Response of a General Circulation Model of the Atmosphere to Removal of the Arctic Ice-cap

R. L. Newson

Editor’s Note

This paper by R. L. Newson of the UK’s Meteorological Office looks highly prescient today. For one thing, Newson shows how the advent of “very high speed computers” was transforming the ability to make predictions about global weather and longer-term patterns of climate change. To illustrate these possibilities, Newson considers how the climate of the Northern Hemisphere would be altered by complete melting of the Arctic sea-ice cap. This reduces reflection of the sun’s rays, and provides a heat reservoir in polar waters. Significant temperature changes occur at least as far as mid-latitude Europe and North America. Newson presents this merely as a test case, but today the disappearance of the Arctic ice due to global warming is a distinct possibility. 中文

OVER the past few years, numerical models have been developed for investigating the general circulation of the atmosphere and studying its long term behaviour. (See, for example, Smagorinsky et al. ¹ , Kasahara and Washington ² .) These models are firmly based on the equations of fluid motion and thermodynamics and simulate in mathematical terms the chief physical processes which are thought to be of importance in determining large scale atmospheric motions over long periods of time (a month or more). The advent of very high speed computers has made numerical experiments with these models reasonably easy. 中文

One model has been described in detail by Corby et al . ³ . The mode of computation essentially consists of integration forward in time from an initial state for a considerable number of model days using physical parameters appropriate to a given season and obtaining the mean seasonal state by averaging the later computed states; in the experiments described here the model was integrated forward in time for 80 days under winter conditions and the mean seasonal state obtained as the average of the last 40 days. 中文

It is possible to vary the model to determine its reaction to substantial changes in the assumed physical properties. In one of these variants the region of winter arctic ice, as defined by the mean climatological position, was replaced by open ocean maintained at freezing temperature; what had been a surface with a temperature determined by radiational balance became a surface of constant temperature. The results of computations with ice and with open ocean differ quite considerably. The most marked difference is, of course, the great warming of the lower layers of the troposphere over the arctic basin, where, in the ice-free arctic experiment, there was an oceanic reserve of heat to balance radiative losses; this was to be expected. It might also have been expected that in mid-latitudes temperatures in the lower atmosphere would be raised. In fact, there is a lowering of mid-latitude continental temperatures near to the surface but with little change over the sea where the mean sea surface temperature had been left unchanged. This result is illustrated in Fig. 1, which shows the temperature difference between the two computed seasonal means at the level in the model most representative of the surface. The changes are considerable. The mid-latitude cooling may be compared with the temperature anomaly of –5℃ in the very severe January of 1963; a temperature anomaly of –2℃ would indicate a very cold winter month. 中文

中文

Other results from these experiments show a distinct southward displacement and weakening of the prevailing mid-latitude westerlies when there was no polar ice-cap, and this is similar to the result obtained from another ice-free arctic general circulation experiment conducted with the Mintz-Arakawa model ⁴ . What seems to happen is that the general decrease of temperature gradient between equator and pole reduces the strength of westerly flow in mid-latitudes and, there, the circulations become more blocked. Consequently regions which at present benefit climatically from the westerlies might be cooler in an ice-free arctic regime. 中文

I do not claim that these results necessarily indicate what might happen in the real atmosphere, because the model itself has several shortcomings, and one should not dwell on the results in particular areas because they are almost certainly not quantitatively correct. The sign of temperature changes in mid-latitudes is perhaps rather unexpected, and the results illustrate that qualitative arguments may be deceptive and unreliable in considering a physical system as complicated as the atmospheric circulation which includes so many non-linear interactions and feed-back mechanisms. 中文

Considerable refinement and elaboration of numerical models will be required before their answers to questions of the kind raised in this letter can be accepted with any confidence. Nevertheless the potential superiority of the numerical approach for the investigation of problems of climatic modification is beyond doubt. 中文

( 241 , 39-40; 1973)

R. L. Newson

Meteorological Office, Bracknell, Berkshire

Received October 23, 1972.

---

References: 3epCTzeMSmTRsn2wSk064Qw215ZP5SWeNw/236ALsJdwzsMC96WmKKUt22xye9Ak

1. Smagorinsky, J., Manabe, S., and Holloway, J.L., Monthly Weather Review , 93 , 727 (1965).
2. Kasahara, A., and Washington, W. M., Monthly Weather Review , 95 , 389 (1967).
3. Corby, G. A., Gilchrist, A., and Newson, R. L., Quart. J. Roy. Met. Soc. , 98 , 809 (1972).
4. Warshaw, M., and Rapp, R. R., R-908-ARPA (1972).