This study focuses on the hydraulic behavior of potential water repellent soils. The temporal dynamics of water repellence in soils have a strong influence on both hydraulic properties and preferential flow occurrence. During the seasons, the porous system of the soil changes its physical behavior from wettable to water repellent and reverse. To quantify these effects on the soil water flux, field investigations were conducted at a sandy site near Berlin, Germany. The numerical model HYDRUS 2D was enlarged for the calculation of water flow in water repellent soils. The site is a former waste water disposal field, which contains a potentially water repellent top soil (to 40-60 cm depth) with increased organic matter content. The subsoil with lower contents of organic matter is wettable. Soil samples were collected over a 3-yr period. Additionally, a TDR-array with 63 probes measured the water content hourly on a transect of 130x60 cm. Based on the sampling campaigns the fraction of water repellent soil regions was measured. Water content changes were observed using TDR probes in high spatial and temporal resolution. To quantify the heterogeneity (i.e. the degree of the preferential flow), we propose a new parameter, the Effective Cross Section of water flow (ECS). This parameter was calculated by fitting the beta function to the cumulative values of the water content change over a horizontal cross section after single rainfall events. ECS determines the horizontal area, on which 90% of the water fluxes takes place. The parameter can be calculated easily with field TDR measurements and can be compared with results of numerical models. In our study the index reached values between 0.26 and 0.73 (90% of the water content changes on 26% resp. 73% of the area). During the summer time the effective cross section showed values in the range of 0.3. During winter time the ECS was in the range of 0.7, except during periods with frozen soil. The soil water content is the main influencing factor on the effective cross section. Additionally, rainfall intensities, rainfall durations, and potential evapotranspiration influence the effective cross section and the preferential flow. Water retention curves and unsaturated hydraulic conductivity were measured in the laboratory experiments on undisturbed soil columns. The measurements were conducted for wettable and water repellent conditions for the same soil. The numerical code of HYDRUS 2D was extended to calculate the water fluxes under water repellent conditions in a way that the concept of the critical water content was integrated in the code. For a water content below this critical water content, the soil is water repellent, above it the soil is wettable. Field investigations of our site showed a critical water content of about 0.12 m³/m³. It depends slightly on the amount of organic matter. Different hydraulic functions for wettable and repellent conditions were developed and used for the wetting cycle. For drying conditions, the same hydraulic functions can be used for water repellent as for wettable conditions. This was a result of the laboratory measurements. The critical water content was used in the numerical model to switch between the different hydraulic functions. As a result, the wettable soil fraction and, therefore, the ECS increases during the rewetting phase in autumn. It also became necessary to consider hysteresis for the calculation of these complex processes. The approach of Lennards and Parker was used to avoid the pumping effect during wetting-drying cycles. The 2-dimensional numerical model allows the calculation of the upturning and vanishing of the water repellent areas during the seasons. The effective cross section can also be applied to a one dimensional model as a fraction of the water repellent area.