, atmosphere-ocean eastern tropical Atlantic. models that describe them. We must determine For the ENSO-like mode a dynamic interaction whether variability simulated in models is fun- between the equatorial trades and SST (the damentally coupled, whether one component Bjerknes mechanism) is believed to be operat- drives variability in the other, whether modeled ing to a certain extent. Although the prelimi- variability bears any relation to observed variabil- nary coupled model studies point to the poten- ity, and whether modes of variability once iden- tial importance of regional air-sea feedbacks in tified are predictable. The improvement of atmo- tropical Atlantic, many fundamental issues spheric and oceanic circulation models, especially remain unsolved. For example, pertinent ques- boundary layer formulations, is a central part of tions are: ACVE. • To what extent do local air-sea feedbacks Our specific objectives are to: contribute to the decadal hemispheric SST • Describe and model coupled atmosphere- gradient variability and ENSO-like ocean interactions in the Atlantic Sector and interannual variability in the tropical Atlan- quantify their influences on the regional and tic? global climate system, and investigate their • How much tropical Atlantic SST variability predictability. is attributable to “remote” influences of • Understand and model the response of the ENSO in the Pacific and NAO in the North atmosphere to air-sea exchanges in the Atlan- Atlantic? tic (in particular the anomalous exchanges • What are the oceanic and atmospheric associated with observed tropical and mid- processes responsible for the coupling in latitude SST anomalies) as a function of hori- the tropical Atlantic? In particular, what is zontal pattern, scale, amplitude and frequency. the relationship between the off-equatorial • Provide an improved description and under- SST anomaly and surface heat flux anoma- standing of the ocean’s response and poten- lies? tial feedback onto mid-latitude, NAO-like atmospheric forcing as a function of ampli- 5. A Proposed Research Program tude and frequency, and the associated time- The structure of this chapter is as follows: varying fluxes of heat, fresh water, carbon 5.1 Elements of the basin scale ACVE dioxide and other gases, momentum and en- 5.2 Analysis of Historical and Proxy Data ergy. 5.3 Sustained Observations on the Basin Scale • Assemble a data set that may be used to test 5.4 Observing the meridional overturning cir- and improve models of coupled Atlantic cli- culation mate variability. 5.5 Modeling, Theory and Ocean State Estima- • Investigate the sensitivity of the meridional tion overturning circulation to changes in surface 5.6 Regional aspects of ACVE: Process Stud- forcing and assess the likelihood of abrupt ies climate change.

20 Some of the processes involved in climatic between the observed ice-core records and cli- modulation of the North Atlantic will require a mate variability. broad-scale study of the entire gyres or basins in Addressing the key scientific issues discussed the context of the global atmosphere; other top- in Section 4 requires an interdisciplinary research ics can be examined in focused process experi- effort. Unfortunately such efforts are too often ments. Some investigations must extend over a limited or hindered by the availability of, or easy decade, but clever treatment of existing histori- access to, historical data. Investigators are fre- cal observations can also provide an important quently unaware of existing data sets, handi- long-time-scale perspective. capped by data access difficulties, or unaware of documented problems with certain observations. 5.1 — Elements of the basin scale ACVE New, updated or improved data sets are also dif- Key basin-scale elements of ACVE will be: ficult to identify and locate. We suggest that a • analysis of historical and proxy data, major step forward would be to establish an At- • basin scale observational networks, lantic Climate Variability Data Center to improve • modeling, theory and state estimation (data data accessibility and documentation, and foster assimilation), “data archeology”. • process studies embedded in the observa- The primary purpose of the data center will tional networks. be to catalog and provide easy access to basic observed historical data sets for interdisciplinary 5.2 — Analysis of Historical and Proxy Data studies of Atlantic climate variability. The data Continued analysis of historical instrumental center could be as simple as a world wide web and proxy data records is an essential component. page, pointing toward existing, publicaly avail- Only through such study can the spatial and tem- able data, but maintained by a person or group of poral properties of recurring interannual to people with a basic understanding of the goals of decadal climate variability be characterized. In the ACVE. To enhance productivity, the data cen- hand with modeling approaches, data analysis will ter should utilize the knowledge of “expert us- be essential to evaluate emerging hypotheses con- ers’’, soliciting advice on the advantages and dis- cerning the mechanisms of Atlantic climate vari- advantages of individual data sets from the re- ability. Data analysis must also play a key role in searchers who process and/or use them. Such in- optimizing the design and assessing the expected formation could be included as meta data, orga- errors of future observing networks. nized by data set name and linked to the data sets Proxy records of Atlantic climate variability themselves. are provided by tree ring and ice core stable iso- Recommendations: tope chronologies and marine and lake sediment • Creation of an Atlantic Climate Variability records. The use of such records should continue Data Center to improve data accessibility and to be actively pursued. Such data provide the only documentation, and foster “data archeology” window on climate fluctuations before the instru- • Close collaboration with activities of the mental period. They furthermore are particularly paleoclimate community organized under attractive for studying continental climate and its PAGES relation to oceanic conditions. Interpretation of proxy climate records is an area where models, 5.3 — Sustained Observations on the Basin suitably equipped with tracer or other specialized Scale capabilities, will be of substantial utility. For ex- Two categories of measurements are needed ample, atmospheric models with isotopic water- to form the observational basis of ACVE: (i) sus- vapor/tracer capabilities can be used to syntheti- tained, Atlantic-wide observing systems in the cally generate ice core records, potentially lead- atmosphere and ocean and (ii) shorter-term pilot ing to a better understanding of the relationship network studies and process experiments. The

21 design of the pilot and sustained observing sys- A. Atmospheric Observing Networks tems must be guided by examinations of the his- Our general knowledge of seasonal-to- torical record, in combination with numerical decadal atmospheric variability in the Atlantic model studies. region will be based primarily on analyses from The goal of the sustained observing systems the operational weather-prediction centers and on is to provide comprehensive descriptions of the re-analyses of historical data. The backbone of oceanic and atmospheric circulations believed to such work is the World Weather Watch (WWW/ be associated with Atlantic Climate Variability. GCOS) upper-air sounding network on the pe- This includes in particular the upper-ocean mixed riphery of the Atlantic and those few island sta- layer, the atmospheric boundary layer, and the tions. Surface and upper-air observations from coupling between them. In the off-equatorial re- ships and commercial airliners provide additional gions where SST exhibits prominent decadal data over the open ocean, largely concentrated variation, it is important to measure the time-vary- along shipping lanes and air corridors. ing air-sea heat flux, and establish the atmospheric Although since the 1970s satellite observa- and oceanic processes that determine it. It is also tions provide a significant constraint on the analy- important to measure subsurface thermal struc- ses, the Atlantic sector remains a data-void area. tures, since ocean circulation and SST anomalies This is demonstrated in Figure 5.1 which shows are typically associated with large changes in ther- the WMO’s global upper-air network that is be- mocline depth. In the following we outline ex- ing proposed under the Global Climate Observ- plicitly which measurements will be the central ing Systems (GCOS). Figure 5.2 shows the ini- elements for testing hypotheses of coupled cli- tial GCOS Surface Network (GSN). Regrettably, mate variability in the Atlantic sector. the WWW upper-air network has deteriorated

Figure 5.1 — The initial GCOS Upper-Air Network (GUAN) (courtesy of GCOS Joint Planning Office, 1998).

22 since the 1960s in terms of the number of sta- able cost are underway. Atmospheric profilers and tions and the frequency of observations. The At- containerized radiosonde launchers are mature lantic basin data-void areas will seriously impact field instruments that can be operated from re- ACVE process studies in which detailed knowl- search and commercial ships of opportunity. More edge of the atmosphere’s vertical structure is re- systematic measurements may become possible quired. The lack of in situ soundings over data- with autonomous aircraft and robotic balloon plat- sparse regions also hinders efforts at the opera- forms, which under development and will prob- tional weather analysis and climate reanalysis ably become operational during ACVE. centers to identify and remove regional biases in It is important to realize that there is increas- their data assimilation and analysis systems. ing evidence that atmospheric radiative transfer Attempts to calibrate satellite instruments in equations transmit too much energy through the field and to develop improved remote sens- clouds. In the TOGA COARE region for example, ing retrieval algorithms benefit fundamentally models show as much as a 30 W/m2 bias in cloudy from new high-quality in situ soundings. Satel- conditions compared to observations. Globally, lite retrievals of the vertical temperature and hu- the average bias in insolation may be about 10 midity structure are unable to resolve the details percent, which, if not corrected, has serious con- of boundary layer structure, and the weighting sequences for ocean climate models. Technical functions of the radiative flux channels measured developments are underway to acquire direct flux by the instruments place fundamental limits on measurements autonomously from volunteer ob- the maximum resolution that can be expected. serving ships (VOS). In combination with the Wind information from space is obtained at only profiling instruments, these measurements will a few levels in the vertical. provide important new information crucial for Efforts to increase atmospheric upper-air sam- enhancing our obviously limited understanding pling in remote marine environments at a reason- about the atmospheric radiation transfer and as-

Figure 5.2 — The initial GCOS Surface Network (GSN) (courtesy of GCOS Joint Planning Office, 1998).

