Date of Award
Doctor of Philosophy (PhD)
Professor Ioannis K. Tsanis
A geographical information system has been creatively utilized to assist in the study and better understanding of the spatial variation and distribution of certain processes in water resources and coastal engineering. Hydroinformatics is the contemporary term that best describes the interactive knowledge-based non-expert decision support systems (DSSs) developed in this study in the form of GIS-based algorithms.
The spatial distribution of the annual orographic rainfall on the island of Crete (Greece) was modelled by using the multiple linear regression (MLR) method. The MLR models, developed and applied through the GIS interface, provided better estimates/predictions of rainfall at un-gauged locations than the conventional spatial interpolation techniques. This, in turn, resulted in more realistic spatial distribution of rainfall. A new DSS was developed that was applied to the island of Crete and Switzerland to asses rain gauge worth within an established network. This system provided the tools needed to reduce the number of gauges, if necessary, in an existing network, which means eliminating redundant gauges while maintaining an efficient network. Another tool was also developed to provide a ranking system that is based on performance evaluation of a number of spatial interpolation techniques in large and small rain gauge networks. Recommending the usage of certain techniques over others for a certain set-up of gauges and their records should improve the gridded precipitation input to distributed hydrological models. The same procedure may be followed for any other spatial variable (hydrological, meteorological, etc). Switzerland and the municipality of Hamilton-Wentworth (Ontario, Canada) were used as test cases for system evaluation.
A new GIS module was developed to estimate reference evapotranspiration based on the Penman-Monteith and Class A Pan Evaporation methods and using the station- and grid-based approaches. Although the combination of the grid-based approach and the Penman-Monteith method is recommended, the collection of significant amount of meteorological data from a relatively dense meteorological network is required for better results.
GIS was also used to facilitate the estimation of irrigation water requirements for different scales. Zooming in from large-scale (in the island of Crete) with limited number of meteorological stations to smaller scales with more "representative" stations provided support for the usage of the grid-based approach for calculating irrigation requirements.
For coastal engineering, a GIS pre-processor was developed for pollutant transport modelling in coastal areas. It provides accurate input to the models in the form of user-specified well-descritized bathymetry and shoreline, which are also used for post-processing purposes (i.e. graphical display). In addition to saving time in generating input files, this GIS pre-processor provides higher spatial accuracy, since all themes are geographically-referenced, which eliminates the possibility of overlapping or misplacing. It also shows great flexibility in generating grids of different cell sizes, interpolation techniques, and spatial extent. Transferability of the project between users and terminals makes it more accessible and convenient.
Naoum, Sherif, "A Hyrdoinformatic Approach to Basin/Coastal Water Management" (2003). Open Access Dissertations and Theses. Paper 809.