meeting summary

Wadati Conference on Global Change and the Polar Climate, 7-10 November 1995, Tsukuba, Japan

John E. Walsh,*+ Hiroshi L. Tanaka,# and Gunter WeIIer+

1 .Introduction mate change have served as scientific stimuli for the recent surge of polar climate research. At the same The Wadati Conference on Global Change and the time, technology has provided an unprecedented ar- Polar Climate was held in Tsukuba, Japan, on 7- ray of new research opportunities through advances 10 November 1995. The conference brought together in satellite measurements, ice-coring capabilities, au- approximately 60 scientists from North America, Eu- tomated measurement techniques, and more power- rope, Japan, and Australia to review and assess recent ful computers. The results of recent research on polar polar climate variations in the context of global climate have accumulated at an increasingly rapid rate change. The conference consisted of oral presenta- in disciplinary as well as interdisciplinary outlets, cre- tions, a poster session, and a plenary discussion ses- ating a need for the synthesis and assessment that were sion in which an attempt was made to synthesize the the objectives of this conference. results of polar climate research into a consistent pic- In order to achieve the goal of a synthesis and as- ture of recent changes. sessment of information on recent polar climate varia- The need for the current assessment of polar cli- tions, the Wadati Conference's Organizing Committee mate in the global context arises, in part, from sev- brought together an international group of experts to eral generations of global climate model experiments address four topics that are central to the role of po- indicating that the strongest greenhouse effects might lar climate in global change. The conference's four occur in the polar regions. The recent (1995) assess- sessions were devoted to these topics. Session 1 ad- ment of the Intergovernmental Panel on Climate dressed the observational evidence of climate change. Change indeed cites the polar amplification of near- The emphasis of this session was on instrumental mea- surface warming, at least in the Arctic, as a charac- surements, thereby creating a focus on changes over teristic of greenhouse experiments with global models. the past century or two. The presentations in session Feedbacks involving sea ice, snow cover, polar 2 addressed the processes underlying the interactions clouds, the hydrologic cycle, and the stratification of and feedbacks within the polar climate system. These the polar oceans have been proposed as mechanisms feedbacks include the albedo-temperature feedback, by which the polar regions might amplify global ocean-atmosphere exchanges, cloud-radiative inter- change. However, these feedbacks are poorly under- actions, and other polar processes contributing to glo- stood, and their treatments in global models are largely bal change. Session 3 dealt with the longer timescales, unexplored. The uncertainties surrounding polar cli- over which paleoclimatic reconstructions provide a window on climate variations. The polar climate re- constructions draw heavily upon ice cores. Finally, the *Department of Atmospheric Sciences, University of Illinois, Urbana-Champaign, Urbana, Illinois. increasing use of models to examine polar climate +Geophysical Institute, University of Alaska, Fairbanks, Alaska, variability was the subject of session 4. Papers in this institute of Geoscience, University of Tsukuba, Tsukuba, Japan. session spanned the range of regional and global mod- Corresponding author address: Dr. John E. Walsh, Department els, including the coupled ocean-atmosphere global of Atmospheric Sciences, University of Illinois, Urbana- models of several major modeling centers. The ple- Champaign, 102 Atmospheric Sciences Building, 105 S. Gregory Ave., Urbana, IL 61801. nary discussion session represented an attempt to rec- In final form 2 February 1996. oncile the model results with the observational and ©1996 American Meteorological Society process-based studies of the first three sessions.