23 sociated differential heating. • availability of upgraded reanalysis products In addition to VOS, well-instrumented sur- from ECMWF and NCEP, and face moorings are recommended at select sites. • the free and open exchange of observational These will return accurate reference data to verify data sets. and/or calibrate surface meteorological fields None of the above can necessarily be taken from models, remote sensing platforms and other for granted but require significant attention to in-situ measurements. Work is in progress to up- assure their continuation. In addition, special at- grade the quality of the surface meteorology and mospheric observations (localized in space and air-sea fluxes on a global basis from the VOS fleet time) are required to address specific issues re- is dependent on such fixed reference sites. The lated to improvements of the parameterizations reference sites will serve as primary standards for of essential boundary layer processes in coupled VOS data quality control and help define regional atmosphere - ocean - land models or for choices of air-sea flux formulae. The present drift- intercomparison of measurements made with dif- ing buoys provide a good example of how in situ ferent kinds of instrumentation. observations are used to calibrate satellite data; in their case they provide the means to calibrate B. Ocean Observing Networks the AVHRR fields of SST. The reference site sur- Permanent upper-ocean networks are needed face moorings will additionally provide surface to document low-latitude anomalies and the gen- wind, humidity, surface short wave and long wave eration and subsequent evolution of higher-lati- radiation, and rainfall estimates. Possible sites tude features such as propagating SST and SSS include the mid-subpolar gyre near historical anomalies. Some of the required information can Ocean Weather Station “Juliet’’ or “Lima’’ and be obtained by high-quality remote sensing data the area east of the Grand Banks, the New En- sets: gland Bight region that experiences polar air out- • Sea surface temperature — SST is one of the breaks, the center of the Azores-Bermuda high fundamental climate parameter which can be and the North Atlantic trade wind region to its observed from space. ACVE will strongly south, the Gulf of Lyons in the Mediterranean Sea, benefit from the operational SST network. and the area southwest of Cape Town in the east- • Sea surface elevation — Altimetry has shown ern South Atlantic. to provide crucial information about changes Recommendations in the geostrophic currents and heat content The primary source of atmospheric observa- as a function of space and time. Continuous tions for ACVE will be the GCOS, which is built altimeter coverage is mandatory in the con- on the observing system that supports operational text of ACVE, since SSH data uniquely pro- numerical weather prediction. This implementa- vide global information about variability in tion plan requires: the ocean flow-field near the surface, but • maintenance of the critically important strongly reflecting interior dynamics. They are ground and space-based components of crucial to interpolate in time and space many GCOS, of the interior measurements which are nec- • the existing observational network to be essarily sparse. optimized and modernized, in particular • Surface winds — Scatterometer data are not enhancing the technologies to recover routinely available. However, when they were atmospheric profile and boundary layer data available they have greatly improved the sur- in remote ocean regions, face wind stress fields and their variability. • continuing improvements in the perfor- • Sea ice — Sea ice concentration is one of the mance of 4-dimensional variational data few global measurements that we have which assimilation schemes that operate in near- directly relates to upper ocean fresh water. real-time, SSMI data also allow to estimate sea ice drift

24 velocities. water. At present, PIRATA is only structured In situ networks capable to capture climate as a pilot project and operated jointly between features and their development in time will con- Brazil, France and the United States with sist of the following elements, some of which are funding ending after 2000. It is essential to operational at the time of writing or can be an- augment the PIRATA array with five extra ticipated to be so in the near future: buoys (Figure 5.3) to measure surface fluxes • Tropical moored array — Continuation of the and the upper-ocean thermal structure. A key present PIRATA moored array in the tropical area for improved flux estimates is under the Atlantic with an additional five stations is re- ITCZ where winds are weak. quired to better cover anomaly patterns in the • PALACE float network — A proposal has southeastern and northeastern upwelling re- been made to deploy a global network of gimes as well as the northern tropics. West- PALACE floats, and we embrace this concept ern boundary observations are also needed to for the Atlantic. A deployment of many hun- measure the equatorward flow of thermocline dreds floats, covering both the north and south

Figure 5.3 — A strawman ACVE mooring network. Shown are the present PIRATA surface mooring array in the tropics and its possible enhancement, potential buoy reference stations for improving air-sea flux estimates, current and proposed hydrographic time-series stations to be instrumented with autonomous systems, and sites where transport monitoring arrays might be located.

25 Atlantic, enables the evolution of T, S field to salinity sensors whenever possible. The prob- be mapped autonomously (Figure 5.4). In the lem of floats vacating some areas and clus- tropics, the array will observe the variability tering in others could be overcome in the fu- associated with the southern and northern el- ture by new types of glider floats that can ac- ements of the tropical ‘dipole,’ as well as the tively navigate to maintain position or carry interior structure of the subsurface, out local surveys. equatorward-flowing branch of both the • VOS-XBT lines northern and southern STCs. Floats should 1) Ocean measurements — To complement also populate the northward-flowing branch the float array, particularly in regions char- of the MOC and the paths of the propagating acterized by strong currents and small mid-latitude SST anomalies. Salinity obser- spatial scales, continuation of the volun- vations are particularly important in the sub- teer observing ship XBT program at the polar North Atlantic, so the floats should have WOCE intensity is recommended (Figure

Figure 5.4 — A random realization of a 700 profiling float network. Design studies are needed to determine the number and regional distribution of an efficient float network for ACVE.

26 5.5). Design studies are needed to estab- base. Salinity observations are particularly lish where such high-density sampling is important in the subpolar North Atlantic, required. In the tropics, the TOGA lines and as reference data for floats and remote should be maintained as a complement to sensing instruments globally; the use of the PIRATA and PALACE arrays until any thermosalinographs on volunteer ships redundancy is established. A proposed should have high priority. new line along 60°N (see below) should 2) VOS meteorological measurements — contribute valuable information on sub- The North Atlantic is particularly well polar exchanges with the Nordic Seas. sampled by VOS meteorological obser- Some technical issues must also be ad- vations. However, an improved accuracy dressed. In the subpolar North Atlantic, of the data is required. For example the the depth-range of T7 XBT probes is not present VOS data may underestimate the always sufficient to reach the mixed-layer heat fluxes during winter time cold air

Figure 5.5 — The present XBT/VOS network developed under WOCE, TOGA and ACCP/ACCE. High-density lines are marked with bold lines, low-density with thin. Design studies are recommended for determining the optimal mix of high- and low-density XBT lines within ACVE in light of the proposed profiling float array, remote sensing tools, and the envisaged network of Eulerian stations.

27 outbreaks over the western North Atlan- onal baroclinic transport variability there. tic. Improved instrumentation (which is One station near the eastern boundary being developed both in the U.S. and Eu- might be part of the PIRATA array; its rope) should be implemented on a subset western counterpart should be located of the Atlantic VOS. close to the boundary (e.g., near the Lesser • Time series stations — Continuous records Antilles). at a relatively small set of stations have proven - Other stations are under consideration for to be very effective in documenting annual- observing the temporal evolution of heat to-inter-decadal variability in the ocean. It is content and water mass properties in as- vitally important to continue these important sociation with the PALACE array, and/or stations. While originally conducted using providing baroclinic transport time series research vessels, time series stations in future for key advection elements in the subpo- will be autonomous utilizing moored or free- lar and subtropical domains. floating profiling vehicles. - Expansion of the time series network into - It is essential to continue stations “Bravo” the South Atlantic should be considered. and Panulirus/Bermuda in the centers of • Flux monitoring systems — At key sites, di- the subpolar and subtropical cells, respec- rect measurement of ocean currents is envis- tively, which already demonstrated sub- aged. For example, long term Eulerian mea- stantial value in providing information on surements of the dense overflows of Nordic decadal heat storage and circulation Sea waters through the Denmark Straits and anomalies. Faeroe Bank Channel will help quantify the - Station “Mike” in the Norwegian Sea has changing climate system of the northern At- given insight into decadal warming of the lantic. Similarly, measurements in the Strait Norwegian Sea Deep Water and propa- of Gibraltar would provide integral measures gation of the Great Salinity anomaly. This of air-sea exchange with the Mediterranean station, which monitors the properties of and the supply of saline intermediate water the northward flowing Atlantic waters that to the Atlantic. Monitoring of the Florida ultimately feed the deep water formation Current heat and volume transport would fa- sites, must be continued. cilitate study of the meridional overturning - The ESTOC station was begun in 1994 circulation’s warm limb at subtropical lati- north of the Canary Islands in the eastern tude; a companion array spanning the deep subtropical Atlantic. Together with Ber- western boundary current could track the cold muda, hydrographic data from this station limb of the circulation. Developments now yields a baroclinic circulation index across underway for long-term (5-year), low-cost the subtropical gyre. Its continuation is current meter moorings with data telemetry encouraged. will make such programs economically fea- - Companion station(s) on the western mar- sible. gin of the basin would return baroclinic • Drifters Surface — Drifters have proven to circulation indices for both the Gulf be a valuable resource for SST ground truth Stream and the net MOC, and would to satellite-based observing systems as noted monitor the meridional spreading of wa- above. Drifter data have additionally yielded ter mass anomalies originating in the high- information on upper-ocean circulation pat- latitudes. terns and Ekman transports. Sensor suites on - Similarly, two bounding stations at the drifters may be enhanced to measure winds, latitude of the tropical- subtropical ex- atmospheric pressure and near-surface salin- changes (i.e., near 10-15°N) could pro- ity. On interannual time scales, the horizon- vide integral measurements of the meridi- tal transport of heat in the upper ocean be-

