|
|
This page contains highlights from our working group meeting in Boulder, March 29-30, 2005. Complete reports from all of our meetings can be found on the Polar Climate Working Group home page. If you are interested in working with PCWG members on these projects or any others, please contact the working group co-chairs.
The Dansgaard-Oeschger (D-O) events that punctuated the last glacial period are abrupt warming episodes recorded in Greenland ice cores. The leading
class of theories invokes a shift in the mode of ocean thermohaline
circulation; however, most climate models subjected to such THC changes
are unable to recreate the striking 7-10oC increase in surface temperature
observed at the Greenland Summit. We performed a sensitivity study using
CCM3 to investigate the idea that sea ice, with its sensitivity to forcing
from the atmosphere and ocean and its threshhold behaviour, plays a key
role. Results from our simulations show that removal of sea ice (about 4
million square km, less than the Northern Hemisphere seasonal cycle in the
modern climate) in the North Atlantic can explain the D-O warming signal.
Furthermore, the response is consistent with the accumulation and oxygen
isotope records from Greenland ice cores and is in agreement with other
observation-based evidence that D-O events are primarily a wintertime
phenomena. The two figures which follow compare the LGM control experiment with the reduced sea ice scenario.
Figure 1: Annual mean of sea surface temperature boundary conditions
(degrees Celsius) for the LGM scenario (left) and reduced sea
ice scenario (middle). Maximum (February) and minimum (August)
sea ice extents are indicated with the solid and dotted lines,
respectively. Scenario I has a maximum sea ice extent equivalent
to the LGM perennial ice cover, and a minimum sea ice extent
equivalent to the modern day minimum ice cover. The ice thickness
is 2 m, which is a typical value for the Arctic today.
Figure 2: The difference in surface air temperature between the two experiments (degrees Celsius).
The ice cover of the Arctic Ocean has declined in recent decades and is predicted to continue to decrease throughout the coming century. The potential for ice reduction to impact the climate system is large, and sensitivity studies conducted with CSM1 have indicated that reducing Arctic ice cover could shift winter storm tracks over western North America and drive a drying trend in the western United States with wetter conditions in British Columbia and Alaska. New work with CCSM3 duplicates this precipitation response to a specified change in Arctic ice cover. Analysis of the fully coupled system further supports the likelihood of this precipitation response; the reduction in Arctic ice that corresponds to a quadrupling of CO2 drives a significant drying (up to 30 cm less precipitation annually) of the western United States and significant wetting (up to 30 cm increase in annual precipitation) over Alaska and British Columbia. These three results suggest very strongly that a decline in Arctic ice cover will have a large impact on water resources in the American West.
We investigate causes for a strong high latitude imposed ice (land or sea) influence on the marine Intertropical Convergence Zone (ITCZ) in the Community Climate Model version 3 coupled to a 50-m slab ocean. The marine ITCZ in all ocean basins shift meridionally away from the hemisphere with fixed imposed additional ice cover, altering the global Hadley circulation with increased subsidence in the tropical hemisphere with imposed ice, and uplift in the other. The effect appears independent of the longitudinal position of the imposed ice. The anomalous ice induces a rapid cooling and drying of the air and surface over the entire high and mid latitudes; subsequent progression of cold anomalies occur in the Pacific and Atlantic northeasterly trade regions where a wind-evaporation-SST feedback initiates progression of a cold SST "front" towards the ITCZ latitudes. Once the cooler SST reaches the ITCZ latitude, the ITCZ shifts southwards, aided by positive feedbacks associated with the displacement. The ITCZ displacement transports moisture away from the colder and drier hemisphere and into the other hemisphere, resulting in a pronounced hemispheric asymmetric response in anomalous air humidity. We speculate that this tropical asymmetric response largely prevents cold and dry air anomaly from progressing into the southern hemisphere. From an energy balance viewpoint, the increased outgoing radiative flux at the latitudes of the imposed ice is compensated for by increased radiative energy flux at the tropical latitudes occupied by the displaced ITCZ, and subsequently transported by the altered Hadley and eddy circulations to the imposed ice latitudes. Our mechanism may be applicable to past climates like the Last Glacial Maximum where hemispheric asymmetric changes to ice cover occurred. Major caveats include omission of interactive sea ice physics and ocean dynamical feedback, and sensitivity to atmospheric physics parameterizations across different models.
I have participated in collecting large turbulent-flux data sets over both Arctic and Antarctic sea ice. The Arctic experiment was SHEBA, the experiment to study the Surface Heat Budget of the Arctic Ocean, which ran from October 1997 to early October 1998. The Antarctic measurements were on Ice Station Weddell, which we manned from February until June 1992. The objectives of both of these experiments involved measuring the turbulent surface fluxes of momentum and sensible and latent heat and devising bulk flux parameterizations from these data for use in modeling air-surface coupling in a sea-ice environment.
In my talk, I reviewed our data and the resulting parameterizations for the drag coefficient and the roughness lengths for wind speed, temperature, and humidity. Parameterizations for these roughness lengths allow estimates of the turbulent surface fluxes through Monin-Obukhov similarity theory. Our measurements of the drag coefficient and the roughness lengths generally differ significantly from parameterizations that CCSM currently uses for a sea-ice environment.
