other theses

To view abstract, click on author

WJ home

Any questions or comments?

©1997 W James

Last modified 07/16/97

Wall, R.C. 1968. Ph.D. An Analysis of the Current Status of The Schistosome Dermatitis Problem in Michigan.

Burges, S.J. 1970. Ph.D. Use of Stochastic Hydrology to Determine Storage Requirements of Reservoirs -- A Critical Analysis.

Smith, R.L. 1972. M. Eng. "The Design of Compound Critical Sections for Open Channel Flow Measurement".

El Hyashab, A.M.M. 1975. M.Sc. "Hydraulics of Flow Over Side Weirs".

Froise, S. 1975. Ph.D. Least Cost Control Strategies in Urban Drainage Design: A Dynamic Programming Approach.

Gupta, P.K. 1977. Ph.D. "Dynamic Optimization Applied To Systems With Periodic Disturbances".

Munro, D.M. 1978. "Stormwater Management Modelling: An Application of SWMM to the Carling Street Catchment, London, Ontartio".

Tucci, C.E.M. 1978. Ph.D. Hydraulic and Water Quality Model for a River Network".

Ammon, D.C. 1979. M.Eng. "Urban Stormwater Pollutant Buildup and Washoff Relationships"

MacRae, C.R. 1979. M.Sc. "Hydrologic And Water Quality Modelling in a Small Urban Catchment in London, Ontario".

Thompson, L.R.1979. M.A.Sc. "A Comprehensive Subcatchment Hydrologic Simulation Model for Urban and Rural Applications".

Williams, K.K.1979. Ph.D. Oklahoma City Urban Storm Runoff Quantity: Comparison and Calibration of Predictive Methods.

Book, D.E.1980. M.Sc. "A Dynamic Hydraulic Model for Simulation and Evaluation of Complex Drainage Systems".

Donald, S.1980. Ph.D. Uncertainty Analysis of a Lake Eutrophication Model.

Jewell, T.K.1980. Ph.D. "Urban Stormwater Pollutant Loadings".

Kelton IV, W.D.1980. Ph.D. The Startup Problem in Discrete-Event Simulation.

Maunder, D.E.1980. M.Eng. "Modelling of Regional Sanitary Sewerage Networks for Minimum Cost".

Uchrin, C.G. 1980. Ph.D. "Mathematical Modeling of Suspended Solids and Associated Pollutant Transport".

Iman, E.H.H.1981. M.A. Sc. "Numerical Modelling of Rectangular Clarifiers".

Lee, K.-B.1981. M.Sc. M.Eng. "Development of a Versatile Storm Water Quantity Model".

Watson, M.D.1981. M.Sc. "Appication of ILLUDAS to Stormwater Drainage Design in South Africa".

Burn, D.H.1982. M.A.Sc. "Optimization Techniques Applied to Water Quality in an Uncertain Environment".

Cheung, P.W.K. 1982. M.A. Sc. "A Standard Hydrograph Method for the Preliminary Analysis of Stormwater Management Projects".

Hamam, M.A.M. 1982. B.Sc. (Honour), M.A. Sc. "Transition of Gravity to Surcharged Flow in Sewers".

Kassem, A.M.A. 1982. Ph.D. "Development and Application of A Simultaneous Routing Model For Dual Drainage Systems".

Nix, S.J.1982. Ph.D. Analysis of Storage/Release Systems in Urban Stormwater Quality Management.

P'NG, C.E. 1982. M.A. Sc. "Conceptual Hydrologic Modelling for Master Plans in Urban Drainage".

Southerland, E. 1982. Ph.D. A Continuous Simulation Modeling Approach to Nonpoint Pollution Management.

Hamilton, D.L. 1983. M.Sc. "Microcomputer Simulation of Canal Operation ".

Goodrich, J.A. 1983. Ph.D. Predicting Toxic Waste Concentrations in Community Drinking Water Supplies From Upstream Industrial Discharges: A Vulnerability Analysis.

Lee, C.J. 1983. Ph.D. A Method for Estimating Human Risk Due to Toxicants.

Patry, G.G. 1983. Ph.D. Real-Time Forecasting of Water Quantity and Quality in Urban Catchments.

Qaisi, K.M. 1985. Ph.D. Uncertainty Analysis of Stream Dissolved Oxygen Models.

Gallagher, D.L. 1986. Ph.D. The Application of Data Based Transformations to Parameter Estimation in Water Quality Models.

Sircar, J.K. 1986. Ph.D. Computer Aided Watershed Segmentation for Spatially Distributed Hydrologic Modeling.

Wilson, L.C. 1986. Ph.D. A Continuous Spaciotemporal Stochastic Model for Short-Time-Increment Precipitation.

de Roulhac, D.G. 1987. M.S. Application of Computer Graphics in the Selection of Rainfall Frequency Models for Environmental Engineering.

Dilks, D.W. 1987. Ph.D. Analysis of Water Quality Model Uncertainty Using a Bayesian Monte Carlo Method.

Johansen, N.B. 1987. Ph.D. Discharge to Receiving Waters From Sewer Systems During Rain".

Lansey, K.E. 1987. Ph. D. "Optimal Design of Large Scale Water Distribution Systems Under Multiple Loading Conditions".

Melching, C.S. 1987. Ph.D. A Reliability Analysis on Flood Event Forecasting With Uncertainties.

Paulfalcone, J. 1987. M.Eng., P.Eng. "MIDUSS: The Development of an Interactive Model for the Design of Stormwater Drainage Systems ".

Pitt, R.E. 1987. Ph. D. "Small Storm Urban Flow and Particulate Washoff Contributions to outfall Dishcarges (Ontario, Canada, Milwaukee)".

Tarboton, D.G. 1987. M.Sc. "Hydrological Sampling: A Characterization in Terms of Rainfall and Basin Properties".

Yim, C.S. 1987. Ph. D. "Development and Testing of Granular Filter Design Criteria for Stormwater Management Infiltration Structures (SWMIS)".

Baffaut, C. 1988. Ph.D. "Knowledge Techniques for teh Analysis of Urban Runoff Using SWMM (Storm Water Management Model)".

Duan, N. 1988. Ph.D. "Optimal Reliability Based Design and Analysis of Pumping Systems for water Distribution Systems".

Holub, L.G. 1988. Ph.D. "Uncertainty Modeling of Water Distribution Systems".

Rivera Santos, J. 1988. Ph.D. A Methodology for Parameter Estimation in Conceptual Precipitation-runoff Models with Emphasis on Error Analysis.

Corsi, R.L. 1989. M.S., Ph.D. "Volatile Organic Compound Emissions from Wastewater Collection Systems".

Holbert, S.B. 1989. M.L.Arch. "Development of a Geographic Information System Based Hydrologic Model for Stormwater Management and Landuse Planning".

Hotchkiss, R.H. 1989. Ph.D. Reservoir Sedimentation and Sediment Sluicing: Experimental and Numerical Analysis.

Moys, G.D. 1989. Ph.D. "The Prediction of Rainfall Excess in Urban Catchments".

Thompson, D.B. 1989. Ph.D. Determining Parameters for a Continuous Simulation Model by Estimation, Synthetic Calibration and Analytic Calibration.

Zheng, S. 1989. Ph.D. Stochastic Modeling of Urban Drainage Systems".

Abbott, H.A. 1990.Ph.D. "Permeability and Suction Characteristics of Compacted Unbound Aggregates (Pavement Drainage)".

Au-yeung, W.K. 1990. Ph.D. "Flow Routing in the Parallel Pipes of a Combined Sewer System".

Brion, L.M. 1990. Ph. D. "Methodology for Optimal Operation of Pumping Stations in Water Distribution Systems".

Buck, D.R. 1990. M.S. "Evaluation of a Percolation Basin to Remove Heavy Metals and Particulate from Highway Runoff.

Lin, L.1990. Ph. D. "Transport of Heavy Metals Through Acumulated Sediments in Detention/Retention Ponds (Sediments)".

Xu, C. 1990. Ph.D. "Optimal Operation and Reliability Evaluation of Water Distribution Systems".

Yan, J. 1990. Ph.D. Parameter Estimation for Multipurpose Hydrologic Models.

Gonzalez, G.S. P. 1991. Ph.D. "Bedload Transport in Storm Sewers Stream Traction in Pipe Channels".

Graham, E.I. 1991. M.A.Sc. "An Urban Runoff Infiltration Basin Model".

Li, J.Y. 1991. Ph.D. Comprehensive Urban Runoff Control Planning

Rezek, S.E. 1991. MS. "The Effects of Urbanization on a Portion of the Portage Creek Watershed: A Case Study on Surface Water and Groundwater Quality".

Seo, D.-I. 1991. Ph.D. Optimal Complexity Analysis of Total Phosphorus Models.

Young, B.C. 1991. M.Sc. "Dynamic Modelling of Combined Sewer Systems with Radial Gates and Siphon-Weirs".

Michaud, J.D. 1992. Ph.D. "Distributed Rainfall-Runoff Modeling of Thunderstorm-Generated Floods: A Case Study in a Mid-Sized, Semi-Arid Watershed in Arizona".

Paturel, J.-El. 1992. Do. "Study of Rain Phenomena at Small Scale: contribution to Risk Analysis in Urban Hydrology".

Xu, Patterson Ping. 1992. " A Planning Level Model For Calculation of Toxic Pollutant loadings in Urban Runoff ".

Culp, Dwayne Edward. 1993. Urban Stormwater Treatment Pond Volume Optimization .

Spangberg, A. 1994. "Stormwater Quality Measured by Turbidity".

Bishop, Brian Edward. 1995. " HSP-F Simulation of a Constructed Wetland Stormwater Best Management Practice for Urban Highway Runoff ".

Dent, M. C. 1996. Individual and Organisational Behavioural Issues Relating to Water Resources Simulation Modelling and its role in Integrated Catchment Management in Southern Africa

Eskandari, Abdollah. 1997 Ph.D. Decision Support System in Watershed Management Under Uncertainty

ASCE Task Committee on Sustainability Criteria, Water Resources Planning and Management Division. 1997
Sustainability Criteria for Water Resource Systems

A planning-level model PMSQ was developed for calculation of hourly and event toxic pollutant loadings in urban runoff. The model allows the user to specify the pollutant concentrations or use the stormwater runoff quality data base stored in the model. From the model menus, concentrations in the data base can be selected and processed according to the associated locations, land use types and pollutant names. Three methods were used for estimating the mean concentrations from censored quality data using the lognormal distribution: (1) extrapolating the regression equation; (2) replacing the data below the detection limit by half of the detection limit; and (3) using the maximum likelihood estimation method. Runoff quantities are calculated by the STORM model using the catchment characteristics and the AES precipitation data. A case study of selected toxic pollutant loadings in urban runoff in the City of Sarnia is given to illustrate the applications of the model.

This project describes the use of the HSP-F program methodology to model and design Stormwater Best Management Practices (SWMPs) for a proposed freeway interchange in Hamilton, Ontario within the Red Hill Creek watershed.

The objective of modelling the freeway interchange is to develop a system of stormwater management controls that will adequately address fisheries and water quality criteria for this freeway interchange. Since the wetlands' pollutant removal capabilities are directly linked to the biological interactions and aquatic vegetation in the wetlands, control of the operational hydraulics of the wetlands are critical. This project focuses on the use of HSP-F, in continuous simulation mode, to assess the fluctuations in water levels within the proposed wetland. It has been shown how the information developed through this analysis is used to design the wetland bathymetry and aquatic planting species distribution.

In order to satisfy regulatory agency criteria for stormwater quality, three wetlands have been designed to treat the runoff volume from a 13 mm storm event over the entire contributing drainage area. The wetlands have been designed to retain the water quality volume for 24 - 48 hours. The wetlands have been designed as a multi-functional system, with such functions as sedimentation, nutrient removal, heavy metal uptake, and aquatic and wildlife habitat enhancement.

The reinstatement of Red Hill Creek as part of the Dartnall Road Interchange implementation has been proposed to adhere to natural channel design guidelines. The proposed creek reach has been designed to provide a stable conveyance system which would also serve as a wildlife corridor. The creek has been designed with a meandering alignment stabilized, vegetated slopes, pools, riffles, overhanging embankment cover and fisheries habitat structures.

It is proposed to construct the stormwater management system as documented herein as part of the Dartnall Road Interchange design subject to the following:

1. Wetland vegetation plan should be prepared by a vegetative biologist with possible input from a landscape architect.
2. Geotechnical input should be provided on the contaimnent embankments (i.e. configuration, selected source of material, etc.).
3. An assessment of the effectiveness of the wetlands ability to enhance water quality through HSP-F modelling. This information will be useful to establish the post-construction monitoring program.
4. Wetland water budgets should be calculated, including the increase in local evaporation rates which could be expected at a wetland.
5. A sensitivity analysis which examines the spatial dependence and variability of both precipitation and evaporation.

Urban stormwater runoff is a significant collector of nonpoint source pollution. Its treatment is therefore an important aspect of water quality management. Historically stormwater has been managed primarily with respect to quantity. Herein, a hydrologic model is developed to determine the optimum stormwater pond volume required to manage both the quantity and quality of urban runoff. Thirty years of hourly rainfall data from Lester B. Pearson International Airport in Toronto is used to implement and develop the model. Hourly runoff is estimated using the Soil Conservation Service-Curve Number (SCS-CN) approach. Pond volume and dimension design curves are developed for both rectangular weirs and circular horizontal orifices. The optimization is based on a capture ratio versus relative detention ratio methodology. Pond emptying times (i.e. time to drain) of 6, 12, 24, 36, and 48 hours are examined. Optimum pond designs are presented for CN=95.

M.C. Dent 1996. Individual and Organisational Behavioural Issues Relating to Water Resources Simulation Modelling and its role in Integrated Catchment Management in Southern Africa

The aim of this research is to review individual and organisational behavioural issues relating to water resources simulation modelling and its role in integrated catchment management in southern Africa.

Key findings are that :

• current approaches to develop integrated water resources simulation modelling systems for integrated catchment management are not at all adequate;
• excellent communication in the fullest sense of the word is essential;
• there are numerous barriers to communication and to integration;
• transcending technical, organizational and psychological barriers to communication is central to the solution;
• there are no quick fixes;
• the problem is multi-dimensional and therefore requires multi-dimensional approaches to resolve;
• environmental driving forces are assisting the process greatly;
• there are definite strategic actions which will greatly improve the current process;
• many challenges with an organizational and cultural change dimension are present in the problem;
• supporting inter-disciplinary, inter-organizational efforts to develop integrated computer simulation modelling systems is probably the most important process which scientists and stakeholders can become involved in.

Watershed ecosystems consist of numerous resources which have important environmental, social, cultural, and economic values. The mutual existence and interaction among different resources within the watershed ecosystem calls for a multiobjective watershed resources management analysis. These objectives are often uncertain since they are based on estimation and/or measurement data. Probabilistic methods or fuzzification are usually the methods used in modelling these uncertainties. Selection of the best decision alternative is based on using some Multiple Criterion Decision Making (MCDM) technique. Through simulation in this dissertation, we examine the probabilistic model to address the watershed management problem. In particular, the distance-based methods, which are the most frequently used MCDM techniques, are employed in the problem analysis. In most cases, several interest -roups with conflicting preferences are willing to influence the final decision. In our study, a new method is suggested to incorporate their preference orders into the DM's final preference. The application of MCDM techniques combined with stochastic simulation and conflicting preference orders is new in the watershed management literature. Detailed analysis and comparison of the numerical results will help to decide on the suitability of the MCDM technique in watershed resources management. In particular, our numerical results indicate that in practical applications the best alternative selection is significantly influenced by the uncertainties in the payoff values. Hence. in situations where suitable data are available, our methodology is highly recommended.