28 comes an important process. At the present Recommendations: time, there are no systematic, basin-scale Design studies will help to determine the most measurements of upper-ocean circulation in efficient mix of the above mentioned networks the tropical Atlantic (unlike the Pacific and for sustained Atlantic climate observations. The Indian Oceans) leaving this area void of any networks will build on existing and planned glo- data set by which to verify modeled ocean bal remotely-sensed data sets, those time-series circulations. The meridional component of stations remaining in operation, existing VOS- ageostrophic flows (Ekman flow) is signifi- XBT, drifter and float programs, and the capa- cant in the tropical region, showing speeds bilities offered by newly developed floats, moor- there that can be twice as large as the geo- ings and VOS technologies. strophic currents. Moreover, the sparse drifter data which exists (Richardson et al., 1988), 5.4 — Observing the meridional overturning indicate a net poleward transport, whereas the circulation (MOC) surface dynamic topography indicates a net Coarse-resolution climate models have sug- equatorward mass flux. It is therefore recom- gested scenarios for global thermohaline circula- mended that a basin-scale array of drifters be tion collapse. While these model oceans are of- deployed and maintained for at least one cycle ten unrealistic, the results deserve scientific at- of the TAV on a resolution of 2° in latitude tention. It is important to determine how realistic and 6° in longitude and within 18° of the equa- model simulations of large thermohaline circula- tor. tions variations are at short (decadal) time scales • Tide-gauges— Although the Atlantic does not under possible anthropogenic changes. For com- allow for an extensive tide-gauge network, as parison with model simulations, essential param- does the Pacific, a few stations located in the eters include observations of the meridional over- eastern, equatorial region are useful for moni- turning circulation rate and the associated heat toring the Atlantic El Niño-like phenomenon. and fresh water transports. Coverage at a particu- Stations in the Cape Verde and Ascension Is- lar latitude might involve elements of sustained lands are near the centers of action of decadal observations (see above) in combination with: dipole variability, and thus are useful indica- • Occasional repeated high-resolution transoce- tors of net steric (storage) processes associ- anic hydrographic ship-based sections includ- ated with SST changes. The CPACC stations ing tracers and current measurements. along the western boundary and in the Antilles • Moored current-meter measurements at the Islands are useful for monitoring changes in western boundary and additional stations for advection/export from the tropics via west- monitoring integrated baroclinic transports in ern-boundary currents on interannual time the interior. scales. • Moored bottom devices to interpolate be- • Acoustic Tomography — The combination of tween full-depth stations, such as barotropic altimetry and long range acoustic tomogra- flow meters and inverted echosounders. phy has been shown to be an effective means A feasibility study for monitoring the MOC of tracking basin-scale changes in ocean heat needs to be carried out, analyzing variability pat- content (ATOC Consortium, 1998). Tomog- terns in existing hydrographic sections and the raphy can average over the noisy mesoscale output of high-resolution numerical models. A eddy field, and thus greatly enhance the sig- 48°N section (the division line between the sub- nal to noise ratio of climate signals. ACVE polar and subtropical gyres) would be a priority needs to explore how a few acoustic sources candidate for a MOC monitoring section because can be used to maximize the amount of new (i) anomaly patterns suggest a dipole of North- information. South heat storage anomaly across this latitude, (ii) a number of WOCE sections have already

29 been obtained and can serve for pattern analysis, models. Despite obvious shortcomings, oceanic and (iii) intent has been expressed by BSH (Ger- general circulation models have achieved a suf- many) to re-occupy this section at WOCE qual- ficient degree of realism that they can now be ity as a contribution to GOOS every three years. employed to address process oriented or experi- Other potential sections for repeat top-to-bottom ment design questions, and to synthesize diverse hydrographic coverage, possibly combined with observations into ‘best-estimates’ of the quanti- moored boundary arrays, are shown in Figure 5.6 ties of interest. and should include: An active program in data-model synthesis • Near 25°N, close to the heat transport maxi- must be a component of ACVE to both help plan mum, where considerable prior knowledge on it and to interpret the observations. The combi- the variability has been built up. nation of model and data fields, presuming that •10°S, to measure the combined effects of the both contain elements of the same variability, al- shallow tropical cell (STC) and the top-to- lows for a better definition of what took place in bottom MOC the Atlantic Ocean than either by itself. If pre- •30°S, to measure the net heat and freshwater dictability can been established, these analyses exchange of the Atlantic with the global cir- will serve as the initial conditions for forecasts culation system . of the coupled system. A number of assimilation • Two meridional sections to quantify the efforts are underway as part of WOCE AIMS and changes in water mass inventories of the west- are anticipated as part of GODAE. Although tech- ern and eastern Atlantic basins. Sampling nology and computer power will certainly im- along 52°W longitude, which would repeat prove over the next decade, it is fair to state that measurements that extend back as far are the basic “know-how” has been developed and is 1950s, is strongly suggested. The optimal now available for ACVE. Additionally, the basic track line in the east has yet to be identified. infrastructure which is being proposed by In conjunction with the MOC section mea- GODAE, including active management of data surements, inventory observations of water mass and forcing fields, will be of benefit to the ACVE changes and anthropogenic CO2 uptake need to research activities. be carried out in appropriate time intervals (to be Numerical sampling studies are needed to determined from the WOCE/AIMS analysis). help determine effective and efficient observa- Recommendations: tional networks. High-resolution forward mod- Quantitative design studies for an MOC moni- els can be used to investigate sampling strategies, toring program including measurement of the and the rigorous framework of data assimilation associated meridional property transports and can lead to objective criteria for efficiently as- storage changes are required. Such studies needs sessing sampling alternatives. to determine the sampling frequencies of the zonal In the following, we provide some specific lines required to quantify variability in the At- recommendations for modeling studies that are lantic MOC and associated meridional transports. highly relevant to Atlantic Climate variability:

5.5 — Modeling, Theory and Ocean State Es- Mechanisms responsible for atmospheric vari- timation ability in the Atlantic sector. Theory and models will play a central role in Many atmospheric GCMs still have major the design, implementation and synthesis of an problems simulating synoptic and lower fre- ACVE. They will facilitate refinement of work- quency variability in the North Atlantic region. ing hypothesis, which can then be tested using Rather than curving northeastward, simulated observations. The complexity of the models will storm tracks are often quite zonal. The occurrence vary from idealized studies of particular physical of Atlantic blocking events is often underesti- mechanisms to fully coupled ocean- atmospheric mated. Lower-frequency variability such as that

30 of the stationary waves is often poorly repre- using GCMs seasonal, decadal and anthropo- sented. The difficulty experienced in trying to genic-induced variability can be credible. remedy these model problems reflects poor un- Recommendations: derstanding of the processes involved. The biases • Studies using a hierarchy of atmosphere in atmospheric GCMs are particularly evident in models to investigate the mechanisms the Atlantic sector, and could lead to poor repre- responsible for atmospheric variability on sentation of interannual and interdecadal variabil- time scales from days to decades, and the ity in such models and of the fluxes used to force processes that determine the spatial coher- ocean components in coupled models. These ence of this variability uncertainties need to be reduced before forecasts • Studies to investigate the mechanisms

Figure 5.6 — A schematic drawing of a potential ACVE repeat-section program. Hydrographic and tracer proper- ties would be sampled on these lines on an interannual time scale to quantify net changes in water mass properties and heat content. Thick lines should be repeated more often.

31 behind atmospheric teleconnections into be determined. Additionally the sensitivity and out of the Atlantic sector. For example, of atmospheric models to SST anomalies as how can a diabatic heating anomaly in the a function of spatial resolution needs to be tropics lead to dynamical changes in the investigated. Atlantic sector? Do changes in the Atlantic • Studies that determine the scales and pat- subtropical jet lead to changes in the Atlan- terns of Atlantic SST to which the atmo- tic inter-tropical convergence zone and vice sphere is most sensitive, and the atmo- versa ? spheric predictability that derives from changing SST. Ensemble integrations of Influence of the Atlantic Ocean on the atmosphere atmospheric models forced with both Long-range forecasting relies heavily on the observed and idealized SST anomalies are influence that SST anomalies exert on the atmo- needed. Particular attention should be sphere. Yet the nature of this control, especially focused on: a) whether the observed outside the tropics, is very poorly understood. multidecadal fluctuation in the NAO index Experiments in which extratropical SST anoma- can be reliably reproduced and if so, which lies are imposed in GCMs generally indicate weak regions of SST are important; b) The influ- response by comparison with the internal vari- ence of tropical and subtropical Atlantic ability. But these results could be misleading. It SST anomalies on the tropical and extratro- appears that the response is unlikely to be through pical atmosphere. Carefully controlled a simple locally enhanced heating, as is often the studies in which several different models case in the tropics. Rather, anomalies may im- are forced with the same SST anomalies pact the development of individual weather sys- would be useful to check the robustness of tems which tend to grow over the western North results and increase ensemble sizes. Atlantic and decay over the eastern Atlantic/West- • Studies to investigate the usefulness and ern Europe. It is unknown to which space and limitations of atmosphere model experi- time scales the atmosphere respond most strongly, ments with prescribed SST anomalies. nor the extent to which the structure of the Gulf Under what circumstances are the results of Stream is important to the atmosphere. These such experiments misleading? To what questions are central to possible interactive modes extent can the results be related to the of the atmosphere-ocean system and to coupled variability that arises when the same model modeling of the North Atlantic in general. An is- is coupled to an ocean mixed layer or to a sue of particular importance is whether or not the dynamical ocean model? What role do recent trend in the NAO index arose in response atmosphere-land surface interactions play? to changes in SST. Further development of data sets, for example Recommendations: reanalysis such as NCEP and ECMWF, to im- To address these issues we need: prove characterization of the observed atmo- • Studies that investigate the sensitivity of sphere and ocean variability. individual weather systems to SST, initially using high resolution forecasting models The mechanisms responsible for variability in the and cases for which much additional obser- Atlantic Ocean vational data exist (e.g., FASTEX IOPs). Recent work suggests that a large fraction of The structure of the boundary layer, the North Atlantic Ocean variability can be under- magnitude of the surface fluxes, and the stood as the ocean’s response to atmospheric fluc- development of the whole system should be tuations. These fluctuations, which are dominated compared with observations, and the sensi- by large scale patterns such as the North Atlantic tivity to parameterizations, to the specified Oscillation, influence the wind and buoyancy SST, and to an oceanic mixed layer need to driven ocean circulations as well as the convec-

