Papers by ARIS GEORGAKAKOS
Joint Variable Spatial Downscaling (JVSD): A New Downscaling Method with Application to the Southeast US
AGU Fall Meeting Abstracts, Dec 1, 2011
ABSTRACT Joint Variable Spatial Downscaling (JVSD) is a new downscaling method developed to produ... more ABSTRACT Joint Variable Spatial Downscaling (JVSD) is a new downscaling method developed to produce high resolution gridded hydrological datasets suitable for regional watershed modeling and assessments. JVSD differs from other statistical downscaling methods in that multiple climatic variables are downscaled simultaneously to produce realistic and consistent climate fields. JVSD includes two major steps: bias correction and spatial downscaling. In the bias correction step, JVSD uses a differencing process to create stationary joint cumulative frequency statistics of the variables being downscaled. Bias correction is then based on quantile-to-quantile mapping of these stationary frequency distributions probability space. The functional relationship between these statistics and those of the historical observation period is subsequently used to remove GCM bias. The original variables are recovered through summation of bias corrected differenced sequences. In the spatial disaggregation step, JVSD uses a historical analogue approach, with historical analogues identified simultaneously for all atmospheric fields and over all areas of the basin under study. Analysis and comparisons with 20th Century Climate in Coupled Models (20C3M) data show that JVSD reproduces the sub-grid climatic features as well as their temporal/spatial variability in the historical periods. Comparisons are also performed for precipitation and temperature with the North American regional climate change assessment program (NARCCAP) and other statistical downscaling methods over the southeastern US. The results show that JVSD performs favorably. JVSD is applied for all A1B and A2 CMIP3 GCM scenarios in the Apalachicola-Chattahoochee-Flint River Basin (southeast US) with the following general findings: (i) Mean monthly temperature exhibits increasing trends over the ACF basin for all seasons and all A1B and A2 scenarios; Most significant are the A2 temperature increases in the 2050 - 2099 time periods; (ii) In the southern ACF watersheds, mean precipitation generally exhibits a mild decline in early spring and summer and increases in late winter; For the northern ACF watersheds, mean precipitation decreases in summer and increases mildly in winter (as in the south); (iii) In addition to mean trends, the precipitation distributions stretch on both ends with higher highs (floods) and lower lows (droughts). The downscaled temperature and precipitation scenarios are the basis of a comprehensive hydrologic and water resources assessment (reported elsewhere) assessing significant water, agricultural, energy, and environmental sector impacts and underscoring the need for mitigation and adaptation measures.
Improved Management of the Nile River Basin Through Modeling the Sudd, a Wetland with Vital Socioeconomic and Environmental Services
AGU Fall Meeting Abstracts, Dec 1, 2017
Stochastic Control of Hydropower Systems
Decision Support Systems, 1989
A two-level stochastic control method is presented for the real-time operation of hydropower proj... more A two-level stochastic control method is presented for the real-time operation of hydropower projects. The first level assumes fixed turbine discharges and utilizes probabilistic inflow forecasts in connection with the Extended Linear Quadratic Gaussian reservoir control method to determin optimal generation schedules. The purpose of the second level is to find the turbine discharges which either corresponds to best turbine efficiency or result in maximum power. The model distinguishes between peak and off-peak generation periods and attempts to generate maximum energy during the former and only when necessary during the latter. This stochastic control scheme is tested and implemented for the real time operation of a Georgia Power reservoir.
New reservoir control approach with application to the management of lake Lanier. Technical completion report
... A major complicating factor in water resources systems management is handling unknown inputs.... more ... A major complicating factor in water resources systems management is handling unknown inputs. ... This approach is applied to the operation of Lake Lanier, a US Army Corps of Engineers multiobjective reservoir and is shown to be an effective management tool. ...