Water resource professionals have an obligation to design and manage water resource systems so that they can fully contribute to an improved quality of life for all humans. Water resource systems that are able to satisfy the changing demands placed on them, now and on into the future, without system degradation, can be called "sustainable". Sustainability is a unifying concept that emphasizes the need to consider the long term future as well as the present. This includes the future economic, environmental, ecological, impacts that will result from decisions and actions taken today. While we cannot know with certainty what all these impacts will be, or what future generations of individuals or societies will want or value, we can attempt to predict what we think might happen and what future generations may want or value as we develop our current plans, designs and management policies. Admittedly we can only guess at what future generations would like us to do now in our generation for them in their generations. We must take these guesses into account as we make our decisions or take actions to satisfy our immediate demands and desires. Because sustainability is a function of various economic, environmental, ecological, social and physical goals and objectives, analyses must inevitably involve multi-objective tradeoffs in a multi-disciplinary and multi-participatory decision-making process. No single discipline, and certainly no single profession or interest group, has the wisdom to make these tradeoffs. They can only be determined through a political process involving all interested and impacted stakeholders. The participants must at least attempt to take into account the likely preferences of those not able to present in this decision-making process namely those who will be living int he future and who will be impacted by current resource management decisions. Sustainability is intimately related to various measures of risk and uncertainty about a future we cannot know but which we can surely influence. Clearly our guess about the future will, with certainty, be wrong. Hence they will need to be revised periodically. Recognizing that some management objectives will change over time, we must consider the adaptability or robustness of the systems we design and operate today to this management uncertainty and to the inevitable changes in the quantity and quality of the resource being managed. This document examines many of the major issues and challenges raised by the concept of sustainability applied to water resource system design and management. It reviews various guidelines that have been suggested for achieving a greater degree of sustainability and the extent to which they have been applied int he development and management of water resource systems. It outlines some approaches for measuring and modelling sustainability and illustrates ways in which these measures and models might be used when evaluating alternative designs and operating policies. This monograph is written for those interested and involved in the planning and management of water resource systems. This includes students as well as practitioners. It was written by individuals sharing a background in environmental and water resources systems planning and management, but having different experiences and opinions. Hence different points of view are presented with the hope that they will stimulate thinking about just how water resource systems should be developed and managed, not only for those living today but also for those who will be dependent on these systems in the future. This document is the product of two groups of professionals: a working group of the International Hydrological Programme of the United Nations Scientific, Educational and Cultural Organization (UNESCO) and members of the Task Committee of the Division of Water Resources Planning and Management of the American Society of Civil Engineers (ASCE). Both groups were formed to explore ways in which the concept of sustainability might be used as a measure of system performance when evaluating alternative water resource plans and management policies. We begin Chapter I with a discussion of the definition of sustainability that is commonly used, but which we think is not very helpful for water resources planning and management. Nevertheless, this common definition is often cited, used and discussed in the considerable literature that exists today on the subject of sustainability and sustainable development. In Chapter 2 we review some of the major issues and challenges posed by this commonly accepted definition of sustainability and try to identify why sustainability has been so difficult to quantify and to define very precisely. The discussion of the challenges and issues associated with this broad concept of sustainability (as applied to various water resources purposes) leads us to the particular definition we propose for water resource systems planning and management. In Chapter 3 we define sustainability in a way that seems appropriate to those of us involved in water resources planning and management. This definition allows us, in Chapter 4, to identify and examine in more detail some ways of measuring sustainability for selected water resources functions or purposes. These measures rely on the inputs and judgments of those having an interest in such systems. The interests of different stakeholders may differ. While we recognize computer analyses lie behind most decisions involving facility design, construction and operation, it is their development and use toward achieving a common (shared) vision among all stakeholders that is important with respect to system sustainability. This growing concern over the need to achieve increased levels of system sustainability (however defined) has lead to the creation of a number of guidelines for its achievement. The guidelines, summarized in Chapter 5, were created by various professional engineering organizations in various countries. The primary purpose of these guidelines is to help those in the engineering profession, and especially the practising engineer, design and manage systems that are sustainable. Systems can be designed and managed so as to be sustainable even though particular projects and even institutions making up those systems may not be. There are many aspects of sustainable systems, including change, but all should lead toward the development and use of more sustainable technologies, to more sustainable environments and ecosystems, to more sustainable economic and financial policies, to more sustainable institutions and societies, and to improved long-term health and welfare. Chapter 5 also includes a series of brief descriptions - case studies - of some water resource development and management projects. They serve to illustrate the extent to which sustainability criteria have or have not been achieved in particular situations. These real-world examples also illustrate the difficult tradeoffs that must be made among various goals and interest groups when designing water resource systems and implementing policies for managing them. Chapter 6 explores some economic criteria and associated models that consider economic objectives. The discussion examines the issue of appropriate discount rates - the weight we put on our current assets compared to the weights we assign to assets of future generations. One of the major needs still unmet is our ability to value non-monetary goods (such as those derived from our environment and our ecosystems) in monetary terms. Since we have not yet learned how this can be one, we are forced-to make comparisons and tradeoffs between economic and environmental or ecological criteria expressed in different metrics (units of measure). Chapter 7 identifies some of these environmental and ecological criteria and associated modelling and implementation approaches aimed at achieving a greater degree of sustainability. Through a series of examples emphasis is placed on the importance of communication and public participation, hopefully leading to a shared vision in what the outcome should be. The adaptive approach to planning is shown to be extremely useful guidelines given for its implementation. Individuals and societies manage themselves through their institutions. While this monograph is not focused on that aspect of sustainability in any great detail, Chapter 8 briefly reviews some of the social and institutional aspects of sustainability. Institutional capabilities as well as conflicts and constraints are of major importance for water resources planners and managers. It is, after all, through our institutions that decisions are made regarding our water resources development and use. They are the clients for all of our technical planning and design models and decision support systems. Institutions can foster and encourage increasing sustainability, or they can do just the opposite. Three examples are given in support of that observation. Chapter 9 examines how a variety of modelling technologies can and should contribute toward higher levels of sustainability. Emphasis is given to ways of improving the planning and management process and the information upon which recommendations are made and decisions are based. The potential role of Decision Support Systems (DSSs) is presented and illustrated. Chapter 10 examines how economic, environmental and hydrologic risk and uncertainty impacts on our attempts to define and work with sustainability criteria. Anyone involved in water resources planning and management must contend with risk and uncertainty and this is especially so when required to look into a distant future. No one can look into even the near future with precision. Future supplies of water and future demands for the services provided by water resource systems are unknown at the time system design and operating decisions are made. Professionals are asked to provide for this uncertain future. As a result, system robustness and other risk-based measures of system performance become important considerations, and are intimately tied into any measure of sustainability. Chapter 11 addresses some equity, education and technology transfer issues related to sustainability. While the discussion is brief, the subject is as important as any discussed in this monograph. Different individuals will have different views as to what is equitable or ethical. The correct view is not always obvious. Particularly for those working in educational institutions, it is important to consider capacity building and the technology transfer issues with respect to sustainability. In Chapter 11 some of the education, training and technology transfer aspects of sustainability are examined. Also discussed are the important roles professional societies as well as educational institutions play in producing and providing the expertise needed to continue into the future the efforts being made today toward achieving more sustainable systems. The monograph concludes with Chapter 12 highlighting some key points concerning the planning and management of sustainable water resource systems. It emphasizes that in our search for sustainable development, the effectiveness of any mechanism derived to reach that goal depends, in the end, on the quality of the individuals interested in pursuing it.

Wall, R. C. 1968. Ph.D. An Analysis of the Current Status of The Schistosome Dermatitis Problem in Michigan.

This study analyses the schistosome dermatitis problem in Michigan to determine its current status and to make recommendations regarding how the problem should be handled now and in the future.
Three aspects of the problem were analyzed, (1) the distribution pattern of lakes from which reports of the disease have come since 1939; (2) the State's control program since 1939; and (3) the impact the disease has had and in having on people who own or operate facilities on infested lakes.
The number of lakes from which reports of the disease have come has increased and the geographical area from which these reports have come has spread.
Treatment chemicals and techniques have remained virtually unchanged since the inception of the program in 1939. The majority of the people contacted in a mail survey conducted an part of the study, believe that copper sulfate treatment is effective in controlling the disease.
Nearly 90 per cent of the survey respondents are aware of the disease and over two-thirds of them have had it at their lake facilities. The majority of the cottage owners, camp directors, and park managers contacted do not feel that schistosome dermatitis is having an economic impact upon them, although 20 per cent believe it is affecting property values on their lakes. Nearly half of the resort operators contacted said they have lost business because of outbreaks of the disease in their lakes. Twenty per cent said the disease is influencing property values on their lakes and 30 per cent said its presence in their lakes has influenced decisions they have made concerning expansion of their resort facilities.
Resort operators and park managers believe experiencing schistosome dermatitis or even hearing about it influences guests' and visitors' attitudes about returning to their lake and to Michigan for a vacation.
Over 75 per cent of the respondents feel that schistosome dermatitis is a problem in their lakes and in Michigan, 12 per cent "don't know" how serious a problem it is, while the remaining 12 per cent feel that it is merely a nuisance and not a problem.
The schistosome dermatitis problem in Michigan is increasing in magnitude and the presently utilized control method apparently is not adequate to cope with the problem now and in the future. The evidence gathered in this study indicates that the most logical course of action to follow relative to this problem is to; (1) continue the present treatment program until the effectiveness of copper sulfate treatment has been determined; (2) initiate a comprehensive research program designed to fill the present gaps in information available about the dermatitiscausing parasites, the impact the disease is having on people who own facilities on infected lakes, the effectiveness and safeness of copper sulfate treatment, and the identity of new chemicals and treatment techniques to use in the control program; and (3) as newer molluscicides and techniques are identified, and if copper sulfate treatment is shown to be ineffective, phase out the old chemical and technique in favor of the newer ones.
The most realistic and practical approach to this problem is to carry on a combination research and treatment program aimed at achieving effective, efficient, safe, and dependable control.

Burges, S.J. 1970. Ph.D. Use of Stochastic Hydrology to Determine Storage Requirements of Reservoirs -- A Critical Analysis.

This study resulted from an initial attempt to identify components that contribute to uncertainty in the results from a decision flow optimization study for a major water resource project. The simplest possible case (Linsley et. al., 1969) considering urban water supply had so many components in the decision flow process that it was not possible at current technology levels to examine the impact of all of them on the final optimization result. A simplified version of the same problem was studied and errors in demand estimates as well as errors in hydrology were shown to significantly influence the distribution of storage needed to satisfy demand.
From this simple exploratory investigation two major avenues of research opened. The first was an attempt to estimate errors in demand projections (Whitford, 1970) for urban water supply and the second, which forms the body of this thesis, examined possible procedures for estimating the storage requirements for a storage reservoir.
In all sections of this report the demand sequence is assumed to be known so that storage could be determined as a function of the demand sequence and inflow model characteristics. For the purpose of this study reservoir reliability means that for all the possible inflow sequences to a reservoir, the storage S having reliability x per cent is such that x per cent of the inflow sequences will meet the demand sequences with no shortage. The reservoir problem to be investigated is: What capacity should a reservoir have to satisfy demand with reliability x per cent over the economic life of the structure. Stochastic generation of streamflow sequences was proposed by Fiering (Kaas, et. al., 1962) as a solution to the problem and this study explores the uncertainty in such stochastic methods.
The Markov stochastic generation model was studied extensively because it is the model most commonly employed. The basic procedure used in stochastic generation was thoroughly examined to determine if procedural or data errors contributed significantly to errors in determining reservoir capacity.
Chapter 3 describes an exploratory study to locate problem areas in typical applications of Markov generation models. Parameter estimation and generation techniques were examined and components that have a large impact on the required storage distribution were identified.
Chapter 4 examines the basis of stochastic generation while Chapter 5 is an analysis of the application of a normal Markovian generator to reservoir storage distribution determination using precisely known generator parameters. Chapter 6 considers the validity of correlations in annual data. Chapter 7 examines the possibility of using multiple lag models as runoff generators. Chapter 8 compares monthly and annual generation for reservoir studies.
The log-normal Markov model is examined in Chapter 9 where approximate and mathematically correct generation procedures are studied for both monthly and annual analysis.
Demand is varied linearly and by known amounts in Chapter 10 to determine the influence of growth in demand, particularly near the end of the economic life, on the required storage distribution. Chapter 11 studies factors that influence reservoir reliability. For traces where deficiencies occur, the deficiency pattern is studied as a guide to economic analysis.
The principal conclusions of the study are presented in Chapter 12 together with recommendations for further studies on methods using stochastic processes in water resources project analyses.

Smith, R.L. 1972. M. Eng. "The Design of Compound Critical Sections for Open Channel Flow Measurement".

Weirs of the conventional shape are amenable to analysis based on an assumption of one-dimensional flow and a number of computational routines have been developed for this type of transition problem. When critical flow occurs in a highly non-uniform section, a more sophisticated approach is necessary.
In conjunction with laboratory tests on a typical compound control, a mathematical model was formulated for the development of the stage-discharge relation. It is felt that this model will allow an accurate prediction for water quantity from fluctuating sources.

El Hyashab, Ahmed M.M. 1975. M.Sc. "Hydraulics of Flow Over Side Weirs".

This thesis reviews the present state of knowledge of the hydraulics of low over side weirs, and describes an experimental investigation into the lfow of water over side weirs using a laboratory flume.
A detailed examination is made of the assumption conventionally used by many investigators that the total energy of the flow remaining in the channel does not vary along the length of the side weir (excluding friction loss), and following consideration of earlier works in this subject. Some new conclusions are suggested concerning the weir coefficient, the variation of the velocity and pressure distribution coefficients and the end controls.
The secondary flow observed along the side weirs in the side weirs the experimental work and the changes it underwent on progressing along the side weir are described in detail. Some degree of confirmation of the analysis was achieved for the exact way in which the lfow moves along the side weir as it leaves the channel section and enters the side weir section.
The background of the side weir theory is given in detail, preceding an examination of the predictions of the this theory in terms of the results of the present and other experimental investigations. It has been found that even when the energy of te secondary flow and the non-uniform velocity distribution are allowed for, the observed Total Energy Line along the weir length falls considerably below the horizontal. From observations it could be seen that the longitudinal component of velocity of water flowing over the side weir is usually considerably higher than the average velocity of the water remaining in the channel. In this situation it was found necessary to use a momentum approach, taking into account the momentum of the water flowing over the side weir.
The derived equation for the water surface profile along the side weir based on the momentum principle has been discussed in detail, some relationships for different cases are derived depending on the ratio between the spill flow over the side weir to the lfow in the main channel upstream of the side weir, and alos on the type of flow along the side weir.
A computer programme based on the deprived momentum equation predicted water surface profiles which compared well with the observed profiles. Also, a satisfactory comparison was made between the observed water surface profile based on the technique developed in this thesis.

Froise, S. 1975. Ph.D. Least Cost Control Strategies in Urban Drainage Design: A Dynamic Programming Approach.

This dissertation concerns a problem of considerable engineering importance; an effective approach to its solution is given. The problem identified is one part of the total problem of storm sewer design. The analytical methodology developed is of considerable importance in the evaluation of technical alternatives for handling stormflow conveyance.
The author has taken existing techniques and combined them to yield solutions to problems involving least cost combinations of conveyance and storage elements needed for managing storm water flows. The well known techniques of dynamic programming and unsteady flow routing were used. The novel approach used to couple conveyance elements, unsteady flow routing, design cost information and a dynamic programming optimization scheme exhibited mature academic and professional understanding of complex and difficult problems. Demonstration of a feasible scheme for handling multiple state-multiple decision problems as well as unsteady flow eminently qualifies this work for acceptance as a Ph.D. dissertation.
Careful attention to detail by the author resulted in his investigation into the usefulness of an approximate unsteady flow routing scheme in certain instances in lieu of a more precise but considerably more expensive solution scheme. The overall results of this research effort provide a workable tool that can be used in part to assess how well the storm water runoff process needs to be modelled. Furthermore, the optimal design scheme developed is able to offer cost savings up to approximately 30% of conventionally designed systems.

Gupta, P.K. 1977. Ph.D. "Dynamic Optimization Applied To Systems With Periodic Disturbances".

The primary objective of this research work was to develop a mathematical method of dynamic optimization for systems having periodic disturbances. The motivation for the work was the use of a marsh for tertiary treatment of wastewater. Four problems of varying orders of difficulty, were selected and solved. The objective function was an integral of a continuous function of control variables, Constraints on state and control variables were included.
The method of finite differences is used to transform the dynamic optimization problem into a constrained multivariable optimization problem. An approach, called ASPE (Algorithm to Satisfy Periodicity Equalities), is used to satisfy equality constraints representing periodicity conditions. ASPE is based on the Newton-Raphson's method of solving nonlinear algebraic equations. For a system in which the derivatives of final states with respect to initial states remain constant, ASPE exactly satisfies the periodicity equalities at each point during the optimization process, but for a nonlinear system, it approximates the exact solution. As the optimum is approached, many small moves toward the optimum are made. Because of this characteristic ASPE gives a very good approximation to the exact solution.
Two methods, mixed penalty function and Powell's multiplier, were tested for the constrained multivariable search and Powell's multiplier method was selected.. The inequality constraints are included in the penalty function by transforming them into a moving loss function that maintains continuity of the modified objective function with respect to the control variables. Three efficient unconstrained optimization methods were tested for unconstrained search: 1) Hooke and Jeeves direct search, 2) Powell's method, and 3) the Davidon-Fletcher-Powell method. In the D-F-P method, the derivatives were calculated by the difference scheme suggested by Stewart. For moderate accuracies, the D-F-P method was found to be most efficient. Coggin's method was used for the one-dimensional search.
The first two problems, four stirred tank reactors, were solved to maximize the amount of feed without violating certain constraints. The disturbance was an input stream whose flow was periodic and preset. Two marsh problems were attempted, based on a simulation of water movements and nutrient dynamics in the marsh. A simulator REBUSII was written for this purpose. The whole marsh is divided into a small number of three dimensional blocks with uniform properties. REBUSII consists of construction and simulation phases. The construction phase reads the dimensions for the problem and builds a main program and BLOCK DATA for the simulation. The simulation phase reads the rest of the input and performs simulation and/or optimization.
Problem #3 was solved to maximize the amount of water which can be introduced into the marsh without flooding it. The water was introduced only in the summer months and water levels were kept annually periodic. It was found that 6.4X10 m water could be introduced to a 4 km marsh without violating a 2 5 cm maximum water depth constraint. There was little change in spring water levels. It was concluded that the water should be added at maximum pumping rate till water depths reach maximum allowed values, followed by constraint maintenance.
Problem #4 was solved to maximize the amount of wastewater which can be treated by the marsh. Nitrogen (in nitrate and ammonium) and phosphorus (total dissolved phosphorus) nutrients were considered. The concentrations of nutrients remain much below the maximum permissible values, and hence the pumping rate obtained in problem #3 were the solutions for this problem too. Slow accumulation of nutrients indicate 100 years to reach the concentration constraints.

Munro, D.M. 1978. "Stormwater Management Modelling: An Application of SWMM to the Carling Street Catchment, London, Ontartio".