1268 Vol. 77, No. 6, June 7 996

Unauthenticated | Downloaded 10/06/21 10:25 AM UTC 2. Evidence of climate change T. Yasunari showed that subarctic features of the re- cent (1964-93) temperature changes extend from the Session 1 began with an overview by J. Walsh of surface to 500 mb. Yasunari's EOF analysis revealed global climate model simulations of greenhouse two leading modes of Northern Hemisphere tempera- changes in the Arctic, with an emphasis on the sea- ture variability, one with a time series corresponding sonality and regionality of the changes. The largest to hemispheric snow cover and having an amplitude greenhouse warming projected by the models is in the maximum near the surface, and the other with a time central Arctic Ocean during autumn and early win- series corresponding to equatorial SST and having a ter; this warming is largely a response to a thinning maximum amplitude at 500 mb. Yasunari also showed and/or disappearance of sea ice during the summer and that wintertime soil temperatures (120 cm) at Yakutsk early autumn. While the Arctic warming is generally in Siberia have increased since 1973. H. Tanaka pre- diminished somewhat when the models include sul- sented evidence of several jumps in climate variables, fate aerosol effects and coupling to a deep ocean, a in particular in Alaskan temperatures in 1989 and in polar amplification is a characteristic feature of the vorticity over the Arctic Ocean in 1988-89. Tanaka model-projected greenhouse warming in the North- argued that these and other more transient shifts may ern Hemisphere. In contrast to the model projections be triggered by abnormal persistent blocking. of a maximum warming over the Arctic Ocean, ob- Several speakers in session 1 also addressed recent servational data for the past 30-40 years show a variations of sea ice. R. Barry discussed the impor- warming that appears to be strongest over the north- tance of monitoring changes in cryospheric variables ern land areas and during spring and winter, although (particularly snow cover, sea ice, glaciers). Remote there has not yet been a systematic integration of es- sensing of spatial characteristics complemented by timates of recent temperature changes over arctic sea surface measurements of snow water equivalent, sea ice and land areas. ice thickness, and glacier mass balance, in particular, As noted by C. Folland, much of the wintertime is required. J. Zwally noted that the passive microwave warming over northern Eurasia is due to an increase database points to an absence of trends in Antarctic of westerly airflow. Moreover, the signal-to-noise sea ice and a very small (>1% per decade) decrease ratio of the greenhouse warming in GCM simulations of Arctic sea ice. This decrease is concentrated in the is smaller in polar regions than in tropical latitudes. eastern Siberian sector and is associated with strong P. Jones showed that the spatial scale of summer tem- low pressure systems in 1990 and 1993, according to perature anomalies is much smaller in the northern Barry. Zwally also stressed the importance of consis- polar regions than in lower latitudes. Jones also tent observations over long intervals; the use of dif- showed that interannual variations of temperature ferent passive microwave sensors for 1978-87 and over Antarctica are divorced from those of the rest 1987-94 has created uncertainty over a recently re- of the Southern Hemisphere during the period of in- ported decrease of Arctic ice area in summer. N. Ono strumental data. More comprehensive integrated as- showed that there is considerable interannual variabil- sessments of ocean surface temperature and sea ice ity in sea ice extent as well as glacier mass balances, margin variations are now becoming possible with the while P. Wadhams showed that submarine measure- newly available Global Ice and Sea Surface Tempera- ments of ice draft (thickness) may be useful in trend ture (GISST) dataset, version 2.2, compiled at the detection if sampling errors are estimated from a se- Hadley Centre of the U.K. Meteorological Office ries of short (-50 km) transects obtained in a particu- (UKMO) by Folland and colleagues. A cooling of the lar area and time. subpolar North Atlantic over the past several decades is a consistent result of temperature trend analyses based on sea surface temperatures as well as land sta- 3. Processes, interactions, and feedbacks tion data. In the stratosphere, temperature trends reported by Session 2 included a report by K. Aagaard on re- K. Labitzke show a negative trend of approximately cent Arctic/subarctic ocean variations obtained from -0.6°C per decade at 50 mb over the central Arctic ship cruises. In the Arctic Ocean, a warming of the over the past several decades. A link to the solar cycle upper part of the water column is approaching 1°C; is present in the upper stratosphere, and the link also in the vicinity of the Lomonosov Ridge, signs of appears to be present in the lower troposphere. warming extend down to 1-2 km. In the Greenland

