Summary of the Joint CCSM
Climate Variability and Paleoclimate Working Group Meeting
February 2-3 2004, NCAR

Working Group co-Chairs: Bette Otto-Bliesner, Paleoclimate
Michael Alexander and Clara Deser, Climate Variability

Summary of Talks
Variability in CSM simulations for four climates: present, glacial, Eocene, and Cretaceous- Part 1 (M. Huber, Purdue University) Matt Huber found tropical interannual ENSO-like variability in all climate simulations but with varying amplitudes. This argues against a 'Permanent El Nino' condition for warm climates. The warm climates simulated showed enhanced tropical variability and weaker polar variability. (Huber.Caballero, pdf)

Variability in CSM simulations for four climates: present, glacial, Eocene, and Cretaceous- Part 2 (R. Caballero, University of Chicago) Rodrigo Caballero discussed specific differences in the pattern, amplitude, and frequency of the variability. He speculated that the differences may be related to mean state and polar amplification changes. (Huber.Caballero, pdf)

Changes in ENSO global teleconnections since the Last Glacial Maximum (C. Morrill, NCAR ASP) Carrie Morrill analyzed the global teleconnections of surface temperature, sea level pressure, and precipitation to ENSO in CSM simulations for the present, Holocene, and Last Glacial Maximum (LGM). She found increased (decreased) sensitivity of sea level pressure and surface to tropical Pacific SST anomalies over Asia (North America) at LGM compared to present.

How the extra tropical atmosphere has been altered by changes in the tropical climates during the last glacial-interglacial cycle (M. Khodri, LDEO) Myriam Khodri compared CSM simulations for the LGM and early Holocene (11 ky BP). She presented results documenting changes in the latitude of the jet streams and transient eddies and a change in the Walker circulation. (Khodri, pdf)

Dynamical constraints of the tropical Pacific sea surface temperatures on the mid-Holocene hydroclimate changes over North America (S-I. Shin and co-authors: R. S. Webb and Prashant D. Sardeshmukh, NOAA) Sang-Ik Shin presented sensitivity experiments with CCM3.10 that show the proxy-indicated climate for the mid-Holocene of wet in north Africa and dry in North America can be reproduced if the tropical Pacific mean state is "La Nina-like".

External forcing and internal modes of climate: New perspectives on natural climate variability on decadal-to-centennial time scales (C. Ammann, NCAR) Caspar Ammann presented results from transient simulations of the climate of the last millennium with CSM1. Decadal-to-century time scales of solar influence are identified in the simulations. Additional results shown from collaborations with Mike Mann, as well as recent simulations by Mann and Cane, suggest that the El Nino-Southern Oscillation system might respond to external forcing such as volcanic eruptions and changed solar forcing. Ocean dynamic feedbacks could be responsible to modify the mean state at different time scales. (Ammann, pdf)

Variations of the Hadley Circulation in CCM3 and CAM2 (X. Quan, CDC) Xiao-wei Quan examined changes in the Hadley Circulation using an ensemble of CCM3 and CAM simulations forced with the observed evolution of SST for the period from 1950 to 1999, station observations of precipitation, and the NCEP reanalysis data. The changes in the Hadley Circulation appear to be seasonal dependent: the simulated Hadley Circulation is intensified during the northern winter since 1950 while no long-term trend exists for the northern summer. About half of the winter intensification in the Hadley Circulation can be explained by the linear response to the increased amplitude of El Nino in the central and eastern tropical Pacific Ocean. Quan attributed the remaining part to i) the nonlinear response to ENSO, i.e., the magnitude of the response is larger for El Nino than La Nina events, and ii) warming in the tropical Indian and western Pacific Oceans, which exhibits an interdecadal trend rather than increase in the magnitude of the interannual variability.

