Interception is the part of the rainfall that is intercepted by the earth’s surface and which subsequently evaporates. In this definition the earth’s surface includes everything that becomes wet after a rainfall event and that dries out soon after. It includes: vegetation, soil surface, litter, build-up surface, etc. How much of the precipitation evaporates depends on land cover characteristics, rainfall characteristics, and on the evaporative demand. Interception can amount up to 15-50% of precipitation, which is a significant part of the water balance. One can distinguish many types of interception, which can also interplay with each other. For example canopy, forest floor, fog, snow, and urban interception. This study we focus on canopy and forest floor interception. We measured interception of three dominant European vegetation types at three locations. In the Huewelerbach (Luxembourg) a beech forest has been investigated, in Westerbork (the Netherlands) grasses and mosses, and in the Botanical Garden (Delft, the Netherlands) a Cedar tree. Canopy interception is determined by the difference between gross precipitation and the sum of throughfall and stemflow. To measure forest floor interception a special device has been developed. It consists of two aluminium basins which are mounted above each other. The upper basin is permeable and contains the forest floor. By weighing both basins simultaneously, evaporation from interception can be calculated. For the beech forest we found that canopy interception has a clear seasonal trend ranging from 15% of rainfall in summer to 7% in winter. On the other hand, forest floor interception appears to be constant over the year and evaporates on average 22% of precipitation. Evaporation from the Cedar needle floor is only a bit lower: 18%, although the storage capacity is significantly lower: 1.0 mm for the needle floor compared to 1.8 mm for beech leaves. Both interception thesholds have a coefficient of variation as high as ±100%. However, the interception process is not sensitive to this variability, resulting only in 11% variation of evaporation estimates for the beech forest. Hence the number of raindays and the potential evaporation are stronger drivers of interception. Furthermore, the spatial correlation of the throughfall and infiltration has been investigated with semi-variograms and time stability plots. Within 6-7 m distance throughfall and infiltration are correlated and the general persistence is weak. The effect of spatial variability of interception on subsurface storm flow has also been investigated with a virtual experiment. A virtual experiment is a numerical experiment driven by collective field intelligence. It provides a learning tool to investigate the effect of separated processes in a complex system. We used this approach to better understand the generation and behaviour of subsurface stormflow (SSF) at the hillslope scale, because this is still poorly understood. Interactions between the permeable soil and the less permeable bedrock may cause non-linearity in subsurface flow depending on several hillslope attributes such as soil depth, slope angle, and bedrock permeability. It is known that the size of storm events also controls subsurface flow generation. The objectives of this study were three-fold: 1) to investigate if and how different configurations of throughfall patterns change the SSF behaviour; 2) to investigate the interplay between the spatially variable input and the hillslope attributes (slope angle and soil depth) on the generation of SSF; and 3) to investigate a geo-statistical tool that uses semi-variogram characteristics to analyse if soil moisture patterns during an event are dominated by throughfall patterns or by bedrock topography patterns. In our virtual experiment we combined spatial throughfall data from the Huewelerbach catchment in Luxembourg with the topography characteristics of the Panola hillslope in Georgia, USA. We used HYDRUS-3D as a modeling platform. The effect of the spatial throughfall pattern appears to be large on both SSF generation and the spatial variability of SSF along the hillslope, but only marginal on total SSF amounts. The spatial variability of SSF along the hillslope appears to be closely related to the drainage pattern of the bedrock. The geo-statistical analysis indicates that during the event soil moisture distribution reflects throughfall patterns, whereas after the event, during the drainage of the hillslope, the bedrock topography increasingly dominates soil moisture patterns. Furthermore, we found that on a daily time scale, interception is a typical threshold process. We used this characteristic to upscale daily interception to an annual evaporation model and found similarities with the Budyko curve. The Budyko curve is often used to estimate the actual evaporation as a function of the aridity index in a catchment. Different empirical equations exist to describe this relationship; however, these equations have very limited physical background. Our model concept is physically based and uses only measurable parameters. It makes use of two types of evaporation: interception and transpiration. Interception is modeled as a threshold process at a daily time scale. If multiplied with the rainfall distribution function, integrated, and multiplied with the expected number of rain days per month, the monthly interception is obtained. In a similar way, the monthly interception can be upscaled to annual interception. This results in a Budyko-type equation. Analogous to the interception process, transpiration can be modeled as a threshold process at a monthly time scale and can be upscaled by integration and multiplication with the expected number of rain months. The expected rain days per month are modeled in two ways: as a fixed proportion of the monthly rainfall and as a power function based on Markov properties of rainfall. The latter is solved numerically. It appears that on an annual basis the analytical model does not differ much from the numerical solution. Hence, the analytical model is used and applied on 10 locations in different climates. We show that the empirical Budyko curve can be constructed on the basis of measurable parameters representing evaporation threshold values and the expected number of rain days and rain months and, in addition, a monthly moisture carry-over amount for semi-arid zones. Overall, we can conclude that interception has different roles in the hydrological cycle. The most important role is as a rainfall reducer, causing a significant amount of rainfall to be directly fed back to the atmosphere which is not available for infiltration. Second, interception influences the spatial distribution of infiltration. This has large influences on the soil moisture pattern and on subsurface flow paths. Finally, interception redistributes the water flows in time. Due to the filling of the spatial variable storage capacity and rainfall, the delay time is not homogeneous in space. This thesis shows that interception is a key process in the hydrological cycle. It involves significant fluxes in the water balance and influences the subsequent processes both in quantity and timing. It is an important cause for non-linear behaviour of catchments. The role of interception in the hydrological cycle is crucial.