25.1 Introduction
25.2 Decision Support Systems in Relation to Groundwater
25.2.1 Aquifer Performance
State conditions | Inflows |
Storage
| Outflows | Model considerationsf |
---|---|---|---|---|
Aquifer typeb (m) | Natural recharge spatial component [Rn(x,y)] | Specific storage (Ss) | Natural discharge spatial component [Qn(i,j)] | Planning Horizon (g) |
Boundary conditionsc | Natural recharge rate [Rn(t)] | Saturated thickness (b) | Natural discharge rate [Qn(t)] | Stress Period (p) |
Areal extent of aquiferd (A) | Artificial recharge spatial component [Ra(x,y)] |
Storage
(ST) | Pumping well spatial component [Qa(i,j)] | Time Step (t) |
Porosity (ø) | Artificial recharge rate [Ra(t)] | Specific yield (фeff or Sy) | Pumping well discharge rate [Qa(t)] | Cell (i,j,k,z) |
Hydraulic conductivity (K) | Return flow (α)e | Storativity [−] | Evapotranspiration [Qe(t)] | Zone (z) |
Land Surface Elevation (mij) | Lateral or vertical influx (V)e | Hydraulic head [h(x,y,z)] | Lateral or vertical outflux (V)e | Bottom confining unit elevation (nij) |
Drain elevation (d) | Unrecoverable Storage (Su) | Minable Storage (Sm) | Replenishable Storage (Sr) | Diffusivityg (T/S) |
Transmissivity (T) | Acceptable variance (X) |
25.2.2 GroundwaterGovernance
25.2.3 Decision Support Systems and Processes
25.3 Data and Modeled Attributes for Aquifer Performance
25.3.1 Natural HydrogeologicAttributes and Uncertainty
25.4 Addressing Stakeholder Perspectives for GroundwaterGovernance
25.5 Decision Support Systems: Background and Types
25.5.1 The Emergence of Decision Support
25.5.1.1 Passive
Source | Problem | Scale | GW Simulation | Optimization & larger DSS | Objective function | Decision variables |
---|---|---|---|---|---|---|
Fienen et al. (2013) | Forecasting changes in sea level rise and groundwater | SEAWAT model aggregated to Bayesian network | Bayesian network to emulate groundwater response/uncertainty | Propagate uncertainty efficiently for use in forecasts for decision makers | Focus on model performance and calibration; decision model not developed | |
Molina et al. (2013a) | Evaluating climate change impacts over time | Regional: Serral-Salinas aquifer, Spain | Used Post-process to evaluate groundwater response; MODFLOW model | Scenarios tested with an Object Oriented Bayesian Network (OOBN) | Comparative analysis across scenarios and time windows; Extensive list of performance measures based on : Agricultural net profits and aquifer storage; Maximizing Total income, employment rates | Intervention actions, such as water rights purchase, land sale, sale of water for irrigation, |
Hadded et al. (2013) | Water management generally | Local to Regional: Zeuss Koutine aquifer, Tunisia | MODFLOW | WEAP-MODFLOW link | Demand satisfaction, cost and drawdown minimization | Limited by salinity levels and flow capacity |
Molina et al. (2013b) | Regional: El Salobra aquifer, Spain | Lumped parameter representation of the aquifer within a linked hydro-economic model | Object Oriented Bayesian Network (OOBN) for stochastic modeling | Assess groundwater quality control with uncertainty; Minimize nitrate concentration and recovery times | Fertilizer quotas Fertilizer prices | |
Le Page et al. (2012) | Water allocation | Regional: Haouz-Mejjate plain, Morroco | MODFLOW | WEAP-MODFLOW link | Evaluate impacts to regions and identify mitigation options | Principally used to validate modelled aquifer response and sensitivities to parameter change |
Moura et al. (2011) | Assess groundwater quality control with uncertainty | Local to Regional: farm and aquifer for case studies; Upper Guadiana Basin; Altiplano, Spain | Lumped parameter representation of the aquifer within a linked hydro-economic model | General Algebraic Modeling System (GAMS) and Object Oriented Bayesian Network (OOBN) for stochastic modeling | Maximize gross margin at the farm level as a function of crop prices and yields; the OOBN added response levels in groundwater | Crop surface Irrigation method Soil type |
Triana et al. (2010) | Evaluate feasibility and performance of water mangemetn strategies | Regional; Lower Arkansas River Basin | Canal seepage and infiltration to groundwater estimated from a MODFLOW/MT3DMS simulation | Based on River GeoDSS with an Artificial Neural Network (ANN) used to distribute recharge to groundwater | Comparative analysis of estimated performance with a prioritization structure based on performance with Total Storage Water shortages Compliance with legal compact Impacts to water quality | Water strategy choices include: Total water diverted Use of storage Weighting of priorities (shown in Objective column) |
Van Cauwenbergh et al. (2008) | Ranking alternative water management options with multi-criteria | Local aquifer to regional watershed scale | Mike-SHE Lumped cell structure | Not clearly described; a simplified water transfer model with limited cells | Minimize pumping costs, recharge, and water transport | Not clearly stated, penalty functions are included in the formulation |
Quantifying Sustainable Yield | Local to Regional: Central Texas, Barton Springs aquifer | MODFLOW or an aggregated Systems Dynamics Model of the same system | Link to TABU global search algorithm and systems dynamics model of ancillary systems | Six Objectives defined with stakeholders Max water allocation and location of pumping; two formulations for maximizing minimum spring flow; saturated thickness; total storage | Pumping (location and rate) drought policy levels for alarm and critical stages Impervious cover and land use | |