Modes of variability in a changing climate (G. Branstator, NCAR) Grant Branstator presented results from the "Dutch Challenge Project" that consists of a 62-member ensemble integration of CSM1 at T31 resolution conducted by the KNMI and University of Utrecht. Each integration differs only in its initial conditions and consists of observed greenhouse gas, sulfate aerosols, and solar and volcanic forcing from 1940 to 2000 and "business-as-usual" IPCC greenhouse gas scenario forcing for 2001-2080. Branstator focused on the simulation of the NAO in the 62 member ensemble and whether it showed any systematic response to the imposed forcings by the end of the integration period. He found no significant response of the NAO, either in terms of its polarity or amplitude. He briefly documented the barotropic wave guide and showed differences in this feature between the first and last 30 years of the integration; these differences showed some linkages to rainfall changes in the Tropical Indo-Pacific. (Branstator, pdf)

Mean-state or variability: Using numerical models to decipher ambiguous records of paleoenvironmental change (N. Diffenbaugh, UCSC) Noah Diffenbaugh used results from a regional climate model simulation at 6000 years ago to help interpret ambiguous proxy records that suggest both colder SSTs and weaker upwelling. The model results show a longer and weaker upwelling season compared to present, which can be linked to orbital forcing and agrees well with proxy records. (Diffenbaugh, pdf)

Tropical Atlantic-Pacific interaction (Saravanan, NCAR) Saravanan examined to what degree the Tropical Atlantic influences the Tropical Pacific, noting that the dominant signal (ENSO) is the Tropical Pacific affecting the Tropical Atlantic. His motivation was in part the large SST biases in CCSM2 in both the Tropical Atlantic and Pacific. To examine this question, two experiments were compared: in one, a slab ocean was coupled to CAM2 with the required "Q-flux" needed to maintain the SST climatology simulated by CCSM2; in the other, this "Q-flux" was set to zero in the Tropical Atlantic. The differences in SSTs over the Tropical Pacific were noticeable (~ 2K cooling without the Atlantic Q-flux) between these two experiments, demonstrating that the Tropical Atlantic does affect the mean state of the Tropical Pacific. He showed further that the speed at which the Atlantic influence "propagated" into the Pacific was sensitive to the mixed layer depth used in the slab ocean. (Saravanan, pdf)

Characterization of millennial scale climate variability (G. Roe, University of Washington) Gerard Roe presented results from statistical models that characterize the millennial-scale climate variability in Greenland and Antarctica ice core records. He finds, using a first-order autoregressive model, that the Byrd record from 80-20 ky BP can be modeled as red noise while the GISP2 record contains more than red noise with a peak at 1500 years. The characteristic time scales for the two ice cores vary between glacial and Holocene, suggesting significantly different mechanisms of climate variability.

Developing a millennium-length record of the Asian Monsoon (D. Anderson, NOAA) Dave Anderson constructed a new proxy record of monsoon variability for the last 2000 years from indicators of upwelling in Arabian Sea sediments. This record shows monsoon winds stronger during the Medieval Warm Period (1200-1400AD), weaker during the Little Ice Age several centuries ago, and an increase in the last century.

Megadroughts in the Indian monsoon and southwest U.S. regions in a 1350 year global coupled model simulation (G. Meehl, NCAR) Jerry Meehl discussed the occurrence of "megadroughts" in the Indian Monsoon and southwest U.S. in a 1350 year coupled model simulation using the Parallel Climate Model at T42 resolution (CCM3 atmosphere, POP ocean, CSM land surface model). Meehl began with proposing a definition of "megadrought": an 11 year running mean of regionally-averaged precipitation anomalies less than zero for at least 20 consecutive years. The Indian Monsoon (SW U.S.) exhibited 10 (9) megadroughts over the period of integration, roughly one every 140 to 150 years. Meehl showed that these megadroughts are linked to global-scale patterns of precipitation, SST and SLP anomalies, and that droughts in the two regions tend to be anticorrelated. He suggested that the strength of the oceanic subtropical meridional overturning circulation may be a regulator of megadroughts in these two regions. (Meehl, pdf)

Twentieth century climate change: Understanding the effect of Indian Ocean warming (M. Hoerling, NOAA/CDC) Marty Hoerling presented a talk entitled "A pre-mature indictment of Indian Ocean warming as a cause for Sahelian Drought?". Hoerling examined the role of the oceans in the increased sub-Saharan aridity since 1950 by means of AMIP ensemble integrations with CCM3. In particular, he examined the influence of the recent positive SST trends in the Tropical Indian Ocean that coupled model simulations show are a response to increasing greenhouse gas concentrations. He found that the Indian Ocean SST warming led to rainfall increases (not decreases) over the Sahel. Instead, the inter-hemispheric SST contrast in the Atlantic sector appeared to be a primary cause for the increased Sahelian aridity, with a cold NH/warm SH favoring reduced precipitation. (Hoerling, pdf)

