Palaeoclimate Modelling with the UM

Much of the palaeoclimate work at Reading is now beginning to move to the use of the UM. The model has a number of advantages over the IFS, not least the existence of a slab ocean and coupled atmosphere-ocean version including a reasonable parameterization of sea-ice. However, at the time of writing, all we have done is to run a couple of years of the present day simulation and so the following is a description of our plans rather than what we have done. The work can be separated into the recent Quaternary (100,000 years ago) and the more distant past.

Quaternary work will be in close collaboration with Chris Hewitt and John Mitchell at the Hadley Centre. Initial work will be focusing on three major areas.

1. Matthew Woodman (see later) is starting to look at the effects of fresh water influxes at the Younger Dryas.

2. Chris Hewitt and myself will be attempting a coupled ocean atmosphere simulation for the last glacial maximum. This has a number of problems to be overcome (like spinning up the deep ocean) but will be the first attempt at a coupled simulation for this period.

3. With Richard Betts at the Hadley Centre, we will be looking at the interaction between vegetation and past climate change. We will be doing this in the context of the global and limited area models.

The pre-Quaternary work will initially focus on two aspects.

1. New simulations for the late Cretaceous and the Miocene. This is related to work we've been doing with some Russian researchers who have compiled a large amount of geological data for these periods.

2. Investigations of vegetation/climate interactions for the Mesozoic. Initial experiments with the UGCM have shown that including vegetation may go along way towards reconciling the model predictions with the geological indicators of past climate.

As ever with the palaeoclimate group, these plans are not so much limited by people time as by computer time. One of the additional advantages of the UM is that it can run on work stations and we hope to start making heavy use of our local computing facilities, rather than using all of the computer time at RAL!

Paul Valdes
(Reading)

Palaeoclimate modelling using the Coupled Hadley Centre Model

I have just begun a PhD supervised by Dr. P.J. Valdes. I shall be running the UK Meteorological Offices' Unified Model GCM in its fully coupled Ocean-Atmosphere state (HADCM3). As a preliminary experiment this shall be used to model a Younger Dryas (YD) type event. The YD is a cooling event which occurred about 11Kyrs BP and lasted about 1.3Kyrs. The YD is superimposed onto the general warming of climate during the deglaciation of the Last Glacial Maximum (LGM) ice sheets. This cooling was so strong that much of the Northern Hemisphere returned to a glacial climate almost as extreme as that of the LGM (14-22Kyrs BP). Evidence suggests that this is a global event. The experiment will consist of the instantaneous influx of low salinity 'meltwater' into the North Atlantic Ocean. The presence of fresh water should inhibit the production of North Atlantic Deep Water (NADW) and a reduction in the intensity of the conveyor belt circulation. The Climate of the North Atlantic should cool enabling the re-advance of the ice sheets. The experiment will look at the time response of the climate and ocean.

Suggested reading:
Broecker et. al., 1988, Paleocenog. Vol. 3, pp1-19

Ruddiman W.F. and A McIntyre 1981, Paleogeog. Palaeoclimate. Palaeocol. Vol 35, pp145-214.

Mathew Woodman
(Reading)

The Last Glacial Maximum Climate simulations

The climate during the Last Glacial Maximum (LGM) has been simulated using the U.K. Universities Global Atmospheric Modelling Programme (UGAMP) General Circulation Model (GCM) with both prescribed sea surface temperatures (SSTs) based on the CLIMAP data set and computed SSTs with a simple thermodynamic slab ocean. The model includes a simple thermodynamic treatment of sea ice and the ocean heat transports are assumed to be the same at the LGM as they are in the present day. Consistent with the Palaeoclimate Modelling Intercomparison Project (PMIP), the other boundary conditions include the large changes in ice sheet topography and geography, a lower sea level, a lower concentration of CO₂ in the atmosphere, and a slightly different insolation pattern at the top of the atmosphere. The decrease in the globally averaged annual mean surface air temperature is 3.9^oC in the LGM simulations with both prescribed SSTs and computed SSTs. The cooling is larger over land (7.6 and 7.6^oC) than over ocean (1.6 and 1.4^oC). Over land the cooling is larger in JJA (8.6 and 7.8^oC) than in DJF (7.0 and 7.5^oC), which reflects the impact of the large difference in surface albedo during the northern hemisphere summer due to the presence of ice sheets at the LGM. It has been shown that the computed SSTs are generally in agreement with the CLIMAP reconstruction. However, there are significant regional differences (Fig. 30). The sea ice extent around the Antarctic is underestimated although there is the possibility that CLIMAP overestimated the sea ice extent at the LGM. The other major discrepancy occurs in the Pacific warm pool in low latitudes where CLIMAP estimated a warming as opposed to the cooling in the model. It has been suggested that this warmth of LGM subtropical ocean is a dynamical response of the ocean to changes in atmospheric circulation. Such effects are not included in the simulations. The Asian and African summer monsoon circulations decrease due to the decreased land-sea thermal contrast and much reduced summer convective activities over the monsoon regions during the LGM simulations. As a result, precipitation and soil moisture are reduced. These changes in hydrology mimic the geological evidence of drier and colder climates during the LGM. The coarser horizontal resolution model suggests that both the global averaged climate changes and the regional climate changes due to the imposed ice age boundary conditions are sensitive to the horizontal resolution. The decrease in the globally averaged annual mean surface air temperature is 4.4^oC in the simulation at T21 resolution. it is 8.5^oC over land and 1.8^oC over ocean. The regional climate changes simulated in a lower resolution model differ significantly from the those in a higher resolution model due to the poor simulations in planetary waves and storm track activity. The results indicate that the lower resolution model cannot reproduce the present day climate as accurately as the high resolution model. Therefore, the simulated climate change using the lower resolution model is questionable. We concluded that a higher resolution model, at least T42, is needed, for the UGAMP GCM, to simulate climate change. The results shown in this study imply that the model horizontal resolution is an important factor which should be borne in mind in exploring the model-model and model-data differences in the PMIP. Some of these differences may arise from the difference in the horizontal resolution.

References

Dong, B.-W. and Valdes, P. J., 1996: PMIP simulations of the Last Glacial Maximum climates using the UGAMP GCM, Part I: Prescribed versus computed SSTs. Draft available from authors. Dong, B.-W. and Valdes, P. J., 1996: PMIP simulations of the Last Glacial Maximum climates using the UGAMP GCM, Part II: Sensitivity to horizontal resolution. In preparation.

Buwen Dong and Paul J. Valdes
(Reading)

© 1997 Centre for Atmospheric Science/UGAMP. All scientific articles are unpublished. No text or graphics may be copied or used without permisson. Newsletter Editor: Glenn Carver, Cambridge University.