Carrera-Hernandez and Gaskin (2006) | Spatially explicit groundwater modeling | Any | MODFLOW | Link to GRASS for geospatial groundwater modeling | Pure simulation capabilities | Not Applicable |
Letcher (2005) | Water allocation for a watershed basin | Regional to Large: | Network-nodes linked with surface water sites | WaDSS based on ICMS Applied to Namoi & Gwydir River Basins, Australia | Max water allocation | Not clearly stated, but variable options |
Recio et al. (2005)a | Link hydrogeologic model with econometric for agricultural decisions | Regional: Eastern Mancha aquifer, Spain | MODFLOW, possibly 3-D (not clear) steady state | GESMO | Land allocation for crops; Crop yield maximization | Pumping Head levels Electricity costs |
Mysiak et al. (2005) | Water resource management (general) | Local to regional | Not specified | MULINO | Multi-criteria weighting applications | Varies |
Lanini et al. (2004) | Participatory Integrated model for basin study | Local to regional: Herault Middle Valley, France | Lumped parameter model of socio-hydrosystem | (no optimization) Matlab/Simulink | No clear description; Stock and flow/steady state system | Head – drawdown Pumping natural discharge |
Quintana et al. (2005)b |
Groundwater
governance
| Local to regional: Herault Middle Valley, France | Not clear, but indicates that a groundwater module included | GOUVERNe or TIDDD (Tool to Inform Debates, Dialogues & Deliberations) | Exploratory decision support with stakeholder participants | Not clearly defined |
Fredrick et al. (2004) | Contaminant susceptibility | Local: single aquifer, NY | 2-D Steady state AEM | (no optimization) Spatial indexing Drastic method | Minimize pollution potential | Water table levels Drastic scores |
Aziz et al. (2003) | Optimization link for groundwater monitoring plans | Local: contaminant plume various sites | Linear regression for plumes, empirical data, and simplified models | MAROS | Minimize the number of sampling sites and frequency | Monitoring location and time |
Fatta et al. (2002) | Landfill leachate impact | Local: Ano Liosia landfill, Greece | MODFLOW/MT3D | ECOSIM : Pilot version / local client–server architecture | Linked simulation models, GIS | No decision problem results reported |
Nalbantis et al. (2002)c |
Conjunctive use
management
| Regional: Athens, Greece | MODFLOW Multi-cell and Lumped parameter models | HYDRONOMEAS: Multi-reservoir system management | Stochastic optimization (limited solution algorithm description) | Pumping |
Oxley et al. (2002) | Land degradation in the Mediterranean | Regional: Argolida, Greece Marina Baixa, Spain | MODFLOW | MODULUS DSS: 9 sub-models for integrated assessment modeling | Solution algorithm and specific objectives not defined: General problem environmental problem scopes | Mentions as possible: Crop choice subsidy change water management and others |
Naveh and Shamir (2000) | Groundwater level management | Local: Hula Lake, Israel | MODFLOW with GMS | Spreadsheet model | Microsoft Excel solver optimization add-ins | head levels canal flow rates |
Demetriou and Punthakey (1999) |
Sustainable
groundwater
management
| Regional: Wakool, Murray Darling Basin Australia | MIKE SHE, 3-D flow | MIKE SHE No optimization | Scenario modeling | mainly crop and vegetation related defined with historic data for scenarios |
Sophocleous and Ma (1998) | Saltwater intrusion (estimate parameters) | Local: Great Bend Prairie aquifer | 3-D density dependent flow/solute transport (SWIFT II) | Linear regression (forward, backward, stepwise) | Minimize saline intrusion | Hydraulic conductivities pumping rate Distance to – saline interface Layer thickness |
McKinney et al. (1997) | GIS-based DSS for River Basin Management (prototype level) | Local to regional: Hypothetical | No groundwater component described | GAMs (General Algebraic Modeling system) | Maximize supply; downstream flow; Minimize salt concentrations; power; import sources | Not clearly stated |
Latinopoulos et al. (1996) | Engineering supply & remediation | Small : Hypothetical | 2-D Method of Characteristics (1 year) | Monte Carlo; Stochastic programming | sum of Total costs + risk | Broken into costs, failure risks, tolerance |
Andreu et al. (1996) | River basin planning & operational management | Local and Regional: | Eigen value aquifer response flow module – Segura & Tagus basins, Spain | AQUATOOL | Not clearly stated | Not clearly stated |
Datta and Peralta (1986) | Alternative selection (Surrogate Worth Tradeoff) | Regional: Grand Prairie, AR | 2-D Steady state Flow | Dynamic Multi-objective optimization (Quadratic & Linear) | Min Cost of Water And Max total supply | Pump location & volume Head drawdown Vol. surface water diverted |
25.5.1.2 Active
25.5.1.3 Proactive
25.5.2 Applications of Decision Support to Groundwater Cases
25.6 Factors Related to Adoption of DSS
-
Financial costs – because implementing a DSS system limits groundwater management districts frequently requires software licenses and staff or consultant time.
-
Knowledge to implement – use of a DSS system requires the technical capacity to operate and use advanced software products.
-
Adaptability of DSS – every decision situation has contextual elements and situation-specific considerations. DSS systems must be easy to adapt to each case before use.
-
Multi-disciplinary team – the range of knowledge and expertise necessary to represent a groundwater problem can be very broad and requires expertise across domains.
-
Adequate governance structures – without appropriate authority to manage the resources or infrastructure to support a DSS long-term the likelihood of adoption and use drops
-
Trust – DSS deployment depends on trust among collaborators.