The role of statistical analysis in the process of establishing and utilizing climate and other environmental
models is discussed. The role of empirical evidence is made explicit by having not only dynamical equations but observational equations as well. In ``quasi-realistic models'', statistical thinking is encoded in the  parameterizations, and is required for extracting experimental evidence and for validation. Data assimilation techniques are used to systematically combine observational evidence and quasi-realistic models. While quasi-realistic models serve as complex substitute reality, is dynamical knowledge represented trough simplified models. These ``cognitive'' idealized models have to be fitted to observational data when adapted to  real situations.

Literature:
von Storch, H., and F.W. Zwiers, 1999: Statistical Analysis in Climate Research, Cambridge University Press, ISBN 0 521 45071 3, 494 pp.

von Storch, H., and A. Navarra (Eds.), 1999: Analysis of Climate Variability: Applications of Statistical
Techniques, Springer Verlag,  2nd updated extended edition (ISBN 3-540-66315-0), 342 pp,

von Storch, H., 2000: Statistics - an indispensable tool in dynamical modelling. In: H. von Storch and G. Flöser (Eds): Models in Environmental Research. Springer Verlag (in press)