32 tion and subduction processes that determine the higher latitudes. properties of the ocean thermocline. We are, how- • Studies with ocean GCMs are needed to ever, far from understanding the detailed relation- investigate how well the observed evolution ships between the surface fluxes of heat, fresh of the Atlantic Ocean can be simulated water and momentum and temporal variation of given observed surface fluxes or related the ocean state. quantities. (Careful thought should be given Of particular importance is an improved un- to the formulation of the surface boundary derstanding of the oceanic pathways through condition). Multidecade-long model simula- which fluctuations in surface exchange in one tions should be rigorously evaluated against region can influence the ocean, and especially available surface and subsurface observa- SST, in another region at a later time. For ex- tions, with particular attention to poorly ample, fluctuations in wind stress curl and flux understood phenomena such as the develop- anomalies over the subtropical gyre may promote ment, propagation and decay of heat content the development of heat content anomalies that anomalies, and to basic features such as the propagate northward to affect deep convection in distribution of principle water masses. A the Greenland or Labrador Seas. Changes in deep comparison between different GCMs would convection may in turn impact the overturning help to differentiate between the effect of circulation, including its contribution to Gulf surface flux errors and model errors in the Stream transport. Fluctuations in wind or buoy- solutions. Model experiments could be ancy fluxes in the subtropics may also influence designed to address which features in the equatorial SST via meridional circulation cells surface forcing are essential to reproduce that feed equatorial upwelling. key features of the observed variability. Recommendations: Specific points include: • Model studies to investigate the local and - Investigating the nature and the role in non-local relationships between temporal the large-scale circulation of key physi- variations in the surface fluxes and tempo- cal processes that are poorly represented ral variations in the ocean state. In the in ocean GCMs, in particular the role of design of models and experiments special topography (especially “choke points”) attention should be paid to the following: and of turbulent mixing. - Key processes in the ‘warm water - Development of ocean data assimilation transformation’ pathway including the systems to provide best estimates of the mechanisms responsible for the devel- seasonal-to-decadal variability in the opment, propagation, and decay of heat Atlantic Ocean (in close collaboration content anomalies, the interplay of with GODAE). processes that influence oceanic con- vection, the processes that modulate the Variability and Predictability of the Atlantic overturning circulation, and the impact Ocean-Atmosphere System MOC variations have on SST. Coupled ocean-atmosphere GCMs are pow- - The mechanisms that determine SST erful tools for understanding the climate system. variability in the tropical Atlantic, However, due to biases in both the atmospheric particularly the role played by dynami- and the oceanic components, such models often cal ocean processes. Also the mecha- have unrealistic mean states, annual cycles, and nisms linking the tropical Atlantic with interannual variability. Much work remains to higher latitudes including the modula- document, understand and alleviate such prob- tion of the STC’s by subtropical winds lems. or buoyancy fluxes and the influence of Coupled GCMs are the central tools for sea- the tropical Atlantic variability on sonal and longer time scale forecasting. There is

33 evidence of seasonal predictability over Europe circulation. In the following we highlight some in spring and summer but the mechanisms that regional processes that might deserve special at- give rise to this predictability are not understood. tention and a targeted effort. Such process experi- High priority must be given to research directed ments are anticipated to be of shorter duration at improving the realism of coupled forecast mod- (one to five years) and embedded in the sustained els, improving understanding of the basis for pre- observing networks. Focused pilot studies may dictability, and improving forecasting techniques. also be needed to help design the sustained ob- Recommendations: servational networks. ACVE will need: • Studies that elucidate the mechanisms Tropical Atlantic responsible for variability in the Atlantic AGCM experiments, in which the evolving Sector in coupled GCMs, and evaluate the pattern of tropical Atlantic SST is prescribed, will realism with which observed seasonal to be instrumental in examining the underlying dy- decadal variability is simulated. Insight into namics that govern the atmospheric response. mechanisms may be gained through experi- Questions to be addressed include: What are the ments in which ocean-atmosphere interac- atmospheric responses to dipole-like SST anoma- tions are restricted to certain regions (such lies? Are the anomalous circulations due to a as the Atlantic). boundary-layer response to SST or to a whole- • Studies of ENSO influence on the Atlantic scale rearrangement of the circulation in the tropi- Sector. What SST anomalies patterns are cal Atlantic troposphere? How does the anoma- directly forced by ENSO force in the lous circulation pattern affect the surface heat Atlantic? Do these anomalies in turn influ- flux? What is the role of land heat sources in driv- ence the atmosphere? How predictable is ing the anomalies? the Atlantic’s response to ENSO? Several oceanic processes are believed to play • Studies to investigate the mechanisms roles in tropical Atlantic variability such as oce- responsible for variability in the tropical anic advection including cross-gyre/cross-equa- Atlantic, including the ‘Atlantic ENSO’. torial exchanges and vertical mixing/subduction • Research into the use of coupled models for processes. However, the relative importance of forecasting climate fluctuations in the these processes in decadal TAV is unknown. Ex- Atlantic Sector on seasonal and decadal periments using models with varying degrees of time scales, including the problems of dynamical complexity must be carried out to de- initializing and assimilating data into termine which oceanic processes regulate off- coupled models, and the sensitivity of equatorial SST variability. For example, ocean forecasts to model formulation. models forced only with winds can be compared • Studies to investigate how changing levels to those forced only with surface buoyancy fluxes. of greenhouse gases may influence climate Emphasis should be placed on the following is- variability in the Atlantic Sector, and how sues: How can subsurface thermal anomalies in- natural variability influences the response fluence SST changes? What is the relative im- of climate system to such changes. portance of horizontal heat transport and vertical mixing/subduction processes? How important are 5.6 — Regional aspects of ACVE: Process cross-equatorial and cross-gyre exchanges? What Studies is the relative importance between western- Regional aspects of ACVE will encompass boundary-current transports and interior Ekman observationally-based process studies and focused transport? Do oceanic Rossby waves play an im- modeling activities. Both will serve to enhance portant role in determine SST fluctuations in the insight into mechanisms which, although local, off-equatorial SST variability? effect the entire Atlantic atmospheric or ocean It has been demonstrated that tropical ocean-

34 atmosphere interactions can lead to decadal TAV midlatitude variability on equatorial thermocline in relatively simple coupled models (e.g., Chang structure and SST. Experiments with forcing con- et al., 1997). In these models, the atmosphere is fined to different hemispheres will also tell us assumed to be largely determined by SST in such which hemisphere is more important. Detailed a manner that the tropical atmospheric circula- analyses of models constrained by observations tion and associated surface heat flux adjust in- will indicate which oceanic processes (e.g., sub- stantaneously to changes in the SST. This “slaved’ duction, wave propagation and/or changes in atmosphere may exaggerate the atmospheric feed- strength of shallow meridional cells) are likely to back to SST changes. Additionally, there are large contribute to equatorial SST variability. Result- uncertainties in surface heat flux estimates. Such ant SST anomaly patterns from such experiments models should be regarded as being near the bot- will then be fed into an atmospheric model to tom of a hierarchy of coupled systems, ranging identify possible feedbacks. For the atmosphere- from intermediate coupled models to fully model experiments, one could specify SST in the coupled GCMs, that are needed in order to un- tropical Atlantic to examine the resulting extra- derstand air-sea coupled processes in the tropical tropical variability. Particular attention should be Atlantic. Specifically, dynamically based atmo- paid to the relationships between the Hadley cell spheric models must be coupled to a variety of and storm-track variability in midlatitudes. Re- ocean models, ranging from mixed-layer models verse experiments can also be carried out by forc- to oceanic GCMs, in order to diagnose the im- ing the atmosphere in midlatitudes to examine portant feedback loops in TAV and to identify key possible NAO connections to tropical and south- delays in these loops. The coupled model experi- ern-hemisphere variability. Coupled model ex- ments will be crucial for addressing important periments can also be carried to determine the questions such as: Is the dipole-like SST variabil- relation between NAO and TAV. For example, one ity a truly coupled phenomenon? If so, where are can drive a coupled model with NAO-like forc- the regions in which the key air-sea feedbacks ing to see how much NAO influence there is on take place? What are the fundamental oceanic and TAV. atmospheric processes that control these feed- backs? Much more modeling work is needed to Subpolar Atlantic and Subtropical-subpolar con- fill in this hierarchy. It is not likely that a process nections study, focused on critical issues, can be designed Among the most pressing problems to be ad- until this modeling activity is well underway. Fi- dressed during ACVE is the complex of questions nally, it is important to emphasize that coupled associated with the atmospheric response to models will be instrumental in evaluating the pre- midlatitude SST/heat flux anomalies, with the dictability of the climate phenomenon. Existing Gulf Stream position, and with propagating observations already point to a tight coupling SSTAs. Although many of those questions will between rainfall variability in Northeast Brazil be addressed in a modeling framework, some and Sub-Saharan Africa, as well as Caribbean and enhancement of measurements in the subpolar tropical-Atlantic SST variability. Coupled model Atlantic is required. The unknown consequences experiments must be conducted to examine the of heat content anomalies along the boundary predictability of tropical Atlantic SSTs, and between the subtropical and subpolar gyre might whether seasonal-to-interannual SST forecasts justify a targeted process experiment: can improve seasonal rainfall forecasts. An anomaly process experiment After identification by the observing system, Subtropics and tropical-subtropical connections developing upper ocean heat content anomalies We propose to study tropical-subtropical in- will be sampled intensively to determine their teractions using model-data combinations. Ocean evolution and propagation characteristics. models will be used to study the effect of Anomalies in the upper layers of the ocean, cir-

