Book Description

Surface water and groundwater systems consist of interconnected reservoirs, rivers, and confined and unconfined aquifers. The integrated management of such resources faces several challenges: High dimensionality refers to the requirement of the large number of variables that need to be considered in the description of surface water and groundwater systems. As the number of these variables increases, the computational requirements quickly saturate the capabilities of the existing management methods. Uncertainty relates to the imprecise nature of many system inputs and parameters, including reservoir and tributary inflows, precipitation, evaporation, aquifer parameters (e.g., hydraulic conductivity and storage coefficient), and various boundary and initial conditions. Uncertainty complicates very significantly the development and application of efficient management models. Nonlinearity is intrinsic to some physical processes and also enters through various facility and operational constraints on reservoir storages, releases, and aquifer drawdown and pumping. Nonlinearities compound the previous difficulties. Multiple objectives pertain to the process of optimizing the use of the integrated surface and groundwater resources to meet various water demands, generate sufficient energy, maintain adequate instream flows, and protect the environment and the ecosystems. Multi-objective decision models and processes continue to challenge professional practice. This research draws on several disciplines including groundwater flow modeling, hydrology and water resources systems, uncertainty analysis, estimation theory, stochastic optimization of dynamical systems, and policy assessment. A summary of the research contributions made in this work follows: 1. High dimensionality issues related to groundwater aquifers system have been mitigated by the use of transfer functions and their representation by state space approximations. 2. Aquifer response under uncertainty of inputs and aquifer parameters is addressed by a new statistical procedure that is applicable to regions of relatively few measurements and incorporates management reliability considerations. 3. The conjunctive management problem is formulated in a generally applicable way, taking into consideration all relevant uncertainties and system objectives. This problem is solved via an efficient stochastic optimization method that overcomes dimensionality limitations. 4. The methods developed in this Thesis are applied to the Jordanian water resources system, demonstrating their value for operational planning and management.