1999

NCAR | UCAR | NSF | ASR 98

Climate Change Research Section

CGD ASR99 Home | ASR99 Home | CGD Home

CCR is also involved with Los Alamos National Laboratory (LLNL) and the Naval Postgraduate School (NPS) in developing and using a high-peformance coupled model, the Parallel Climate Model (PCM). This collaborative effort between UCAR and the Department of Energy is managed under a Cooperative Agreement that is aimed at using massively parallel processor (MPP) computers.

Paleoclimate Model Development and Applications

PaleoCSM

Bette Otto-Bliesner (Climate Change Research Section, CCR), Esther Brady (CCR), and Christine Shields (CCR) continued work on the PaleoCSM. The PaleoCSM is a version of the CSM that uses a T31 resolution for the atmosphere and land models and the new x3' grid for the ocean and sea ice models. The ocean model version incorporates a spatially varying, three-dimensional eddy viscosity coefficient (anisotropic viscosity). In addition, this version of the model allows specification of the solar constant, atmospheric trace gases, Milankovitch orbital variations of the solar insolation, continental configuration and ocean bathymetry, vegetation, and soil characteristics. The new ocean grid enhanced the latitudinal resolution to 0.9° from 10°S to 10°N and 1.8° at middle and high latitudes. This version better resolves tropical currents than the previous version with the x3 resolution and isotropic viscosities. The equatorial undercurrent at 140°W now has zonal velocities of 80cm/s and a realistic latitudinal width. A version of the PaleoCSM is now being developed with reduced background vertical diffusivity (fkph) to improve the simulation of interannual variability in the tropical Pacific Ocean.  Results show an improved signal with standard deviation of the NINO3 sea surface temperatures (SSTS) of 0.68°C in the simulation with fkph=0.15 compared to 0.71°C for the observational record from 1950-1979.

This figure (49K) showsthe annual-mean ocean zonal velocity at 140°W in PaleoCSM simulations with the x3- isotropic viscosity (top) and x3'-anisotropic viscosity (bottom). A realistic estimate of the observed undercurrent is about 100 cm/s.

This figure (11K) shows the NINO3 (red) and NINO4 (blue) standard deviations as a function of the background vertical diffusivity (fkph) in 40-year PaleoCSM simulations.

A tutorial for the PaleoCSM can be found on the Paleoclimate Model Working Group web page.

Present and Past Climate Coupled Simulations

Otto-Bliesner, Brady, and Shields have used the PaleoCSM (with fkph=0.5) to complete three fully coupled simulations for present-day, for pre-industrial times, and for the mid-Holocene (6000 years before present). The present-day simulation was forced by modern-day insolation (1367 W m^-2) and atmospheric trace gas concentrations (CO₂= 354.4 ppm, CH₄=1722.3 ppb, N₂O=308.4 ppb, CFC11=513.95 ppt, CFC12=462.67 ppt). The pre-industrial simulation was forced with insolation and trace gas concentrations appropriate to 1800 A.D. (1365 W W m^-2, CO₂= 280 ppm, CH₄=700 ppb, N₂O=275 ppb, CFC11=CFC12=0). The mid-Holocene simulation was forced with Paleoclimate Modeling Intercomparison Project (PMIP)-specified boundary conditions for orbital parameters but used the same insolation and trace gas concentrations as the pre-industrial simulation. The pre-industrial simulation was run for 400 years, the present-day was run for 200 years, and the mid-Holocene was run for 50 years.

Trends in the volume mean ocean potential temperatures were small: -0.06°C/century in the pre-industrial run and 0.003°C/century in the present-day run. Annual-average, global surface temperatures were 1.3°C cooler in the pre-industrial simulation compared to the present-day simulation with the largest cooling in the middle and high latitude oceans in regions of sea-ice formation. Ice areas were 10% greater in both polar regions in the pre-industrial run. Increased summer and early fall insolation at northern middle latitudes during the mid-Holocene led to delayed sea-ice formation during the fall and a 5% decrease of sea-ice annually compared to pre-industrial times.

Interannual variability of tropical Pacific SSTs has been better defined compared to previous versions with increased variability in the central and eastern Pacific. Atmospheric modes, such as the Arctic Oscillation and North Atlantic Oscillation were reproduced realistically, as was their expression on northern continental climates. A one-standard deviation positive change in the NAO index led to warming over Europe and extending into northern Asia and into central North America and cooling over the northwest Atlantic, Greenland, and eastern Canada.