35 culating in both the subtropical and subpolar mechanisms and inherent time scales) to those gyres, should be sampled intensively over sev- localized changes. At these latitudes, the surface eral annual cycles, to determine the manner in and deep waters are related through the vertical which their associated SST and thermal anoma- stability which in turn is strongly influenced by lies are modified by air-sea interaction, and the sea surface salinity. The time-dependent fresh cycle of mixing and restratification. Fluid pass- water budget appears to be the clearest diagnos- ing through the Gulf Stream system is exposed tic, if not the controlling factor for deep convec- to the influence of strong NAO forcing at the start tion and the water types that result. of the Atlantic storm track. Its properties, changed Strategies need to be developed to investigate by convective mixing, then flood the subpolar and the following topics: subtropical gyres. We need to determine the life- • The export of waters from the Greenland- time and characteristic pathways of these anoma- Iceland-Norwegian (GIN) Seas including lies, and their potential influence on the atmo- the densest Denmark Strait Overflow Water sphere above. (DSOW) and the shallower boundary To carry this out, the upper ocean observing currents (including ice) that are important to network would be augmented with additional the fresh-water balance. Close cooperation floats and surveys in key areas (e.g., the region with ACSYS (joint memberships in Imple- of18 degree water under the influence of strong mentation WGs) is recommended. Monitor- NAO forcing). When studied in combination with ing the overflows over the Greenland- data assimilating models, the fate of the anoma- Iceland-Scotland sills is considered very lies, their longevity and characteristic pathways important as these measurements may be can be mapped out. We would also like to quan- applied as northern boundary conditions in tify, as a function of time-scale, the role of the restricted-domain Atlantic Ocean modeling ocean in the advection and transfer of heat to and efforts. from the surface, through subduction, Ekman lay- • The properties of those Atlantic waters that ers, geostrophic eddies etc. enter the northern seas and subsequently An anomaly process experiment to investi- involved in dense water mass formation and gate the generation, propagation and fate of tem- modifying the Arctic pycnocline. This perature anomalies at the boundary between the might include a proposed 60N VOS pro- subpolar and subtropical gyres should include: a gram on a 3-week repeat schedule between locally augment float and drifter array and Denmark to Greenland. Eulerian time-series stations at key sites of SSTA • The actual formation mechanism(s) of the evolution, and an instrumented monitoring line dense overflow waters to quantify how at the southern outflow of the subpolar gyre much of the newly ventilated deep and (roughly 47°N), extending eastward to the mid- intermediate water is formed in the gyre Atlantic ridge, to sample the deep and intermedi- centers versus the boundary current sys- ate southward flows as well as the northward tems. Exploratory Lagrangian programs North Atlantic Current transport and T/S proper- might follow the conversion of Atlantic ties. Water into DSOW in or about the GIN Seas. Study of the entrainment mixing The Polar Basins and Subpolar-polar connections processes active within the descending The central questions here include under- plumes of dense overflow waters is also standing the variations in low-salinity surface needed. water and deep water mass formation/export, their • The shallow boundary currents along the relationships to atmospheric and ocean processes, Labrador and Greenland continental shelves and the adjustment of the basin-to-global scale that, despite their small mass transports, thermohaline circulation (in terms of physical carry very low-salinity waters and hence

36 have a large fresh water transport that bines into a coherent effort, the elements (D1- appear able to cap off the deep convection D3) of the International CLIVAR program sites. that deal with the Atlantic Sector. By doing • The energetics/boundary layer physics of so, it is hoped that a more natural and focused ocean/atmosphere coupling at high latitude. study will result. The Atlantic is surrounded To address this, the high-resolution ETA by many developed countries which will con- atmospheric model needs to be extended tribute to ACVE scientifically and financially. eastward and a reanalysis program of the Given the partial overlap of physical pro- standard NCEP model organized. A general cesses and scientific questions, ACVE will goal is improved E-P and heat flux predic- maintain close scientific ties with our Pacific tions. The dipolar response of sea-ice to the counterpart: PBECS. NAO, in the GIN and Labrador Seas. • GODAE/GCOS/GOOS (Global Observing • The ice edge provides a possibly strong System) — The proposed Global Ocean Data ocean-atmosphere feedback through surface Assimilation Experiment (GODAE) will pro- heat/moisture flux. This suggests a process vide a convincing demonstration of real-time study of upper ocean/ice measurement and assimilation systems, with core operational atmospheric energetics. activities to be carried out 2003-2005 on a Recommendations: global basis. The aim is to assimilate data on Detailed planning for these imbedded process space-time scales of 100 km and a few days, studies within the ACVE sustained observational and produce integrated global analyses that networks must begin immediately, building on the can be used to support local-scale studies and modeling work and observational program results initiate model runs. This study has been re- of ACCE. ceived enthusiastically by the WCRP. Plans call for a multi-node operation, with differ- 6. Programmatic context ent groups focusing on different oceans, al- though all will use a common model and for- The Atlantic Climate Variability Experiment has mat so as to produce consistent products. At been conceived as a major contributor to the present, it appears that GODAE jointly with CLIVAR program. As such, ACVE will provide CLIVAR UOP will foster a global profiling- real time assessments of the state of the Atlantic float network, ARGO, as part of its sustained Ocean in relation to climatic variations of the at- subsurface ocean observing system. The At- mosphere. These will be of great utility to research lantic component of this array will contribute programs which seek to understand the role of substantially to ACVE. ocean biogeochemistry in climate, and the im- • PIRATA PACS — The international Pilot pacts of climate variability on fisheries. There- Research Moored Array in the Tropical At- fore strong connections exist with many other on- lantic (PIRATA) will maintain an array of up going or planned programs. These are summa- to 14 surface moorings in the equatorial At- rized below. lantic in the 1997-2000 period. The acquired • International CLIVAR Program — An data from this network will help guide plan- important objective of both the GOALS and ning for the augmented array envisioned for Dec-Cen components of CLIVAR is to iden- ACVE. The Pan-American Climate Studies tify and exploit incremental elements of cli- program (PACS, the US component of mate predictability provided by quantified VAMOS) aims to improve the skill of opera- modes of ocean-atmosphere variability. By tional seasonal-to-interannual climate predic- developing improved understanding of the tion over the Americas, with the emphasis on NAO and TAV, the ACVE directly addresses forecasting warm season rainfall. Focal points these CLIVAR program goals. ACVE com- include the relationship between boundary

37 forcing and climate variability; processes gov- past environment. We anticipate that long erning SST variability in the tropical Atlantic proxy records of climate variability in the and Pacific; and the development of the mixed Atlantic sector will come out of this effort layer. In its second phase, there are plans that which will contribute to the ACVE goals. PACS will shift its activities into the tropical • GLOBEC — International GLOBEC (now Atlantic which will allow close collaboration an official IGBP program) is engaged in plan- with PIRATA-ACVE. ning for the Small Pelagics and Climate • WOCE AIMS — The World Ocean Circula- Change (SPACC) program. Involved coun- tion Experiment (WOCE) is now entering in tries include Japan, U.S., Mexico, Peru and its Analysis, Interpretation, Modeling and Chile. Retrospective studies, monitoring, Synthesis phase (AIMS). In particular the re- modeling and process studies are being con- sults and experience gained from the Atlantic sidered. There are clear connections between Circulation and Climate Experiment (ACCE) the physical elements of these efforts and will benefit the planing and final design of ACVE. ACVE. The analysis phase of WOCE will also provide the opportunity to inspect the histori- Timetable cal and WOCE records in the light of the A definitive time table for ACVE will be de- ACVE goals. veloped in the coming six months. Tentative plans • ACSYS — The Arctic Climate System Study call for core field activities beginning 2002. How- (ACSYS) aims to answer two related ques- ever, planing and implementation activities will tions: start as early as FY 1999, these will include de- - Is the Arctic climate as sensitive to glo- sign studies and error estimates based on histori- bal changes as models seem to suggest? cal data and numerical studies. The ACCE data - What is the sensitivity of global climate set will be a central element in these studies. to Arctic processes? Hence ACSYS will focus on the climate of Budget the polar oceans and will be an important part- A draft budget for ACVE will be assembled ner in the northern part of the Atlantic where in advance of the December international variability in sea ice distribution, fresh water CLIVAR meeting. It can be anticipated that a large export from the Arctic and the rate of warm fraction of the experiment will be covered by to cold water formation might hold the key to Euro-CLIVAR. In the U.S., the National Science Atlantic variability. (http://www.npolar.no/ Foundation and National Oceanic and Atmo- acsys/) spheric Administration are actively involved at • PAGES — The International Geosphere-Bio- this time. NASA is also involved with remote sphere Program project charged with provid- sensing elements of the program. Contributions ing a quantitative understanding of the Earth’s from ONR are unclear at this point.

