GIS for Coastal Zone Management - Chapter 9

CHAPTER NINE Application of a Decision Support System in the Development of a Hydrodynamic Model for a Coastal Area Roberto Mayerle and Fernando Toro 9.1 INTRODUCTION Recent advances in numerical modelling of physical processes and field survey technology nowadays allow the development of numerical models with extensive data sets. As a consequence, model developers are facing new challenges to handle the increasing amount of data and its analysis. Furthermore, model development and application concerning coastal areas are heavy time demanding tasks that need tools to assist the researcher. Most of the time they involve analysis of field measured data and its comparison with numerical model outputs. Even though there are good tools for specific tasks, there is a lack of them for integrating measured and modelled data. An application of a Decision Support System (DSS) in the development of a hydrodynamic model for a coastal area is presented. Description of the DSS components, their interaction, and its GIS capabilities to handle spatial data are also given. 9.2 DEVELOPMENT OF HYDRODYNAMIC MODELS A hydrodynamic model is the mathematical representation, with the assistance of numerical algorithms, of flow and wave situations to study them more effectively. However to obtain reliable information, development of numerical models must include evaluation of results. This process is time-consuming and involves comparison of numerical model outputs with measured data. Moreover, a unique model development procedure can be restrictive due to the large amount of variables involved in the physical process. The number of variables and their levels of significance depend on the objective of the study. In spite of these difficulties a general trend for the development of hydrodynamic models can be defined and is presented next. Figure 9.1 shows the phases and steps in the development of a hydrodynamic model. The model development consists of several phases (i.e. sensitivity analysis, calibration, validation and application) in which the modelling experts fine-tune © 2005 by CRC Press LLC the numerical and physical parameters to reproduce the field conditions. Each of these phases has several steps that share information with the other phases. Phases Sensitivity analysis Calibration Validation Application Data correction and selection Preparation of measured and model data Numerical model runs Model data analysis Reports Figure 9.1: Phases and steps in the development of hydrodynamic models The “sensitivity analysis” evaluates the response of the model to changes of the physical and numerical parameters. The numerical parameters are set and the physical parameters having an effect on the results are selected. Physical parameters are adjusted to represent the conditions in the studying area during the “calibration”. During the “validation” the model ability in reproducing the field conditions, within certain level of accuracy, is checked. The model “application” is the last phase in the model development and is performed by the end users to make decisions. The validated model is then used in hindcasting, nowcasting and forecasting situations. Regardless of the model development phase, each of them has several steps as indicated in Figure 9.1. The first one is the “data correction and selection” in which the measured data is corrected, selected and stored in a database. This is usually data that needs to be processed to fit the model formats. In situ measurements are in the form of time series at a certain location or recordings at a certain area for a given time. The “preparation of measured and model data” includes the creation of the model grid and definition of some default (but logical) numerical and physical parameters. The measured data preparation includes reformatting of the data to be used as initial conditions and boundary conditions for the model and to enhance easy comparison with the model results. “Numerical model runs” is the core of each phase and refer to determination of the change in the numerical model outputs caused by changes in input data. “Model data analysis” is the step in which the model results are analysed by means of the experience and judgement of the model developer, and compared with other model results or measured data. The data analysis is an iterative process with the model runs and gives feedback to the model developer whether or not to choose new © 2005 by CRC Press LLC parameters to run the model again. This is a procedure in which software tools are needed to speed up and facilitate model development. Therefore, a “Decision Support System” (DSS) that assists the model developers is of vital importance. And finally, the “reports” document the work that has been done and summarises the experience and conclusions obtained. These written reports are used as reference in the future phases of the model development. 