The Paleoclimate Group (Otto-Bliesner, Brady, and Shields)   is also involved in a multi-model intercomparison of coupled model simulations for the mid-Holocene as a subproject of PMIP and PAIN (Pan-Arctic Initiative), an international consortium of ecologists, paleoecologists, paleoclimatologists, and climate modelers studying changes in the diversity of tundra vegetation for past and future climates.

This figure (90K) shows annual-average SSTs for the pre-industrial simulation (top), present-day simulation (middle), and difference map, pre-industrial minus present-day.

This figure (21K) shows daily Northern and Southern Hemisphere ice areas smoothed by a 365-day running mean for present-day (green), pre-industrial (red), and mid-Holocene (blue) simulations. Present-day observed values are approximately 1x10¹³ m² for the Northern Hemisphere and 0.85x10¹³ m² for the Southern Hemisphere.

This figure (27K) shows the time series of winter values of the NAO simulated for years 100 to 199 in the present-day simulation and corresponding changes in surface temperature (°C) corresponding to unit deviations of the NAO index.

Visiting student Caspar Ammann from the University of Massachusetts, Amherst, in collaboration with Jeffrey T. Kiehl (Climate Modeling Section, CMS) and Charles Zender (ACD, now University of California at Irvine), added stratospheric aerosol into the radiation code of CRM/CCM/CSM. Single eruption simulations of Pinatubo aerosol are currently underway. The experiments include sensitivity analysis of the effects of aerosol size, spatial and temporal resolution of the forcing, and vertical resolution of the stratosphere. The following three projects are being performed with the volcanic aerosol code:

Ammann, Kiehl, and Zender are involved in a model intercomparison of the Pinatubo eruption among several GCMs, coordinated by G. Stenchikov of the Rutgers University.

A volcanic forcing run will be performed at T42 for the 20th century Climate Change working group of CSM to complement the set of other natural external forcing experiments. The forcing dataset is based on Sato et al. (1993) and Stothers (1996).

Otto-Bliesner, Brady, Shields, and Ammann are performing a set of forcing experiments for the pre-industrial time period with the T31 paleo model (see above). By applying different time series of external forcing, like solar variability and volcanic eruptions, their relative contributions to interannual to decadal and centennial climate
variability in CSM is investigated. Comparisons with observations, mainly based on proxies and recent multi-century global reconstructions, help determine the time evolution and spatial representation of variability in CSM compared with the real world. These experiments are an important test of the model's ability to simulate future climate scenarios.

Ocean Spinup Experiments for Paleoclimate Studies

Brady and Otto-Bliesner continued investigating the effect initial conditions have on ocean model spinup. For past climate periods, temperatures and salinities of deep and surface oceans are not known in sufficient detail to specify initial conditions for the ocean spinup phase of the coupled integration. Ocean simulations were done for the warm climate period of 80 million years ago when the continental configuration and ocean bathymetry were significantly different than present, atmospheric CO₂ may have been as high as 6 times pre-industrial levels, and proxy evidence suggests deep ocean temperatures 8-12°C warmer than present. Four experiments were done to examine the effect of different temperature profiles and atmospheric forcing on the spinup of a paleo-ocean. Simulations with high CO₂ (1680 ppm) and low CO₂ (1120 ppm) atmospheric forcing combined with warm (12°C) and cold (10°C) initial deep water temperatures were compared. The ocean component was integrated forward 100 surface years and 5000 deep years. In terms of volume-integrated ocean temperature, the simulations with initially warm deep ocean temperatures equilibrated after approximately 1000 deep years. A choice of cold initial deep ocean temperatures require significantly longer spinup times. Atmospheric forcing also affected the solution with higher CO₂ levels resulting in higher salinity levels in the South Atlantic and a more intense Southern Hemisphere meridional overturning cell.

This figure (12K) shows the time series of the globally and vertically integrated temperature over the 100 surface years of the simulations.

This figure (37K) shows the meridional overturning stream function in Sverdrups (10⁶ m³s^-1) averaged over the last 10 years of the spinups.

More results are described on the Paleoclimate Ocean Spinup Experiments web page.

