The use of bioclimatic indices could be a major step forward in the methodology of pollen forecasting. The basis for this proposal is that simple meteorological parameters do not reflect the global status of the atmosphere, but merely some static measurements. However, pollen dispersal is, above all, a dynamic phenomenon, and this fact should be reflected in the variables we used to explain it. Here, we test the two methodologies for routine pollen forecasting by comparing correlation coefficients using the same daily Poaceae airborne pollen data base from León (6 years, from 1994 to 1999) as the dependent variable and either simple daily meteorological variables or compound daily bioclimatic indices as independent variables. Both simple and compound indices reproduced the same profile of evolution of plant eco-physiological requirements, as the length of the study period during the pollen season increased. However, for time frames larger than the main pollen period, bioclimatic indices gave superior coefficients, which seems to indicate that these could be more valuable for pre-season pollen forecasting. The continentality index produced the highest mean coefficient, higher than those generated by any meteorological variable. Furthermore, at least for a Mediterranean climate, site location and evapotranspiration in relation to precipitation seem to be the most promising factors for increasing success when forecasting Poaceae airborne pollen concentration.