What is stormwater management modelling? To the model user it is a collection of techniques which are difficult to ascribe to any one body of theory; although this new approach to solving urban stormwater problems does incorporate ideas from several scientific methodologies, within a systems framework. To Planning agencies storrawater modelling is a useful tool which can provide information to the planning process.
What are the problems facing stormwater management modelling? At present, the development of technology is not one of the major problem The lack of data, or to be specific, applicable and utilizable (i.e. quantifiable) data is a considerable Droblem. Information collection system and agencies will have to be enormously improved if they are to be effective counterparts to present computer technology. At the same time, modellers must be familiar with and have confidence in their models otherwise coefficients and numbers will be pulled from statistical "cookbooks" without a feel for the underlying purposes.
Vlachos and Flack (1975) have pointed out two limitations of empirical approaches to urban analyses: (1) social and psychological phenomena cannot be analyzed within a rigorous mathematical framework; and, (2) intangibles such as environmental amenity and quality of life are difficult to define. Stormwater modelling cannot cope with problems of this type, however, it can provide useful data which may be used in the decision~making process in water resources management. Legislation problems are evident when government enforced regulations and penalties apply to industrial and municipal polluters but not to the construction industry which discharges huge amounts of sediment from subdivision and highway developments into local streams and rivers every year. What contributions can geographers make towards solving stormwater management problems?
Overton (1976) stresses that the immediate problem facing urban drainage management is the gap between theory and practice. New technology should be placed within a humanistic context to close the gap, help solve the problems and increase our understanding of the urban environment. To date, the majority of professionals using stormwater management models have been engineers and technicians with backgrounds in pure and applied sciences. This has introduced an engineering bias to stormwater modelling which Ilas neglected the contributions of the social scientist. There has been a lack of skilled personnel with multi-disciplinary backgrounds. Perhaps, as Kasperson (1969) suggests,geographers with their integrative capabilities are best qualified to fill the gap between theory and practice.
Stormwater management models have shown their usefulness in solving urban drainage problems. To cite Waller (1976, 401) it may be concluded that: "Urban drainage models are an important new tool for examining the alternatives for design and management of urban drainage systems. As they develop and improve, as their benefits become better recognized, and as the confidence of Canadian engineers in the applications of these tools increases, these models may prove to be the most significant single product of current programs that are aimed at improving our competence to solve urban drainage problems."

Tucci, C.E.M. 1978. Ph.D. Hydraulic and Water Quality Model for a River Network".

A river network system consisting of branches and loops is sometimes complicated by downstream effects from tides, lakes, and control, and because of this, management of water quality, sediment in such rivers is a difficult task. Development of tools to floods aid in the management decision-makinc process is an important area of research; ultimately resulting in more reliable results. River behaviour can be modelled in detail (one-,two-,or three-dimensional models) using numerical methods. Usually the level of detail is determined by the size of the system. Large system models are restricted in size and detail due to the high cost and storage requirements of the computer.
A model was developed to simulate the hydraulic behavior and water quality of a river network on a one-diniensional representation. The two complete St. Venant equations and the transport equation were solved by the finite difference method. The transport equation utilizes the advection, dispersion, and source and sink terms. The system of equations resulting from use of an implicit scheme was solved by a modified Gauss elimination procedure.
The model can simulate biochemical oxygen demand, dissolved oxyaen, or any other conservative substance. The basic equations are solved; thus, the simulation of other substances can be added to the model by including the mathematical description of the reaction processes in the source and sink terms of the transport equation.
The hydraulic module of the model was adjusted and verified with data from the Jacui Delta, Brazil. Good agreement between the calculated results and the observed data resulted. The water quality model was tested under hypothetical conditions for the same Delta in order to demonstrate te utility of the mathematical model in making decisions at the management level. This model is a mathematical method that can be used in large systems of variabel comlexity to help in understanding their processes, controlling data measurements, and reaching sound management decisions.

Ammon, D.C. 1979. M.Eng. "Urban Stormwater Pollutant Buildup and Washoff Relationships"

Urban stormwater quality predictions in the Environmental Protection Agency Storm Water Management Model (SWMM) are based on linear accumulation of solids, and on washoff rates which are proportional to the runoff rate and the amount of pollutant remaining on the ground. Recent studies indicate that the accumulation of solids on impervious areas, such as streets, is nonlinear.
Alternative quality algorithms are developed and incorporated into a deterministic mathematical model, an extension of SWMM. The effects of the various quality alternatives are demonstrated using both continuous and single event simulations. The pollutant accumulation options include linear, power and exponential decay relationships. An additional washoff relationship for impervious areas is based on sediment transport theory. For pervious areas, quality predictions are based on sediment rating curves or on detachment of particles by raindrops and subsequent transport by overland flow.

MacRae, C.R. 1979. M.Sc. "Hydrologic And Water Quality Modelling in a Small Urban Catchment in London, Ontario".

The hydrologic and water quality characteristics of urban runoff were examined for a small (18.2 ha) highly impervious (65%) separate storm sewer catchment located in the older portion of central London, Ontario. Landuse in the basin is mixed but dominated by industrial (27%) and commercial (19%) activities.
Hydrologic data were collected from August 16, 1976 to November 10, 1976 and again from May 24, 1977 to June 1, 1977. Twenty-nine rainstorms and 5 snowmelt events occurred during this time. A survey of the hydrologic record shows that stream flow may be differentiated into four distinct types; baseflow (average 0.049 cfs), anthropogenic inputs (maximum 0.089 cfs), stormwater runoff (maximum 21.84 cfs) and snowmelt (maximum 0.70 cfs). Baseflow is derived primarily from groundwater seepage and sump pump activity. Anthropogenic inputs occurred as frequent (mean 10 per work day), small and random pulses ranging from a few minutes to several hours in duration. The temporal pattern of these inputs showed a strong correlation with the standard work day. Storm runoff was found to be the principal flow component contributing 51% of the total runoff volume (1.013 x 10 6 ft 3 ) recorded from all sources during the study period.

Thompson, L.R.1979. M.A.Sc. "A Comprehensive Subcatchment Hydrologic Simulation Model for Urban and Rural Applications".

A Subcatchment Hydrologic Simulation model which attempts to satisfy a set of design criteria discussed in Chapter 1, has been developed.
The model in its present form estimates the runoff response of a single subcatchment and therefore is meant to be a component in an overall hydrologic model.
The model separates estimated runoff response into its various components and is intended to be applicable to both urban and rural, discrete and continuous simulations, utilizing variable time steps.
The model may be run efficiently on a micro computer thus greatly reducing the cost of its application.
The model has been tested for discrete simulation on an urban subcatchment of 18.2 hectares and a rural subcatchment of 19.46 sq. km.
By varying only initial soil moisture content, the model was able to simulate observed runoff within an acceptable degree of accuracy, from four storms ranging in size from 1.96 cm. to 5.2 cm.
The model simulations also indicate that under intense storm conditions the presence of substantial shallow, agricultural tile drains significantly affects runoff hydrographs, tending to increase peak flows.

Williams, K.K. 1979. Ph.D. Oklahoma City Urban Storm Runoff Quantity: Comparison and Calibration of Predictive Methods.

Six discrete event urban rainfall-runoff quantity models commonly used by federal agencies were calibrated on twenty-three events recorded by the U.S. Geological Survey on three urban basins during 1974-1975 in Oklahoma City. The models were the Rational Method (Department of Housing and Urban Development), TR-20 (Soil Conservation Service), HEC-1 (Corps of Engineers), Urban Flood Hydrograph Synthesis model (Geological Survey), SWMM (Environmental Protection Agency), and MINICAT (National Weather Service, River Forecast Center). All the models were calibrated for peak discharge on the recorded floods, and all except the Rational Method were calibrated for runoff volume. It was found during the calibration process that antecedent soil wetness was not an influence on runoff from the basins and storms used in the study. The calibrated models were compared on how accurately they reproduced the recorded hydrographs, engineering applications, and relationships between various hydrograph parameters. It was found that each model calibrated nearly as well as the others, except that HEC-1 was a little more reliable in reproducing the recorded events than the other models, and TR-20 tends to bias, making the larger floods too large and the smaller hydrographs too small.
It was found that the models vary greatly in complexity, resource requirements, and usefulness to various applications. The Rational Method and regression equation developed by the Geological Survey from its model are simplest and most suitable as aids in sizing small numbers of hydraulic structures, such as individual roadway culverts. The Geological Survey's regression equations are best suited for flood plain boundary studies, provided the basin is not regulated by reservoirs. TR-20 and HEC-1 are computer models requiring more resources and are suitable for use in flood control project design, while SWMM and MINICAT are the largest models requiring the most resources and are suitable for analyzing and designing large complex sewer systems.
Formulas, tables, and graphs were derived so that if one knows the unit hydrograph share parameters for one of the models such as Snyder's Unit Hydrograph, Clark's Unit Hydrograph (as computed by the Corps of Engineers), Clark's Unit Hydrograph (as computed by the Geological Survey), or the Soil Conservation Service Unit Hydrograph, then he can convert to another model with its parameters such as to get the same shape unit hydrograph.

Book, D.E. 1980. M.Sc. "A Dynamic Hydraulic Model for Simulation and Evaluation of Complex Drainage Systems".

As population concentrates in urbanizing areas and development continues to take place rapidly, urban storm drainage can become a significant problem. Present methods for evaluating such problems do not always take advantage of existing knowledge and capabilities. Many of the simplified procedures now used are unnecessary due to the increased accessibility to high speed computers. The purpose of this thesis is to present a physically-based computer model which gives increased accuracy and flexibility at reasonable costs.
The first component is a rainfall-runoff model. Presented here is a modified version of the San Francisco runoff model, which uses kinematic wave routing for overland flow and minor sewers. A Green-Ampt. infiltration scheme is incorporated to add physical relevance to the program. Emphasis in this section is on model application with a comparison made to a unit hydrograph procedure. It was shown that use of the rainfall runoff model required little additional effort while providing increased flexibility.
The second component is a routing model for the conveyance system. The model UNSTDY is presented which incorporates the full St. Venant equations. A linearized, fully implicit solution scheme which makes use of an efficient double-sweep technique to solve the system of equations is utilized. This formulation shows favourable results when compared to a nonlinear scheme. Provision is made for branched systems as well as other properties of urban systems such as overflow weirs, control gates, and pump stations. For demonstration purposes, the rainfall-runoff model is combined with UNSTDY to model the Oakdale drainage basin.
In the final section, an evaluation is made to determine under what conditions it is necessary to employ dynamic wave rather than kinematic wave routing. The results from a completely kinematic wave routine model, KINEROS, are compared to those obtained from UNSTDY. A hypothetical watershed is considered in which several parameters are varied over a wide range of values. The models are also applied to an existing rural watershed to check the results of the tests. Criteria are presented to assist in choosing the appropriate model for a system. The results show that the slope can be used as a preliminary indication of the applicability of kinematic wave routine. A more precise indication results when a dimensionless term relating channel slope, normal flow conditions for the peak discharge, and the time to peak is considered.
The result of this work is a set of documented computer pro-rams which can be used by practising engineers. The formulation, documentation, and examples of use presented here provide the user with adequate information to apply the models to a variety of drainage systems.

Donald, S. 1980. Ph.D. Uncertainty Analysis of a Lake Eutrophication Model.

Mathematical modeling of aquatic ecosystems has advanced to the stage of relatively common use in management contexts. In many cases, it has also become a useful tool for suggesting research needs, synthesizing extant information, and analyzing these ecosystems in ways that are not tractable, and often impossible, through field and laboratory studies alone. The models used most often in these contexts have similar attributes; they are generally time-dependent, often nonlinear, differential equation models based on parameterized physiological processes and mass conservation.
These models, whether from the management or the research milieu, have another common thread - they are generally deterministic. That is, although stochastic properties of model initial conditions, parameters, and forcing functions are often recognized, these properties are seldom accounted for. Moving beyond acknowledgment of these stochastic elements to accounting for their effects is important because these stochastic properties affect the confidence that can be placed on the model output.
Analysis of errors is important in a management context to establish error bounds on predictions. Eutrophication models have been developed to generate deterministic predictions of water quality based on present and expected scenarios of system inputs and controls (e.g., DiToro et al., 1971; Chen and Orlob, 1975; Thomann et al., 1975, 1976, 1979; Canale et al., 1976; Bierman, 1976). Output from these deterministic models often influence decisions affecting many thousands of people and millions of dollars; yet quantitative limits of confidence are lacking for these models, unlike many of their simpler counterparts (see Reckhow, 1979). This was demonstrated clearly at a recent conference on verification of water quality models (Mueller, 1980). There it was revealed that only qualitative evaluations of calibrations and verification of eutrophication models have been carried out to date. The track record for even these tests is quite short. Eutrophication models are crude representations of highly variable, stochastic systems; ignoring such important attributes results in naive confidence in the models' solutions. For these models to become more generally accepted and used in the decision making process, they must be placed in their proper perspective. Performing quantitative evaluations and establishing error bounds and confidence limits on model output provide some of the needed perspective.
Analysis of errors is also important in a research context where a model's ability to simulate must be evaluated prior to investigation of specific system properties. Models have recently received more use as data synthesizers and as tools for detailed ecological systems analysis (e.g., Lehman, 1978; Robertson and Scavia, 1979; Scavia, 1979b; Halfon, 1979; Scavia and Bennett, 1980). Output from these models is often used to assess the relative importance of various system compartments or processes and to thus focus additional effort on key problems. Prior to using a model in this context it is important to evaluate its ability to function as a synthesizer or interpolator. Traditionally this evaluation is done simply as a comparison of model and measurement trajectories with no quantitative assessment of other model errors. It has been demonstrated that comparison of modelled and measured state variables alone is not sufficient for this purpose (Scavia, 1980). Calculation of errors associated with model state variables and of correlations among state variables and parameters will assist in evaluation of these models for use in research contexts.
Sources of error whose effects may propagate through models in both management and research contexts can be grouped into four categories 1) initial conditions, 2) parameters, 3) inputs or forcing -functions, and 4) model equations. Errors In the first and third category are often measured and quantified (i.e., mean and variance). Statistical distributions of parameter values have not been generally determined. However, recent works have examined some parameters, and at least their ranges can be determined from the literature. There has not been a thorough examination of uncertainties inherent in the model equations themselves. Experience with the models and knowledge of the adequacy of individual theories and assumptions upon which they are built allow a relative assessment of faith in model equations, at least qualitatively.
Given the need to assess overall model errors and given the potential sources of those errors, two problems can be posed. The first is to establish a method by which errors can be propagated through the model to estimate error bounds on the output. The second problem deals with identifying how output errors could be reduced to within some acceptable bound. Closely related to the problem of model uncertainty are two aspects of model development, model structure identification and parameter estimation. The former is used to build a model structure used to calibrate the model to a particular sat of observations. With this in mind, the following goals and objectives were established for the work reported herein: 1) To evaluate a method of error propagation for a relatively complex eutrophication model. 2) To apply the method and establish output error bounds based on uncertain initial conditions, parameters, loads, and equations. 3) Given those error estimates, explore ways by which they can be reduced most effectively. In the following paragraphs, methods of error propagation, parameter estimation, and model structure identification that have been applied in water resource problems are reviewed. Following that, a rationale for selecting the procedures used herein is offered.

Jewell, T.K. 1980. Ph.D. "Urban Stormwater Pollutant Loadings".

Unless stormwater runoff can be treated or controlled, many receiving waters will not become "fishable and swimmable" as proposed by the Water Pollution Control Act Amendments of 1972. In order to determine where and when to control stormwater runoff, it is necessary to assess its impacts and estimate the effects of various control measures, tasks that would be extremely difficult using measured data alone.
Therefore, researchers have used simulation models to predict the quantity and quality of stormwater runoff and to determine how it would affect the environment, under both controlled and uncontrolled conditions. The quality portions of these models, however, have not fostered confidence in their predictive capabilities. Lack of confidence has been engendered by the use of unverified predictive algorithms and by the variability of measured stormwater pollution data. This study was designed to produce improved stormwater pollutant washoff prediction techniques which would be verifiable and would take into account data variability.
To assess the variability of stormwater pollutant data and to form a data base for developing improved predictive algorithms, available storm event pollution data was examined and the best of the data catalogued into a data storage and retrieval system. The catalogued data included 261 storm events from 26 basins in 12 geographic areas.
The data file was used to show that it was not possible to derive general pollutant washoff functions containing a given set of independent variables that would give reasonable results for most areas. This was true for either storm event total loadings or instantaneous fluxes. It was concluded that data should be gathered for each basin to be studied and models developed from this data to predict stormwater pollution washoff.
A methodology was outlined for developing these models. It included suggested trial formulations, and guidelines for applying linear and intrinsically linear multiple regression analyses, to pick the best model for predicting the washoff of each pollutant. Recommended data gathering and reporting procedures were also developed.
This methodology represents a distinct improvement over existing stormwater quality predictive techniques. Researchers will now be able to predict stormwater pollutant loadings with a measurable degree of certainty.

Kelton IV, W.D. 1980. Ph.D. The Startup Problem in Discrete-Event Simulation.