Progress on the Parallel Climate Model (PCM)

Ocean Component Development

Albert Semtner (Affiliate Scientist, Naval Postgraduate School, NPS), Anthony Craig (CCR), Warren Washington (CCR), John Weatherly (visitor, U.S. Army Cold Regions Research Engineering Laboratory, CRREL), Gerald Meehl (CCR), and Frank Bryan, (Oceanography Section, OS) continued to evaluate, improve, and optimize a modified version of the Parallel Ocean Program (POP). The Parallel Climate Model (PCM) uses a 2/3 degree (on average) displaced pole, global configuration of  POP, while the Climate System Model (CSM) is developing a 2 degree version with a different numerical pole point. During the past year there was an effort to have closer collaboration between the CSM and PCM efforts. An attempt is being made to introduce the Gent-McWilliams diffusion scheme and K Profile Parameterization (KPP) vertical upper ocean scheme. This latter activity is being done in collaboration with scientists at Los Alamos National Laboratory. The PCM ocean model includes increased latitudinal resolution near the equator to resolve the strong tropical current systems. This version of POP was integrated for over a century with acceleration at the bottom of the ocean of a factor of 10 to bring the ocean simulation closer to equilibrium. The ocean model component of PCM continues to be in production mode on the 128 Silicon Graphics Inc. (SGI) Origin 2000 at  NCAR’s Climate Simulation Laboratory (CSL) and on the SGI machines at Los Alamos National Laboratory. Recently, the PCM components have started production on the NCAR IBM SP2. The ocean simulations revealed remarkable eddy structure that compared quite favorably with higher resolution versions and limited observations. Also, the narrow flows along the coastal regions were in good agreement with observations. The meridional overturning in the North Atlantic and worldwide conveyor belt circulation were well represented in the simulations. This version of POP is an integral component of the PCM, a coupled model developed as part of a multi-institutional, distributed research effort for use on current and future generations of Massively Parallel Processors (MPP). A discussion of the ocean model simulation in the coupled model can be found in two papers by Semtner (2000) and Washington et al.(2000) to be published in  Climate Dynamics.

This figure (93K) shows a simulation of the Arctic and the North Atlantic upper ocean flow. In both ocean basins there are realistic eddie structures. The Arctic simulation shows the strong flow across the pole showing no evidence of numerical difficulties near the North Pole.

This figure (106K) of Gulf Stream simulation shows a relatively warm and narrow flow that cannot be captured by coarser resolution models. We do note that in the simulations of the 1/10th degree version of POP that the eddies are even more realistic and the 2/3 degree resolution is able to capture some of the eddy details.

The PCM version of POP was optimized by Vincent Wayland (CCR) for the National Energy Research Scientific Computing Center (NERSC) Cray T3E and demonstrated excellent scalability using 8, 16, 32, 64, and 256 processing elements (PEs). Semtner and colleagues provided ferret analysis tools for POP that allows for diagnosis of the simulations in comparison with observations. This series of plots has allowed for a better analysis of the ocean simulation characteristics.

Examples of the ocean simulation of the model can be found on the Parallel Climate Model web page.

Coupled Model Simulation

The PCM makes use of the NCAR Community Climate Model (CCM3) and Land Surface Model (LSM) for the atmospheric and land surface components, respectively, the Los Alamos National Laboratory POP, and the NPS sea-ice model. The PCM executes on several distributed and shared memory computer systems. The coupling method is similar to that used in the CSM in that a flux coupler ties the components together, with interpolations between the different grids of the component models. Flux adjustments or corrections are not used in the PCM. The ocean component has 2/3° average horizontal grid spacing with 32 vertical levels and a free surface that allows calculation of sea level changes. Near the equator, the grid spacing is approximately 1/2° in latitude to better capture the ocean equatorial dynamics. The North Pole is rotated over northern North America, thus, producing resolution smaller than 2/3 degree in the North Atlantic where the sinking part of the world conveyor circulation largely takes place. Because this ocean model component does not have a computational point at the North Pole, the Arctic Ocean circulation systems are more realistic and similar to the observed. The elastic viscous plastic sea-ice model has a grid spacing of 27 km to represent small-scale features, such as ice transport through the Canadian Archipelago and the East Greenland current region.