38 References

Alexander, M.A, and C. Deser, 1995: A mechanism for the recurrence of wintertime SST anomalies. J. Phys. Oceanogr., 25, 122-137. ATOC Group, The, 1998: The general circulation and heat budget of the North Pacific Ocean from Satellite altimetry, acoustic tomography, and a general circulation model, Science, 281, 1327- 1332 Bacon, S., and D.J.T. Carter, 1993: A connection between mean wave height and atmospheric pres- sure gradient in the North Atlantic. Int. J. Climatology, 13, 423-436. Barnston, A.G., and R.E. Livezey, 1987: Classification, seasonality and persistence of low-frequency atmospheric circulation patterns. Mon. Wea. Rev., 115, 1083-1126. Battisti, D. S., U. S. Bhatt, and M. A. Alexander, 1995: A modeling study of the interannual variabil- ity in the wintertime North Atlantic Ocean. J. Climate, 8, 3067-3083. Belkin, I.M., S. Levitus, and J. Antonov, 1998: “Great Salinity Anomalies” in the North Atlantic. Prog. Oceanogr., in press. Beniston, M., 1997: Variations of snow depth and duration in the Swiss Alps over the last 50 years: Links to changes in large-scale climate forcings. Climate Change, 36, 281-300. Bjerknes, J., 1964: Atlantic air-sea interaction. Adv. Geophys., 10, 1-82. Carton, J.A., X. Cao, B.S. Giese, and A.M. da Silva, 1996: Decadal and interannual SST variability in the tropical Atlantic. J. Phys. Oceanogr., 26, 1165-1175. Cayan, D.R., 1992a: Latent and sensible heat flux anomalies over the northern oceans: the connec- tion to Monthly Atmospheric Circulation. J. Climate, 5, 354-369. Cayan, D.R., 1992b: Variability of latent and sensible heat fluxes estimated using bulk formulae. Atmos.-Ocean, 30, 1-42. Chang, P., L. Ji, and H. Li, 1997: A decadal climate variation in the tropical Atlantic Ocean from thermodynamic air-sea interactions. Nature, 385, 516-518. Cook, E.R., R.D. D’Arrigo, and K.R. Briffa, 1998: A reconstruction of the North Atlantic Oscillation using tree-ring chronologies from North America and Europe. The Holocene, 8, 9-17. Covey, D.L. and S. Hastenrath, 1978: The Pacific El Niño phenomenon and the Atlantic circulation. Mon. Wea. Rev., 106, 1280-1287. Delworth,T.L., S. Manabe, and R.J. Stouffer, 1997: Multidecadal climate variability in the Greenland Sea and surrounding regions: a coupled model simulation. Geophys. Res. Lett., 24, 257-260. Deser, C., and M. Blackmon, 1993: Surface climate variations over the North Atlantic Ocean during winter: 1900-1989. J. Climate, 6, 1743-1753. Dickson, R., J. Lazier, J. Meincke, P. Rhines, and J. Swift, 1996: Long-term coordinated changes in the convective activity of the North Atlantic. Prog. Oceanogr., 38, 241-295 Dickson, R.R., J. Meincke, S.-A. Malmberg, and A.J. Lee, 1988: The “Great Salinity Anomaly” in the northern North Atlantic 1968-1982. Prog. Oceanogr., 20, 103-151. Drinkwater, K.F., 1994: Environmental changes in the Labrador Sea and some effects on fish stocks. ICES CM 1994/MINI:4, 19 pp. Enfield, D. B. and D. A. Mayer, 1997: Tropical Atlantic SST variability and its relation to El Niño- Southern Oscillation. J. Geophys. Res., 102, 929-945. Folland, C.K., J.A. Owen, M.N. Ward, and A.W. Colman, 1991: Prediction of seasonal rainfall in the Sahel Region using empirical and dynamical methods. J. Forecasting, 10, 21-56.

39 Folland, C.K., T. Palmer, and D. Parker, 1986: Sahel rainfall and worldwide sea temperatures: 1901- 85. Nature, 320, 602-606. Frankignoul, C., 1985: Sea surface temperature anomalies, planetary waves and air-sea feedback in midlatitudes. Rev. Geophys., 23, 357-390. Friedland, K.D., D.G.Reddin, and J.F.Kocik, 1993. Marine Survival of North American and Euro- pean Atlantic salmon: effects of growth and environment. ICES J. Mar. Sci.,50, 481-492. Fromentin, J-M, and B.Planque, 1996. Calanus and environment in the eastern North Atlantic. II. Influence of the North Atlantic Oscillation on C.finmarchicus and C.helgolandicus. Mar. Ecol. Prog. Ser., 134, 111-118. Gray, W.M.,1990: Strong association between West African rainfall and US landfalling intense hurricanes. Science, 249, 1251-1256. Hall, M.M, and H.L. Bryden, 1982: Direct estimates and mechanisms of ocean heat transport. Deep- Sea Res., 29, 339-359. Hameed, S., K.R. Sperber, and A. Meinster, 1993: Teleconnections for the Southern Oscillation in the tropical Atlantic sector in the OSU coupled ocean-atmosphere GCM. J. Climate, 6, 487-498. Hansen, D. V., and H. F. Bezdek, 1996: On the nature of decadal anomalies in North Atlantic sea surface temperature. J. Geophys. Res., 101, 9749-9758. Hastenrath, S., and L. Greischar, 1993: Circulation mechanisms related to northeast Brazil rainfall anomalies. J. Geophys. Res., 98, 5093-102. Hastenrath, S., and L. Heller, 1977: Dynamics of climatic hazards in Northeast Brazil. Quart. J. Roy. Meteor. Soc., 103, 77-92. Hastenrath, S., 1984: Interannual variability and annual cycle: mechanisms of circulation and cli- mate in the tropical Atlantic sector. Mon. Wea. Rev., 112, 1097-1107. Hastenrath, S., L.C. Castro, and P. Aceituno, 1987: The Southern Oscillation in the tropical Atlantic sector. Contrib. Atmos. Phys., 60, 447-463. Houghton, R.W., 1996: Subsurface quasi-decadal fluctuations in the North Atlantic. J. Climate, 9, 1363-1373. Houghton, R.W., and Y.M. Tourre, 1992: Characteristics of low-frequency sea surface temperature fluctuations in the tropical Atlantic. J. Climate, 5, 765-71. Hurrell, J.W., and H. van Loon, 1997: Decadal variations in climate associated with the North Atlantic Oscillation. In: Climate Change. Proc. Int. Workshop on Climatic Change at High Elevation Sites, Wengen, Switzerland. Hurrell, J.W., 1995: Decadal trends in the North Atlantic Oscillation: regional temperatures and precipitation. Science, 269, 676-679. Hurrell, J.W., 1996: Influence of variations in extratropical wintertime teleconnections on Northern Hemisphere temperature. Geophys. Res. Lett., 23, 665-668. Kushnir, Y., and J. M. Wallace, 1989: Low-frequency variability in the Northern Hemisphere winter: geographical distribution, structure and time-scale dependence. J. Atmos. Sci., 46, 3122-3142. Kushnir, Y., 1994: Interdecadal variations in North Atlantic sea surface temperature and associated atmospheric conditions. J. Climate, 7. 142-157. Kushnir, Y., V.J. Cardone, J.G. Greenwood, and M. Cane, 1997: On the recent increase in North Atlantic wave heights. J. Climate, 10, 2107-2113. LaRoche, J., R. Nuzzi, R. Waters, K. Wyman, P.G. Falkowski, and D. W. R. Wallace, 1997: Inci- dence of brown tide blooms in Long Island’s coastal waters linked to interannual variability in groundwater flow. Global Change Biology, 3, 397-410. Lamb, H.H. and R.A. Pepler, 1987: North Atlantic Oscillation. Concept and application. Bull. Amer. Meteor. Soc., 68, 1218-1225.

40 Lamb, P. J., 1978: Large-scale tropical Atlantic circulation patterns associated with Subsaharan weather anomalies. Tellus, 30, 240-251. Lau, N.-C., 1988: Variability of the observed midlatitude storm tracks in relation to low-frequency changes in the circulation pattern. J. Atmos. Sci., 45, 2718-2743. Levitus, S., 1989: Interpentadal variability of temperature and salinity at intermediate depths of the North Atlantic Ocean, 1970-1974 versus 1955-1959. J. Phys. Oceanogr., 24, 1480-1493. Lough, J.M., 1986: Tropical Atlantic sea surface temperature and rainfall variations in Sub-Saharan Africa. Mon. Wea. Rev., 114, 561-570. Mann, M.E., and J. Park, 1996: Global scale modes of surface temperature variability on interannual to century time scales. J. Geophys. Res., 99, 25, 819-25,833. Markham, C.G., and D.R. McLain, 1977: Sea surface temperature related to rain in Caera, North- eastern Brazil. Nature, 265, 320-323. Marshall, J.C., and F. Molteni, 1993 Towards a dynamical understanding of weather regimes. J. Atmos. Sci., 50, 12, 1792 - 1818 Marshall, J.C., A.J.G. Nurser, and R.G. Williams, 1993 Inferring the subduction rate and period over the North Atlantic. J. Phys. Oceanogr., 23, 7, 1315 - 1329 McCartney M.S., R.G. Curry, and H.F. Bezdek, 1996: North Atlantic’s transformation pipeline chills and redistributed subtropical water. Oceanus, 39, 19-23. Mehta, V.M., 1998: Variability of the tropical ocean surface temperatures at decadal- multidecadal time scales, Part I: The Atlantic Ocean. J. Climate, in press. Molinari, R.L., D.A. Mayer, J.F. Festa, and H.F. Bezdek, 1997: Multi-year variability in the near- surface temperature structure of the midlatitude western North Atlantic Ocean. J. Geophys. Res., 102, 3267-3278. Moura, A. and J. Shukla, 1981: On the dynamics of droughts in Northeast Brazil: observations, theory, and numerical experiments with a general circulation model. J. Atmos. Sci., 38, 2653- 2675. Nakamura, H., 1996: Year-to-year and interdecadal variability in the activity of intraseasonal fluctua- tions in the Northern Hemisphere wintertime circulation. Theor. Appl. Clomatol., 55, 19-32. Nobre, P., and J. Shukla, 1996: Variations of sea surface temperature, wind stress, and rainfall over the tropical Atlantic and South America. J. Climate, 9, 2464-2479. Perlwitz, J., and H-F. Graf, 1995: The statistical connection between tropospheric and stratospheric circulation of the Northern Hemisphere in winter. J. Climate, 8, 2281-2295. Philander, S.G.H., 1986: Unusual conditions in the tropical Atlantic in 1984. Nature, 322, 236-238. Post, E., M. C. Forchhammer, and N. C. Stenseth, 1997: Climate change and ecosystem function. Science, submitted. Reverdin, G., D.R. Cayan, and Y. Kushnir, 1997: Decadal variability of hydrography in the upper northern North Atlantic, 1948-1990. J. Geophys. Res., 102, 8505-8533. Rogers, J. C.: 1990, Patterns of Low-Frequency Monthly Sea Level Pressure Variability (1899— 1986) and Associated Wave Cyclone Frequencies, J. Climate, 3, 1364—1379. Rossby, C.G., et al., 1939: Relation between variations in the intensity of the zonal circulation of the atmosphere and the displacments of the semi-permanent centers of action. J. Mar. Res., 2, 38-55. Servain, J., 1991: Simple climatic indices for the tropical Atlantic Ocean and some applications. J. Geophys. Res., 96, 15137-15146. Shaeffer, J.D., 1996: Tropical cyclone activity as a diagnostic climate indicator. Proceedings of the 20th Annual Climate Diagnostics Workshop, October 1995. University of Washington. Sutton, R.T., and M.R. Allen, 1997: Decadal predictability in Gulf Stream sea surface temperature. Nature, 388, 563-567.