Paleomodeling of decreased aridity in northern Africa in the Holocene (B. Otto-Bliesner, NCAR) Bette Otto-Bliesner presented results from CCSM2 simulations for 8500 yr BP compared to present. Milankovitch orbital variations in solar radiation for this past period enhance the North African monsoon with a more northern position of the ITCZ, a stronger tropical easterly jet, a displacement north of the African easterly jet, and stronger low-level westerlies at low latitudes. (Otto-Bliesner, pdf)

Response of the Mediterranean Seas circulation to orbital forcing in CCSM2: Implications for sapropel formation (E. Brady, NCAR) Esther Brady presented results from the Mediterranean region of the ocean component of CCSM2 for 8500 and 130000 years ago when African monsoons were enhanced relative to present day. Results show reductions in Mediterranean outflow transport and salinity at Gibraltar and a reversed circulation in the Eastern Mediterranean are forced by changes in the surface hydrological cycle of the basin related to increased Nile discharge. Model results agree well with proxy salinity data. (Brady, pdf)

Marine productivity changes associated with the Permian-Triassic boundary (A. Winguth, University of Wisconsin) Arne Winguth discussed results from simulations at the Permian/Triassic boundary that show meridional climate zones in the model in good agreement with geological evidence. Ocean meridional overturning and horizontal gyre circulations are vigorous and sensitive to different levels of atmospheric CO₂ and high latitude freshwater input. (Winguth, pdf)

The South Atlantic Circulation at the LGM in the NCAR Coupled Model (I. Wainer, University of Sao Paulo, Brazil) Ilana Wainer compared South Atlantic ocean circulation in CSM1.4 simulations for the Last Glacial Maximum to a present-day control. Results show the intensification in the southward Brazil Current transport at LGM is related to changes in the global thermohaline overturning circulation, which shows a decrease in NADW and an increase in AABW at LGM.

North Pacific decadal variability in CCSM2 control experiment (Y.-O. Kwon, NCAR) Young-Oh Kwon presented ongoing research on the nature and origin of decadal (10 to 20 year period) variability in SSTs over the North Pacific and related climate anomalies in the 1000-year control run of CCSM2. He showed that over the Kuroshio Current Extension region, the net surface heat flux damps rather than drives the decadal SST variations, and that anomalous Ekman heat transport acts as a positive feedback response to the SST anomalies via the mean current advecting the anomalous SST gradient. Precipitation was also shown to respond to SST anomalies. The SST fluctuations originate off Japan and are advected eastward, along with the net surface heat flux and precipitation response. The subsurface thermal structure of the decadal variability was not yet investigated. (Kwon, pdf)

CAM coupled to a mixed layer ocean model: Model physics and climate (M. Alexander, NOAA/CDC) Mike Alexander presented an overview of a coupled model, developed by Christophe Cassou, in which CAM (version 2) is coupled to an ocean mixed layer model (MLM) and the NCAR thermodynamic sea ice model. The ocean model, which has the same resolution as CAM, predicts the temperature, salinity, and depth of the ocean mixed layer. Once a flux correction is applied, to accommodate the absence of ocean heat transport, the grid values of the monthly mean SSTs from an 80-year control run are generally within a degree of the corresponding observed average. Preliminary analyses indicated that the model exhibited variability in the Indian Ocean that influenced precipitation in the surrounding land areas and the circulation over the Northern Hemisphere. The model's NAO exhibits interannual and decadal variability, which in turn affects the SST, sea ice, and mixed layer depth. These North Atlantic SST anomalies are very similar to the observed "tripole pattern," and recur from one winter to the next via the reemergence mechanism. (Alexander, pdf)

Development status and plans for the new GFDL coupled climate model (T. Delworth, GFDL) Tom Delworth presented an overview of the development of the new GFDL coupled climate model, including the model formulation and simulation characteristics. The coupled model includes a new gridpoint atmosphere model (2.5 deglon by 2 deglat by 24 levels), along with a new ocean model based on the MOM4 code (varying horizontal resolution from 1/3 deg to 1 deg, with 50 vertical levels). There was also a description of the new land and sea ice models employed. The new coupled model will be used for a wide range of applications, from ENSO forecasts to multicentury global climate change. Model biases, including a tendency for a cold Northern Hemisphere, and a fresh bias in the upper ocean, were discussed. It was also shown that the model has very good skill in ENSO forecast mode, based on a suite of hindcast experiments. Initial experiments are underway with the new model for the upcoming IPCC report. (Delworth, pdf)

CCSM3 (W. Collins, NCAR) Bill Collins, chair of CCSM, discussed the CCSM3 simulations that will be released to the public on May 18, 2004, as well as plans for a Special Issue of Journal of Climate devoted to CCSM3 (see CCSM web page at NCAR - www.cesm.ucar.edu).