Results from a 300-year present-day coupled climate control simulation by Washington, Weatherly, Meehl, Semtner, Thomas Bettge (CCR), Craig, Gary Strand (CCR), Julie Arblaster (CCR), Wayland, Rodney James (Scientific Computing Division,  SCD) and Yuxia Zhang (NPS) showedthat the PCM gave a very stable simulation with approximately the observed interannual and decadal variability.  Five transient 1% per year CO₂ increase experiments that showed a global warming of about 1.26°C for a 10 year average at the doubling point of CO₂ have been carried out. One of the experiments was allowed to go to the quadrupling point and the global average warming was 2.89°C. There was a gradual warming beyond the doubling and quadrupling points. A 0.5% per year CO₂ increase experiment was also performed showing a global warming of 1.49°C and a similar geographic warming pattern to the 1% per year doubling experiment. Globally averaged sea level rise at the time of CO₂ doubling was approximately 17cm and at the time of quadrupling it was 15 cm. The regional sea level changes were much larger and reflect the adjustments in the temperature, salinity, internal ocean dynamics, surface heat flux, and wind stress on the ocean. El Niño and La Niña events in the tropical Pacific Ocean show approximately the observed frequency distribution and amplitude, which leads to realistic variability on interannual timescales of tropical and extratropical planetary wave patterns.

Presently, historical future climate model ensemble simulations are being conducted.  The simulations make use of the same forcing as the CSM. They simulate from 1870 to year 2100.  Additional simulations with the added effect of solar forcing on the climate system are also being conducted.

This figure (93K) shows the near surface Arctic Ocean pattern for one instance in time.  Ocean eddy structures are shown along with a transpolar flow.

This figure (111K) shows three transient climate change simulations using 0.5% CO₂ per year and 1% CO₂ per year increases.  December, January, and February (DJF) and June, July, and August (JJA) are surface temperature 10-year difference means. The top figure shows a warming pattern (°C) for the 0.5% per year simulation at the doubling point (140 years), the middle figure shows the warming pattern for a 1% CO₂  per year simulation at 70 years, and the bottom is the warming at 1% CO₂  per year at the quadrupling point (140 years).  As expected the warming is largest in the winter hemisphere season and half the rate of warming, 0.5% per year, over 140 years gives a pattern very similar to that of 1% CO₂  per year over 70 years.  This warming is proportionally amplified at the quadrupling point.

The following are scientists and programming staff involved in the PCM effort or its components in alphabetical order: J. Arblaster (NCAR), T. Bettge (NCAR), A. Craig (NCAR), J. Dennis (NCAR), J. Dukowicz (LANL), J. Hack (NCAR), S. Hammond (NCAR), E. Hunke (LANL), R. James (NCAR), P. Jones (LANL), R. Loft (NCAR), R. Malone (LANL), M. Maltrud (LANL), W. Maslowski (NPS), G. Meehl (NCAR), A. Middleton (NCAR) A. Semtner (NPS), R. Smith (LANL), G. Strand (NCAR), W. Washington (NCAR), J. Weatherly (CRREL), V. Wayland (NCAR), D. Williamson (NCAR), and Y. Zhang (NPS).

Climate System Ice Model (CSIM)

Weatherly analyzed the sea ice and polar climate in the control and transient CO₂ PCM simulations. The doubled CO₂ run showed significant thinning of Arctic sea ice and decreased summer ice concentrations, similar to the recent observational evidence from submarine- and satellite-based data. The ice and snow albedos used in PCM and CSM were compared with observations from the year-long Surface Heat Budget of the Arctic (SHEBA) field program by Weatherly and Donald Perovich (CRREL).

Craig and Elizabeth Hunke (LANL) modified the LANL sea-ice model
(CICE) to include parallel message-passing (MPI and Open MP) and have it
running on the new parallel IBM. Cecilia Bitz (University of  Washington) has implemented her improved ice thermodynamics and multiple ice-thickness category model into the CICE code for use in both PCM and CSM. Craig, Bitz, and Weatherly are working to couple the new CICE/Bitz code into the developing PCM-2.

Coherent Decadal Climate Variability

Arblaster and Meehl analyzed observations and results from the DOE PCM. They showed that coherent decadal climate variability extended over the entire Pacific basin, and affected  El Niño teleconnections between equatorial Pacific SSTs and Australian rainfall. Such time-varying links between decadally-varying tropical Pacific SSTs and El Niño teleconnections with Australian rainfall noted in observations were shown to arise in the model due to a combination of decadal modulation of  El Niño amplitude and longitudinal shifts of the Walker Circulation.