41 Thompson, D. W. J., and J. M. Wallce, 1998: The Arctic Oscillation signature in the wintertime geopotential height and temperature fields. Geophys. Res. Letts., 25, 1297-1300. Tourre, Y.M., Y. Kushnir, and W. B. White. 1997. Evolution of quasi-cyclic interdecadal variability in the Pacific ocean- atmosphere system. J. Phys. Oceanogr. (in press). Tourre, Y. M., B. Rajagopalan, and Y. Kushnir, 1998: Dominant patterns of climate variability in the Atlantic Ocean region during the last 136 years. J. Climate, in press Trenberth, K.E., and A. Solomon, 1994: The global heat balance: Heat transports in the atmosphere and ocean. Clim. Dyn., 10, 107-134. Walker, G.T., and E.W. Bliss, 1932: World weather V. Mem. Roy. Meteorol. Soc., 4, 53-84. Wallace, J.M., and D.S. Gutzler, 1981: Teleconnections in the geopotential height field during the Northern Hemisphere winter. Mon. Wea. Rev., 109, 784-812. Ward, M.N., and C. K. Folland, 1991: Prediction of seasonal rainfall in the north Nordeste of Brazil using eigenvectors of sea-surface temperature. Inter. J. Climatology, 11, 711-743. White, J.W.C., D. Gorodetzky, E.R. Cook, and L.K. Barlow, 1996: Frequency analysis of an annu- ally resolved, 700 year paleoclimate record from the GISP-2 ice core: Climatic variations and forcing mechanisms of the last 2000 years. Eds.: P.D. Jones R.S. Bradley and J. Jouzel, NATO ASI Series, 1, 41 Springer Berlin, 193-212. Williams, R.G., Spall, M.A., and J.C. Marshall, 1995 Does Stommel’s mixed layer “Demon” work? J. Phys. Oceanogr., 25(12): 3089- 3102. Wunsch, C., 1998: The interpretation of short climate records, with comments on the North Atlantic and Southern Oscillations, BAMS, submitted for publication. Zebiak, S. E., 1993: Air-sea interaction in the equatorial Atlantic region. J. Climate, 6, 1567- 1586. van Loon, H., and Rogers, J. C., 1978: “The Seesaw in Winter Temperatures between Greenland and Northern Europe. Part I: General Description,” Mon. Wea. Rev.: 106, 296- 310.

Annex

List of Participants a) Dallas Workshop, February 1998

David Battisti Jim Carton University of Washington University of Maryland/JCESS Dept. Of Atmospheric Sciences Department of Meteorology P.O. Box 351640 Computer & Space Sciences Bldg. Seattle, WA 98195 College Park, MD 20742 [email protected] [email protected]

A. J. Busalacchi Ping Chang NASA Department of Oceanography Laboratory for Hydrospheric Processes Texas A&M University GSFC College Station, TX 77843-3146 Greenbelt, MD 20771 [email protected] [email protected]

42 Clara Deser Bill Large Climate and Global Dynamics Division NCAR/CU National Center for Atmospheric Research P.O. Box 3000 P.O. Box 3000 Boulder, CO 80307-3000 Boulder, CO 80307-3000 [email protected] [email protected] Susan Lozier Randall Dole Division of Earth and Ocean Sciences NOAA/ERL/CDC Nicholas School of the Environment Mail Code: R/E/CD Duke University 325 Broadway Box 90230 Boulder, CO 80303 Durham, NC 27708-0230 [email protected] [email protected]

Isaac M. Held John Marshall Geophysical Fluid Dynamics Laboratory MIT (EAPS) NOAA 77 Mass Ave P. O. Box 308 Cambridge, MA 02139 Princeton, NJ 08542 [email protected] [email protected] Michael McCartney Nelson Hogg WHOI WHOI Woods Hole, MA 02543 Woods Hole, MA 02543 [email protected] [email protected] Jay McCreary Jim Hurrell Oceanographic Center NCAR Nova Southeastern University P.O. Box 3000 800 N. Ocean DR Boulder, CO 80307-3000 Dania, FL 33020 [email protected]>, [email protected]

W. E. Johns Robert Molinari RSMAS/MPO NOAA/AOML University of Miami 4301 Rickenbacker Causeway 4600 Rickenbacker Causeway Miami, FL 33149 Miami, FL 33149 [email protected] [email protected] P. Niiler Yochanan Kushnir Scripps Institution of Oceanography Lamont-Doherty Earth Observatory La Jolla, CA 92093-0230 of Columbia University [email protected] Route 9W Palisades, NY 10964 Brechner Owens [email protected] WHOI Woods Hole, MA 02543 [email protected]

43 R. W. Reynolds Detlef Stammer Coupled Model Project Massachusetts Inst. of Technology NMC 77 Mass. Av. NOAA Cambridge, MA 02139 Camp Springs, MD 20747 [email protected] [email protected] John Toole Peter Rhines WHOI Box 357940 Woods Hole, MA 02543 University of Washington [email protected] Seattle, WA, 98195 [email protected] Martin Visbeck Lamont-Doherty Earth Observatory of Walter Robinson Columbia Unversity Dept. of Atmospheric Sciences RT 9W University of Illinois Palisades, NY 10964 105 S Gregory Street [email protected] Urbana, IL 61801 [email protected] John Walsh Dept. of Atmospheric Sciences Tom Rossby University of Illinois Graduate School of Oceanography 105 S Gregory Street Narragansett Bay Campus Urbana, IL 61801 University of Rhode Island [email protected] Narangansett, RI 02882 [email protected] Carl Wunsch MIT 54-1524 R. Saravanan Cambridge, MA 02139 NCAR [email protected] P.O. Box 3000 Boulder, CO 80307-3000 Steve Zebiak [email protected] Lamont-Doherty Earth Observatory of Columbia Unversity Michael Spall Route 9W WHOI Palisades, NY 10964 Woods Hole, MA 02543 [email protected] [email protected] b) Florence workshop (May 1998)

D. Anderson Fax: + 44 1189 986 9450 EMCWF E-mail: [email protected] Shinfield Park READING RG2 9AX C. Appenzeller United Kingdom University of Bern Climate and Env. Physics Tel: + 44 1189 949 9706 Institute

44 Sidlerstrasse 5 a CH - 3012 S. Corti Bern Switzerland CINECA (Bologna) Via Tel: + 41 31 631 4871 Magnanelli 6/s Casalecchio di Reno Fax: + 41 31 631 4405 Bologna 40033 Italy E-mail:[email protected] Tel: + 39 51 617 1581 E-mail: [email protected] A. Bamzai National Science Foundation M. Davey Climate Dynamics Program, Room 775 Hadley Centre 4201 Wilson Boulevard Met Office London Road Bracknell Arlington VA 22230 USA Reading RG12 2SY UK Tel: + 1 703 306 1527 Tel: + 44 1344 85 4648 Fax: + 1 703 306 0377 Fax: + 44 1344 85 4898 E-mail: [email protected] E-mail: [email protected]

D. Battisti P. Delecluse University of Washington LODYC Dept. Of Atmospheric Sciences 351640 Univ. Pierre & Marie Curie 4, Seattle WA 98195 USA Place Jussieu 75005 Paris, Cedex 04 France Tel: + 1 206 543 2019 Tel: + 33 1 44 27 7079 Fax: + 1 206 543 0308 Fax: + 33 1 44 27 7159 E-mail: [email protected] E-mail: [email protected]

G. Boccaletti T. Delworth IMGA Area della Geophysical Fluid Dynamics Lab./NOAA Ricerca CNRVia Gobetti 101 P.O. Box 308 Princeton University Bologna Italy Princeton NJ 08542 USA Tel: + 39 51 639 8014 Tel: + 1 609 452 6565 Fax: + 39 51 639 8132 Fax: + 1 609 987 5063 E-mail:[email protected] E-mail: [email protected]

G. Branstator M. Deque NCAR Meteo-France P.O. Box 3000 42 Av. Coriolis Boulder, CO 80307 USA 31057 TOULOUSE France Tel: + 1 303 497 1365 Tel: + 33 5 61 07 9382 Fax: + 1 303 497 1700 Fax: + 33 5 61 07 9610 E-mail:[email protected] E-mail:[email protected]

A. Capotondi L. Dilling NCAR/CU NOAA/OGP P.O. Box 3000 1100 Wayne Av. Suite 1210 Boulder, CO 80307 USA Silver Spring MD 20910 USA Tel: + 1 303 497 1353 Tel: + 1 301 427 2089 (ext. 106) Fax: + 1 303 497 1700 Fax: + 1 301 427 2073 E-mail: [email protected] E-mail:[email protected]