Ocean model results from the CCSM3 present-day integrations with a T42 atmosphere (G. Danabasoglu, NCAR) Gokhan Danabasoglu discussed aspects of the ocean model simulation in CCSM3. He listed salient features of the ocean model that is based upon the POP 1.4 code, including physical parameterizations. One novel aspect is the inclusion of a diurnal cycle in shortwave heat flux into the ocean, although the model is coupled to the atmospheric model only once per day. He also characterized the slow drifts in heat and salt over the 300 year (500 year) integration of the T85 (T42) version. He also compared the SST climatologies from the T85 and T42 versions with observations and each other, presented the climatology of the upper ocean thermal structure in the equatorial Pacific Ocean, and documented the Meridional Overturning Circulation and associated heat transports. (Danabasoglu, pdf)

CCSM3 T42 variability (C. Deser, NCAR) Clara Deser presented a brief overview of some aspects of climate variability in the T85 and T42 control integrations of CCSM3 that have run out to more than 500 years at T42 and 300 years at T85. This material is documented in the CCSM web pages (see www.cesm.ucar.edu). In particular, she showed the patterns of SLP, SST, air temperature, and precipitation variability associated with the Arctic Oscillation and "Pacific Decadal Oscillation." She also described the ENSO variability, noting that the model's tropical precipitation and SLP response to SST anomalies is too localized and therefore may not provide a realistic spatial pattern for forcing atmospheric teleconnections to higher latitudes. (Deser, pdf)

Report on the CCSM Isotope Workshop (N. Mahowald, NCAR) Natalie Mahowald summarized the outcome of the CCSM Isotope Workshop held 13-15 January 2004 at NCAR. This workshop identified science questions, priorities, and a work plan  for modeling water and carbon isotopes. Ongoing collaborative efforts and difficulties in implementation were identified.

Climate Variability Working Group Breakout Session
Proposed Model Simulations
The group discussed what additional runs are needed for papers in the Journal of Climate Special Issue and what runs are needed in the next 6 months to 1 year time frame.

Runs for J. Climate Special Issue: 5-member T85 AMIP integration of CAM3, 1950-present, forced with observed global time-varying SSTs plus greenhouse gas, aerosol, and volcanic and solar forcing. These runs would allow for the best comparison with observations, and would provide a useful benchmark for comparing with the CCSM3 "Climate of the 20th Century" runs, especially given the deficiencies in ENSO precipitation patterns in CCSM3 (see talk by Deser). The CVWG also requested that selected atmospheric fields are saved at 4x daily resolution from a 50-year segment of the T85 "Climate of the 20th Century" run for use in stormtrack diagnoses, weather regime analyses, and analyses of extreme events.

Runs in the next 0.5 to 1 year (in order of priority):

Proposed Papers for the J. Climate Special Issue:

All are relevant to the 5th main theme of the Special Issue identified by the SSC ("Major modes and manifestations of climate variability") and make use of the century-long control integrations at both T85 and T42 resolution, as well as the proposed T85 AMIP integrations and the T85 "Climate of the 20th Century" run. Four papers were proposed:

Paleoclimate Working Group Breakout Session

Discussion and Recommendations:

The Paleoclimate Working Group members met on the 2nd afternoon to discuss the Journal of Climate Special Issue and CCSM3 Paleoclimate Working Group simulations for this issue, as well as possible simulations under a new CSL allocation that will be submitted this summer. The latter includes a number of time slices from deep-time through the Quaternary, ensembles for the last millennium, and transient simulations for the Holocene, deglaciation, and initiation of glaciation. The members felt that full implementation (atmosphere-land-ocean-sea ice) of water isotopes into CCSM should be a high priority. Some members also were interested in investigating the utility of adopting M. Alexander's mixed-layer ocean model for deep-time applications. Finally, the recommendation was made that a 2 to 3 day CCSM user workshop be held.

Proposed J. Climate Papers and Model Simulations: