A few case studies will be presented involving both radioactive and chemical pollution at small, medium, and large space-time scales. Reported are recent advances in the field of environmental pollution involving the use of fractals and multifractals. The mathematical tools proposed here may offer new perspectives for investigating many of the problems of nonlinear variability which commonly arise when dealing with pollutants, such as the presence of outliers and the sparseness of the sampling networks. They may also lead to a simplification of the models adopted for studying natural phenomena, thanks to a scaling approach. Finally, they may provide parameters whose values are directly related to the nonlinear dynamics involved in the pollutant distribution in the environment which, in turn, may be relevant for computer simulation and epidemiological or risk assessment purposes.