Gerd H. Schmitz

Towards an integrated arid zone water management using simulation-based optimisation

For ensuring both optimal sustainable water resources management and long-term planning in a changing arid environment, we propose an integrated Assessment-, Prognoses-, Planning- and Management tool (APPM). The new APPM integrates the complex interactions of the strongly nonlinear meteorological, hydrological and agricultural phenomena, considering the socio-economic aspects. It aims at achieving best possible solutions for water allocation, groundwater storage and withdrawals including saline water management together with a substantial increase of the water use efficiency employing novel optimisation strategies for irrigation control and scheduling. To obtain a robust and fast operation of the water management system, it unites process modeling with artificial intelligence tools and evolutionary optimisation techniques for managing both water quality and quantity. We demonstrate some key components of our methodology by an exemplary application to the south Al-Batinah region in the Sultanate of Oman which is affected by saltwater intrusion into a coastal aquifer due to excessive groundwater withdrawal for irrigated agriculture. We show the effectiveness and functionality of a new simulation-based water management system for the optimisation and evaluation of different irrigation practices, crop pattern and resulting abstraction scenarios. The results of several optimisation runs indicate that due to contradicting objectives, such as profit-oriented agriculture versus aquifer sustainability only a multi-objective optimisation can provide sustainable solutions for the management of the water resources in respect of the environment as well as the socio-economic development.

OCCASION: New Planning Tool for Optimal Climate Change Adaption Strategies in Irrigation

To sustain productive irrigated agriculture with limited water resources requires a high water use efficiency. This can be achieved by the precise scheduling of deficit irrigation systems taking into account the crops’ response to water stress at different stages of plant growth. Particularly in the light of climate change with rising population numbers and increasing water scarcity, an optimal solution for this task is of paramount importance. We solve the corresponding complex multidimensional and nonlinear optimization problem, i.e., finding the ideal schedule for maximum crop yield with a given water volume by a well tailored approach which offers straightforward application facilities. A global optimization technique allows, together with physically based modeling, for the risk assessment in yield reduction considering different sources of uncertainty (e.g., climate, soil conditions, and management). A new stochastic framework for decision support is developed which aims at optimal climate change ada...

Optimal planning and operation of irrigation systems under water resource constraints in Oman considering climatic uncertainty

In this contribution, we introduce a stochastic framework for decision support for optimal planning and operation of water supply in irrigation. This consists of (1) a weather generator for simulating regional impacts of climate change on the basis of IPCC scenarios, (2) a tailor-made evolutionary optimization algorithm for optimal irrigation scheduling with limited water supply, (3) a mechanistic model for simulating water transport and crop growth in a sound manner, and (4) a kernel density estimator for estimating stochastic productivity, profit, and demand functions by a nonparametric method. As a result of several simulation/optimization runs within the framework, we present stochastic crop-water production functions (SCWPF) for different crops which can be used as a basic tool for assessing the impact of climate variability on the risk for the potential yield for specific crops and specific agricultural areas. A case study for an agricultural area in the Al Batinah region of the Sultanate of Oman is used to illustrate these methodologies. In addition, microeconomic impacts of climate change and the vulnerability of the agro-ecological system are discussed.

An integrated approach to conceptualise hydrological and socio-economic interaction for supporting management decisions of coupled groundwater–agricultural systems

The management of complex interacting hydrosystems is challenging if in addition to the physical processes also socio-economic and environmental aspects have to be considered. This causes conflicts of interests among various water actors with mostly contradicting objectives and uncertainties about the consequences of potential management interventions. The objective of this paper is to present a methodological framework to support decision making under uncertainties for the management of complex hydrosystems. The proposed framework conceptualises hydrological and socio-economic interactions by constructing a Bayesian network (BN)-based decision support tool for a typical management problem of agricultural coastal regions. Thereby, the paper demonstrates the value of combining two different commonly used integrated modelling approaches. Coupled domain models are applied to simulate the nonlinearities and feedbacks of a strongly interacting groundwater–agriculture hydrosystem. Afterwards, a BN is used to integrate their results together with empirical knowledge and expert opinions regarding potential management interventions. A prototype application is performed for a coastal arid region, which is affected by saltwater intrusion into a coastal aquifer due to excessive groundwater extraction for irrigated agriculture. It addresses the issues of contradicting management objectives such as sustainable aquifer management vs. profitable agricultural production and the problem of finding appropriate management interventions or policies. Several policy combinations have been analysed regarding their efficiency within different management scenarios in a probabilistic way, which enables decision makers to assess the risks associated with implementing alternative management strategies. In addition, efficient metrics for evaluating performance and uncertainty of the developed task-specific BN are used which underline the reliability of the results.