We are concerned with methodological issues arising when the steady-state mean of a discrete-time stochastic process is estimated by means of digital computer simulation. Our focus is on statistical properties of point and interval estimators and how these properties react to various tactical approaches to the simulation study. The principal overall methodology examined is that of simply making independent replications to serve as a basis for statistical analysis. The difficulty is that artificial initial conditions used to start the simulation are generally unrepresentative of steady-state conditions, with the result that the early output data will be biased relative to the steady-state mean. We examine the effect of this bias, and propose and test a new method for dealing with it.
The first part of the thesis derives exact measures of performance of alternative simulation strategies when the process being simulated is first-order autoregressive (AR(1)). Under a fixed total budget constraint, we allow choice of the number of replications, as well as choice of initial deletion amounts from the beginning of each replication, to mitigate the startup bias. We examine three properties of the point estimator obtained by taking the average of all the undeleted data from each replication, namely mean-squared error, mean absolute error, and probability of an excessive error. We also derive and compute exact, true coverage probabilities of confidence intervals formed from the averages from within the replications, which are stated to have a given, nominal probability of covering the steady-state mean. Our conclusion from these theoretical investigations is that the technique of initial data deletion can be an effective and efficient approach to diminishing the startup bias in replications.
We then develop and empirically test a practical, data-based procedure for identifying appropriate deletion amounts in practice. This procedure, which also determines appropriate total replication lengths, utilizes a time series regression technique from the econometrics literature to identify when the output data appear to have stabilized. Results of testing the procedure on a wide variety of stochastic models with known steady-state means indicate that the procedure largely eliminates the bias owing to the artificial initial conditions. The ability to eliminate bias in replications would allow the simulator to use the basic method of replication as an approach to solving statistical inferential problems when the basis for study is steady-state behavior. We give two examples of such uses of our deletion procedure in concert with the method of replication. The first is a sequential confidence interval for the steady-state mean, which must satisfy a smallness condition. The second example is a multiple selection problem where the criterion for selection is the steady-state mean of each of the several alternative systems designs.

Maunder, D.E. 1980. M.Eng. "Modelling of Regional Sanitary Sewerage Networks for Minimum Cost".

A computer model is developed for the solution of regional sanitary sewerage systems on a minimum cost basis. The regional system is considered to consist of a substantial number of alternative transportation (collection) routes, several potential treatment facilities and a receiving body of water which is able to accept the effluents from these facilities. The model, using a steady state analysis, facilitates the selection of a good collection network from the large number of possible networks, the selection of good depth-diameter combinations for each link in this final network and the selection of the size and number of treatment facilities which should. be involved in the regional system. In addition, the required treatment efficiency of each facility will be determined so that a specified water quality goal may be satisfied.
A major point to be considered in attempting to select a least cost system is the tradeoff in costs between constructing a few, large, centralized treatment facilities (thereby benefitting from the economies of scale which generally exist) and the extra cost associated with transporting the wastes an extra distance to these centralized plants. The possibility of higher treatment efficiencies being required at these centralized treatment facilities (due to a greater stress being exerted on the water quality of the receiving body of water) should also be considered. The piping costs are a function of the wasteflow. Fixed costs may be included in the cost function. The treatment plant costs are a function of both wasteflow and treatment efficiency. Fixed costs may also be included when determining the treatment plant costs.
Linear programming and dynamic programming techniques are used as a basis for the computer model. Additional subroutines were required to ensure that the technical constraints (for example, the minimum and maximum velocity constraints) were satisfied and to provide the missing components which are necessary to develop a complete computer package.

Uchrin, C.G. 1980. Ph.D. "Mathematical Modeling of Suspended Solids and Associated Pollutant Transport".

The transport dynamics of particulate solids have been of long-standing concern from the perspectives of sedimentation operations in water and wastewater treatment, and bed-load movement and siltation processes in rivers, streams, and dredged channels. Only recently, however, has attention focused on detailed descriptions of the behavior of suspended solids in natural water systems from the perspective of water quality transformations. This interest is predicated largely on increased awareness of the role of suspended solids in the transport of certain pollutants in natural water systems.
Pollutants adsorbed on, or contained within, suspended solids constitute a separate phase in a heterogeneous system, and can be expected to behave quite differently chemically, biochemically, hydrodynamically, and toxicologically -- from dissolved pollutants. Yet most water quality models presently used to describe the behavior and fate of pollutants account only for the transport and distribution dynamics of the homogenous phase, or dissolved pollutants. It has become increasingly clear that accurate description of the environmental distribution and accumulation of solids-associated pollutants, of the impact of these pollutants on the food webs of the aquatic environment, and of their toxicologic implications to. man must take account of the dynamics of transport of suspended solids.
At the commencement of this research, an acceptable approach to modelling the transport of suspended solids in water systems was lacking. Conventional approaches dictated the use of a bed-load function. However, while Einstein (1950) found that particles smaller than 62 microns, commonly called "wash-load" comprised an insignificant mass portion of the bed, and concomitantly an insignificant portion of the bed-load in alluvial streams, Weber, et al. (1979) found that montmorillinite clays of median diameter approximately 0.7 microns demonstrated a greater capacity for the adsorption of PCB's than kaolinite clays of median diameter approximately 25, microns, and that very small organic sediments likewise exhibited large adsorption capacities. These findings suggested that particles less than 62 microns in diameter may not be insignificant from the pollutant transport perspective.
The wash-load/bed-load distinction becomes critical when the case of a river discharging into a bay is examined. Due to a decrease in water velocity, greater deposition of suspended solids occurs as smaller particles are able to settle out. It is commonly assumed that in slow moving waters, the silt and clay particles completely settle out and stick to the bottom; the equilibrium condition is therefore defined as a zero concentration. Partheniades among others, has demonstrated this to he incorrect, as equilibrium concentrations of suspended sediments in flumes do not always reach zero. There is evidently a continuum of situations, based on particle and channel characteristics, in which particles can be expected to either reach the bottom and stop, reach the bottom and continue moving, or resuspend tack into the fluid. one of the major aspects of this research was to develop a formulation to describe this continuum.
The second major aspect of this research was to develop a method to describe the settling behavior of small, cohesive types of materials. Approaches based upon classical equations fail due to the inability to precisely define the required parameters. The method developed is stochastic in nature and is independent of said parameters.
The nature of environmental systems modeling is such that a true assessment of success can only be made by applying laboratory verified submodels to actual natural systems. Hence the two-part layout for this dissertation. Part One entitled "Conceptual Model Development" consists of the development of an overall system modeling approach describing the transport dynamics of suspended solids and associated pollutants. Part Two entitled "Case Study, Ford and Belleville Lakes System," describes the application of the conceptual model to construct an overall system model describing the suspended solids-coliform bacteria transport in the Ford Lake-Belleville Lake system.

Iman, E.H.H. 1981. M.A. Sc. "Numerical Modelling of Rectangular Clarifiers".

A numerical model has been developed to simulate the settling process of discrete particles in rectangular clarifiers operating at neutral density conditions. First, the stream function-vorticity version of the equations of motion in the 'conservation form' are solved numerically to establish the velocity field in the clarifier using a constant eddy viscosity turbulence model. Then a transport equation is solved for the spatial distribution of suspended solids concentration. When the steady-state conditions are reached, the concentration distribution yields the desired removal rate of the clarifier.
The numerical model employs a finite-difference scheme in which the unsteady term of the transport equation is replaced by a three-time level approximation, the convective terms are approximated using a weighted upwind-centred difference formulation, the diffusion terms are differenced using 'second-order' accurate expressions. The resulting algebraic finite difference equations are then solved using the Alternating Direction Implicit scheme. The computational domain is discretized using a variable size mesh to reduce the number of required nodes, without sacrificing the high resolution required in regions where flow gradients are expected to change rapidly. A partial slip boundary condition is used for the clarifier bottom.
The proposed numerical model was verified and numerically tested prior to its calibration and was found to be stable and convergent to the "exact solution". A truncation convergence criterion was derived and confirmed by numerical experimentation. The dominant factors in selecting the mesh size and time increment were the local Courant and grid or cell Reynolds number. The same two factors were also found to control the computational stability. Sensitivity analyses were carried out to investigate the effects of bottom and baffle-lip boundary conditions, entrance velocity distribution, degree of upwinding, eddy viscosity and initial conditions on the ultimate steady-state solution. The hydrodynamic submodel was found to be satisfactory in reproducing the main flow features and a reasonable agreement was observed between predicted and measured velocity fields.
An unsteady version of the transport submodel was used to simulate the flow-through characteristics of a neutral density tracer. A fair agreement with the experiment was observed but the need for a more sophisticated turbulence model is indicated. The steady-state transport submodel was successful in simulating the secrete suspended solids. The model was applied to predict .the removal rate of a non-uniform size mixture of discrete particles in a hypothetical tank. The simulation results were consistent with the results of Camp and Hazen. The model was also used to investigate the effect of scour and relative baffle submergence on the solids removal.

Lee, K.-B. 1981. M.Sc. M.Eng. "Development of a Versatile Storm Water Quantity Model".

The objective of this study Is the development of a new storm water drainage model which allows the user increased flexibility with respect to the choice of methods or mechanisms used to simulate the. various components of a watershed model. In order to achieve this goal, four currently popular models such as Rational Method, ILLUDAS, SWMM and HYMO are studied in great detail. Also, previous research is reviewed with respect to the following subjects: (1) Selection and/or Synthesis of Design Storm (2) Estimation of Rainfall Runoff and/or Rainfall Loss (3) Computation of Overland Flow Hydrograph (4) Routing of Flow through Drainage Systems A new model is developed on the basis of study results of other models and previous research.
The model consists of six program modules called RAINFALL. DATAGEN, OVLAND, PROFILE. ROUTE, and REPORT. Each module has Its own unique function. RAINFALL generates various design storm hyetographs. DATAGEN creates and/or modifies watershed data. OVLAND estimates the overland flow hydrograph. PROFILE computes the profile (s) of water level along the drainage system. ROUTE performs the routing of the inflow hydrograph. Finally, REPORT allows the user to select the format of the output results. The model is versatile in terms of: (1) Unit system (either metric or imperial units can be used). (2) Methodology of computation (e.g. 5 types of design storms, 2 ways of estimating effective rainfall, 6 different Resistance Laws. etc.). (3) Applications (it can be applied to both rural and urban watershed). (4) Modelling watershed (i.e. it has an ability to model any type of drainage system including reservoirs, bridges, culverts, streams, sewers, etc.). (5) Output format (the User can select the format). The model has been tested on two monitored watersheds, i.e. the Oakdale Catchment, Chicago. Illinois. and the Malvern Catchment, Burlington, Ontario. The hydrographs produced by the model are in agreement with the measured hydrograph.
Since the model is written in ANSI FORTRAN in a structured programming fashion and requires only 64K core memory, it can be implemented by any medium size computer where a FORTRAN compiler is available. However, minor modifications may be necessary with respect to Input/Output file manipulation.

Watson, M.D. 1981. M.Sc. "Appication of ILLUDAS to Stormwater Drainage Design in South Africa".

The Illinois Urban Drainage Area Simulator (ILLUDAS) is evaluated for the purpose of stormwater drainage design in South Africa. The historical development of the model is described and published verification results are reviewed. The principles of the model are described, as also are the features of the computer program which was made available during the early part of 1980 by the Illinois State Water Survey (ISWS) .
ISWS recommendations for parameter were tested with rainfall-runoff data from two local catchments and in general found to be adequate. Improvements to the method of accounting for grassed area losses in sub-catchments having long flow paths were found to be necessary.
Problems associated with design inputs are discussed and use of the Chicago design storm is demonstrated. A modified metric version of the program is presented and tentative recommendations are made for design parameter estimation.

Burn, D.H. 1982. M.A.Sc. "Optimization Techniques Applied to Water Quality in an Uncertain Environment".
Methodologies are presented to apply linear and nonlinear optimization techniques to the solution of water quality planning problems wherein uncertainty is of concern. The principal sources of uncertainty for the management problems investigated are considered to be embodied in transfer coefficients which describe the pollutant transport in the water body. The uncertainties present are characterized using first-order uncertainty analysis.
The linear optimization technique employed is chance constrained programming in which probabilistic constraints are replaced by their deterministic equivalents. A form of duality theory was invoked to manipulate the problem into a format amenable to the application of chance constrained programming.
The nonlinear optimization techniques include a minimize variance model and fractile programming. The intent in using the nonlinear programming techniques is to provide an alternative interpretation of the stochastic aspects of the problem. The alternative formulations also allow a comparison of the resulting solutions with those obtained by the linear programming technique thus providing a means of evaluating the implications of choosing a particular model formulation.
A total of three case studies are examined, two of which involved the interaction between biochemical oxygen demand (BOD) and dissolved oxygen (D.O.) concentration, while the third is concerned with coliform bacteria. The case studies are used to verify the assumption that a lognormal distribution is appropriate for the characterization of the distribution of the transfer coefficients. As well, the case studies provide an illustration of the trade-offs amongst the many facets of a multiobjective water quality management problem.

Cheung, P.W.K. 1982. M.A. Sc. "A Standard Hydrograph Method for the Preliminary Analysis of Stormwater Management Projects".

Current practices of runoff computations in urban drainage designs employs different levels of analyses. At the first level or preliminary stage, manual or desk-top methods are used and at the master plan and detailed design levels, complex computer models are used. Although the sue of desk-top methods at the preliminary stage of the design seems appropriate, a situation is being created under which inconsistent and incompatible results are obtained when compared with the computer simulations used in the final stage of design. A review of presently applied desk-top methods for drainage designs, including the Rational Method, the Modified Rational Method Hydrograph, and the SCS TR-55 Method shows that they have many limitations and weaknesses. Runoffs estimated by these various desk-top methods are inconsistent and are not compatible with the final design model. In order that the urban runoff control criteria can be properly defined, a standard method for all runoff computations must be sued. This research was imitated by the need for such a standard and desk -top method for the preliminary design of urban drainage systems and stormwater control facilities.
The first part of the research was carried out to investigate the possibilities of interfacing the commonly used manual Rational Method with the complex computer model SWMM for designing the pipe system. An extensive assessment of the Rational Method for peak discharge estimations using real and conceptual watersheds was carried out. It is concluded that the Rational Method gives results significantly different from the more complex computer model. Although a methodology is developed by which the rational method peak estimations can be adjusted, it is concluded that the Rational Method is basically an inadequate method to sue for the present state-of-the-art of urban drainage design.
The second part of the research is therefore devoted to the development of an improved desk-top method entitled the Standard hydrograph Method (SHM). The method developed is based on computer simulations using design storms. It is simpler to apply than the Rational Method but is provides both peak discharges and runoff hydrographs. The SHM procedure can be based on any design storm distributions. An example is given to develop a set of SHM relations using the computer model SWMM and Keifer and Chu's design storm distribution because they are common methods of design in Canada. As a further demonstration of the applicability of the SHM procedure, the SCS 24-hour Type II distribution design storms are also used.
A selective runoff control criterion is proposed instead of the zero runoff increase criterion. It is demonstrated, on the basis of a study conducted with a real rural watershed, that the latter is not only impractical but is also difficult to implement. The SHM procedure is next extended to develop two desk-top methods for the estimation of detention storage for stormwater runoff control. The first method can be used for estimating volumes in conventional designs; the second method can be used for estimating volumes in a "dual storage" design which is presently being employed in new developments.
While most of the research is based on homogeneous watershed, it is proposed that for the preliminary analysis of non-homogeneous watershed desk-top programs for use on programmable calculators can be interfaced with SHM for routing hydrographs through channel and storage reservoirs.

Hamam, M.A.M. 1982. B.Sc. (Honour), M.A. Sc. "Transition of Gravity to Surcharged Flow in Sewers".

This Thesis deals with transients that occur when gravity flow is suddenly changed to pressure flow by the occurrence of a surge in the form of a travelling hydraulic jump in the line. The pressure head fluctuations associated with this transient have been studied. The transition is complicated by the mixture of air and water in the pipe. The transients in this two-phase air-water flow have been considered in this study. Some of the factors affecting these pressure transients are: pipe size, pipe shape, flow velocity, Froude number, relative depth of flow, alignment of the pipe, pipe material, venting arrangements and boundary conditions such as pumps, interceptors, and drop pipes.
A mathematical model is developed to describe the mechanics of surcharging of a sewer; the model predicts the surge velocity, the surcharge at the surge front relative to sewer crown, the water level in the sump well or manhole, the velocity of the water surface in the sump well, the distance travelled.by the surge front and, the transient pressure rise in the air-water twophase flow.
The mathematical model results are compared with the experimental results from the present study. The mathematical model results are in general agreement with the experimental data but the experimental pressure fluctuations were approximately 50% of the theoretical predictions. The random nature of the instability of the air-water interface and the extent of the blockage contributed to disagreement between the theory and the model prediction.

Kassem, A.M.A. 1982. Ph.D. "Development and Application of A Simultaneous Routing Model For Dual Drainage Systems".

A comprehensive mathematical model (Dual Drainage Simultaneous Routing Model DDSRM) has been developed for the design analysis of urban storm drainage systems considering infrequent as well as frequent storms. The model is based on the dual drainage principle (major minor systems). Surface runoff is routed simultaneously in the two interconnected networks formed by the storm sewers system and the streets. Special arrangements proposed for the design of the dual drainage system, and considered in DDSRM, include limitation of flow access into the sewers (inlet control) for the purpose of limiting sewer surcharge, and separate storage facilities for the minor and the major systems (dual storage). DDSRM has several options which allow the flexibility to handle a variety of design conditions. Flow charts for various sub programs and a user's manual have been provided.
The application of DDSRM for the design of dual drainage storage systems is demonstrated through an example for a residential subdivision. Input data and computer output for the numerical example are explained. The model has also been utilized to investigate the voice of inlet efficiency and spacing in the design and operation of dual drainage systems and recommendations for the practical implementation of inlet control have been derived. DDSRM is also suitable for application in master drainage studies.
Another aspect of this study is the development of improved routing methods for unsteady flow under both free surface and pressurized (surcharged) conditions. For free surface flow, a numerical solution scheme for a specially weighted finite difference formulation of the kinematic wave model, with discharge dependent wave celerities, has been developed for flow routing in the sewer and street networks. Sewer surcharge simulation is based on a modified Hardy Cross method, incorporated into an existing dynamic routing model for sewer networks. The validity of the routing sub models of DDSRM has been investigated through several numerical experiments. In all, the experiments conducted (with slopes greater than 0.1 %), the simplified routing method for free surface flow has been found to give results (time-variation of flows, peak flows, time to peak and mass conservation) which are not significantly different from those obtained by the complete solutions of the Saint Venant equations for unsteady gradually varied flow. The method for the simulation of sewer surcharge has been verified under transient and steady surcharge conditions. It has been found to predict surcharge levels and durations, and sewer flows with reasonable accuracy for practical purposes.
DDSRM, lie other urban drainage models, does not simulate energy losses at the sewer junctions, nor can it simulate hydraulic jumps. The kinematic wave routing and design sub-model cannot simulate backwater effects. However, the surcharge sub-model is capable of representing backwater in the sewer network. The model also assumes a continuous major system and cannot simulate systems with low points where water ponding may occur.

