Stemflow is of hydro-ecological importance in forested and agricultural ecosystems because it is a spatially localized input of water in the soil at the foot of the plant stem and has a significant influence in the groundwater recharge or in nutrient leaching (Levia and Herwitz, 2000). Lysimetry and tensiometry experiments were carried out to quantify the spatial distribution of drainage in an andosol after rainfall interception by the banana plant. HYDRUS 2D solved numerically the Richards equation for variably distributed saturated water flows (Simunek et al., 1999). The hydraulic properties of the andosol were measured (field and laboratory methods - Wind, 1968) and estimated by numerical inversion from tensiometric data, and employed as parameters to numerically simulate water movement, allowing to van Genuchten (1980) and Mualem (1976) equations. Distributed drainage obtained with the model was validated with field drainage data and statistical evaluation (Nash and Sutcliffe, 1970). Experimental bias were assessed, wick lysimeters could just simulate soil pressure head superior to -1.3 m, and corrected thanks to simulations without lysimeter. The transferred volumes under the banana stem were 2 times higher than downstream from the stem in the row (to 1,20 m from the stem as shown in the figure 1) and between the rows since these zones were sheltered by the banana leaves and received essentially throughfall. Analysis of others scenarios of simulation showed the generic aspect of an extreme situation and helped in understanding what would expect without slope, neither or after tillage practise and during the banana growth. The generation of the stemflow (evolution during the growth and localisation), was the principal distributive source of the drainage under banana plant. Soil hydraulic properties, tillage pan, soil inclination and initial boundary conditions had also a moderate impact on distributed drainage. These last results could question the concept of fertiliser localisation at the plant foot all the more that abundant stemflow may rapidly leach the nutrients, particularly after flowering, when rainfall interception was significant. References: (i) Levia, D.F. and Herwitz, S.R., 2000. Physical properties of water in relation to stemflow leachate dynamics: implications for nutrient cycling. Canadian Journal of Forest Research, 30(4): 662-666. (ii) Mualem, Y., 1976. A new model for predicting the hydraulic conductivity of unsaturated porous media. Water Resources Research, 12(3): 513-522. (iii) Nash, J.E. and Sutcliffe, J.V., 1970. River Flow Forecasting through Conceptual Models - Part I: A Discussion of Principles. Journal of Hydrology, 10: 282-290. (iv) Simunek, J., Huang, K., Sejna, M. and Genuchten, M.T.v., 1999. The HYDRUS-2D Software Package for Simulating Two-Dimensional Movement ofWater, Heat and Multiple Solutes in Variably-Saturated Media. Version 2.0. IGWMC-TPS-53, Int. Ground Water Modeling Center, Colorado School of Mines, Golden, CO. 251 pp. (v) van Genuchten, M.T., 1980. A closed form equation for predicting the hydraulic conductivity of unsaturated soils. Soil Science Society of America Journal, 44: 892–898. (vi) Wind, G.P., 1968. Capillary conductivity data estimated by a simple method, P.E. Rijtema and H. Wassink (ed.) Water in the unsaturated zone. Vol. 1. Proc. Wageningen Symp. June 1966. Int. Assoc. Scientific Hydrol., Gentbrugge, Belgium., pp. 181-191. Figure 1: Flux velocity vectors under banana plant in the row.