This paper investigates the effect of sub-grid rainfall variability on the simulation of land surface hydrologic processes of three regions (Europe, Africa and Amazon) with contrasting precipitation and vegetation characteristics. The sub-grid rainfall variability is defined in terms of the rainfall coverage fraction at the model's grid cells, and the statistical distribution of rain rates within the rain-covered areas. A statistical-dynamic approach is devised to incorporate the above variability properties into the canopy interception process of a land surface model. Our results reveal that incorporation of sub-grid rainfall variability significantly impacts the land-atmosphere water vapor exchanges. Specifically, it alters the partitioning between runoff and total evapotranspiration as well as the partitioning among the three components of evapotranspiration (canopy interception loss, ground evaporation and plant transpiration). This further influences the soil water, and to a lesser effect surface/vegetation temperatures and surface heat fluxes. It is shown that, overall, rainfall variability exerts less of an impact on the land-atmosphere flux exchanges over Europe compared to Africa and Amazon.