Emergency closure of dike breaches

Abstract

Floods cause, even in modern countries, a lot of victims and damage every year. Generally, countries invest in prevention and protection, yet, also emergency response measures could be a (last) option. Regarding the latter, a distinction is made between preventive emergency measures (e.g. sand bags to raise the crest level of a flood defence) and curative emergency measures (e.g. measures to combat a breach in a flood defence). It can be stated that the knowledge about the closure of a dike breach and the implementation of emergency measures is not at the desired level. Besides, closing a breach is very difficult and it is rarely performed successfully. There is a need for research regarding emergency closure of dike breaches indicated by among others Dutch Water Boards. The problems of a dike breach can broadly be categorized in three fields: organizational problems, logistical problems and technical problems. This thesis focusses mainly on the technical problems, distinguishing breach characteristics and type of emergency measures. Therefore, in this thesis the research question is: What is the effectiveness and applicability of curative emergency measures, applied in developing breaches in a Dutch river or lake dike? To get a better understanding of the effect of emergency measures, simulations of developing breaches including emergency measures can be made with a numerical model. Simulations have several advantages. It allows identification of breach characteristics like duration and breach stages during closure attempts. Furthermore, it can be used to optimize closure strategies and measures and it can help flood managers to prepare for emergency situations. However, such a model is non-existent. Yet, XBeach is capable of simulating emergency measures as so-called non-erodible layers in developing breaches. With some modifications XBeach seems promising for simulating breach development with emergency measures. The modifications are based on XBeach’s limitations. First, emergency measures are implemented as non-erodible layers, these are however not adjustable in time. Running multiple simulations after each other representing the different closure phases, solves this problem. Second, the non-erodible layer is always stable. This means that the emergency measure can not move due to high flow velocities or scour during the simulation. This is the main limitation of XBeach for this research. To still be able to check the stability, separate calculations are done. Third, XBeach is not a 3D model. Therefore, emergency measures where a 3D effect (piping) plays an important role are not modelled. Fourth, XBeach is only capable of simulations with non-cohesive sand. To deal with this, a scaling from non-cohesive to cohesive timespans was made. For a breach closure in a sand dike, there is (likely too) little time. To estimate the time available in cohesive dikes, the timing is scaled to realistic proportions. Various curative measures exist. Simulated measures in this study in XBeach are a scaffold, Big Bags, a vessel and an emergency dike. The emergency dike seems to be the most promising measure. It makes use of the smaller flow velocities upstream of the breach. The breach dimensions stay smaller and the polder water level is lower. Point of attention is the static stability after the closure, with piping as the most important threat. Logistically seen, a complete closure with an emergency dike for a breach in a clay dike is plausible, using trucks to bring in Big Bags and helicopters to place them at the desired location. Next to these simulations, case studies of performed closures are evaluated, a field test is carried out and recommendations to set up a Decision Support System are done.