Of all climatic parameters, rainfall has the greatest variability in space as well as time. It also has the greatest influence on the breeding and behaviour of migrant pests, supplying the moisture needed both for their development and for the growth of vegetation to sustain a population. Soil moisture controls both processes directly but cannot be adequately surveyed even by remote sensing techniques, so we rely upon water balance models to interpret rainfall measurements and to forecast pest populations. Both terrain and rainfall are very inhomogeneous on the kilometric scale, a scale which is not matched by observations either from conventional raingauges or current meteorological satellites. Modelling the effects of rainfall must take account of these inhomogeneities, and the processes involved will be discussed. However, it is clear than more detailed studies of rainfall-habitat interaction are needed to derive soil moisture from rainfall estimates by using knowledge of the microtopography. Such relationships must be capable of being generalized so that future monitoring by satellites, essential to give complete and uniform coverage, can be realistically interpreted.