The spatial distribution of soil hydrological properties and associated flow pathways within a single 0.01 to 10 km2 catchment are usually poorly defined by typical programmes of point-scale measurements. This is due not only to "measurement uncertainty" caused by instrument error and inadequate spatial sampling but also to "conceptual uncertainty" resulting from a deficiency in the Darcy-Buckingham-Richards description of subsurface flow. This study examines these two sources of uncertainty in the identification of subsurface flow beneath the Bukit Timah forest, Singapore. The two distinct approaches of water-tracing and Darcy-Buckingham calculations are compared. Flow estimates resulting from quasi-steady Darcy-Buckingham calculations were found to be very sensitive to the magnitude and pattern of the saturated hydraulic conductivity, which itself was sensitive to the permeametry technique used. The use of a Hydro-Physics ring permeameter in the estimation of subsurface flows gives rates that are far greater than the mean propagation rates of tagged water though the approach identifies the relative difference between the behaviour of the two slopes studied. Whilst the lumped differences between the two tropical slopes may be identified with the Darcy method, the poor correlation with the absolute velocity and direction of the tracer plume means that catchment or even slope scale simulation using parameters of the Darcy-Buckingham-Richards equation will be highly uncertain.