Optimal real-time of hydropower systems
This article describes a climate change and hydrological impact assessment for several basins in ... more This article describes a climate change and hydrological impact assessment for several basins in Georgia. First, a new statistical technique, Joint Variable Spatial Downscaling (JVSD), is developed to produce high resolution gridded hydrological datasets for the Southeast US from 13 different Global Circulation Models (GCMs). A lumped conceptual watershed model (Georgakakos et al., 2010) is then employed to characterize the hydrologic responses under the historical climate and the future climate scenarios. The historical (baseline) assessment is based on climatic data for the period 1901 through 2009. It consists of running the hydrological models under historical climatic forcing (of precipitation and temperature) for the 109 year period from 1901 to 2009 (in monthly steps). The future assessment consists of running the Georgia watershed models under all A1B and A2 climate scenarios for the period from 2000 through 2099 (100 years) in monthly time steps. For the baseline scenarios and each of the 26 future climate scenarios (i.e., 13 A1B scenarios and 13 A2 scenarios), this study assesses the changes of both climate variables (i.e., precipitation and temperature) and hydrologic variables (i.e., soil moisture, evapotranspiration, and runoff) for each watershed. The results show that: (1) the 26 IPCC future climate scenarios (2000-2099) do not indicate any long term change in average precipitation; (2) the precipitation distribution is expected to "stretch" becoming wetter and drier than that of the historical climate; (3) temperature and potential evapotranspiration (PET) show consistently increasing historical and future trends; (4) soil moisture storage exhibits a declining trend historically and for future climates; and (5) watershed runoff, and thus river flow, exhibits a similar historical decline across all Georgia watersheds.
Control models for hydroelectric energy optimization
Water Resources Research, Oct 1, 1997
The optimization of hydroelectric energy is addressed via a new multilevel control model, which i... more The optimization of hydroelectric energy is addressed via a new multilevel control model, which is used to derive estimates of system firm energy with or without dependable capacity commitments. The model is able to optimize individual turbine operation as well as overall system operation on an hourly and daily basis. The mechanism by which the various models are linked and exchange information ensures full compatibility among the control levels and guarantees operational consistency across all timescales. The model is applied to the Lanier‐Allatoona‐Carters system, located in the southeastern United States, and is suitable for planning as well as operational applications.
Control Model for Hydroelectric Energy-Value Optimization
Journal of Water Resources Planning and Management, 1997
ABSTRACT
Extended linear quadratic Gaussian control: Further extensions
Water Resources Research, Feb 1, 1989
The extended linear quadratic Gaussian (ELQG) control method (Georgakakos and Marks, 1987) is a s... more The extended linear quadratic Gaussian (ELQG) control method (Georgakakos and Marks, 1987) is a stochastic control algorithm for the optimal operation of multiobjective reservoirs. Mathematically, this method optimizes a general functional of a stochastic system in state‐space form with upper and lower release constraints and probabilistic storage bounds. ELQG is a sequential algorithm which accounts for stochastic effects by preserving the first two statistical moments of the system's inputs and storages. In this paper, the method is first extended to handle nongaussian features which frequently characterize reservoir inputs. Second, ELQG's efficiency with respect to reliability storage constraints is discussed, and a new barrier function method is researched. These modifications are tested in case studies with the Savannah river system.
A new method for the real-time operation of reservoir systems
Water Resources Research, Jul 1, 1987
This paper introduces a new method for the real‐time operation of reservoir systems. The system i... more This paper introduces a new method for the real‐time operation of reservoir systems. The system is represented by a set of stochastic differential equations describing the reservoir and river dynamics in state space form. The formulated reservoir operation problem calls for finding policies which maximize the expected benefits of one system's objective while satisfying the remaining objectives at prespecified reliability levels. The solution is obtained by a new method named extended linear quadratic Gaussian (ELQG) controller. ELQG draws on and extends stochastic control theory results, and it is well suited for the optimization of constrained dynamical systems. It is a trajectory iteration algorithm theoretically expected to exhibit reliability and computational efficiency. The new method is tested and evaluated in a real‐world case study.
