Scenarios of the Arctic Climate in the 21si Century 1121 The GCM Method

Two types of GCMs can be distinguished: equilibrium and transient models. The first group simulate changes of climate for C02 doubling occurring rapidly, while the second group compute the same for a gradually in creasing CO, (most often of a !% compounded increase per year). The second approach is more realistic and resembles the contemporary changes of CO, concentrations.

The first equilibrium experiments conducted using atmospheric general circulation models simulated a very large increase of temperature in the Arctic (up to 10-16°C in winter), especially in the northern parts of the Atlantic and Pacific Occans (see Washington and Meehl 1984, or Mechl and Washington 1990)- In the rest of the Arctic the projected rise of temperature varies between 4"C and 6°C in winter and about 2nC in summer. This geographical pattern of temperature changes is in disagreement with recent observational changes of temperature in the Arctic (see e.g. Chapman and Walsh 1993, their Figure 1, or Przybylak 1996a).

More sophisticated high-resolution equilibrium simulations of the 2 x CO, climate, which use GCMs couplcd with mixed layer occans, give significantly more reliable results for some parts of the world, but probably not for the Arctic (see Figure 5.4 in Houghton et al. 1990). Results from the three models presented in this Figure (CCC, GHHI, and UK-HI) show that the warming in the Arctic is highest in late autumn and winter. The projected warming in winter in the case of the first two models exceeds more than 10"C. I lowever, the regional and local differences between winter surface air temperature responses simulated for the doublcd-CO, climate by these models show a 10"C range. A review of current model results (see sub-chapter 11.1) shows that the largest disagreement between couplcd climatc model simulations of present-day climate is still in the Polar regions.

Limited space prevents me from providing a more detailed presentation of the changes of other meteorological elements simulated by the equilibrium GCMs for the 2 x CO, case. Generally, an increase in the cloudiness and precipitation and a decrease in the sea ice extent and its thinning in the Arctic is predicted. For more information see, for example, Washington and Meehl (1984); Schlesinger and Mitchell (1987); Houghton et al. (1990); Mcchl and Washington (1990); Przybylak (1993) or Kozuchowski and Przybylak (1995).

The main weakness of these kinds of models (except for the leap rise in CO, concentration) is the fact that in simulations they neglect the thermal inertia of the deep ocean, and therefore give an exaggerated response, particularly in regions of deep mixing. A comparison of transient and equilibrium responses of surface air temperature with the doubling of CO, (Figure 11.5) using the same model (GFDL) entirely confirms the above conclusion. From this figure, it can be seen that the transient response of the surface air temperature is particularly low over the northern North Atlantic and over the circumpolar ocean of the Southern Hemisphere, where the deep vertical mixing of water is the greatest and. as a consequence, the effective oceanic thermal inertia is very large. According to Manabe et al. (1991), a relatively small surface wanning in the northern North Atlantic Ocean is caused also by the reduction in the near-surface advection of warm water from South (the excess of precipitation over evaporation in high latitudes leads to a weakening of the thermohaline circulation). In the Arctic, the transient model experiment shows about 1.4-2.0 times lower wanning than in the case of the equilibrium response (Figure 11.5c). According to the GFDL model results (Figure 11.5a), the annual air temperature in the Arctic should increase by 4-5°C for the period of CO, doubling. In winter and summer (not shown) the warming should be equal to 5-8°C and 0 2°C, respectively.

Figure 11.5. (a) The transient response of the surface air temperature of the coupled ocean-atmosphere model to the 1%/year increase of atmospheric carbon dioxide. The response (°C) is the difference between the 20-year (60lh to 80,h year) mean surface air temperature (1%/year increase of CO,) and 100-year mean temperature (CO, constant), (b) The equilibrium response of surface air temperature to the doubling of atmospheric carbon dioxide, (c) The ratio of the transient to equilibrium responses (after Manahe et al. 1991).