Global optimization of deficit irrigation systems using evolutionary algorithms

Water is a limited resource and the dramatically increasing world population requires a significant increase in food production. For improving both crop yield and water use efficiency, the usual optimization strategy in furrow irrigation at the field level considers scheduling parameters, i.e. when and how much to irrigate, as well as control parameters, i.e. the intensity and the irrigation time, for each water application. Optimizing control and schedule parameters in irrigation is considered as a nested problem. The objective of the global optimization is to achieve maximum crop yield with a given, but limited water volume, which can be arbitrary distributed over the number of irrigations. It is difficult to solve the global optimization problem, because the target function has many locally optimal solutions and the number of optimization variables, i.e. the number of irrigations is unknown a-priori. For this reason, a made to measure evolutionary optimisation technique (EA) is employed to find a near-optimal solution of the global optimization problem within acceptable computational time. The results provided by the new optimization strategy are compared with the popular SCE-UA optimization algorithm and Mesh-Adaptive Direct Search (MADS). The comparison demonstrated a striking superiority of the new tool with respect to both the achieved irrigation efficiency and the required computational time.

Flash Flood Forecasting Combining Meteorological Ensemble Forecasts and Uncertainty of Initial Hydrological Conditions

Abstract Flood forecasting for fast responding catchments encounters problems, especially in terms of short warning periods and a very limited reliability. Within a new stochastic framework based on rigorous rainfall-runoff modelling and Monte Carlo simulations, we consider uncertainties of two sources: (i) uncertainty from the estimation of initial hydrological conditions, and (ii) the uncertainty of the meteorological rainfall forecast. We avoided the high computational demand of extensive Monte Carlo simulations by using a symbiosis between physically-based hydrological modelling and computationally highly efficient artificial intelligence techniques. The new PAI-OFF methodology (Process Modelling and Artificial Intelligence for Online Flood Forecasting; see Schmitz et al, 2005, and Cullmann, 2006) employs a physically-based hydrological/hydraulic model of the considered catchment for generating, in a first step, the complete range of realistic possible flood scenarios on the basis of a catchment specific meteorological analysis. The resulting database of corresponding input/output vectors – supplemented by generally available hydrological and meteorological data for characterising the catchment situation prior to a storm event – serves, in a second step, for setting up a set of task-specific artificial neural networks (ANN), which finally portray both the rainfall-runoff process and the hydrodynamic flood wave propagation in the river. We subsequently use this tool for investigating the global uncertainty of flash flood forecasting in a small-to medium-sized catchment on the basis of a comprehensive Monte Carlo analysis. Along these lines, the computationally highly efficient PAI-OFF technique allowed performing an extensive number of runs for obtaining ensembles of predicted stream flow that can be used to evaluate probabilities of exceedance of critical river stages/flows via an integration of both the hydrological uncertainty and the meteorological uncertainty. This approach was then implemented and applied to the Freiberger Mulde catchment in the Ore Mountains in Eastern Germany (with an area of about 3000 km2). The results of the overall ensemble predictions in the form of the ensemble mean values unveiled an astonishing underestimation of the recorded flood peak – most due to the bias of the considered initial hydrological conditions.

A conservative 2D subsurface flow model for curvilinear coordinate systems

Abstract A mass conservative finite difference scheme to describe 2D unsaturated/saturated subsurface flow is developed. The development takes into account important findings of Celia et al. (Water Resour. Res., 26(7) (1990) 1483–96) and extends their mass conservative 1D approach to two space dimensions. Moreover, using the theory of metric coefficients, the numerical model is not restricted to cartesian coordinates but is generally formulated for arbitrary orthogonal coordinate systems, thus accounting for curved boundaries of a computational domain. For flux boundary conditions convenient formulae avoid the cumbersome use of imaginary points outside the computational domain without reducing the order of magnitude of the approximation error. A flexible iteration strategy caters, on the one hand, for accuracy and stability when simulating sharp wet fronts and, on the other hand, for an economical calculation. The model was compared with the outcome of a laboratory experiment which investigated the transient development of the infiltration pattern in initially dry sand as a result of filling a circular cavity.

Design of Artificial Neural Networks for Flood Forecasting

This chapter provides an overview of currently available neural network design with the purpose of a timely warning for operational flood risk management, considering the need to evaluate the uncertainty of the forecast. Neural network models are very effective with regard to their computational requirements and provide new options for operational scenario analysis and ensemble forecasts. Here the widely used “multi layer feed forward network” is compared to an alternative, the “polynomial neural network”. A new training strategy permits to discriminate between input vectors. This method opens a way to reflect physical facts by means of input vectors in neural models, i.e. the neural model is portraying the rainfall runoff process on the basis of process understanding and physical boundary conditions of the considered catchment.

About the stationary inverse groundwater problem

The usual groundwater problem is to compute the stationary groundwater surface z = f(x, y) from a given recharge/drainage rate N(x, y). We treat the inverse problem (f N) in the two dimensional case.