Ecological Applications, Dec 1, 2015
The western United States is a region long defined by water challenges. Climate change adds to th... more The western United States is a region long defined by water challenges. Climate change adds to those historical challenges, but does not, for the most part, introduce entirely new challenges; rather climate change is likely to stress water supplies and resources already in many cases stretched to, or beyond, natural limits. Projections are for continued and, likely, increased warming trends across the region, with a near certainty of continuing changes in seasonality of snowmelt and streamflows, and a strong potential for attendant increases in evaporative demands. Projections of future precipitation are less conclusive, although likely the northernmost West will see precipitation increases while the southernmost West sees declines. However, most of the region lies in a broad area where some climate models project precipitation increases while others project declines, so that only increases in precipitation uncertainties can be projected with any confidence. Changes in annual and seasonal hydrographs are likely to challenge water managers, users, and attempts to protect or restore environmental flows, even where annual volumes change little. Other impacts from climate change (e.g., floods and water-quality changes) are poorly understood and will likely be location dependent. In this context, four iconic river basins offer glimpses into specific challenges that climate change may bring to the West. The Colorado River is a system in which overuse and growing demands are projected to be even more challenging than climate-change-induced flow reductions. The Rio Grande offers the best example of how climate-change-induced flow declines might sink a major system into permanent drought. The Klamath is currently projected to face the more benign precipitation future, but fisheries and irrigation management may face dire straits due to warming air temperatures, rising irrigation demands, and warming waters in a basin already hobbled by tensions between endangered fisheries and agricultural demands. Finally, California's Bay-Delta system is a remarkably localized and severe weakness at the heart of the region's trillion-dollar economy. It is threatened by the full range of potential climatechange impacts expected across the West, along with major vulnerabilities to increased flooding and rising sea levels.
Water Resources Research, Nov 1, 1993
Reservoir operation decisions require constant reevaluation in the face of conflicting objectives... more Reservoir operation decisions require constant reevaluation in the face of conflicting objectives, varying hydrologic conditions, and frequent operational policy changes. Optimality is a relative concept very much dependent on the circumstances under which a decision is made. More than anything else, reservoir management authorities need the means to assess the impacts of various operational options. It is their responsibility to define what is desirable after a thorough evaluation of the existing circumstances. This article presents a model designed to generate operational trade-offs common among reservoir systems. The model avoids an all-encompassing problem formulation and distinguishes three operational modes (levels) corresponding to normal, drought, and flood operations. Each level addresses only relevant system elements and uses a static and a dynamic control module to optimize turbine performance within each planning period and temporally. The model is used for planning the operation of the Savannah River System. 1.
Journal of Hydrometeorology, Apr 1, 2015
Agricultural models, such as the Decision Support System for Agrotechnology Transfer cropping sys... more Agricultural models, such as the Decision Support System for Agrotechnology Transfer cropping system model (DSSAT-CSM), have been developed for predicting crop yield at field and regional scales and to provide useful information for water resources management. A potentially valuable input to agricultural models is soil moisture. Presently, no observations of soil moisture exist covering the entire United States at adequate time (daily) and space (;10 km or less) resolutions desired for crop yield assessments. Data products from NASA's upcoming Soil Moisture Active Passive (SMAP) mission will fill the gap. The objective of this study is to demonstrate the usefulness of the SMAP soil moisture data in modeling and forecasting crop yields and irrigation amount. A simple, efficient data assimilation algorithm is presented in which the agricultural crop model DSSAT-CSM is constrained to produce modeled crop yield and irrigation amounts that are consistent with SMAP-type data. Numerical experiments demonstrate that incorporating the SMAP data into the agricultural model provides an added benefit of reducing the uncertainty of modeled crop yields when the weather input data to the crop model are subject to large uncertainty.