12 57 Bulletin of the American Meteorological Society

Unauthenticated | Downloaded 10/06/21 10:25 AM UTC Sea, 1994 was the warmest of the past 30 years, prob- to the aggregate of ice-core evidence from around the ably because the cooling from deep convection has world. Evidence of the Little Ice Age is spotty, as the been weak during much of this period. Aagaard con- cooling of that period is not detectable in ice cores cluded that "we are beginning to accumulate evidence from West Antarctica, confirming Jones's finding of a change in ocean properties in the Arctic Ocean from instrumental temperature records that Antarctica and the Greenland Sea." is divorced from the rest of the Southern Hemisphere. E. Augstein reported on aircraft measurements On longer timescales, a cooling over the past 4000 made over Fram Strait, where the greatest water mass years, particularly around 2000 BP, has been inferred and heat exchanges between the Arctic and Atlantic from <5180 measurements of ice from the Antarctic Oceans appear to occur. The measurements showed Plateau. The <5lsO data are ambiguous and may be that the net vertical radiative flux at the ice-covered indicative of temperature changes or of changes in the air-sea interface significantly exceeds the turbulent source region of precipitation. The ice-core data also sensible heat flux but that the latter is up to 10 times indicate that precipitation has been increasing over all larger than the net radiative flux over open water, parts of Antarctica. The increase is especially large particularly in an off-the-ice cold air flow. Exchange over the Antarctic Peninsula and the plateau/South coefficients for the surface momentum and heat fluxes Pole region, where there has been a 20% increase in in the marginal ice zone have been derived from the accumulation since 1954. The percentage increase is aircraft observations; they are essential for the flux even greater over some Antarctic regions if longer computation in ocean and sea ice models. time periods are considered. T. Yamanouchi, describing a study of the role of R. Koerner reviewed evidence obtained from ice polar regions in the earth radiation budget, reported cores in the Northern Hemisphere. The complete dis- that the effect of clouds on the earth-atmosphere appearance of the Laurentide ice sheet did not occur energy budget is reduced over sea ice; conversely, until about 6000 BP, which is shortly after the time the effect of sea ice is reduced when clouds are when a long cooling began. This cooling is apparent present. K. Yamazaki described a study of the atmo- in circumpolar cores but not in the Greenland sum- spheric water budgets of the two hemispheres, show- mit cores. The long cooling was ended only by the ing that the poleward moisture flux convergence recent warming, which may be related to anthropo- is larger in the Northern Hemisphere poleward of genic factors. The Northern Hemisphere ice-core 70° and in the Southern Hemisphere equatorward of records also suggest a large release of freshwater 70°. A. Ohmura showed that a state-of-the-art global around 8500 BP, which may represent the last of the circulation model (ECHAM3) is able to reproduce Dansgaard-Oeschger events. M. Nakawo addressed quite accurately the precipitation fields over the evolution of the Shirase drainage basin of the East Greenland and Antarctica, provided that the model is Antarctic ice sheet. A rapid lowering of the surface run at high (T106) resolution. When run in a green- elevation has been observed recently in this region. house (2 x C02) mode, the same model predicts that A dynamical response to an event in the past few thou- the increase of sea level resulting from increased melt sand years is suggested. over Greenland will be largely offset by increased ac- J. Svoboda addressed the susceptibility of the east- cumulation over Antarctica. ern Canada-Greenland region to glaciation under present conditions, even if the near-term scenario in- cludes global warming. Much of the vegetation pres- 4.Paleoclimatic reconstructions ently found in the high Arctic is not older than 150 years, implying a rapidly varying ecology in this re- Session 3 emphasized the recent information de- gion. The underlying deep and cold permafrost rep- duced from ice cores. The Antarctic ice core records resents a concealed glaciation, which could be ready surveyed by E. Mosley-Thompson suggest a recent to support new glaciation if increased precipitation is warming over most of Antarctica, including a strong delivered. Interestingly, the onset of the last ice age warming over the Antarctic Peninsula consistent with occurred when the temperature and C02 concentra- recent station data. However, parts of the East tion were high. The recent pattern of global "warm- Antarctic region appear to have experienced a recent ing" shows a cooling over northeastern Canada. The cooling, as shown by both Jones and Mosley-Thomp- possible abruptness of a future onset of glaciation in son, and thus are somewhat anomalous with respect this region remains an open question.

