The HyMeX White Book often mentions scale multiplicity, intermittency and limited predictability of hydro-meteorological fields as key issues. They are indeed rather indispensable to get both a global and a detailed understanding of the variability of these fields to know how to observe and to simulate them. This calls for a quantitative assessment of these related behaviours both in data and in modelling. We will illustrate how this can be achieved in space and/or time with the help of a few studies performed on Nîmes extreme events, TRMM satellite radar data and large scale model downscaling, as well as on the large CEATI/DSIG data base of river discharges, by determining characteristic scales, scaling ranges, critical exponents, probability distributions, etc.