Water Resources Research, Apr 1, 1990
In this paper a systematic methodology for making real-time irrigation decisions is presented. A ... more In this paper a systematic methodology for making real-time irrigation decisions is presented. A physically based representation of the dynamics of the soil-crop-atmosphere system is used. The variables characterizing the crop and soil status are concurrently simulated with an integrated state space model. Soil moisture and salinity conditions, which synergistically control the plant water uptake, are obtained by using lumped parameter mass balance models for the root zone. Crop yield is predicted by explicitly modeling the plant growth processes, such as assimilation, respiration, and transpiration, which are driven by the climatic inputs. The control model is an analytical optimization method for multistage multidimensional sequential decision-making problems. It is suitable for systems with nonlinear dynamics and objective functions. The method is based on local iterative approximations of the nonlinear problem with a linear quadratic problem. This approach is evaluated in a series of case studies, where optimal irrigation schedules are obtained on an hourly basis over the growing season. 1. INTRODUCTION Agricultural activities are important for the economies of most countries. The demand for agricultural products increases continuously as a result of growing populations, higher incomes, and new uses of traditional products. It is estimated that production will increase by expansion of arable land and, mainly, by intensified use and better management of the production factors. There is therefore an immediate need for efficient use of resources, in particular of water, in the production process. Modem agriculture is a specialized and highly mechanized industry, in which solar energy is transformed to useful organic products. During the growing season the farmer makes important operational decisions which affect the final yield. Competitive and efficient agriculture requires such decisions to be made optimally. In the short run, maximization of a performance index is sought by exhaustive allocation of the limiting production factors, that is, water, land, and nutrients among different crops. The index is usually net benefits, although in some cases, when water is scarce, the objective may shift to maximizing produced crop weight per unit of water used. In the long run, objectives are related to societal issues such as welfare, malnutrition, deforestration and climatic change, environmental pollution, income redistribution and employment, soil conservation, and energy consumption. Even if optimal use of resources can lead to short-term gain in crop yield, the long-term effects can be immeasurable. Yet, in current practice the farmers act with the short-term objective of maximizing the benefits of the production process by prudent use of the available resources. This is also the philosophy of our study, which aims in introducing modern decision-making methodologies for water use in agriculture. •Now at Metcalf and Eddy, Incorporated, Wakefield, Massachusetts.
Chlorophyll a estimation in lakes using multi-parameter sonde data
Water Research, Oct 1, 2021
Algae blooms are of considerable concern in freshwater lakes and reservoirs worldwide. In-situ Ch... more Algae blooms are of considerable concern in freshwater lakes and reservoirs worldwide. In-situ Chlorophyll a (Chl-a) fluorometers are widely used for rapid assessments of algae biomass. However, accurately converting Chl-a fluorescence to an equivalent concentration is challenging due to natural variations in the relationship as well as nonphotochemical quenching (NPQ) which occurs commonly in surface waters during daytime. This study is based on water quality data from a freshwater lake from October 2018 to December 2020. Initial analysis of sonde Chl-a fluorescence and laboratory extracted Chl-a concentrations shows that the two data sets exhibit a nonlinear relationship with positive correlation and significant errors. A bias correction method was next developed based on (1) concurrent sonde measurements of other water quality parameters (to account for nonlinearities) and (2) a bias correction approach for nonphotochemical quenching effects in surface waters. The new Chl-a model exhibits much improved accuracy, with a root mean square error (RMSE) less than 0.95 µg/L. The new method facilitates accurate Chl-a characterization in freshwater lakes and reservoirs based on readily obtainable in-situ fluorescence sonde measurements.
Water Resources Research, Aug 1, 1991
In groundwater management, uncertainty mainly stems from imprecise parameters and boundary condit... more In groundwater management, uncertainty mainly stems from imprecise parameters and boundary conditions. This paper first formulates a stochastic groundwater management problem and subsequently proposes an appropriate solution approach. The equations of flow are converted to a dynamical state-space system using finite element and difference techniques. Parameter and boundary condition uncertainty is incorporated using the small perturbation method. Management objectives are expressed as a composite performance index which may be used to minimize pumping costs, maintain hydraulic heads and pumping rates in the vicinity of target sequences, or optimally compromise among various system goals. This problem is solved via a numerical optimal control method which exhibits good computational properties. The approach is applied to the management of a two-layer aquifer system with various boundary conditions and uncertainty levels and sources. The results provide useful insights of the system response under uncertainty and quantify the trade-offs between accomplishing average system goals and minimizing uncertainty. 