HydroLogic (Netherlands)

Probabilistic inundation forecasts that consider the uncertainty in rainfall predictions are crucial for flood early warning systems to effectively manage and reduce potential risks posed by pluvial flood events. Timely generation of such forecasts is challenging with physically-based numerical models due to computational demands. In contrast, data-driven models have a relatively low computational cost and can generate results quickly, making them a promising alternative to overcome this issue. This study proposes a long short-term memory (LSTM) neural network that can predict inundation progression over time at a high spatial resolution The network is trained on 1600 hydraulic simulations conducted using a 1D2D hydraulic model. With the trained network, probabilistic inundation forecasts are generated by combining the deterministic inundation predictions of 50 ensemble members of the rainfall forecast. The model is successfully tested for temporally varying rainfall events in a rural study area, and can generate accurate probabilistic inundation forecasts within seconds.

The last 60 years has seen unprecedented groundwater extraction and overdraft as well as development of new technologies for water treatment that together drive the advance in intentional groundwater replenishment known as managed aquifer recharge (MAR). This paper is the first known attempt to quantify the volume of MAR at global scale, and to illustrate the advancement of all the major types of MAR and relate these to research and regulatory advancements. Faced with changing climate and rising intensity of climate extremes, MAR is an increasingly important water management strategy, alongside demand management, to maintain, enhance and secure stressed groundwater systems and to protect and improve water quality. During this time, scientific research—on hydraulic design of facilities, tracer studies, managing clogging, recovery efficiency and water quality changes in aquifers—has underpinned practical improvements in MAR and has had broader benefits in hydrogeology. Recharge wells have greatly accelerated recharge, particularly in urban areas and for mine water management. In recent years, research into governance, operating practices, reliability, economics, risk assessment and public acceptance of MAR has been undertaken. Since the 1960s, implementation of MAR has accelerated at a rate of 5%/year, but is not keeping pace with increasing groundwater extraction. Currently, MAR has reached an estimated 10 km3/year, ~2.4% of groundwater extraction in countries reporting MAR (or ~1.0% of global groundwater extraction). MAR is likely to exceed 10% of global extraction, based on experience where MAR is more advanced, to sustain quantity, reliability and quality of water supplies.

Data collection is a critical activity in the management of water systems because it supports informed decision making. Data are collected by means of monitoring networks in which water level gauges are of particular interest because of their implications for flood management. This paper introduces a number of modifications to previously published methods that use information theory to design hydrological monitoring networks in order to make the methods applicable to the design of water level monitors for highly controlled polder systems. The new contributions include the use of a hydrodynamic model for entropy analysis, the introduction of the quantization concept to filter out noisy time series, and the use of total correlation to evaluate the performance of three different pairwise dependence criteria. The resulting approach, water level monitoring design in polders (WMP), is applied to a polder in the Pijnacker region, Netherlands. Results show that relatively few monitors are adequate to collect the information of a polder area in spite of its large number of target water levels. It is found, in addition, that the directional information transfer DITYX is more effective in finding independent monitors, whereas DITXY is better for locating sets of monitors with high joint information content. WMP proves to be a suitable and simple method as part of the design of monitoring networks for polder systems.

Abstract A method is presented for testing weather forecast products for applications in anticipatory water-system control. The applicability of the ensemble prediction system (EPS) of the ECMWF is tested for flood control in a regional water system in the Netherlands. By performing long-term verification analyses, a full range of probability-threshold-based decision rules to apply anticipatory control actions is evaluated in terms of hits (correct alerts), missed events, and false alarms. The analysis includes forecast horizons from 3 to 9 days for extreme precipitation events and extreme water-level events. The water-level forecasts are prepared by feeding each member of the ECMWF EPS precipitation ensemble into a deterministic water-system control model. The current operational strategy is modeled to forecast when the routine operational control will not be sufficient to prevent high water levels, and therefore anticipatory control actions are needed. The results show that ECMWF EPS precipitation forecasts contain useful information for regional water-system control in the Netherlands. All critical events during the analysis period were forecasted. The applied analysis method allows water managers to define decision rules, on the basis of requirements, in terms of forecast horizon, safety, and allowable costs of adverse effects of false alarms.

Recent developments in the field of computational intelligence (CI) techniques are helping to solve various problems of water resources modeling and management. These techniques have also shown their potential as an alternative approach to conventional controllers. In this paper, artificial neural networks (ANN) and fuzzy systems (FS) are shown to be efficient alternatives to using optimal control algorithms in real-time control of the polder water system of Overwaard in The Netherlands. The relation between the optimal decision or action and the influencing parameters are learned by ANN and FS and then used to derive the decisions and control actions in real-time. It was possible to reproduce the centralized behavior (in terms of water levels and corresponding discharges) of optimal control action by using easily measurable local information. Moreover, it is demonstrated that model simulation with external intelligent controllers is ten times faster than that with the optimal control.

Le test d'ancrage cellulaire a été appliqué au suivi de la qualité d'une eau potable au contact d'un revêtement époxyuréthane. La dureté des eaux testées a nécessité la mise au point d'un protocole particulier pour éviter la précipitation du calcium dans le milieu de culture.Deux séries d'essais ont été réalisées : la première sur l'eau de distribution ayant séjourné au contact du revêtement dans le réservoir, la deuxième sur la même eau aorès contact plus ou moins prolongé avec des éprouvettes au laboratoire.Quelques paramètres physicochimiques et organoleptiques ont été déterminés en parallèle.