Background, Aims, and Scope. Sediment management in coastal zones is taking place in high complex environments. Present management options do not have a sophisticated way of dealing with the actual complexity of the physical and the social systems and with the unpredictability that is inherent with these systems. Therefore, a new approach in both policy making and sediment management is needed that takes this complexity into account. The aim of this article is to explore the dynamics in social and natural systems and to draw the contours of this new approach for policy processes and sediment management that fits to the dynamics of the systems. Methods. Three case studies concerning sediment management in Germany and the Netherlands are analysed in this article, in which unpredictability, or whimsicality, appeared through the occurrence of unexpected events. The case studies are analysed from the perspective of complexity theory. Complexity theory is a systemic theory, which means that it explains empirical phenomena from complex system behaviour. To understand the capricious character of sediment management in coastal zones, we need to frame the issue as the interdependency between the physical system and the social system, causing the two systems to develop in mutual adjustment through feedback. The process of mutual adjustment is one that is charactcrised by a non-linear nature. This is caused by the multiple relationships and the feedback within and between the systems, and the occurrence of chance events. Chance events (surprises) happen suddenly without an apparent cause and are important triggers for change in the systems. In three empirical cases, the occurrence, nature and response to these chance events are analysed as these factors influence the course of sediment management. Results and Discussion. The case studies show that chance events can occur in the biophysical and in the social system. In the three cases, players or actors in the decision process are left with the choice to adapt themselves to the occurring chance events or to refrain from any adaptive behaviour. Chance events can open up new possibilities by activating (new) actors and by coupling to new issues. If the situation is too locked-in (i.e. a stalemate) and is intentional on behalf of the actors, than the chance event will have no effect. There are, however, situations of lock-in that are unintentional, and in such situations a chance event can remove this lock-in, The effects depend largely on the adaptive capacity of the actors to respond adequately and timely to such situations. The adaptive capacity can be increased (and uncertainty reduced) by a better understanding of both the physical and the social system. The case studies show that adaptation is an adequate way of dealing with the occurrence of chance events. Conclusions. We conclude that the way to deal with non-linear developments is through an adaptive policy approach with short feedback loops in order to allow for timely adjustments and learning loops that will progress the understanding of the systems - both social and physical. Besides the instruments that are already available, like modelling and forecasting, instruments like observation and monitoring, stakeholder involvement processes, and learning and adaptation should be developed in this new adaptive approach. Monitoring of the physical system is a key element in this approach as all involved parties and stakeholders can learn how the physical system behaves. In this adaptive approach, whimsicality - that occurs through unexpected events - is an interesting challenge for all concerned with sediment management. Recommendations and Perspectives. The proposed adaptive policy approach should be developed further and should be experimented with in real life situations that are well monitored. © 2007 ecomed publishers (Verlagsgruppe Hüthig Jehle Rehm GmbH).