Forests show, compared to other land uses, in many cases good water retention potential. This is however limited. Whether additional forest area in a catchment leads to a reduction of flooding depends on the pre-event atmospheric conditions, the soil characteristics at the afforested site, the duration and intensity of the rain storm event, and location and size of the afforested area. Further, the forest management, which is only briefly discussed in this thesis, plays an important role. Many water retention related processes occurring during the transformation of a landuse into forest are not yet sufficiently investigated an described. This applies especially to the changes in the hydraulic architecture of the soil. It was shown that after a few years growing forests have already changed the pore distribution, especially in the upper soil horizons. However, further research under different soil and tree type would be desirable. Therefore, a model-based description of land use change towards forest with regard to flood retention comprises uncertainties which should be taken into consideration. Nevertheless, models are the only possibility to assess land use change effects with justifiable expenditure.rnrnIn general, the application of hydrological models comprised sparse useful information about changes in the soil due to a changed land use. Neither the target state nor the progression of the transformation can be predicted with certainty. Further development of models with parallel observations and data gathering is essential. With increasing number of questions regarding modified land use systems, a need arises for innovative forms of parameterisation and model calibration. The increasing degree of process mapping in models may make parameterability difficult, however, adequate process mapping is the key to scenario capable modelling. The communication of results must therefore include a high degree of transparency in the definition of all known uncertainties, because decisions have long lasting consequences.rnrnA qualified prediction of land use changes is a cross-disciplinary task. Ecological, economical, and sociological processes together form the future land use distribution. An important conclusion from this thesis is that the implementation of measures targeting increased water retention requires must result in a consensus with society and economics. Integrated approaches and transdisciplinary assessment of impacts of land use modifications are needed. Although, the uncertainties in model-based land use change assessment are high, there is a need for the definition of “sustainable land use” and “increase of water retention” for the flood risk management plans. Adapted land use as a component of integrated flood risk management has a major constraint: the benefits of water retention in the landscape are mostly not directly noticeable at the place where a measure is implemented. This is highly important for stakeholders and decision makers. However, given that most of the land available for afforestation is a private property, it may be necessary to provide subsidies to encourage landowners to increase the percentage of forested land. Competitive land use system requirements need to be balanced with approaches dealing with different landscape functions. Water retention is part of this functioning. Other protection aims like nature protection, soil protection, aims of the Water Framework Directive, aesthetic land use pattern but also the agrar-economic production play an important role. Well-founded future land use scenarios should use this transdisciplinary view.rnrnFinally, it is also important to keep in mind that floods belong to a healthy river runoff regime. Floods are an important part of the natural hydrological cycle, and therefore the goal of watershed management should not be to eliminate them entirely. Additional forest can help to re-establish the natural water retention potential in anthropogenically disturbed river basins and to decrease the human-made contribution to flood generation.