Environmental science almost invariably proposes problems of extreme complexity, typically characterized by strongly nonlinear evolution dynamics. The systems under investigation have many degrees of freedom - which makes them complicated - and feature nonlinear interactions of several different components taking place on a vast range of time-space scales - which makes them complex. Such systems evolve under the action of macroscopic driving (typically the solar heating) and modulating (e.g. the Earth's rotation and gravitation) agents. The most comprehensive example is the entire climatic system. The description of the macroscopic dynamics of environmental systems is based on the systematic use of dominant balances derived on a phenomenological basis in order to specialize the dynamical equations. Such balances are suitable classes of approximate solutions of the evolution equations which represent a reasonably good approximation to the actual observed fields when sufficiently large spatial or temporal averages are considered. Actually, different balances have to be considered depending on the time and space scales we are focusing our interest on. Such an approach reflects the fundamentally heuristic-inductive nature of the scientific research in environmental sciences, where the traditional reductionistic scientific attitude is not always effective. In order to exemplify this procedure, we consider the very relevant case of the motion of the fluids that permit the existence of life on the Earth, air and water: the so-called geophysical fluids.