In particular, the measured SHEBA drag coefficient has a seasonal
cycle, with values lower than the current CCSM parameterization in the winter but higher in the summer. Furthermore, the measured roughness length for wind speed is a function of wind speed--increasing with wind speed as a consequence of blowing snow--while the CCSM value is constant year-round. In the summer, I base the parameterization for the drag coefficient on the concentration of open water since the edges created by melt ponds and leads foster form drag and, therefore, enhance summer air-ice momentum exchange. Both the Arctic and Antarctic data sets show that the roughness lengths for temperature and humidity are not generally equal to the roughness length for wind speed and, consequently, contradict CCSM's current assumption. Both of these scalar roughness lengths tend to follow Andreas's (1987) model.
Figure: Ten-day averages of the neutral-stability, 10-meter drag
coefficient CDN10 measured at the Atmospheric Surface Flux Group's main SHEBA tower and at five remote SHEBA sites (i.e., named Atlanta, Baltimore,
Florida, Cleveland, and Maui). The error bars are +/- 2 standard deviations in the means. This figure summarizes well over ten thousand hours of data. The dotted horizontal line is the current parameterization for CDN10 in CCSM.
The GFDL atmospheric single-column model (SCM) was integrated through the 11-month annual cycle at the Surface Heat Budget of the Arctic Experiment (SHEBA). The model includes prognostic variables for cloud fraction, cloud liquid content, and cloud ice content. The GFDL stratiform cloud parameterization -- based on the work of Tiedtke (1993) and Rotstayn (1997, 2000) -- is quite different from the CCSM. The SCM was forced with incoming solar radiation at the top of the atmosphere, and albedo and skin temperature at the surface. Lateral forcing consisted of time-dependent vertical profiles of temperature advection, humidity advection and winds obtained from the ECMWF forecast model at the gridpoint closest to the SHEBA site. To control model drift the SCM is run in "forecast" mode, in which the model was integrated for a 36-hour period every 24 hours. Model output from hours 25-36 was compared with observations. The ECMWF forecast analysis assimilated atmospheric soundings from SHEBA, but the next closest sounding was 500-800 km from the site. This will reduce the skill of the ECMWF forecast at SHEBA, but the synoptic (co)variability of the forcing profiles should still be realistic qualitatively. With this forcing, even a perfect SCM would be expected to disagree with observations on timescales less than a day. Thus, the evaluation of the GFDL SCM focuses mainly on statistics on monthly timescales. Hourly time series were also examined for additional insight into these results.
The main results were as follows:
As part of the Arctic Ocean Model Intercomparison Project (AOMIP), the LANL ice-ocean modeling team completed two 55-year, global, ice-ocean simulations forced with reanalysis atmospheric data for 1948-2002. These two simulations differ only in the parameterization used for lateral mixing of tracers (potential temperature and salinity) in the ocean, but the resulting circulation and kinetic energy of the simulated oceans are very different, mainly at high latitudes. The differences can be traced to two effects, (1) scale selectivity, in which the Laplacian form of the Gent and McWilliams (GM) parameterization damps wave energy more quickly than the biharmonic mixing formulation, and (2) grid dependence of the diffusion coefficient, which appears in the biharmonic formulation but not in GM and is particularly important at high latitudes, where the grid scale decreases dramatically on the sphere. Future global simulations using the GM parameterization should include a diffusivity scaling factor given by the square root of the grid cell area, to prevent diffusion from dominating advection in the evolution of high latitude tracers and circulation.
The Los Alamos sea ice modeling team
A new version of the Los Alamos sea ice model, CICE 4.0, is under development. Model development has focused on both improved software and more realistic physics. The major software change is the addition of a block index to horizontal arrays, which allows for cache-friendly grid decomposition, efficient load balancing, and code portability to tripole and geodesic grids. More realistic physics is being incorporated in the treatment of ridging and snow. The code has been rewritten with multiple snow layers to improve the vertical resolution of temperature and optical properties. Later code versions may include snow metamorphosis and liquid water content. The improved ridging scheme should eliminate an unstable dynamic feedback. This feedback can be traced to the ice strength parameterization, which has a sharp cutoff in the participation of open water in ridging. The strength can change dramatically during one time step, causing large stress gradients that lead to excessive strain rates and thicknesses. Modeling the ice strength as an exponential function of the open water fraction appears to remove the instability. We are also testing a new ridging redistribution function to improve the thickness distribution of ridged ice, and are adding a biharmonic viscosity term to reduce small-scale noise in the velocity field. Code release is scheduled for later this year.
Figure: Fractional ice area near the Labrador coast on May 3 during spinup of a standalone sea ice model (CICE) on a 9-km Arctic grid. (a) Control simulation using the ice strength formulation from Thorndike et al. (1975), with a sharp cutoff G* = 0.15 for participation in sea ice ridging. (b) Modified simulation with an exponential dependence of ice strength on open water fraction.
Elizabeth Hunke
Los Alamos National Laboratory
Theoretical Division
MS B216
Los Alamos, NM 87545
Phone: 505-665-9852
Fax: 505-665-5926
e-mail: eclare@lanl.gov
Marika Holland
National Center for Atmospheric Research
P.O. Box 3000
Boulder, CO 80307-3000
Phone: 303-497-1734
Fax: 303-497-1700
e-mail: mholland@ucar.edu