Nix, S.J. 1982. Ph.D. Analysis of Storage/Release Systems in Urban Stormwater Quality Management.

Urban stormwater runoff acts as a transport medium for a variety of wastes. This phenomenon can cause water quality problems and, as a result, corrective measures sometimes are required. Storage/release systems, consisting of a storage basin and a release mechanism, are often used for this purpose. Unfortunately, there are few data to assess their performance in this role and the techniques currently used to design and analyse these systems are generally inadequate. A comprehensive and usable approach, reflecting the dynamic aspects of this problem, is presented in order to fill this need.
This dissertation develops and presents the necessary theories and evaluation techniques to assess the long-term performance of storage/ release systems, the use of the production function to summarize system performance, and several techniques to determine the most cost-effective designs. The focus is on long-term analyses rather than "design storm" or single-event approaches. Several Available computer simulation and statistical models capable of evaluating long-term performance are discussed. However, these models are either too inflexible or simple to be generally applicable. To meet the need for a more sophisticated and flexible model, a computer simulator known as the Storm Water Management Model Storage/Treatment Block is introduced. The production function, a concept taken from economic theory, is used to summarize the relationship between long-term performance and the characteristics of the system. The properties of production functions, as well as several possible mathematical representations, are also discussed. Simple analytical and graphical optimization techniques, which require the information contained in the production function, are presented and developed for use in the determination of cost-effective designs.
A hypothetical case study is used to demonstrate many of the techniques and procedures, illustrate the generally expected results, and suggest an overall methodology for analysing storage/release systems. Emphasis is also placed on the results produced by the Storage/Treatment Block and its sensitivity to changes in the assumptions governing system behavior.

P'NG, C.E. 1982. M.A. Sc. "Conceptual Hydrologic Modelling for Master Plans in Urban Drainage".

In recent years government agencies in Canada have developed policies to implement "stormwater management" techniques in master drainage plans which are designed to regulate watershed development in a manner compatible with watershed needs. In order to do this, the post-development flows need to be assessed long before the detailed design is done. The general objective of this thesis is to develop a planning model for urbanizing watersheds in the framework of master drainage plans. The model should be capable of determining the outflow hydrograph for the post-development watershed using only the limited data available at the planning stage without any calibration.
Lumped models are appropriate where the outflow hydrograph-is required at only a few critical points in the watershed. HYMO, a lumped model which is very popular in Canada for planning purposes, due to its facility of application, is found to underestimate the peak flows for urban watersheds.
A literature review indicates that various conceptual models such as the single linear reservoir (SLR) model have been used by several investigators for urban watersheds. However there has not been a systematic assessment of these lumped models as compared to the detailed simulation using design models under Canadian design conditions. Therefore a methodology is developed using mathematical watersheds to compare the performance of conceptual models under different conditions such as the size and the imperviousness levels of. the urban watersheds. Based on the objective function of minimizing the differences in the peak flows (as compared to those obtained by detailed simulation), the 'optimal' parameters of these conceptual models are obtained by means of Rosenbrock's hill-climbing technique. They are then compared with the parameters predicted by various relations. The one parameter of the-SLR model which is the storage coefficient K has been related to watershed and storm characteristics by means of multiple regression analysis by some researchers while others have related it to the time of equilibrium.
From the analysis of eight relations it is observed that the single linear reservoir model using a relation derived by Desbordes or one derived by Sarma, Delleur and Rao predicts reasonably accurate peak flows for the larger storms for watersheds up to 100 acres (40 ha) in size. A version of this model has also been programmed on a calculator.
For short duration high intensity storms used in design, there is a significant contribution of rainfall excess from the pervious or undeveloped areas. The approach taken in some models is to simulate separately the runoff hydrographs from the impervious and pervious portions of the watershed. Therefore a model consisting of two linear reservoirs in parallel is proposed here and tested.
It is found that the 2 parallel linear reservoirs (2PLR) model is a better tool than the SLR model. For planning purposes the model can predict acceptable peak flows for the whole range of storms tested for watersheds ranging from 2 to 300 acres without calibration. The storage coefficients of the 2 reservoirs are related with the input rate i, length L, slope S and surface roughness n using relations. based on the kinematic wave theory. The model computes the losses in the pervious areas by means of the cumulative form of Horton's infiltration equation and has been programmed as a ready-to-use tool in the HYMO structure. The optimization subroutine with which it has been interfaced can be used for the calibration of model parameters where rainfall-runoff data is available.
Tests with several storms on two residential watersheds, the Gray Haven watershed in Baltimore, Maryland and the Malvern watershed in Burlington, Ontario have validated the performance of the 2PLR model for predicting the outflow hydrographs from urban watersheds.

Southerland, E. 1982. Ph.D. A Continuous Simulation Modeling Approach to Nonpoint Pollution Management.

The purpose of this research project was to demonstrate the types of analyses that should be performed with continuous simulation models in order to maximize their usefulness in evaluating the effectiveness of nonpoint source control strategies and the feasibility of meeting certain water quality goals. A simulation model of a 479 square kilometers suburban Washington, D.C. watershed was used to compare receiving water quality responses to various land use patterns and urban nonpoint pollution management strategies. Individual on-site storage-treatment models were used to simulate the nonpoint pollution removals achieved by best management practices (BMPs), such as multipurpose detention basins and infiltration trenches, which were assigned to all new single family residential, multifamily residential, commercial and industrial land uses projected to occur between 1980 and 2005. Site models were linked with a basinwide continuous simulation model, a predecessor of the EPA Hydrological Simulation Program Fortran (HSPF), in order to predict receiving water impacts of land uses and BMPs. The effectiveness of high, low, and minimal capacity BMPs was based on achievement of the following receiving water goals: 1) allowing no violations of water quality standards or freshwater aquatic life criteria for designated constituents; 2) maintaining the same constituent concentration frequency distributions for the year 2005 land use pattern as for totally forested, undisturbed conditions; and 3) maintaining the same concentration-frequency distributions for the year 2005 land use patterns as for existing 1980 conditions.
Analysis of receiving water goals involved comparisons of constituent concentration-frequency curves simulated for existing land use or forested conditions with those simulated for future land use patterns with and without nonpoint source controls. Biochemical oxygen demand, dissolved oxygen, ortho-phoshate phosphorus, total phosphorus, total nitrogen, lead and were included in the comparisons. Water quality index-frequency curves were also computed from synoptic hourly time series of dissolved oxygen, total phosphorus, lead and zinc concentrations simulated for each day. Index-frequency curves indicated the probability of synoptic water quality criteria violations for the four constituents. Nonparametric multiple comparison tests of daily index values provided a quantitative comparison of receiving water responses to each land use/BMP strategy.
This research effort demonstrated that assessments of nonpoint source control strategies require the concentration-frequency information that only continuous simulation models can provide. For the first time, it was also shown that linkage of individual on-site storage-treatment models with a receiving water model is the preferred method of predicting the effects of these controls on the quality of free-flowing streams. This linked model system revealed that detention basins and infiltration controls may produce more frequent l-hour and 96-hour exceedances of water quality criteria than the uncontrolled watershed due to the capture of highly concentrated first-flush runoff followed by delayed release during dry weather flow periods. Just as hydrologic simulation models are necessary to predict the synergistic flooding effects of runoff controls, this research project found that water quality simulation models are needed to predict the combined effects of stormwater BMPs on downstream water quality.

Hamilton, D.L. 1983. M.Sc. "Microcomputer Simulation of Canal Operation ".

A computer model for the simulation of canal operation was developed. The model is applicable for any non-branching canal system composed of a series of channels separated by control structures. The programming language used was BASIC, a language suitable to most microcomputers. The Delta Field Division Canal of the CalifornJa Aqueduct was used for an example simulation and verification. A comparison of data computed by the model to observed data indicated generally good agreement.

Goodrich, J.A. 1983. Ph.D. Predicting Toxic Waste Concentrations in Community Drinking Water Supplies From Upstream Industrial Discharges: A Vulnerability Analysis.

In February, 1978 the Environmental Protection Agency (EPA) proposed amendments to the National Interim Primary Drinking Water Regulations to deal with the control of chloroform and synthetic organics in drinking water. Under the proposed rules, those utilities that could prove they were not vulnerable to the above parameters would be granted a variance from those regulations that would Increase the cost of service. By virtue of the variance procedure, EPA has concluded that there are variations in vulnerability. In November, 1979 the regulation regarding synthetic organics was dropped because of the difficulty in defining, let alone assessing a community's vulnerability.
This study used QUAL-II, a water quality simulation model, to systematically study many diverse fields of science In a spatial context. Never before have such varied data an organic chemistry, epidemiology and fluid dynamics been brought together and linked areally to produce a rational methodology capable of closing the 'loop' between water pollution control strategies and public health. First order decay coefficients were derived and utilized in QUAL-11 to more accurately simulate the priority pollutants rather than being limited to the traditional water quality parameters usually modeled by QUAL-II in the past. Thus, the main contribution of this research has been the development of an interactive systems analysis approach that can have an impact on public policy.
QUAL-11 simulated toxic waste concentrations in drinking water supplies along the Ohio and Kanawha Rivers between Charleston, West Virginia and Cincinnati, Ohio. The simulated toxic waste concentrations for each community were compared to health guidelines to assess which communities were more "vulnerable" to upstream discharges of industrial waste. The following results are only illustrative of the insight the methodology can provide and should not be considered definitive.
Seven of the ten communities analyzed in the case study area were found to have simulated concentrations exceeding water quality guidelines for taste and odor. All of the utilities were vulnerable to various toxic and/or carcinogenic pollutants. All the communities were vulnerable to spills in addition to the typical daily discharges of pollutants.
The most important factors to consider in identifying vulnerable communities are: flow characteristics of the source of supply, potency and persistence of the pollutants, amount and timing of discharge of pollutants, storage times of utilities and relative location of point sources and community intakes.
There is a great deal of sensitivity between the above factors. Contrary to conventional wisdom, one cannot assume that the most downstream utility faces the greatest risk and therefore be the most vulnerable. The total number of discharges upstream from a utility is not the proper criteria to judge vulnerability. Vulnerability depends on relative locations of intakes and outfalls and flow and persistence )f the pollutants. Thus, it seems that the issue of community vulnerability is not a clear-cut issue of looking for the most downstream utility or modeling a few pollutants at average flow. Very detailed information about flow probabilities, pollutant characteristics, discharge flow and location are required for identifying those communities at risk.

Lee, C.J. 1983. Ph.D. A Method for Estimating Human Risk Due to Toxicants.

A definition of risk in terms of probability function is evaluated and a statistical method is presented for estimating risk. The definition of risk is formulated on the response of well defined population of individuals when subjected to a frequency of dosages of a specified toxicant. The method posed facilitates the evaluation of the accuracy and precision of risk estimates and relates value characteristics with effort or cost in terms of sample size required. Particular implications are discussed in measuring interaction, synergistic and antagonistic efforts, between toxicants.

Patry, G.G. 1983. Ph.D. Real-Time Forecasting of Water Quantity and Quality in Urban Catchments.

Storm-related pollution in general, and combined sewer overflows (CSO) in particular, have been found to contribute significant amounts of pollutants to receiving water bodies, including large quantities of oxygen consuming materials, heavy metals, and petroleum hydrocarbons. Because of the costs involved in controlling pollution from CSO, real-time automatic control of combined sewer systems is now being considered as a possible alternative to the more traditional forms of pollution control. Water quantity and quality forecasts are of critical importance in the successful implementation of any integrated real-time CSO control strategy. Accordingly, the purpose of this investigation is to provide a comprehensive analysis of water quantity and quality forecasts in urban catchments, and to test the proposed forecast models on actual rainfall-runoff events. Both statistical and simple conceptual models are developed based on deterministic and stochastic concepts. Given the availability of real-time data, adaptive error-correcting algorithms are derived and applied to the selected data bases.
The report is divided into four major sections. Section I provides the necessary preliminary information on the need and approaches to real-time water quality and quantity forecasting. Section II is devoted to the problem of urban runoff forecasting while Section III deals with water quality forecasting. Finally, Section IV contains the conclusions of the study and recommendations for future research.
Several urban runoff forecast models are examined, including a nonlinear functional runoff (NLFR) model, a simple hydrograph model, and a group of statistically-based models.
A two-way calibration-verification analysis is developed to provide reliable estimates of the structural characteristics of the NLFR model, including the memory of the system and the order of expansion of the first- and second-order kernels. The NLFR model is subsequently solved recursively for use in real-time runoff forecasting. The accuracy and convergence rate in model parameter estimates are examined under both noise-corrupted flow measurements and time varying runoff conditions. A two-stage NLFR model is formulated and applied to the Curotte-Papineau catchment along with the other NLFR model forms.
A second group of urban runoff models is developed on the basis of a two-stage simple hydrograph model in which first-stage deterministic runoff forecasts are generated through a linear, time-invariant triangular hydrograph model. The second-stage model consists of an adaptive error-forecasting algorithm designed to improve the first-stage forecasts. Parameter estimates of the second-stage model are evaluated recursively at every time step.
A group of statistically based models, of the ARMAX type (autoregressive moving average with erogenous input), are also developed. The accuracy in parameter estimates is examined by using a variety of recursive parameter identification (RPI) algorithms under both input and output noise conditions (white and colored noise models). The time-varying tracking capabilities of the various RPI algorithms is also investigated. The models are applied to the Curotte-Papineau catchment by using a selection of potential model structures under different RPI algorithms.
In general, the best short-term runoff forecasts are obtained by using the adaptive-recursive models, with the two-stage simple hydrograph model ranking first, followed by the ARMAX and the two-stage NLFR models. For longer lead times, the best forecasts are generated by the calibrated NLFR model in simulation mode.
In Section III, water quality forecast models are developed for both suspended solids (SS) and chemical oxygen demand (COD). The modeling strategy followed closely that developed for the flow segment of the study. Two real-time water quality modeling approaches are investigated, including conceptual and statistically-based models.
Statistical models of the ARMAX type are investigated for SS and COD. For both water quality parameters, the analysis lead to ARMAX models of order (2,0,0). In all cases, ARMAX model parameters are evaluated recursively at every time step.
Finally, a two-stage deterministic-stochastic pollutant washoff model is formulated in which a first-order surface pollutant washoff model is used to generate the first-stage water quality forecasts. The second-stage model consists of an adaptive error-forecasting ARMA model based on the observed sequence of first-stage forecast errors. Time-varying model characteristics are incorporated into the algorithm through exponential weighting of the first-stage forecast errors. Both the statistical and the two-stage conceptual models performed relatively well for short-term (5 to 15 min.) forecasts, with mean absolute percent errors ranging from 13 to 24% for SS, and 11 to 23% for COD. In general., the statistical models were found equivalent or superior to the two-stage conceptual model, in terms of both accuracy and computational characteristics.

Qaisi, K.M. 1985. Ph.D. Uncertainty Analysis of Stream Dissolved Oxygen Models.

Uncertainty analysis of models representing water quality systems is a relatively new area of research in environmental engineering. As a consequence, a limited amount of information is available with respect to uncertainty analysis of aquatic system. The present research is an uncertainty analysis applied to a particular case study. Several unique concepts are investigated in the present study and these concepts are applied to a case study. Some of these concepts have been developed from information found in the literature. Some of the ideas have been developed from personal discussions with several associates and analysis after an extensive survey of available literature. As it often happens, unique areas of research require extensive feedback in the form of new ideas and concepts, which form a strong base for further development. Several unique concepts have been developed in the present research that form a new conceptual framework which can serve as a basis for further research.

Gallagher, D.L. 1986. Ph.D. The Application of Data Based Transformations to Parameter Estimation in Water Quality Models.

This research has investigated the usefulness of incorporating power transformations in parameter estimation for water quality models. A Box-Cox power transformation of both sides of the model was performed on a variety of finite section water quality models, using both synthetic data and data taken from existing case studies. The resulting parameter estimates were compared to ordinary least squares estimates. The results showed that significant improvements in the estimates and model forecasting ability were achieved when the observed data were heteroscedastic or skewed. The transformation method was quite robust in the sense that little deterioration in the estimates or forecasting ability was seen when the transformations were included but little heteroscedasticity or skewness was present. The variances of the transformations were generally small for single equation models but increased substantially for coupled system models. Significant improvements were still seen, however, even when large variances were noted.