Figure 11.5. (a) The transient response of the surface air temperature of the coupled ocean-atmosphere model to the 1%/year increase of atmospheric carbon dioxide. The response (°C) is the difference between the 20-year (60lh to 80,h year) mean surface air temperature (1%/year increase of CO,) and 100-year mean temperature (CO, constant), (b) The equilibrium response of surface air temperature to the doubling of atmospheric carbon dioxide, (c) The ratio of the transient to equilibrium responses (after Manahe et al. 1991).

More recently, Cattle and Crossley (1995) published results of the simulation of the Arctic climate change for the doubling of CO, using the UK_MO model. They found that maximum changes of air temperature in winter (more than 10"C) are associated with the marginal ice zone in the Atlantic sector and the regions of the shelf seas (Figure 11.6). The wanning of Greenland lies between 2-5°C. In summer, the change of air temperature over the Arctic is small and mainly oscillates between 0-2°C (see Figure 11.6). The introduction of a simple parametcrisation of the effects of sulphate aerosols significantly reduces the magnitude of the wanning, but changes the overall pattern very little (Figure 11.7). In winter, the predicted wanning varies between 2-5"C. The lowest increase of air temperature should occur in the Atlantic sector of the Arctic. In summer, most Arctic areas show very small warming that oscillates between 0-1 "C. The greatest warming should occur in Greenland and Alaska (l-4°C). Regionally, however, even the cooling may occur (sec Figure 11.7).

Mean annual air temperature changes in the Arctic, according to the model developed by Mitchell et al. (1995), vary from 4nC to 6"C. For temperature simulation, the concentration of aerosols from 1795 to 2030-2050, based on the most probable scenario IS92a proposed by Houghlon et al. (1992), was used. A recent IPCC report (Houghton et al. 2001) also presents annual patterns of air temperature change, but for the period 2071-2100 relative to the period 1961 1990 using a newly introduced set of scenarios (SRES scenarios, for details see Houghton et al. 2001). For two scenarios (SRES A2 and B2) the multi-model ensemble projects the warming ranges from 6nC to HFC and from 5"C to 8"C, respectively for the greater part of the Arctic (see Figure 9.10d and e in Houghton et al. 2001). The greater warming, in comparison to the model of Mitchell et al. (1995), is caused by the fact that the future sulphur dioxide emissions for the six SRES scenarios are much lower, compared to the IS92 scenarios (see Figure 17 in Houghton et al. 2001).

Doubled-CO, climate simulation of the troposphere in the Arctic shows a warming between 1-L5°C. On the other hand, the cooling should occur throughout almost the whole stratosphere. In the lower stratosphere the temperature will be nearly the same as today. In the middle stratosphere the predicted decrease of temperature oscillates between 2 3°C.

Most transient GCMs simulate the increase of precipitation with the doubling of CO, (sec e.g. Manabe et al. 1992; Cattle and Crossley 1995; or Houghton et al. 1996, 2001). However, a careful examination of geographical patterns of precipitation changes shows significant differences between model predictions. Here, I present the results published by Cattle and Crossley (1995). According to their model, winter precipitation should increase slightly over the central Arctic basin with higher local increases over the sun ounding landmasses (see Figure 11.8), A general decrease of precipitation is shown over

Figure ¡1.6. Air temperature change (in °C) over the Arctic for the decade of doubling of carbon dioxide from a run of the Hadley Centre model with transiently increasing greenhouse gases: (a) winter (DJF) and (b) summer (JJA) (after Cattle and Crossley 1995).

the region of the Greenland-Iceland-Norwegian Sea, south-eastern Greenland, Iceland, Spitsbergen and the northern parts of European and the Russian Arctic west of 50°E. In summer (Figure 11.8), a tendency towards reduced precipitation should occur in most parts of the Arctic Ocean, over some fragments of the Greenland and Norwegian seas, over almost whole Barents Sea including its surrounding islands (Novaya Zenilya, Zemlya Frantsa-Josifa and Spitsbergen), as well as over the central part of the Russian Arctic, With the inclusion of aerosol forcing (here is seen an increase in precipitation in the areas of reduced precipitation in winter and more generally in summer (Cattle and Crossley 1995).

