Over geological time, the Martian climate has changed drastically, but Earth's only substantially. Since we cannot explain why in either case, we should keep an open mind about our climatic history and about the factors that can affect it.
This paper is at best a brief summary of a complex subject. For more detail, the recent paper by Pollack (1979) is warmly recommended. For Earth, another excellent source is Appendix A in the United States Committee for the Global Atmospheric Research Program (GARP) report (1975). A perspective over the whole life of Earth is given by Walker (1977).
Figure A.1 in the GARP report (1975) illustrates a notable change in the Argentiere glacier of the French Alps over a period of only 115 years- it has retreated far up the valley from its 1850 position. It is clear that substantial changes in climate occur over the span of a long human lifetime. Larger changes are found-the "Little Ice Age" of 1400-1750 and the ice ages themselves, the latest from about 125,000 to 15,000 years ago. These fluctuations, along with many others, are illustrated in figure A.2 of the same report.
It was demonstrated by Mariner 9, and amply confirmed by the Viking Orbiters, that many areas of Mars have been affected by a running fluid. Although liquid water is unstable on Mars today, all the proposed alternatives are far more objectionable. Therefore, Mars must have once had a very different climate, somewhat warmer and with a considerably greater atmospheric pressure Figure 2 in Masursky et al. (1977) shows two different  channel types, indicating different types of flow. The "fluviatile" channels seem to be ordinary rivers with tributaries, and the simplest way to generate them is by the collection of rainfall over a long period. The more spectacular channels are best explained by a single enormous flow, as by the breaking of a dam. The polar caps show another kind of evidence (Pollack, 1979): a "laminated terrain" that suggests some sort of cyclic variation. The age of these features is unknown, but the channels are thought (through crater counts) to be one to several billion years old.
The existence of ice ages on Earth and the evidence just summarized for Mars have suggested to many people that climate may be hypersensitive to small changes in the forcing factors. In other words, the system may contain positive feedback mechanisms that are significant enough to allow two alternative states, stable to small changes but susceptible to being "flipped" by somewhat larger ones. An obvious mechanism could be an increase in cloudiness, ice cover, or both with reduced temperature. The increased albedo to solar radiation could reduce the net heat input and lead to a further cooling.
Albedo changes fall into the category of internal influences. Others are changes in the ocean circulation, continental drift, veils of volcanic dust, and the evolution of the atmosphere. Models of the climate can be constructed and run on a very large computer, but they must specify all these factors at an observed value. They are essentially useless for predicting a response to changed conditions.
Variations in the output of the Sun are of obvious interest. At very short wavelengths, large variations are well documented and show a strong correlation with solar activity as measured by sunspot area. These wavelengths carry, however, less than 10-4 of the total solar energy, and it is hard to imagine any effect on surface climate even if they disappeared altogether. Rapid variation of the Sun's total output is inhibited by its huge mass-even the outer convective zone has a thermal time constant of 105 years.
 Eddy (1976) has shown that sunspots were absent from 1645 to 1715 and has named this period the "Maunder minimum." Other indirect indicators of solar activity are in agreement and suggest several earlier such periods. There is a rather good correlation with periods of reduced temperature on Earth. As pointed out above, our present understanding of the energetics does not support the idea of a causal connection, but the evidence is nevertheless suggestive.
On a much longer time scale, theories of the Sun's evolution indicate an increase by a factor of 1.37 in output over its life. Curiously, both Mars and Earth appear to have been warmer, if anything, in the remote past. Proposed explanations usually have involved an increased greenhouse effect, perhaps due to NH3 or CO2 in the early atmosphere. This solution has its own problems since NH3 is rapidly photodissociated.
The quasi-periodic changes in orbital eccentricity and axial inclination, generally called Milankovitch variations, comprise a quite different class of external influences. After the polar laminated terrain on Mars was discovered, Ward (1974) computed obliquity changes of 20° peak to peak. Although they are fairly large for this planet, it has been impossible to make a correlation because the time scale for Mars is unknown. The same difficulty held up terrestrial studies until Hays et al. (1976) used deep-sea cores to obtain a chronology over nearly 500,000 years. A Fourier analysis revealed periods of around 22,000, 42,000, and 100,000 years. A similar analysis of the driving functions shows 22,000 year, (precession), 41,000 years (obliquity), and 97,000 years (eccentricity). Although the 100,000-year peak is the largest, there is no obvious reason why the eccentricity of the orbit should affect climate at all! More recently, Kominz and Pisias (1979) did a coherence analysis. They found none at the 100,000-year period, but significant coherence at the two shorter periods. Only about 15% of the variance, however, is accounted for in this way. Nevertheless, there seems to be a clear indication that small external changes do have a measurable effect.
The current rate of escape of hydrogen from Earth, continued over geologic time, would leave behind just the amount of free oxygen now in the atmosphere. Photosynthesis and burial of carbon are now much more important Sources of oxygen. Excess oxygen presumably converts ferrous iron in the crust to ferric. Hydrogen loss at present is limited by the very low stratospheric humidity, which seems to be a sensitive consequence of the particular structure and circulation of the atmosphere. A primitive atmosphere  might support a much larger loss rate and a correspondingly larger generation of oxygen (see, e.g., Hunten, 1973). It can also be argued that crustal iron is directly oxidized by H2O, with H2 coming out in volcanoes and supporting the escape (Walker, 1977; see also Kasting and Donahue, page 149 in this volume). The oxygen would then never appear in the atmosphere.
The principal conclusion is that we should keep an open mind about past climates. They are more likely to be different from than similar to modern ones.
- Eddy, J. A.: The Maunder Minimum. Science, vol. 192, no. 4245, June 18,1976, pp.1189-1202.
- Hays, J. D.; Imbrie, J.; and Shackleton, N. J.: Variations in Earth's Orbit-Pacemaker of Ice Ages. Science, vol. 194, no. 4270, 1976, pp. 1121-1132.
- Hunten, D. M.: The Escape of Light Gases from Planetary Atmospheres. J. Atmos. Sci., vol. 30, Nov.1973, pp.1481-1494.
- Kominz, M. A.; and Pisias, N. G.: Pleistocene Climate: Deterministic or Stochastic? Science, vol. 204, April 13,1979, pp.171-173.
- Masursky, H.; Boyce, J. M.; Dial, A. L.; Schaber, G. G.; and Strobell, M. E.: Classification and Time of Formation of Martian Channels Based on Viking Data. J. Geophys. Res., vol. 82, Sept. 30,1977, pp. 4016-4038.
- Pollack, J. B.: Climatic Change on the Terrestrial Planets. Icarus, vol. 37, Mar.1979, pp. 479-553.
- Understanding Climatic Change, A Program for Action. United States Committee for the Global Atmospheric Research Program, National Academy of Sciences, Washington, D.C., 1975.
- Walker, J. C. G.: Evolution of the Atmosphere. Macmillan, New York, 1977.
- Ward, W. R.: Climatic Variations on Mars. I -Astronomical Theory of Insolation. J. Geophys. Res., vol. 79, Aug. 20, 1974, pp. 3375-3386.