1. INTRODUCTION Groundwater is a vital water resource in many regions of the world. However, continued use of groundwater is threatened by overextraction from increasing water demands or ½0a, tamination from industrial and agricultural wastes. Failare to implement prompt and prudent management measures is bound to increase future restorative costs and threaten the availahlity and usability of this resource. These ramifica-'tins have sustained interest in groundwater management models [Gorelick, 1983; Feinerman et al. , 1985; Willis and Finhey, 1985; Massmann and Freeze, 1987a, b; Makinde-Odusola and Marino, 1989]. Such models use operations research methods with groundwater flow and transport models to assess the relative impacts of various groundwater -'nmagement options. Although the majority of these models are deterministic, the markedly regionalized aquifer properties suggest the need for stochastic groundwater management models. Maddock [1973] analyzed an explicitly stochastic farm irrigation management scheme. His analysis included both economic parameters such as crop prices, pumping costs, interest rate, and consumptive use, and hydrologic parameters such as aquifer transmissivity and storativity; the objective was to maximize discounted net revenues. The economic and hydrologic parameters also were treated as raMore variables. Groundwater flow was modeled as a set of technological (or unit response) functions [Maddock, 1972], with well pumping and drawdown being part of the decision variables. This was a quadratic programming problem because pumping cost depended on both lift and pumping rate. Maddock assessed the relative influence of each parameter based on the regret function which represents net losses in discounted net revenue when other than the actual parameters were used in the management model. His conclusion was that the expected regret is more sensitive to economic parameters, such as crop prices and pumping costs, than to the aquifer transmissivity and storativity. Feinerman et al. [I9851 focused on the effects of spatial variability on irriga-Copyright 1991 by the American Geophysical Union. l•aper number 91WR00763. 0043.1397/91/91 WR-00763 $05.00 tion management. They viewed crop yield as a function of irrigation water quantity and a two-dimensional, normally distributed, random soil function. The objective was to find the optimal irrigation level which maximizes expected profits [(unit yield price) x (spatially averaged yield) -(unit irrigation cost) x (irrigation quantity)] under a deterministic, a risk neutral, and a risk averse decision-making scenario. The approach taken was to express the uncertainty of the yield function in terms of the first and second statistical moments of the soil function and explicitly maximize the profit or expected profit functions. Feinerman et al. concluded that recognizing the spatial heterogeneity of soil and crop yield may substantially reduce irrigation water applications. The reduction percentage depends on the decision maker's risk attitude and the spatial correlation scale of the soil function. Tung [1986] also considered the random effects of transmissivity and storativity in a stochastic groundwater management model. His objective was to maximize pumping under transient conditions from a confined, homogeneous aquifer subject to probabilistic drawdown constraints. Unit response functions were used to quantify aquifer response and relate drawdown to parameter randomness. This nonlinear management problem was solved iteratively, using quasilinearization and linear programming. Sensitivity analysis at many parameter uncertainty levels indicated that optimal pumping rates were sensitive to aquifer transmissivity and virtually independent of the storativity. Wagner and Gorelick [1987] formulated and solved a stochastic groundwater quality management problem whose objective was to minimize pumping while satisfying probabilistic water quality standards. Uncertainty was due to effective porosity, hydraulic conductivity, and longitudinal and transverse dispersivities and was quantified using regression analysis and a first order-second moment system approximation. The iterative solution of this nonlinear problem is discussed in an earlier paper [Gorelick et al., 1984]. The case studies demonstrated that significant pumping or recharge errors (as high as 20%) may result if stochastic effects are not explicitly considered. Wagner and Gorelick [1989] presented a management approach which accounts for spatial parameter variability and combines Monte Carlo simulation with optimization. Their idea was to generate many possible conduc-2077
Water Research, Dec 1, 1996
The Linearized Maximum Likelihood (LML) method for the simultaneous estimation of activated sludg... more The Linearized Maximum Likelihood (LML) method for the simultaneous estimation of activated sludge states and parameters from noisy process measurements (Kabouris and Georgakakos, 1996a, Wat. Res., 30, 2853-2865) is simplified, in terms of its memory storage and computational requirements, for efficient on-line implementation. This is achieved by processing only the four most recent sets of 5-rain on-line measurements at each estimation instance, along with the utilization of simplified estimation equations for tracking state and parameter variations, following the initial convergence period. The algorithm is tested in a computational case study involving a nitrifying activated sludge process, modelled by the IAWQ Activated Sludge Model 1 and incorporating a dynamic settling and clarification model. The on-line LML algorithm is capable of tracking the process states and parameters under dynamic conditions of process inputs and model parameters.
Stochastic Control of the Activated Sludge Process
Water Science and Technology, Sep 1, 1991
A stochastic optimal control method is introduced and applied to the real-time management of an a... more A stochastic optimal control method is introduced and applied to the real-time management of an activated sludge process. The method includes a detailed dynamic process model and accounts for input uncertainty. The objective is to minimize the expected variations of effluent organic concentrations and is met by manipulating process flowrates. The method is applied to a hypothetical case study.
Uploads
Papers by ARIS GEORGAKAKOS