30'N

30'N

30°S

J JA

6TS-

J JA

6TS-

Figttre 11.7. Seasonal change in surface air temperature from 1880-188Q to 2040-2049 in simulations with aerosol effects included, (a) winter (DJF) and (b) summer (J.1A) (after Kattenberg et al. 1996).

Reviewing the literature presenting the transient experiment results using GCMs, I did not find any information about cloudiness changes in the Arctic with the doubling of CO,. What is the current state of knowledge about other meteorological elements and components of Arctic climate system? Wild et al. (1997) computed changes in the zonal mean 10-m wind speed for the whole globe with the doubling of COr In the Arctic, the average wind speed in summer should be higher by about 0.2-0.5 m/s. On the other hand, in winter a reduction of 10-m wind speed up to 0.5 m/s around the Pole and up to 0.4 m/s in the latitude belt 60-75°N should be observable. Winter cyclone frequencies simulated for the Arctic using the CCC GCM (equilibrium model) correspond very well with these results (Lambert 1995). Comparison of Figures 2 and 3 presented in his paper shows that the total number of cyclones decreases in the Arctic in a wanner world. However, one should add that this tendency is limited mainly to weak lows, because intense cyclones show an increased frequency.

Manabe et al. (1991) present surface heat balance components averaged for the "doubled" C02 case simulated by the transient GFDL GCM (Figure 11.9). It can be seen that the net radiation is positive in the whole Arctic (ocean and continental parts). The sensible heat is negative everywhere in the Arctic, except the southern oceanic parts, while the latent heat is negative mainly in the latitude band 60 80"N. The decreases of latent heat fluxes are significantly higher over continental parts of the Arctic (up to 2-3 W/m3) (Figure 11.9b).

A very important component of the Arctic climate system is sea icc. The UKMO GCM predicts that in a warmer world (a "doubled" C03 case) the sea ice thickness should be reduced by over lm in both summer and winter, with maximum changes occurring in area covered by the thickest ice in the control run (Figure 11.10). Similar results have also been found by Manabe et al. (1992) and Ramsdcn and Fleming (1995) using the GFDL GCM and coupled ice-ocean Arctic ocean model forced using an output from the CCC model of the atmosphere, respectively. Ramsdcn and Fleming (1995) concluded that the Arctic ice field appears to act as a regulator of climate change, rather than as an accelerator.

Ramsdcn and Fleming's (1995) model predicts an increase in the ocean surface temperature with the doubling of CO, by about a degree, with the largest increase expected in summer, reflecting the amount of open water. Their model also foresees a slight decrease in surface salinity in the Arctic Ocean.

Figure 11.8. Precipitation change (mm/day) over the Arctic for the decade of doubling of carbon dioxide from a run of the Hadley Centre model with transiently increasing greenhouse gases: (a) winter (DJF) and (b) summer (JJA) {after Cattle and Crossley 1995).

CONTINENTS

A NET RADIATION

CONTINENTS

A NET RADIATION

A LATENT

Figure 11.9. The latitudinal profile of zonal-mean surface heat balance components over oceans (a) and continents (b) between the G (1%/year increase of CO, averaged over the 60,h to 80a year period) and S (CO, constant) integrations (after Manabe et al. 1991).

A LATENT

Figure 11.9. The latitudinal profile of zonal-mean surface heat balance components over oceans (a) and continents (b) between the G (1%/year increase of CO, averaged over the 60,h to 80a year period) and S (CO, constant) integrations (after Manabe et al. 1991).

Figure 11.10. Sea ice thickness change (in m) over the Arctic for the decade of doubling of carbon dioxide from a run of the Hadley Centre model with transiently increasing greenhouse gases: (a) winter (DJF) and (b) summer (JJA) (after Cattle and Crossley 1995).

Was this article helpful?

0 0

Post a comment