Abstract. Transport parameters obtained from laboratory tracer experiments were used to evaluate the stochastic form of the equilibrium convection-dispersion equation (CDE) in describing the transition of scale, i.e. from the column or local scale to a larger field scale. Local-scale solute breakthrough curves (BTCs) were measured in 1-m-long and 0.3-m-diameter undisturbed soil columns by means of time-domain reflectometry at six depths for a 79-h input pulse of chloride. The localscale data were analysed in terms of the equilibrium CDE and the mobile-immobile non-equilibrium transport model (MIM). At the local scale, the MIM transport model better described the observed early breakthrough and the tailing of the BTC than did the CDE. A linear regression analysis indicated that the relationship between the hydrodynamic dispersion D and pore-water velocity v was of the form D = 31v l.92 (correlation r v,D = 0.74). Averaging of the local-scale BTCs across the field produced a large-scale or field-scale mean BTC; at the greatest observation depth (0.8 m) the fieldscale dispersivity / = l equals 0.656 m. The results further showed that for large values of the mean dispersion coefficient, , local-scale dispersion is an important mechanism for field-scale solute spreading, whereas the standard deviation, s D , and the correlation between v and D, r vD , have negligible effects on field-scale transport. Stochastic stream tube models supplemented with statistical properties of local-scale transport parameters provide a practical and computationally efficient tool to describe heterogeneous solute transport at large spatial scales.