It is well recognized that interpretations of Mars must begin with the Earth as a reference. The most successful comparisons have focused on understanding geologic processes on the Earth well enough to extrapolate to Mars' environment (e.g., see reviews by Farr et al.  and Farr ).
The use of terrestrial analogues has a long tradition among scientists who study planetary landscapes (Sharp, 1988). The basic premise is that a planetary feature looks similar to a terrestrial feature, whose properties and origin are known. The known causes of the terrestrial analogue might allow us to infer the causes of the planetary feature under study. This way of analogical reasoning was probably first formally described for the field of geoscience by Gilbert (1886), and the reader is referred to Baker (2008) for a more in-depth discussion of analogical reasoning in planetary geomorphology. It has to be emphasized, however, that analogues do not prove any causal relationships. Instead, they can help to find lines for further reasoning (e.g., multiple working hypotheses) (see also Baker, 1996).
Baker, V.R. (1996) Hypotheses and geomorphological reasoning. In Rhoads, B.L., and Thorn, C.E. (eds.), The Scientific Nature of Geomorphology, pp. 57–85, Wiley, New York. Baker, V.R. (2008) Planetary landscape systems: a limitless frontier. Earth Surface Processes and Landforms, 33, 1341–1353.Farr, T. G. and 46 co-authors (2002) Terrestrial Analogs to Mars. In: Sykes, M.V. (ed.) The Future of Solar System Exploration, 2003-2013, pp. 35-76, ASP Conference Series, 272.Farr, T. G. (2004) Terrestrial analogs to Mars: The NRC community decadal report. Planetary and Space Science, 52, 3-10.Gilbert, G.K. (1886) The inculcation of scientific method by example. American Journal of Science, 31, 284-299.Sharp, R.P. (1988) Earth science field work: Role and status. Annual Reviews of Earth and Planetary Sciences, 16, 1-19.