Sircar, J.K. 1986. Ph.D. Computer Aided Watershed Segmentation for Spatially Distributed Hydrologic Modeling.
Physically based hydrologic models that simulate streamflow in terms of watershed characteristics are important tools in the water resource decision making process. The parameters for physically based models in current use are defined as averages for the watersheds involved. A logical step in improving the quality of the simulated streamflows would be the development of practical, spatially distributed models that are capable of simulating the hydrologic consequences of physical variations within watersheds. A key element in spatially distributed hydrologic simulation is the interdependence between runoff and a three-dimensional flow conveyance network defined by the topography of the watershed. While remote sensing can provide an adequate definition of land cover and, in some cases, even the soil type and soil moisture, an efficient means of defining elements of the topographic network and related catchment areas must be available if practical versions of spatially distributed hydrologic models are to be realized. If the models are to be available for general use on larger watersheds, this topographic analysis must be computer assisted using digital format data extracted from hard copy maps.
There are a number of commercially available scanning systems that can digitize or "capture" the location of black contour lines on a hard copy topographic map to produce a computer compatible image. The objective of the present dissertation is to use this binary image to: 1) define the elevation, slope and aspect of any point on a map surface; and 2) delineate those catchments that drain into any user defined channel segment or point along the channel system.
The technique developed is a hardware/software system that uses a graph-theoretic approach to represent and manipulate raster scanned digital contours in a computer. Segmentation of the watersheds is accomplished through the development of a set of "expert heuristics" that simulate the manual operations of sub-basin delineation on maps. The designed system includes components that: 1) label the digitized contour traces output from a scanner with an elevation attribute; 2) create a corresponding digital matrix of elevations that describe the spatial variation of topography; 3) register digitized drainage networks to the elevation database using a geographic information system framework; and 4) delineate the basin or sub-basins that contribute runoff to user defined points or reaches along the channel network. The performance of the system is tested on a set of topographic scenarios representing a variety of terrain complexity and levels of detail in watershed segmentation. A potentially important practical use of the system is demonstrated through the digital definition of a time-area curve for a small watershed.
The resulting system is computationally intensive and requires the unique features of several different types of computers for efficient implementation. A key objective of the dissertation is to allow the practicing hydrologist to perform the topographic analysis without having to become an expert on the intricacies of computer operations. This objective was accomplished by the development of a prototype "hydrologic workstation". The database management and communications software developed for this workstation allow the hydrologist to use an off-the-shelf personal computer to integrate an array of sophisticated computer systems for the analysis of complex problems such as digital terrain modeling. In addition to running the hardware/software terrain analysis systems developed as part of the dissertation, the performance of the workstation was tested as a component in the National Aeronautics and Space Administration's Pilot Land Data System that allowed the user to access and use a number of remote computers, including Goddard Space Flight Center's Massively Parallel Processor.

Wilson, L.C. 1986. Ph.D. A Continuous Spaciotemporal Stochastic Model for Short-Time-Increment Precipitation.

A spaciotemporal stochastic precipitation model is developed which is capable of simulating precipitation intensity sequences on a continuous basis over an area. The model is based upon five years of breakpoint precipitation recorded by thirteen rain gages over an area of approximately 40 square miles in central Missouri. Data analysis techniques are developed which interpret the spatial intensity patterns to compute a number of raincell parameters including, speed, direction of movement, size, and peak intensity. Temporal characteristics such as cell duration, interarrival time, and cell life cycle are also extracted. Non-cellular precipitation is segregated and modeled using the Matalas multisite generating model.
Model parameters are derived from only precipitation intensity measurements, using no additional meteorologic observations. Nevertheless, agreement between computed cell parameters and published radar-based interpretations indicates that the data analysis techniques are reliable. Peak annual intensities from a forty-five year simulation of continuous precipitation intensities agree with published results for durations of 10, 30 and 60 minutes.
This model, coupled with a digital runoff model which allows spatially distributed precipitation inputs, is suitable for research into the effects of spaciotemporal precipitation on numerous hydrological studies.

de Roulhac, D.G. 1987. M.S. Application of Computer Graphics in the Selection of Rainfall Frequency Models for Environmental Engineering.

This thesis describes graphical techniques in rainfall frequency analysis, and the theory and operation of a personal computer program which implements these methods. The proper choice from among several possible frequency distributions may be determined by visual inspection of annual rainfall maxima plots on the computer display. The criteria for choosing the best frequency distribution is described herein. This thesis provides computer-generated graphical and analytical tools for updating rainfall intensity-duration-frequency (IDF) curves. Such updates are now needed in much of the Western United States. Rainfall data from Tucson and Walnut Gulch, Arizona were used to test the computer programs. The application of rainfall IDF relationships to drainage design and environmental modeling is demonstrated by the use of the U. S. Environmental Protection Agency's Stormwater Management Model.

Dilks, D.W. 1987. Ph.D. Analysis of Water Quality Model Uncertainty Using a Bayesian Monte Carlo Method.

Mathematical models of aquatic systems are used with increasing frequency to determine water quality management strategy. It is now recognized that these models can have a large degree of uncertainty (error) associated with their projections. This uncertainty is caused by a combination of two factors. First, water quality models are only simplistic representations of the real world, and error can be introduced by using a model framework that incompletely describes the true system. Second, model coefficients -also called parameters- cannot be determined exactly and are also a source of uncertainty. The uncertainty in model results caused by uncertain model parameters is often large enough to render the predictive capability of a model useless. These errors have rarely been considered in actual practice, primarily because there is no widely accepted method for quantifying them.
Several techniques have been used in an attempt to quantify the magnitude of model projection uncertainty as its significance has become apparent. Four different techniques have received the majority of this effort. They are: 1) Sensitivity Analysis, 2) First-Order Error Analysis, 3) Monte Carlo Analysis, and 4) Kalman Filtering. Each technique has been shown to contain significant limitations, both theoretical and practical, which restrict its applicability in determining model uncertainty. Monte Carlo analysis exhibits the greatest potential for widespread use of the four techniques, but still contains significant limitations. The primary limitation of the Monte Carlo technique as currently applied consists of the lack of information available to determine input distributions for uncertain model parameters.

Johansen, N.B. 1987. Ph.D. Discharge to Receiving Waters From Sewer Systems During Rain".

Abstract: The project described in this report is divided into two major parts: (1) Measurements. (2) Model development. The measurements included simultaneous recordings of rain, runoff and transport of pollution for three selected catchments with areas in the range of 5-20 ha. The major results from the measurements were: (1) The hydrological processes can for practical purposes be modelled with a simple linear equation. (2) The permeable surfaces gave only an insignificant part of the runoff volume during the measuring period. (3) The time variation of the flow could be computed with an accuracy compatible with the flow measurements by the kinematic wave model ILLUDAS. (4) The quality of the water was fairly similar for the two catchments with combined sewer systems, and the water from the separate sewed catchment was less polluted indicating that the washout of deposits in the pipe-system is a major source of pollution from the combined system. A model was developed especially for the purpose of computing the discharge to receiving waters from the two types of sewer systems (combined and separate). The main philosophy in developing the model was that it should be based on historical rain series. By making statistics on the computed overflow series, the appropriate loading values for the different types of receiving waters could be obtained. The model is based on the time-area principle, modified with respect to an automatic transformation of the time-area curve for each rain event in the series. The hydraulic and hydrological model was supplemented with a simple model for pollutional runoff. Testing of the model, named SAMBA, against a kinematic wave model (ILLUDAS), shows that SAMBA overestimates the overflow volume on a mean basis by 3%. Comparison of the computed annual overflow volumes to measured annual overflow volumes shows that the measured volumes are within the ranges predicted by SAMBA. As SAMBA does not include all phenomena in the runoff process, it is recommended to supplement the SAMBA calculations by a dynamic wave model. These computations should be based on selected rain events. A selection procedure for these events is described.

Lansey, K.E. 1987. Ph. D. "Optimal Design of Large Scale Water Distribution Systems Under Multiple Loading Conditions".

Abstract: Much work has been done in efforts to develop procedures to design least cost water distribution systems. All of the previously published models, however, have drawbacks in one of the following three areas; size of system considered, number of loading conditions analyzed, or the types of components designed. All of these limitations are overcome through problem reduction using reduced gradient techniques and incorporating a network simulator into a nonlinear model. This allows large systems to be analyzed under a number of loading conditions. Along with selecting the optimal pipe sizes the model can analyze the other major components which appear in a distribution system. The model can determine optimal valve settings for both control and pressure reducing valves. The model also has the capability to examine a sequence of demands, either the daily variation of demands or a fire load with specified duration, which permits a true analysis of tanks something no previous model could perform. By employing a nonlinear solution procedure realistic cost functions can be used as well as including the site variation effect on cost since each component will have an independent term corresponding to it in the objective function. Finally, the model can aid in the selection of pumps and provide guidance for the operation of a pump station.

Melching, C.S. 1987. Ph.D. A Reliability Analysis on Flood Event Forecasting With Uncertainties.

A general procedure was developed for applying reliability analysis to consider the uncertainties in hydrologic models used in forecasting and prediction. The reliability analysis approach is capable of considering all sources of modeling uncertainty -- data, model parameters, and model structure -- to produce estimates of the probability that specific hydrologic target levels will be exceeded due to actual, design, or simulated rainstorms. Proper selection of these target levels allows estimation of interesting hydrologic information including assessment of model reliability, supplementation of real-time flood forecasts with flood level exceedance probabilities, etc.
The general procedure describes how reliability analysis may be incorporated with calibrated or physical simulation (non-calibrated) models. However, in this study, the approach was demonstrated only for the case of two calibrated models applied to real-time flood forecasting. In such cases, the data and model parameter uncertainties are assumed to be represented by the variation of the individual event calibration results about the parameter best estimates. The model structure uncertainty is accounted for by a multiplicative correction factor, which is the ratio between the measured peak discharge and peak discharge of the best fit hydrograph. Statistics are estimated and distributions fit for each of the basic variables (the model parameters and the model correction factor). The various reliability analysis methods (Monte Carlo simulation, mean value first-order second moment, and advanced first-order second moment) are then employed to evaluate the reliability of the real-time forecasting schemes utilizing the estimated statistics and distributions. A case study demonstrated the utility and potential benefits of the general procedure. Based on the case study, the following conclusions are made regarding the performance of the various reliability analysis methods; the reliability of the specific hydrologic models used and hydrologic modeling, in general; and the sources of modeling uncertainty. 1. A comparison of system reliability analysis methods indicates: a. Monte Carlo simulation may be used to analyze hydrologic model uncertainty and flood likelihood, but its high computer time requirements make it impractical for real-time forecasting at this time. As computer capabilities increase, the 1,000 to 10,000 simulations necessary for the Monte-Carlo method may be obtained inexpensively and efficiently (in about 30 min.). b. The AFOSM method provides estimates of model uncertainty and flood likelihood comparable to those from Monte Carlo simulation at a considerable savings in computer time. c. The MVFOSM method provides reasonable estimates of model uncertainty and flood likelihood at the decision (accuracy) level. d. The nonlinearity assumptions of the MVFOSM are not valid for the analysis of uncertainties in rainfall-runoff models. e. At this time, the AFOSM method appears to be the best method to assess the reliability of rainfall-runoff models and real-time flood forecasting schemes. 2. Often forecasts of low practical usefulness are "good" stochastic forecasts. In such cases, the forecast variance (uncertainty) is high, indicating reduction of this forecast variance via decreasing the modeling uncertainties is the key to obtaining forecasts with high practical usefulness. 3. "Rules of thumb" may be developed by considering typical forecast reliability to detect likely flood events and aid in flood warning decision making. These "rules of thumb" work fairly well even for the case of somewhat unreliable real-time flood forecasting schemes. 4. The bulk of the forecast uncertainty is due to the uncertainties in the estimation of the rainfall excess via the initial loss-continuing loss rate abstraction scheme. The rainfall data inadequacies apparently make up the majority of these uncertainties. 5. The reliability of forecasts and predictions currently produced by hydrologic models using data from just a few rain gages is questionable at best. 6. The hydrograph parameters and model (structure) correction factor contribute relatively little to the forecast uncertainty. 7. The common assumption that distributed and quasi-distributed models provide greater accuracy and reliability relative to simpler linear, lumped system models is only justified when the input data is sufficiently distributed. For cases where the available input data is sparse, simple linear, lumped system models will suffice and are preferred.

Paulfalcone, J. 1987. M.Eng., P.Eng. "MIDUSS: The Development of an Interactive Model for the Design of Stormwater Drainage Systems ".

Considerable effort has been devoted to defining various hydrologic and hydra processes involved in urban hydrology. Many of the currently popular programs which model the complex rainfall-runoff process with storm water management are intended primarily for analysis and simulation of the drainage system. It would appear that there is a definite discrepancy between the state of modelling art and the design of storm water "stems as practiced in many municipal engineering offices.
This report describes the development of an interactive design program MIDUSS which facilitates the design of conveyance or detention facilities in a drainage network. The model is command driven and allows the design to proceed in a traditional manner by allowing the user to define individual subcatchments to generate overland flow hydrographs. The resulting hydrographs may be manipulated and used to design pipe and pond elements in a drainage network. The model is interactive and allows the user to experiment with alternative trials before processing the final design.
As the design proceeds, a file records the commands, design decisions and data which are input by the user. The design file may he used in subsequent design sessions during which the user may switch from manual (i.e., keyboard) to automatic (i.e., file) input. This allows continuation of a previous design session, testing of a previous design using a different storm, redesign of a specific element, and simulation of a drainage system. The user can monitor the progress of the design at each step and take corrective action in the event of bad input data or an incorrect design decision.
The model employs methods of hydrologic prediction that represents a significant advance over the rational method but do not put unreasonable demands on the designer for data acquisition. The MIDUSS model was developed using a dynamically varying response function which depends on a kinematic wave relationship. The dynamic response function is considered to be non-linea r and synthesizes hydrographs much more accurately both in peak flows and temporal distribution.
The model was originally implemented on a CDC CYBER 170 mainframe computer. More recently, the model was compiled using WATFOR77 and MICROSOFT FORTRAN to run on a microcomputer.

Pitt, R.E. 1987. Ph. D. "Small Storm Urban Flow and Particulate Washoff Contributions to outfall Dishcarges (Ontario, Canada, Milwaukee)".

Tarboton, D.G. 1987. M.Sc. "Hydrological Sampling: A Characterization in Terms of Rainfall and Basin Properties".

This thesis considers the sampling of rainfall and runoff processes both in time and In space and links the sampling problem to basin and rainfall characteristics. Sampling strategies are defined by the number of rain gages, rainfall measurement interval and flow measurement interval. The effectiveness of different sampling strategies is measured by the variance of the error In estimating either the volume or peak of streamflow. This Is related to the rainfall and basin rainfall-runoff properties through parameterizations of these processes. Several rainfall parameterizations, including stationary and non-stationary event based models are used. Runoff from rainfall is parameterized In terms of the fluvial geomorphology of the basin. Linear filtering techniques are used to compute the variance of the estimation error for different sampling strategies. The results are given In the form of quasi-general design aids, which can be used to select appropriate sampling options when doing network design.

Yim, C.S. 1987. Ph. D. "Development and Testing of Granular Filter Design Criteria for Stormwater Management Infiltration Structures (SWMIS)".

Baffaut, C. 1988. Ph.D. "Knowledge Techniques for teh Analysis of Urban Runoff Using SWMM (Storm Water Management Model)".

Abstract: An expert system was built to facilitate and to automate the calibration of the Runoff block of the Storm Water Management Model (SWMM). The expert system is divided in two parts that can be linked together. The first one calibrates the runoff quantity simulation model. The second, that calibrates the runoff quality simulation module, has two versions. One calibrates the model with the only objective of a correct prediction of the pollutant loads. The second calibrates the model on the basis of the loadographs shapes. The expert systems act as a front end to counsel the user on the choice of the parameters. It interprets the results using backward chaining and suggest some useful changes in the values of the relevant parameters using forward chaining. The calibrations achieved by means of the expert systems has been tested on several watersheds. It has been found that, for runoff quality simulation, the accuracy of the prediction results is limited by the quality of the calibration data that are used and by the model itself that does not take into account all the processes involved in the transport of pollutant by urban runoff. A fuzzy set approach is proposed to try to take into account some of the uncertainties related with runoff quality simulation. A simplified fuzzy set version of SWMM is proposed that has been tested on two watersheds, giving some encouraging results.

Duan, N. 1988. Ph.D. "Optimal Reliability Based Design and Analysis of Pumping Systems for water Distribution Systems".

Abstract: Little has been reported in the literature on reliability analysis, particularly the optimal reliability based design of pumps and storage tanks for water distribution systems. This dissertation presents the development of new methodologies for the reliability analysis of pumping stations and the optimal reliability based design and analysis of pumping systems for water supply, including the optimal design of the pipe network. The reliability analysis technique is based upon frequency and duration analysis using a supply model, a demand model, and a margin model. Using pump characteristic curves, the methodology models a continuous-time Markov process, using bivariate analysis and conditional probability approaches in the modifying traditional frequency and duration analysis. A computer code called RAPS has been developed to solve the methodology. A reliability based optimization model has also been developed which is aimed at the following goals: (a) der programming model solved in an optimal control framework. The conservation of flow and energy constraints are solved implicitly for each iteration of the nonlinear optimization procedure using the KYPIPE hydraulic simulation model and the reliability constraints are solved using the RAPS program. The nonlinear programming program is solved using the generalized reduced gradient code, GRG2. Findings of computational results based on the model developed include: (a) optimal solutions obtained from traditional optimization techniques may not be reliable; (b) the number of pumps and tanks have considerable effect on water supply system reliability; (c) reliability requirements have significant impacts on system costs, and (d) the overall reliability based procedure tends to provide optimal pipe networks with more and smaller pipes.

Holub, L.G. 1988. Ph.D. "Uncertainty Modeling of Water Distribution Systems".