1270 Vol. 77, No. 6, June 7 996

Unauthenticated | Downloaded 10/06/21 10:25 AM UTC 5. Climate modeling warming also led to a more drastic weakening of the thermohaline circulation. These model results suggest Session 4 included a summary by A. Lynch of re- that the monitoring of the global ocean must be a high sults from a regional Arctic climate system model. The priority in the greenhouse-detection strategy. model, which includes interactive atmosphere, sea ice, While the North Atlantic exerts a strong control and land/vegetation components, reproduces observed over the thermohaline circulation, 60% of the ocean summer air temperatures over northern Alaska, al- water is colder than 2°C, implying a source in the though the agreement is weaker during winter. Results Antarctic subpolar ocean. Thus, heat fluxes and from global atmosphere-ocean coupled models were watermass formation processes in the Southern Ocean presented by representatives of three major modeling must be modeled realistically if global ocean simula- centers [Meteorological Research Institute (MRI)- tions are to be successful. A. Gordon presented recent Japan, the UKMO, and the Geophysical Fluid Dynam- estimates of the heat fluxes to the mixed layer from ics Laboratory (GFDL)]. T. Motoi showed that the the Weddell Deep Water. These fluxes are much MRI coupled model undersimulates Antarctic sea ice larger (40 W nr2 during winter) than in the Arctic and oversimulates Arctic sea ice in the present climate, Ocean, effectively limiting Antarctic sea ice thickness suggesting that the formulations of polar clouds and to about 0.5 m in this area. Based on the history of albedo require closer investigation. A. Noda described the Weddell Polynya, the Weddell Sea appears to be the results of a transient C02 simulation with the same delicately poised with respect to deep convection. model. The response in the polar regions was found B. Saltzman described the modeling of global ice to be strongly shaped by the inclusion of leads, which variations over the past 5 Myr. When driven by GCM- eliminated the polar amplification in the Antarctic and derived estimates of global ice parameters, this model reduced polar amplification by 50% (relative to indicates that the 100 kyr oscillation of global ice mass GFDL) in the Arctic. is internal rather than driven by Milankovitch cycles. By contrast, H. Cattle showed that the UKMO The model also indicates that ice-age cycles will not Hadley Centre's "unified model" produced a strong continue into the next 100 kyr if C02 concentrations (5°-10°C) warming over the Southern Ocean during are maintained near 350 ppm. A. Abe-Ouchi also de- winter when C02 was slowly increased. Perhaps more scribed an ice model, which when applied to the importantly, the inclusion of sulfate aerosol reduced Greenland ice sheet, predicts ice-sheet disintegration the Arctic warming by some 20% and eliminated the for a warming of only 3°-4°C. decrease of Arctic sea ice in the period correspond- Finally, M. Sugi described experiments addressing ing to 1979-95 (when Zwally had argued earlier in Eurasian snow-monsoon relationships in the global the meeting that the observed decrease has been small model of the Japan Meteorological Agency. The link or nonexistent). At the time of doubling of greenhouse between the monsoon's direct forcing (land surface gas concentrations, changes of ice concentration and temperature) and the late-winter snow fades away by thickness were found to correlate better with areas of June or July, leading to an insignificant correlation maximum warming than did ice extent, suggesting between snow and the monsoon index. These results that the latter may not be the optimum choice of a imply that the more important factor(s) are the inter- variable for greenhouse monitoring. In the ocean com- nal dynamics of the monsoon or influences from out- ponent of the unified model, the increase of C02 was side the monsoon region. accompanied by a warmer core of the North Atlantic inflow to the Arctic, although the model's Arctic Ocean density structure remained stable. 6. Plenary discussions S. Manabe described the use of the GFDL coupled ocean-atmosphere model to examine the role of the The plenary discussion of the detection of the thermohaline circulation in abrupt climate change. greenhouse signal in polar regions revolved around When the surface of the North Atlantic Ocean (50°- two questions. 1) Are we now seeing the greenhouse 70°N) was artificially freshened by surface water in- signal in polar regions? 2) If not, how and when will put, the model's thermohaline circulation rapidly the signal manifest itself? In response to 1), the con- weakened and became more shallow. In addition, it ference papers and discussion led to Table 1, summa- was shown that the disproportionate increase of pre- rizing the observational evidence from the past few cipitation in high latitudes induced by greenhouse decades. A view that emerged on several occasions