This figure (32K) shows the relationship between the interdecadal oscillation index and low pass filtered Australian precipitation in the 300-year control run of the PCM.  The time series are highly anti-correlated at -0.5.  A Lanczos filter with a 13-year cutoff is used on the precipitation.  The interdecadal oscillation index is the principal component time series of the first empirical orthogonal functions (EOFs) of low pass filtered near-global SST.

Meehl, Arblaster, and Strand analyzed versions of an earlier global coupled model and showed that a change in a sea-ice parameter resulted in a reduction in amplitude (of about 60%) and a shortening of the predominant period of decadal low-frequency variability (from about 33 years to roughly 14 years) in the time series of globally-averaged surface air temperature. These changes were also reflected in time series of area-averaged SSTs in the equatorial eastern Pacific Ocean, the principal components of the first EOFs of global surface air temperature and sea level pressure, and other quantities. The alterations in these time series were highly correlated with each other; yet the only change made between the two model integrations was in the sea ice formulation. Thus, coupled ocean-atmosphere-sea ice processes acting on a global scale were altered to produce the change of climate sensitivity and low-frequency decadal timescale oscillations in the model. Global climate sensitivity was reduced when ice albedo feedback was weakened due to the change in sea ice that makes it more difficult to melt. The changes in the amplitude and timescale of the low-frequency variability in the model were traced to changes in the base state of the climate simulations as affected by changes in sea ice. The altered base state then affected the global coupled decadal mechanism and the consequent manifestation as low-frequency decadal timescale variability of the global climate system.

Climate Change Projections

Meehl, Washington, Arblaster, Bettge, and Strand analyzed simulations of
20th century climate and projections of climate into the 21st century
from the earlier version of the PCM. Two climate change experiments were performed, both starting in the year 1900. The first experiment used greenhouse gas radiative forcing (represented by equivalent CO₂) observed during the 20th century and extended greenhouse gas forcing to the year 2035 by increasing CO₂ 1% per year compounded after 1990 (CO₂-only experiment). The second experiment included the same greenhouse gas (equivalent CO₂) forcing as the first but added the effects of time-varying geographic distributions of monthly sulfate aerosol radiative forcing (CO₂+sulfates experiment). These were compared to a 135-year control experiment with no change in external forcing. Climate system responses in the CO₂-only and CO₂+sulfates experiments were marked not only by greater warming at high latitudes in the winter hemisphere but also by a global "El Niño-like" pattern in surface temperature, precipitation, and sea level pressure. A methodology was formulated to evaluate the possible changes in decadal timescale (10 to 20 year period) surface temperature variability associated with the low-frequency fluctuations of anthropogenic forcing and associated changes in climate base state due to that forcing.  An EOF analysis of surface temperature is consistent with previous results in that low-frequency variability with the El Niño-like pattern was introduced into the model coupled climate system with the same timescale as the forcing.

This figure(47K) shows a) EOF1 for annual mean surface temperature from
the coupled climate model CO₂+sulfates experiment, 1900-94, explained variance in parentheses; b) normalized principle component (PC) time series for EOF1 in (a), with anthropogenic forcing superimposed as dashed line (scale for PC time series at left, forcing scale at right); and c) EOF1 for annual mean surface temperature from CO₂+sulfates experiment, 1900-2035, explained variance in parentheses, and b) normalized PC time series of EOF1 in (c), with anthropogenic forcing superimposed as dashed line. Note how PC time series of the response patterns represented by EOF1 for both time periods follow the anthropogenic forcing.

To examine the possible effects of this introduced low-frequency variability and associated changes in base state on decadal timescale variability (10 to 20 year periods), the surface temperature time series were filtered to retain only variability on that timescale. The El Niño-like pattern of decadal variability seen in the observations was present in each of the model experiments (control, CO₂-only, and CO₂+sulfates),
but the magnitude decreases significantly in the CO₂-only experiment. This was associated with changes in the base state climate that include a reduction in the magnitude (roughly 5% to 20% or more) of wind stress and ocean currents in the upper 100 m in most ocean basins and a weakening of meridional overturning
(about 50%) in the Atlantic.  These weakened circulation features contributed to decreasing the amplitude of global decadal surface temperature variability
as seen in a previous sea ice sensitivity study with this model. Thus, the superposition of low-frequency variability patterns in the forcing has less effect on the amplitude of future decadal variability in these experiments than the consequential changes of the
base state climate due to increases of the radiative forcing.

Top of Page | CGD ASR99 Home | ASR99 Home | CGD Home