Abstract: Pipe roughness, total demand, and nodal demand were considered random variables in models of three municipal distribution systems and several elementary networks. The propagation of uncertainties in these input parameters was evaluated by the Monte Carlo method or direct error analysis and resulted in distributions of available discharge for a fixed pressure, residual pressure for a fixed discharge, pressure heads, and head losses. Data from field and laboratory tests of pipes demonstrated statistically significant variations of Hazen-Williams coefficient and Colebrook absolute roughness from constant values for a given pipe. However, both equations were shown to be adequate for modeling distribution systems. Output uncertainties based on pipe roughness for single pipes were larger than those for simple networks and were reduced as the number of pipes in series or parallel increased. Uncertainty of head loss increased with an increase in the head loss. For municipal systems, uncertainties due to nodal demands, except for large concentrated demands, were approximately 10 percent of those due to total demand or the assignment of roughness values to individual pipes. High roughness uncertainty produced coefficients of variation of approximately 10 percent for available fire flows at fixed residual pressures of 20 psi. Pressure head standard deviations were a maximum of 5 feet for a system with a maximum day demand of 2080 gpm and increased to 9 feet with the addition of a fire demand. For a large system supplied by pumps with steep head-discharge relationships, total demand uncertainty was more significant than roughness uncertainty and produced a coefficient of variation for available fire flow twice that of the total demand coefficient of variation. For systems with total demands and fire flows of approximately equal size and supplied by pumps and an elevated tank, roughness uncertainty based on a single of values for all pipes was more significant than demand uncertainty. The use of three roughness zones represented a moderate increase in pipe roughness knowledge and was sufficient to reduce available fire flow coefficients of variation to under 5 percent.

Rivera Santos, J. 1988. Ph.D. A Methodology for Parameter Estimation in Conceptual Precipitation-runoff Models with Emphasis on Error Analysis.

This study addresses the problems of the automatic parameter estimation in conceptual precipitation-runoff models. With the recent publication of hundreds of hydrologic models, the user confronts a hard task when the model is to be fitted to a specific watershed. It is not doubted that to obtain acceptable results from a hydrologic model, a suitable calibration process has to be sought. The problem faced by the casual model user is that there is not an established methodology by which the user may be guided through this complicated process.
The methodology suggested in this work emphasizes on residual analysis and presents various diagnostic techniques, commonly used in regression analysis, to identify and quantify the errors and their sources. Attenuation methods are also presented. Special attention is given to input errors, model structural errors, and parameter estimate errors. A discussion of the different stages of the calibration process is developed, specifically on the selection of the optimization algorithm, the error structure, convergence criteria, calibration and verification data, and sensitivity analysis.
The methodology is applied to a case study using data from the City of Boulder's watershed. The results showed that with the appropriate diagnostic techniques, the identification of random and systematic errors reduces to the inspection and evaluation of plots and some statistics. Knowing what kind of errors are present in the data, a suitable method of calibration can be devised. One important feature of this methodology is the examination of the response surface of the model's parameters in two- and three-dimensional plots. Using this technique, initial and final parameters estimates can be assessed as well as the sources of other related problems such as interdependence between parameters.

Corsi, R.L. 1989. M.S., Ph.D. "Volatile Organic Compound Emissions from Wastewater Collection Systems".

This dissertation was completed to assess the potential for volatile organic compound (VOC) emissions from wastewater collection systems. Work focused on two important phenomena; 1) partitioning of VOCs between sewer atmospheres and wastewater, and 2) convective transport of VOCs by gas exchange with the ambient atmosphere. Tracer studies were completed in operating sanitary sewers., resulting in data used to evaluate methods for predicting VOC partitioning. An oxygen transfer model was suggested, with theoretical-based adjustments for VOCs. Several mechanisms were studied as driving forces for ventilation of sewer atmospheres. Fluid mechanical analyses were applied to hypothetical flow circuits, and the importance of individual mechanisms were compared. It was concluded that no single mechanism dominates sewer ventilation under all situations, but environmental conditions, wastewater flow conditions, and physical characteristics of a collection system were identified under which some mechanisms become dominate. A set of one-dimensional, two-phase finite volume models were developed to predict dynamic mass transport of VOCs in collection systems. Over one thousand hypothetical scenarios were simulated to ascertain the importance of physical characteristics of sewers, flow characteristics, and physico-chemical properties of VOCS. Results provide guidance in identifying potential emission hot-spots, and the relative importance of collection systems compared to wastewater treatment systems.

Holbert, S.B. 1989. M.L.Arch. "Development of a Geographic Information System Based Hydrologic Model for Stormwater Management and Landuse Planning".

Hotchkiss, R.H. 1989. Ph.D. Reservoir Sedimentation and Sediment Sluicing: Experimental and Numerical Analysis.

Reservoir sedimentation and sediment sluicing are simulated in laboratory and numerical experiments.
The laboratory experiments are conducted in a 12.2 meter long, 0.38 meter high, 15 centimeter wide flume. The flume was later modified by adding a 2 meter expanding section at the downstream end. A sluice gate extending completely across the flume width was installed about 9 meters downstream from the entrance for the first set of experiments. Later this sluice gate was replaced with a simulated dam 1.5 meters into the expanding section. The dam was fitted with three sluice gates, each 0.15 meters wide. All gates were manually operated; each gate could be set separately.
All experiments were performed with lightweight sediment consisting of crushed walnut shells with a mean diameter of 0.67 millimeters, gradation coefficient of 1.3, porosity of 0.53, and specific gravity of 1.35.
Steady water discharges ranged from 1.5 to 4.25 liters per second and sediment feed rates (input at the upstream end of the flume) ranged from 50 to 570 grams per minute.
Upon lowering the sluice gate(s) more than about fifty per cent of the flow depth, all incoming sediment deposited near the flume entrance in response to the rise in water surface. Deposition continued until the flow depth at the entrance decreased sufficiently to produce adequate velocity for sediment transport. This process formed a delta, characterized by a steep foreset front with a slope near the submerged angle of repose. Sediment was transported over the delta lip and deposited on the foreset slope, thus moving the delta forward in the flume. Near uniform depth and slope were maintained upstream of the delta lip. The delta stopped moving forward near the sluice gate, and all incoming sediment passed over the delta lip and under the sluice gate(s). At this time the flow field in the flume appeared exactly as before lowering the sluice gate(s), except that the bed had risen as much as the water surface and a delta was in place in front of the sluice gate(s).
Upon raising the sluice gate(s) after a delta had reached equilibrium, the water surface dropped at the dam and sediment transport was increased significantly. The drop in water surface and increase in sediment transport proceeded upstream, causing upstream progressing degradation of the bed. This process continued until a new equilibrium condition was established, consisting of uniform flow with a water surface parallel to the previous condition but lower by an amount equal to the drop in water surface at the dam.
A one dimensional numerical model was developed for aggradation and sediment sluicing in reservoirs. The quasi-steady, one-dimensional St. Venant equations were used to represent the flow field; the sediment continuity equation supplemented by a form of the sediment transport equation due to Parker (1976) described sediment movement. All computations were corrected for flume sidewall and bedform influences. The equations were solved in a decoupled fashion by first computing a backwater profile and then adjusting the bed elevations accordingly. Subcritical flow is maintained at all times in the model. In the numerical model, the delta was simulated by vertical shock face, fitted to the bed at a location where the steepening bed slope exceeded the original slope by at least five times. The shock face was moved forward each time step an amount equal to the volume of sediment passing the face divided by the cross-sectional area of the face. Computed water surface and bed profiles compare very well with the experimental data. The computed shock face moved forward 27 per cent faster than observed, due to a 22 per cent higher-than-observed computed sediment rate for the flume.
Sediment sluicing for the flume was simulated by lowering the water surface at the downstream boundary of the model in discrete steps so as not to cause supercritical flow. Agreement with observed experimental results was good.
The model was applied to the Granite Falls Dam on the Minnesota River as an example of field case application. The dam was first closed in 1871 and was completely full of sediment from 20 to 58 years later. Model results with three different lengths of backwater zones predicted complete sedimentation in 35 to 48 years from closure. Sluicing was modeled by initially drawing down the water surface at the dam to within one per cent of critical depth. Subsequent water surface elevations were lowered as much as the bed at the dam was eroded, thus preserving the new near-critical flow condition at the dam. Results showed that the bed near the dam was eroded to pre-closure conditions in less than three months, and that the upstream progressing degradation was nearly complete after about ten years of simulation.
The equations of motion for gradually-varied flow were examined for steady state equilibrium conditions, and algebraic equations were developed describing the response of the streambed to changes in channel width. The equations, applicable to a rectangular channel, include inertial and pressure effects, but do not account for sidewall or bedform influences.
Equations were also developed for the case of a rectangular channel with a sinusoidal perturbation in width. The first order solution of the perturbation expansion shows the bed response lags the width perturbation as a function of sediment transport and Froude number. As an example of the effects such a channel produces, the equations were applied to the Minnesota River for a low discharge, the mean annual discharge, and bankfull discharge. Perturbation amplitude was equal to ten percent of the uniform channel width. Results show that the location of maximum bed scour migrates downstream for increasing discharge. Specifically, the location of maximum scour for bankfull discharge is at the narrowest section of the channel. This result explains the field observation that scour problems appear to be worst at narrow channel sections where most bridges, crossings, and stream gage stations are located.

Moys, G.D. 1989. Ph.D. "The Prediction of Rainfall Excess in Urban Catchments".

Abstract: Available from UMI in association with The British Library. Urban drainage models are examined in terms of model purpose, type and usage. Model usage for the prediction of flooding behaviour in the UK adopts the design storm approach, which assumes that flood frequency equals rainfall frequency. This equation is shown to be affected by antecedent moisture conditions and rainfall spatial variability. Data collected from three small urban catchments (0.3 ha, 0.7 ha and 3.2 ha) in Southampton are used to calibrate and verify three urban drainage models: namely WASSP-SIM, WALLRUS-SIM and USGS-DR$\sb3$M. A probabilistic approach to model verification is used and demons the neutron probe method are then used to calibrate and test a number of simple soil moisture models including the soil moisture algorithm incorporated in the MORECS procedure and that included in the DR$\sb3$M. These models are found to predict catchment average soil moisture conditions to within $\pm$20 mm, using local rainfall information and standard evapotranspiration curves as inputs. Using the MORECS soil moisture model it is shown that antecedent soil moisture conditions do have an impact on observed percentage runoff values, but that the relationship is confused by seasonal factors. The soil moisture model is then used together with WASSP-SIM to derive discharge-frequency curves using a continuous simulation approach. Comparisons with the design storm approach confirm the assumptions of the approach, although do not prove that either approach produces the correct answers.

Thompson, D.B. 1989. Ph.D. Determining Parameters for a Continuous Simulation Model by Estimation, Synthetic Calibration and Analytic Calibration.

The U. S. Geological Survey's Precipitation-Runoff Modeling System (PRMS) was used to compare results of modeling daily and storm event runoff using estimated parameters, synthetic parameters and analytic parameters.
Estimated parameters for daily-mode simulations were determined by comparing predicted monthly runoff series with the average monthly runoff series from four or five basins (other than the study basin). Mean monthly runoff and the cumulative deviation from mean monthly runoff were used as objective functions for estimating parameters. Analytic parameters were determined by executing a combination of autocalibration and manual adjustment.
Estimated parameters for storm-mode operation of PRMS were determined by "guessing" at appropriate parameter values based on published studies, the PRMS user's manual, and engineering judgement. Synthetic calibration was executed by using regional regression equations for storm peak discharge and storm runoff volume to generate pseudoprobability distributions of peak discharge and runoff volume. Model storm-mode parameters were adjusted to minimize deviations between model-predicted and regression generated pseudo-probability distributions of storm peak discharge and storm runoff volume. Analytic parameters were determined by autocalibration using measured storm events.
Verification was based on a time series and storm events not used for calibrating model parameters. For two of the three study watersheds, use of estimated parameters resulted in better daily simulations than did use of analytic parameters. Although mean storm runoff could be duplicated, runoff from individual storm events was subject to large disparities. Furthermore, simulation of peak discharges was very poor. The model was incapable of satisfactorily executing a storm-mode simulation.

Zheng, S. 1989. Ph.D. Stochastic Modeling of Urban Drainage Systems".

Abstract: During wet weather flow events, bypasses and overflows of raw sewage may occur in many municipalities. It is generally accepted that combined sewer overflows will cause water quality problems such as low dissolved oxygen, lake eutrophication, beach closures, fish kills, and impaired aesthetics. To abate these problems from the existing wastewater collection and treatment facilities, it is worthwhile to consider a computerized control scheme within a wastewater collection and treatment system. The majority of water quantity-quality models describing urban drainage systems developed to date are deterministic. They require considerable data for calibration, verification, and sensitivity analysis. This will limit their applications since most of urban watersheds do not have enough data available. Stochastic models of sewer flows and quality parameters have also been researched. Most of the efforts were based on the application of Box-Jenkins models, which require a tedious three-stage indentification process. This makes the models unsuitable for automatic control. Furthermore the transfer function model identification procedure suggested by Box and Jenkins are not feasible since the rainfall series containing many zeros, can not be "prewhiteninal nonlinear estimation problem is converted into a linear model so that it can be easily estimated using a linear multiple regression technique. A new method for determining the unit hydrograph of an urban drainage system is also developed. The stochastic sewer flow and quality models developed in this study can predict flow and quality values adequately, and can be easily implemented on computers with less external interference. The methodology developed for the unit hydrograph determination may have a wide range of applications such as infiltration-inflow analyses, dry-weather wet-weather flow separation, real time flow forecasting, quality analyses, and classic hydrologic design of storage and conveyance systems.

Abbott, H.A.1990. Ph.D."Permeability and Suction Characteristics of Compacted Unbound Aggregates (Pavement Drainage)".

Abstract: Available from UMI in association with The British Library. Requires signed TDF. The experimental work reported in this thesis was concerned mainly with the study of permeability and suction characteristics of continuous gradings within the Department of Transport Type 1 sub-base grading limits. Three materials, a crushed dolomitic limestone, a sandy gravel and a crushed granite, were selected to cover a range of particle shapes and textures. The maximum particle size was 37.5 mm. A practical experimental technique was developed for measuring the saturated horizontal permeability of compacted aggregates. Darcy's law was only valid at low hydraulic gradients, typically less than 0.01. The horizontal coefficient of permeability was greater than 10$\sp{-3}$ m/s for all the specimens tested. The repeatability of these tests was equivalent to a ratio of 4. The compacted aggregates were anisotropic with respect to permeability. Vertical permeability appeared to be one and a half orders of magnitude less than the horizontal permeability. Estimates of permeability were obtained from a number of empirical correlations, based mainly on particle size, but sometimes including other material properties. These widely differing estimates were compared with the measured coefficients of permeability of nine aggregate/grading combinations. The best correlation was that of Sherard et al (1984), but even this tended to underestimate horizontal permeability by between half and one and a half orders of magnitude. The aggregate suction characteristics of the materials were determined using the capillary rise/fall method and an osmotic suction technique. Curves fitted through the desaturation data were used to estimate the pore size distributions of the materials. The saturated and unsaturated permeabilities of the compacted aggregates were estimated from the desaturation curves using three theoretical capillary models. Relative permeability calculations indicated that a 10% decrease in degree of saturation would result in a reduction of permeability of at least half an order of magnitude. An assessment was made of the theoretical permeability requirements of pavement drainage layers and drainage layers in embankments. For typical applications, a coefficient of permeability of 10$\sp{-3}$ to 10$\sp{-2}$ m/s was required for good drainage. In addition, a survey was conducted to ascertain current British practice in the specification and measurement of permeability. The theoretical permeability requirements were compared both with the measured coefficients of permeability and with typical specified values of permeability.

Au-yeung, W.K. 1990. Ph.D. "Flow Routing in the Parallel Pipes of a Combined Sewer System".

Abstract: Available from UMI in association with The British Library. A parallel pipe model has been constructed and successfully applied to a twin pipe sewer system. The model predicts flow and level for each of the parallel pipes and deals with the complex behaviour which occurs at cross-connections. The model DUPPERS is an enhancement of the single pipe simulation model DUCTS developed at Dundee Institute of Technology. It has been rigorously tested using a variety of artificial parallel pipe system configurations and input discharges. Obtained results showed that the model was operating satisfactorily. The model was applied to the Lyneburn system in Dunfermline, Scotland, with three models being constructed: two sub-models and a full system model. Comparison between observed and predicted hydrographs for both flow and level were in close agreement. Percentage differences for peak flowrates, runoff volumes and levels were all within $\pm$20% of observed and simulated values. A review of flow routing procedures, above-ground hydrological models and the major commercial sewer simulation packages is also included. In addition to the tailored parallel pipe model, single combined pipe models for the system based upon the commercial package WASSP and the in-house model DUCTS were also constructed. The simulation outputs from these two models were found to be close to the observed flow data. However, the lumped pipe models only predicted the combined flow for the twin pipe outfalls. The construction of these models verified that the Sewered Sub-Area model could be applied to the study catchment. Furthermore, the model DUCTS was shown to perform satisfactorily. Catchment data such as contributing areas, sub-catchment type and paved areas used in the lumped pipe models were subsequently used in the parallel pipe model. The data collection techniques are described in detail. A sequential flow logging procedure was used and found to be an effective and economic method of data collection, especially where a limited number of loggers was available. The model was successfully used to examine the storm runoff for the Lyneburn parallel pipe system and also the complicated hydraulic behaviour in cross-connections. The numbering system for the parallel pipe networks, the chosen equations to represent the different flow patterns for cross-connections, together with the level computation procedure form the major enhancements incorporated into DUPPERS.

Brion, L.M. 1990. Ph. D. "Methodology for Optimal Operation of Pumping Stations in Water Distribution Systems".