12 57 Bulletin of the American Meteorological Society

Unauthenticated | Downloaded 10/06/21 10:25 AM UTC is that the "traditional" empha- TABLE 1. Summary of changes observed in the high latitudes over the last few decades sis on sea ice extent and Arctic of the twentieth century. Ocean air temperatures as early indicators may yield to the Parameter Arctic Antarctic search for a broader fingerprint involving information from ice Surface temperature Generally warmer (on land, Slightly warmer, but cores, sea ice concentration and in winter/spring, but some much warmer over cooling; unclear over sea ice) Antarctic Peninsula thickness, subpolar sea surface temperatures, subsurface polar Tropospheric Warmer (lowest layers) Not yet assessed ocean temperatures, and high- temperatures latitude precipitation. The latter set of variables shows some evi- Stratospheric Colder (summer only) Colder temperatures dence of recent changes that are consistent with those anticipated Precipitation Wetter over land, Not certain (wetter in from anthropogenic influences. unknown over sea ice East Antarctica) However, records are long for only surface air temperatures, Extreme weather Not yet assessed Not yet assessed and even these are not long Ocean temperatures Warmer (central Arctic) Weak warming enough to distinguish unam- biguously the natural and an- Snow cover extent Reduced (Eurasia, spring; thropogenic influences. also in Canada/Alaska) A recurring theme of the pre- sentations and discussion was Sea ice extent Slightly reduced No trend (East Siberian Sea) the possibility of an abrupt and surprising climate change in- Sea ice thickness Thinner in some regions Not yet assessed volving the high latitudes. In ad- (short-term record only) dition to the "thermohaline catastrophe" in which deep con- Ice-sheet elevation Higher (South Greenland, Higher (East Dronning no change in Canada) Maud Land) vection in subpolar oceans is the key process, the onset of glacia- Ice-sheet surface Increased Decreased tion is a terrestrial process with melting (South Greenland) (Queen Maud Land) its roots in the subpolar North Atlantic sector. This sector has Ice-shelf extent Reduced (Canada) Reduced (peninsula) indeed cooled in recent decades Permafrost extent Reduced (Alaska, — despite the "global" warming Canada, Siberia) during this period. Given the concomitant increase of precipi- tation in high latitudes, it is quite possible that the polar land areas as well as the sub- attention to the need to monitor the polar climate sys- polar oceans are delicately poised with respect to rapid tem carefully for rapid, complex, and perhaps surpris- climate reversals. ing changes in the near future. Regardless of whether such changes occur in the A preprint volume of the papers presented at the subpolar ocean and land components of the climate conference is available from H. L. Tanaka, Institute system, the results of state-of-the-art GCMs such as of Geoscience, University of Tsukuba, Tsukuba 305, the U.K. unified model and the Goddard Institute of Japan. Space Sciences model, both of which have been run with sulfate aerosol effects, suggest that the enhanced Acknowledgments. We wish to acknowledge the generous greenhouse effect should be detectable in the polar support of the Committee for Energy Policy Promotion, which organized the funding of the Wadati Endowment at the Univer- regions and elsewhere in the next decade or two. The sity of Alaska. The endowment not only provided funds over a Wadati Conference has highlighted the possible early 5-year period for research reported at the conference but also pro- indications of anthropogenic changes and has called vided travel allowances to many of the foreign conference attend-

1272 Vol. 77, No. 6, June 7 996

Unauthenticated | Downloaded 10/06/21 10:25 AM UTC ees. The companies that contributed to the endowment are Chubu vided by several agencies and organizations, including the Uni- Electric Power Company, Inc.; Electric Power Development versity of Tsukuba, the Meteorological Research Institute, the Company, Ltd.; Hitachi, Ltd.; Mitsubishi Heavy Industries, Ltd.; National Research Institute for Earth Science and Disaster Pre- Osaka Gas Company, Ltd.; The Industrial Bank of Japan, Ltd.; vention, the National Institute for Environmental Studies, the The Kansai Electric Power Company, Inc.; The Tokyo Electric National Institute of Polar Research, the Center for Climate Sys- Power Company, Inc.; Tokyo Gas Company, Ltd.; Toshiba tem Research of the University of Tokyo, the Tsukuba EXPO '85 Corporation; and Toyota Motor Corporation. Memorial Foundation, the Meteorological Society of Japan, and We also wish to acknowledge the travel and other support pro- the Central Research Institute of the Electric Power Industry.

% W 1 Wt y

Portland, Oregon Preprint 18-22July 1994 Volume

©I 994 American Please send prepaid Meteorological , orders to: Order Society. Softbound, B&W, 687 pp., m $50.00 list/$35.00 M members (includes shipping and llj handling). ll IDepartment, i American I Meteorological I Society, 45 Beacon I St., Boston, MA f 02108-3693.

12 57 Bulletin of the American Meteorological Society

Unauthenticated | Downloaded 10/06/21 10:25 AM UTC