45 C. K. Folland B. A. King Hadley Centre, UK Southampton Oceanography Centre Empress Met Office London Road Bracknell Dock Southampton Berkshire RG12 2SY United Kingdom SO14 3ZH United Kingdom Tel: + 44 1344 85 6646 Tel: + 44 1703 59 6438 Fax: + 44 1344 85 4898 Fax: + 44 1703 59 6204 E-mail: [email protected] E-mail: [email protected]

C. Gordon G. J. Komen UK Met Office London Road, KNMI Bracknell Berkshire RG12 2SY UK P.O. Box 201 3730 AE De Bilt The Netherlands Tel: + 44 1344 85 6660 Tel: + 31 30 2206 676 Fax: + 44 1344 85 4898 Fax: + 31 30 2202 570 E-mail:[email protected] E-mail: [email protected]

B. Hoskins Y. Kushnir University of Reading Dept. of Meteorology Lamont-Doherty Earth Observatory of Colum- Reading bia Unversity RG6 6BB United Kingdom Route 9W Palisades, NY 10964 USA Tel: + 44 1189 31 8953 Tel: + 1 914 365 8669 Fax: + 44 1189 31 8905 Fax: + 1 914 365 8736 E-mail: [email protected] E-mail:[email protected]

D. R. Cayan N. G. Kvamsto Scripps Institution of Oceanography Geophysical InstituteUniversity of Bergen La Jolla Calif. 92093-0224 USA Allégaten 70 N-5007 Tel: + 1 619 534 4507 Bergen Norway Fax: + 1 619 534 8561 Tel: + 47 55 58 2898 E-mail: [email protected] Fax: + 47 55 58 9883 E-mail: [email protected] J. Hurrell NCAR W. Large P.O. Box 3000 NCAR/CU Boulder, CO 80307-3000 USA P.O. Box 3000 Tel: + 1 303 497 1383 Boulder, CO 80307-3000 USA Fax: + 1 303 497 1333 Tel: + 1 303 497 1364 E-mail: [email protected] Fax: + 1 303 497 1700 E-mail: [email protected] S. Jewson CGAM, University of Reading Dept. of Meteo- M. Latif rology Earley Gate Max-Planck-Institut für Meteorologie Whiteknights P.O. Box 243 Reading Bundesstrasse 55 20146 RG12 2SY UK Hamburg Germany Tel: + 44 1189 87 5123 (x 4281) Tel: + 49 40 411 73 248 Fax: + 44 1189 318 316 Fax: + 49 40 411 73 173 E-mail: [email protected] E-mail: [email protected]

46 H. Le Treut V. Mehta LMD/CNRS 24 rue Lhomond 75231 NASA University of Maryland Paris Cedex 5 France Code 913 NASA/Goddard Space Flight Center Tel: + 33 0144 27 84 06 Greenbelt, MD 20771 USA Fax: + 33 0144 27 62 72 Tel: + 1 301 286 2390 E-mail: [email protected] Fax: + 1 301 286 1759 E-mail: [email protected] D. Marshall University of Reading R. Molinari P.O. Box 243 Reading RG6 6BB UK NOAA/AOML Tel: + 44 118 931 8952 4301 Rickenbacker Causeway Fax: + 44 118 931 8905 Miami, Florida 33149 USA E-mail: [email protected] Tel: + 1 305 361 4344 Fax: + 1 305 361 4392 J. Marshall E-mail: [email protected] MIT Cambridge MA 02138 USA Tel: + 617 253 9615 F. Molteni Fax: + 617 253 4464 ECMWF Shinfield Park E-mail: [email protected] Reading RG2 9AX United Kingdom Tel: + 44 1189 499 601 S. Masina Fax: E-mail: [email protected] IMGA-CNR Via Gobetti 101 40129 Bologna Italy K. Mooney Tel: + 39 51 639 8019 National Oceanic and Atmospheric Administra- Fax: + 39 51 639 8132 tion E-mail: [email protected] 1100 Wayne Avenue, Suite 1225 Silver Spring, MD USA W. May Tel: + 1 301 427 2089 Danish Meteorological Institute Lyngbyvej 100 Fax: + 1 301 427 2073 2100 E-mail: [email protected] Copenhagen Denmark Tel: + 45 39 15 7462 J. Murphy Fax: + 45 39 15 7460 Hadley Centre Met. Office London Road M. McCartney Bracknell Berskhire RG12 2SY United King- MSZ1, WHOI dom Woods Hole, MA 02543 USA Tel: + 44 1344 856 658 Tel: + 1 508 2892 797 Fax: + 44 1344 854 898 Fax: + 1 508 4572 181 E-mail: [email protected]

J. P. McCreary A. Navarra Oceanographic Center Nova Southeastern IMGA Via Gobetti 101 Bologna Italy University Tel: + 39 51 639 8014 800 N. Ocean DR Fax: + 39 51 639 8132 Dania, FL 33020 USA E-mail: [email protected] Tel: + 1 954 920 1909 Fax: + 1 954 921 7764 E-mail: [email protected]

47 P. Niiler Tel: + 33 5 61 33 2926 Scripps Institution of Oceanography Fax: + 33 5 61 25 3205 La Jolla, CA 92093-0230 USA E-mail:[email protected] Tel: + 1 619 534 4100 Fax: + 1 619 534 7931 M. Rodwell E-mail:[email protected] Hadley Centre Met Office London Road Bracknell J. D. Opsteegh RG12 2SY United Kingdom KNMI Tel: + 44 344 856 751 P.O. Box 201 3730 AE De Bilt The Netherlands Fax: + 44 344 854 898 Tel: + 31 30 2206 700 E-mail: [email protected] Fax: + 31 30 2202 570 E-mail: [email protected] R. Saravanan NCAR T. Osborn P.O. Box Climatic Research Unit University of East 3000 Boulder, CO 80307-3000 USA Anglia Tel: + 1 303 497 1329 Norwich NR4 7TJ United Kingdom Fax: + 1 303 497 1700 Tel: + 44 1603 592 089 E-mail: [email protected] Fax: + 44 1603 507 784 E-mail: [email protected] F. Schott Insitut für Meereskunde Düsternbrookerweg 20 V. Pavan 24105 Kiel Germany CINECA via Magnelli 6/3 Tel: + 49 431 597 3820 40033 Casalecchio di Reno Bologna Italy Fax: + 49 431 597 3821 Tel: + 39 51 6171 474 E-mail: [email protected] Fax: + 39 51 6132 198 E-mail: [email protected] F. Selten KNMI G. Philander P.O.Box 201 AE De Bilt The Netherlands Princeton University Tel: + 31 30 2206 761 Dept of Geoscience Fax: + 31 30 2202 570 Gyot Hall Princeton NJ USA E-mail: [email protected] Tel: + 1 609 258 5683 Fax: + 1 609 258 2850 J. Servain E-mail: [email protected] Centre ORSTOM B.P. 70 29280 Plouzané France G. Piovesan Tel: + 33 2 98 22 4506 Disafri University of Tuscia Fax: + 33 2 98 22 4514 Via S.C. de Lellis 01100 Viterbo Italy E-mail:[email protected] Tel: + 39 761 357 389 Fax: + 39 761 357 387 K. Speer E-mail: [email protected] CNRS IFREMER LPO B.P. 70 29280 Plouzané France G. Reverdin Tel: + 33 2 9822 4566 LEGOS Fax: + 33 2 9822 4496 18 Av. E. Belin, 31401 Toulouse France E-mail: [email protected]

48 D. Stammer M. Visbeck Massachusetts Inst. of Technology LDEO Columbia University 77 Mass. Av. RT 9W Palisades, NY-10964 USA Cambridge, MA 02139 USA Tel: + 1 914 365 8531 Tel: + 1 617 253 5259 Fax: + 1 914 365 8157 Fax: + 1 617 253 4464 E-mail: [email protected] E-mail: [email protected] J. M. Wallace D. Stephenson University of Washington Meteo-France P.O. Box 351640 2 Av Coriolis 31057 Toulouse France Seattle Washington 98195 USA Tel: + 33 561 07 9698 Tel: + 1 206 543 7390 Fax: + 33 561 07 9610 Fax: + 1 206 685 3397 E-mail: [email protected] E-mail: [email protected]

R. Sutton H. Wanner University of Reading CGAM University of Bern Institute of Geography Dept. of Meteorology Earley Gate, Hallerstrasse 12 CH-3012 BERN Switzerland Whiteknights Tel: + 41 31 631 88 85 P.O. Box 243, Early Gate Reading RG6 6BB Fax: + 41 31 631 85 11 United Kingdom E-mail: [email protected] Tel: + 44 1189 875 123 (x 7842) Fax: + 44 1189 318 316 R. G. Williams E-mail: [email protected] Liverpool University Oceanography Laborato- ries L. Terray P.O. Box 147 Liverpool L69 3BX United CERFACS 42 Kingdom Avenue Gustave Coriolis 31057 Toulouse Tel: + 44 151 794 5136 Cedex France Fax: + 44 151 794 4090 Tel: + 33 561 19 3015 E-mail: [email protected] Fax: + 33 561 19 3000 E-mail: [email protected]

J. T. Todd NOAA Office of Global Programs 1100 Wayne Avenue Silver Spring Maryland 20910-5603 USA Tel: + 1 301 427 2089 Fax: + 1 301 427 2073 E-mail: [email protected]

U. Ulbrich Institut für Geophysik und Meteorologie Uni Köln Kerpener Strasse 13 D-50923 Köln Germany Tel: + 49 221 470 3688 Fax: + 49 221 470 5161 E-mail: [email protected]

49