Abstract: The determination of optimal pump operation in water distribution systems is an extremely important problem for municipal engineers. An optimal pump operation policy refers to the scheduling of pumps that will result in the minimum operating cost for a given set of operating conditions. More specifically, the optimal operating policy for a pump station denotes a set of rules or guidelines that dictates when a particular pump or group of pumps should be operated over a specified period of time. Also of extreme interest is the real-time operation of water distribution systems for which the model developed herein could be used. The proposed solution methodology for solving the pump operation problem uses the optimal control and augmented Lagrangian methods to reduce the large complex problem to one that is easily solvable by the theory of reduced gradients. The resulting nonlinear optimization problem is solved by a two-level optimization technique wherein at the outer level, the Lagrange multipliers and penalty weights are determined and at the inner level, the decision variables are solved by a nonlinear programming algorithm. If some stopping criteria is not met at the outer level, the augmented Lagrangian terms are updated and a new inner level optimization is performed. This process continues until the stopping criteria in both levels are satisfied.

Buck, D.R. 1990. M.S. "Evaluation of a Percolation Basin to Remove Heavy Metals and Particulate from Highway Runoff..

Abstract: An evaluation of a percolation basin method, to remove heavy metal pollutants from highway runoff was conducted after an initial basin design failed. An analysis of the cause of failure was performed in an effort to produce a design that would remove the heavy metal pollutants. The failure resulted from the effects of clay colloids collecting on the basin surface and clay migration out of the basin soil which collected on the fabric mat which protected the collection drains. Clay percentages averaging 66 and 25% indicate reduced permeability of the basin and the resulting failure. Laboratory analysis showed that Zn, Pb, Cr, and Cd were present in the runoff waters. However, the pollutants were not in suspension, but were attached to the colloidal particulate present in the runoff. Column studies showed that pollutants attached to particulates could be removed by filtering the runoff water through sand.

Abstract: Accumulation of heavy metals (Cu, Pb, Zn) in bottom sediments of detention ponds receiving highway runoff had been identified as a potential source for groundwater contamination. To assess this environmental impact, six detention ponds located in Florida were selected to investigate heavy metals accumulation in their bottom. Core sediment samples were collected from each pond, frozen and divided into six layers. The metal content after digestion with concentrated nitric acid and extractable metal by Toxicity Characteristic Leaching Procedure (TCLP) was determined for each sediment layer. Most of the metals appear to concentrate at the top 10-15 cm of the bottom sediments. The fractions of extractable metals using TCLP exhibited exponential decline with increasing fine particles less than 0.062 mm diameter and organic content. It may be feasible to remove the accumulated sediments saturated with metals from the bottom of ponds for periodic maintenance. An existing model was used to simulate the migration of metals in the sediments. Batch sorption studies and column experiments were conducted to determine the metal transport parameters required to apply and verify the proposed model. The calculated dispersion coefficients of heavy metals through sediments were four orders of magnitude smaller than the measured tracer (NaCl) dispersion coefficients in the same sediments. Predicted concentrations obtained from the transport model appeared to approximate measured values in the laboratory.

Xu, C. 1990. Ph.D. "Optimal Operation and Reliability Evaluation of Water Distribution Systems".

Abstract: Available from UMI in association with The British Library. This dissertation is concerned with various aspects of the optimal operational control and reliability assessment of water distribution systems, including real time demand forecasting, leakage reduction via valve control, optimal valve location and analytical evaluation of network reliability. The problem of real time demand forecasting for water systems is investigated. Time series analysis techniques for the establishment of forecasting models are described. Application of these techniques to develop models for a water distribution system is presented. The actual performance of the models in a real time environment is assessed. Leakage reduction via optimised valve control is investigated. A methodology for the determination of control valve setting is proposed. A solution method based on successive linear programming is described. Two water systems, one of which is a realistic 75 node network, are used to evaluate the potential reduction of leakage. The problems of valve location and its implication on leakage control schemes are examined. Two methods are proposed to identify the best valve location over the network. The first method formulates the problem as a mixed integer program and the second is a heuristic search strategy. The problem of reliability evaluation for water distribution networks is studied. Various methods for reliability assessment are reviewed. A microscopic model for analysis of a water distribution system is developed. A methodology to assess the reliability of a distribution is suggested.

Yan, J. 1990. Ph.D. Parameter Estimation for Multipurpose Hydrologic Models.

Two parameter estimation methodologies for multipurpose hydrologic models have been investigated and developed. The results lead to the following conclusions: 1.The theory for Parameter estimation of multipurpose hydrologic models has been established. 2.This research has demonstrated that the parameter estimation of multipurpose hydrologic models can be improved in many cases using methodologies other thin those developed for single purpose hydrologic models. 3.The estimated parameters of multipurpose hydrologic models can be improved using a weighted multi objective function with MOP methods.

Gonzalez, G.S. P. 1991. Ph.D. "Bedload Transport in Storm Sewers Stream Traction in Pipe Channels".

The hydraulics of sediment transport in part-full pipes was investigated in an experimental study. The sediment bed was permanently deposited and the type of transport observed and measured was bedload transport exclusively. The experiments were performed in a concrete pipe, 225 mm in diameter and 23 m long. Two sand sizes, D50 = 0.9 mm and 2.5 mm, were used. Two sediment bed thicknesses were tested, namely 0.2 D and 0.4 D (D being the pipe diameter). Pipe slopes ranged from 0.002 to 0.006.
Two complementary series of experiments were conducted at Hydraulics Research Ltd. in Wallingford and at The University of Newcastle upon Tyne, both in England. Different pipe diameters, sediment sizes and sediment bed thicknesses were tested. This additional information allowed the present study a wider range of application.
Three procedures, commonly used in this kind of experimental study, were tested to corroborate the validity of the proposed methods: 1) the vertical velocity distribution near the sediment bed can be described by a logarithmic law referred to as the law of the wall 2) the side wall elimination procedures of Einstein and Vanoni-Brooks can be used to compute the hydraul s of the sediment bed with an accuracy which is acceptable for engineering purposes, and 3) the critical shear stress of the sediment particles can be obtained by using Shields' diagram with good approximation.
The methods of Engelund-Hansen and van Rijn for the prediction of flow resistance in alluvial channels were used to compare their predicted values with the experimental data. Both methods gave approximate agreements. A procedure which can be used as a design method for the conditions of the present study was proposed.
Bedform dimensions (height and length) were also measured and compared with the predictors given by the methods of Fredsoe and van Rijn for the prediction of bedform geometry in alluvial channels. The agreement was poor and the methods have to be modified for pipe-channel conditions.
A relationship to estimate sediment transport rates, based on dimensional analysis, was proposed. This relationship is expressed in terms of both flow and particle parameters as well as geometric (shape) factors. The range of application for pipe diameter (D) is 154 mm to 225 mm, for sand size (D50) is 0.5 mm to 2.5 mm, for pipe slope (S) is up to 0.006 and for sediment thickness (t) is up to 0.4 D.
Five available equations for bedload transport computation were used for comparison with the proposed relationship. The equations of Novak & Nalluri and Einstein & Brown, in which transport is exclusively a function of shear stress, gave better estimates than the equations of Meyer-Peter & Muller, Engelund & Fredsoe and van Rijn, which are based on the excess shear.
Both the procedure for the prediction of flow resistance and the proposed relationship for bedload transport estimation can be used in the analysis of storm sewer systems. Further experimental work with larger pipe diameters, mixed gradings, different levels of sediment and unsteady flow conditions is needed before these methodologies can be fully incorporated in a simulation model for sediment transport in sewers.

Graham, E.I. 1991. M.A.Sc. "An Urban Runoff Infiltration Basin Model".

Abstract: A two-dimensional finite element saturated-unsaturated flow and mass transport model has been developed and is coupled to a surface routing model to simulate the transient operation of an urban runoff infiltration basin. Single and multiple runoff events, and extended-period simulations have been performed to study the variations in seepage and pollutant transport. The infiltration rate is significantly affected by the location of the water table. The antecedent soil moisture condition determines the amount of storage volume available in the unsaturated zone. If the pore spaces below the basin are saturated at the onset of an inflow event, the groundwater mound reaches the basin floor and there is a significant decrease in the infiltration rate. The mass transport of chloride indicates a seasonal behaviour of the contaminant beneath, and beyond the basin. The results of the copper simulations indicate that there is a potential for soil-water contamination but further characterization of the metal-soil interactions are necessary before the uncertainties are eliminated.

Li, J.Y. 1991. Ph.D. Comprehensive Urban Runoff Control Planning

With increasing concern for the deteriorating water quality of urban water bodies, planning for urban runoff control should simultaneously consider both quantity and quality management. This thesis presents a methodology for urban runoff quantity and/or quality control planning. The methodology provides a systematic framework for preliminary screening of urban runoff control systems and narrows the range of control alternatives for subsequent design level analysis. It consists of five steps: input rainfall/pollution analysis, assessment of existing runoff/pollution conditions, evaluation of existing control system performance, least-cost analysis of control systems, and sensitivity analysis. The quantity and quality of urban runoff is dependent upon rainfall and catchment and drainage system characteristics. Storm event analyses are conducted to determine the probability distribution and statistics of rainfall characteristics (event rainfall volume, duration, interevent time and annual number of rainfall events) of rainfall series. To model runoff quality, a constant concentration approach is proposed in which runoff pollution load is equal to the product of runoff volume and an expected pollutant concentration. Analytical probabilistic models are then derived to predict long-term average runoff volume and pollutant load controlled. The most promising combination of control measures for achieving a certain level of runoff controlled can be determined by optimizing different kinds of control measures and comparing the cost-effectiveness relationship among different combinations of control measures. Sensitivity analyses are then conducted to assess the impact of uncertainty in parameter estimation on the analysis results. A catchment in the Borough of East York, Toronto, is used to demonstrate the application of the methodology.

Rezek, S.E. 1991. MS. "The Effects of Urbanization on a Portion of the Portage Creek Watershed: A Case Study on Surface Water and Groundwater Quality".

Abstract: The surface water and groundwater quality in the study area of the Portage Creek watershed was analyzed to determine the possible effects of urbanization. Dry weather and wet weather surface water and groundwater samples were collected from June 1986 to November 1986 at selected stations in the watershed. Water samples were analyzed for 26 inorganic and organic parameters. Most concentrations were below ana lytic method detection limits. The analytical data indicate that sodium and chloride are the principal contaminants resulting from urbanization. The probable source of the contamination is deicing salt applied to the City of Portage, Michigan streets. Overall, the surface water and groundwater quality in the study area was good during dry weather sampling events for the selected parameters. Wet weather sampling event results revealed temporary quality degradation to the creek from urban runoff via the Portage Consolidated Drain No. 1.

Seo, D.-I. 1991. Ph.D. Optimal Complexity Analysis of Total Phosphorus Models.

'Eutrophication' occurs when there persists excessive primary productivity in a lake. It is known that phytoplankton is strongly regulated by the presence of phosphorus in most of lakes. Total phosphorus concentration is often used as an indicator of nutrient level for algal species in lakes. Therefore, prediction of total phosphorus concentrations and control of its sources have been primary concerns in the study of lake pollution problem. There have been number of attempts to identify relationships among various components in lakes. However, since they are simply abstractions of the real systems no model is exact or perfect. Therefore, it is necessary to find the systematic way to examine the performance, reliability and errors of phosphorus models.
In this thesis, eight different total phosphorus models are analysed and discussed. The conditions of model use are suggested to help future modelers to chose an appropriate model for a lake. It was possible to fit a model to the system response data with parameters outside the normal ranges. Models with improper mechanism may look reasonable for specific application. However, these models may be misleading in other applications such as lon-term predictions. It is important that the equations be constructed in compliance with the principle of natural and mathematical sciences and coupled with an investigation of the scientific literature, field studies, and experimental research.
In real world, there exist uncertainties in the model parameters and the observed data. These uncertainties can be caused by errors in measurement or by environmental conditions. Many techniques have been developed in the context of specific applications to estimate reliability of models but no single method is suitable for all models. It is also possible that there exists uncertainty in field observed data. However, no attempt has been made to include this effect in model reliability analysis.
In this dissertation, a new approach was developed which considers uncertainties in the model parameters and in the filed observations along with penalty function which is greater for a more complex model. The penalty function was obtained using the concept of the standard of the estimate.
Model errors are not only function of model parameter errors but also function of sensitivity of a model with respect to model parameters. Therefore, the net errors in phosphorus concentration are the combined results of these two. The errors of a model are also functions of model structures. Therefore, it is no great importance that model structure be considered in error and reliability analyses.
However, the judgement on optimal order of complexity should be based on all available information. The extended validation approaches are also necessary for stochastic conditions. Whenever possible, a simpler model is preferred to minimize effort and cost. Draper pointed that if all selection procedures are applied with common sense they can produce useful results; applied thoughtlessly, and/or mechanically they may be useless or even misleading.

Young, B.C. 1991. M.Sc. "Dynamic Modelling of Combined Sewer Systems with Radial Gates and Siphon-Weirs".

Abstract: The purpose of this thesis is to present a hydraulic sewer routing model for simulating combined sewer systems with special regulating structures, such as dynamic radial gates and self-priming siphon-weirs. RGEXTRAN includes all features of the dynamic wave routing Extended Transport (EXTRAN) model, which simulates branched or looped systems, backwater effects, surcharge, flow reversal, and flow transfers by orifices, weirs and pumps. The new code developed for radial gates and siphon-weirs is suitable for handling unsteady flow conditions occurring in real sewer systems. The program also makes provision for dynamic control of the radial gates (or other gate types) based on flow conditions at remote locations in the sewer system. The dynamic radial gate code may also be used for analyzing system optimization in a real-time mode. The radial gate and siphon-weir code has been verified under a wide range of flow conditions. Provision is made for avoiding instabilities under certain flow conditions, by taking special steps when coding the model. RGEXTRAN is also subject to the numerical stability concerns associated with the explicit scheme used in the original EXTRAN program . (Abstract shortened by UMI.)

Michaud, J.D. 1992. Ph.D. "Distributed Rainfall-Runoff Modeling of Thunderstorm-Generated Floods: A Case Study in a Mid-Sized, Semi-Arid Watershed in Arizona".

Abstract: Flash floods caused by localized thunderstorms are a natural hazard of the semi-arid Southwest, and many communities have responded by installing ALERT flood forecasting systems. This study explored a rainfall-runoff modeling approach thought to be appropriate for forecasting in such watersheds. The kinematic model KINEROS was evaluated because it is a distributed model developed specifically for desert regions, and can be applied to basins without historic data. This study examined the accuracy of KINEROS under data constraints that are typical of semi-arid ALERT watershedsme historic data. The accuracy of KINEROS was compared to that of the SCS model. When calibrated, a distributed SCS model with a simple channel loss component was as accurate as KINEROS. Reasons for poor simulations were investigated by examining (a) rainfall sampling errors, (b) model sensitivity and dynamics, and (c) trends in simulation accuracy. The cause of poor simulations was divided between rainfall sampling errors and other problems. It was found that when raingage densities are on the order of 1/20 km$\sp2$, rainfall sampling errors preclude the consistent and reliable simulation of runoff from localized thunderstorms. Even when rainfall errors were minimized, accuracy of simulations were still poor. Good results, however, have been obtained with KINEROS on small watersheds; the problem is not KINEROS itself but its application at larger scales. The study also examined the hydrology of thunderstorm-generated floods at Walnut Gulch. The space-time dynamics of rainfall and runoff were characterized and found to be of fundamental importance. Hillslope infiltration was found to exert a dominant control on runoff, although flow hydraulics, channel losses, and initial soil moisture are also important. Watershed response was found to be nonlinear.

Paturel, J.-E. 1992. Do. "Study of Rain Phenomena at Small Scale: contribution to Risk Analysis in Urban Hydrology".

Abstract: The sanitation field is becoming more and more complex in the last decades. The sewage system is not only urban equipment which guarantees the hygiene and security of the city, but also a means to protect the environment. To face this context, Urban Hydrology develops an important research axis, and our work tries to state some problems and answer questions about risk in Urban Hydrology and the analysis of rain data. (1) In the sanitation field, absolute security is utopic; there is always a risk domain and a non-risk domain. The technicians have fixed norms to delineate them, but the complexity of the water cycle shows limits. To progress, a dialogue between urbanists, technicians, economists, and users is necessary. (2) We also studied the rain-gauging of a watershed of 600 km$\sp2$ area. We showed a variability of the rain phenomena at a small scale (1 measure per 20 km$\sp2$), probably linked to the topography. In conclusion, global and multidisciplinary approaches to the problems faced in urban surroundings must be undertaken.

Spangberg, A. 1994. "Stormwater Quality Measured by Turbidity".

Continuous measurements of storm water quality has been conducted in the run-off from a small urban area by means of turbidity, conductivity, pH and temperature during 158 storms in a period of 6 months.
Rain and discharge was measured simultanously with the quality parameters. All electronic instruments were connected to a data logger to get exact synchronization of the time. Dust collection was made on some small confined surfaces by vacuum cleaning and in a fall-out collector.
The aim of the investigation was to describe and quantify the accumulation and wash-off processes on a single impermeable surface drained by one inlet during a storm.
It was shown that pollution accumulation and wash-off are complex processes on a simple imperrneable surface. These processes depend on the physical characteristics of the surface and its surroundings and several meteorologic parameters related to the time before the actual rain event and rainfall characteristics during the event itself. Results of the investigation indicate that the major part of the material accumulating on the surface comes from surrounding areas; the atmospheric fall-out contributes very little to the accumulated mass. The average pollution wash-off during one event is related to actual rainfall characteristics. Within one event only the first peak of the polutograph is related to preceding rainfall and meteorologic conditions prior to the rainfall event. Time variations of the polutograph for other peaks are only related to rainfall characteristics of the actual rainfall event.