In the past, investigations on surface seals developing under simulated rains usually were performed with uniform rainfall intensities. Recent studies, however, showed that varying-intensity rains affect erosion and volumes of runoff. We conducted a study on surface sealing under varying-intensity rains using physical and morphological methods. Four 1-h-duration rainfall patterns were used: S1, S2, S3 and S4, with mean rainfall intensities of 83, 69, 65, and 51 mm h-1, respectively, and highest intensities of 148, 120, 111, and 100 mm h-1, respectively. Two experiments were performed in triplicate: (i) in packed soil columns, pressure heads at 1-cm depth were measured, and runoff (Rt) and wash sediment (St) collected at short time intervals during S1, S3, and S4 patterns; and (ii) packed shallow soil flumes were successively exposed to 10, 15, 30, 40, and 60 min of S2 and S4 patterns, and the surface was then characterized by micromorphological analysis of thin sections (S2) and top-view photographs (S2 and S4). The largest differences between rainfall patterns occurred during the intensity peaks (10–30 min): at the end of these periods, measured surface seal hydraulic conductances (β) attained values of 0.80, 0.25, and 0.51 h-1 for S1, S3, and S4, respectively. These relative β values are thought to be due to partially open surfaces as a result of microrill erosion (S1), complete sealing without erosion (S3), and incomplete sealing (S4). Presented Rt, St, and morphological data support these conclusions. After 5 min of the intensity peaks of S2, surface seals attained maximum observed thickness and minimum macroporosity (>30 µm). Planar voids (>500 µm) were eliminated at this point. Under continuing rain, erosion reduced the seal thickness and increased macroporosity, mainly as planar voids.