9.3 A DECISION SUPPORT SYSTEM FOR MODEL DEVELOPMENT The handling of information for hydrodynamic model development is a difficult task as a consequence of different sources, types and formats of the data and the increasing amount of data involved in such activities nowadays. An efficient way to retrieve, analyse and store information accurately and speedily is crucial for the good performance in the model development process. A Decision Support System (DSS) for development of hydrodynamic models is an interactive computer-based system intended to help modellers use data from measurements and models to simplify and speed up model creation. It helps to retrieve, summarise and analyse data. It comprises three integrated components: “interface”, “database” and “models”. Figure 9.2 shows the components of a DSS for development of hydrodynamic models, with the input and output data. Input data – Field measurements – Numerical model results Models of the DSS – Data simplification – Analysis Interface – Format convertors Database Output data – Visualisation – Statistical results Figure 9.2 Components, input and output data of the DSS © 2005 by CRC Press LLC The proposed DSS centres on the model development of depth-integrated (2DH) flow models with field data applied to coastal areas. The DSS has been tailored to handle primarily data such as water levels and current velocities. The input data to the DSS, from measurements and numerical model outputs, is fed to the system through the interface and is stored in the database component of the system. The GIS application, as a DSS generator, is used to handling spatial data in conjunction with the database. The instructions to the system are given by the decision-maker through the interface that passes the type of analysis to be made by the model components of the DSS. Finally, the output data is presented in the form of statistical tables or visualised in plots and Figures. The “interface” component is the part of the DSS that interacts with the user. It should be user-friendly and usually includes graphical components and it is then called a “Graphical User Interface” (GUI). The front end of the system is a graphical interface created customising ArcView 3.2a GIS software (ESRI, 1996). The “database” is the collection of data that is organised so that its contents can easily be accessed, managed and updated. The size and effort spent in the database depend on the amount of available information. It is the data source for the DSS that processes and presents the information. The importance of the database in the system relies on the versatility for the selection of the data for analysis rather than in the storage capacity. In the case of the DSS, its capabilities are implemented using the GIS software generator, with tables connected under the concept of relational database. The “models of the DSS” use the data provided by the database and allow the user a direct opportunity to explore the data relevant to the problem. The model components of the DSS take care of evaluating, analysing and correlating the information from different sources available in the database and reduce the information to make it easy for the user to understand the physical process. The distinction between the models of the DSS and the numerical models should be made clear here. Whereas the latter ones are software tools used to solve partial differential equations for the hydrodynamics in coastal areas, models of the DSS are software tools to link and analyse numeric output and in-situ measurements. The model components are mainly programmed as Matlab (The MathWorks, 2000) functions grouped in a software library that makes the DSS a versatile system open for new types of data analyses, new types of data sources and makes their maintenance easier. Figure 9.3 makes a parallel between the model development steps in which the DSS assists the researcher and the DSS components. The DSS interface is used in all steps of model development and even in earlier stages during the data acquisition to visualise information. The database component is used all along the development steps and it is queried according to the model development phase that is being analysed. The data simplification, analysis and conversion are performed by the DSS models and is made during the data preparation to run the model and during the evaluation of the numerical model outputs. © 2005 by CRC Press LLC Data acquisition Data correction and selection Preparation of measured and model data Numerical model runs Model data analysis Reports Model development steps DSS components Figure 9.3 Model development phases and DSS components 9.4 APPLICATION This section presents the application of the DSS during the development of a hydrodynamic model for an area on the German North Sea coast. The multiple possibilities of the DSS for data handling and its versatility to the needs of the user are shown. The in situ measurements presented herein were collected as part of the project “Predictions of Medium Scale Morphodynamics” (Promorph) (Zielke et al., 1998) and were used for the development of the flow model (Palacio et al., 2003). The domain of interest is the Meldorf Bight and the adjacent tidal channels covering an area of about 600km2. Around 50% of the area of investigation is dry during the ebb phase and almost the whole domain is submerged during the flood phase. Figure 9.4 shows the area of investigation, model boundaries and location of in-situ measurements. The tidal channels have a maximum depth of about 20m. Tidal range is around 3.0 to 3.5m. The calibration of the flow model is carried out © 2005 by CRC Press LLC ... - tailieumienphi.vn