Geographic features change over time, this change being the result of some kind of event or occurrence. It has been a research challenge to represent this data in a manner that reflects human perception. Most database systems used in geographic information systems (GIS) are relational, and change is either captured by exhaustively storing all versions of data, or updates replace previous versions. This stems from the inherent diffculty of modelling geographic objects in relational tables. This diffculty is compounded when the necessary time dimension is introduced to model how those objects evolve. There is little doubt that the object-oriented (OO) paradigm holds signi cant advantages over the relational model when it comes to modelling real-world entities and spatial data, and it is argued that this contention is particularly true when it comes to spatio-temporal data. This thesis describes an object-oriented approach to the design of a conceptual model for representing spatio-temporal geographic data, called the Feature Evolution Model (FEM), based on states and events. The model was used to implement a spatio-temporal database management system in Oracle Spatial, and an interface prototype is described that was used to evaluate the system by enabling querying and visualisation.

A model that emphasizes possible alternative sequences of events that occur over time is presented in paper 1 (chapter 2) of this dissertation. Representing alternative or branching events captures additional semantics unrealized by linear or non-branching approaches. Two basic elements of branching, divergence and convergence are discussed. From these elements, many complex branching models can be built capturing a perspective of events that take place in the future or have occurred in the past. This produces likely sequences of events that a user may compare and analyze using spatial or temporal criteria. The branching events model is especially useful for spatiotemporal decision support systems, as decision-makers are able to identify alternative locations and times of events and, depending on the context, also identify regions of multiple possible events. Based on the formal model, a conceptual framework for a branching events model for flood disasters is presented. The framework has five parts, an event handler, a query engine, data assimilator, web interface, and event database. A branching events viewer application is presented illustrating a case study based on a flood response scenario. A spatiotemporal framework for building evacuation events is developed to forecast building content evacuation events and building vulnerabilities and is presented in paper 2 (chapter 3) of this dissertation. This work investigates the spatiotemporal properties required to trigger building evacuation events in the floodplain during a flood disaster. The spatial properties for building risks are based on topography, flood inundation, building location, building elevation, and road access to determine five categories of vulnerability, vulnerable basement, flooded basement, vulnerable first-floor, flooded first-floor, and road access. The amount of time needed to evacuate each building is determined by the number of vulnerable floors, the number of movers, the mover rate, and the weight of the contents to be moved. Based upon these properties, six possible evacuation profiles are created. Using this framework, a model designed to track the spatiotemporal patterns of building evacuation events is presented. The model is based upon flood forecast predictions that are linked with building properties to create a model that captures the spatiotemporal ordering of building vulnerabilities and building content evacuation events. Applicable to different communities at risk from flooding, the evacuation model is applied a historical flood for a university campus, demonstrating how the defined elements are used to derive a pattern of vulnerability and evacuation for a campus threatened by severe flooding. Paper 3 (Chapter 4) of this dissertation presents a modeling approach for representing event-based response risk. Surveys were sent to emergency managers in six states to determine the priorities of decision makers during the response phase of flood disasters. Based on these surveys, nine response events were determined to be the most important during a flood response, flooded roads, bridges closed, residential evacuations, residential flooding, commercial flooding, agricultural damage, power outage, sheltering, sandbagging. Survey participants were asked to complete pairwise comparisons of these nine events. An analytic hierarchy process analysis was completed to weight the response events for each decision-maker. A k-means clustering analysis was then completed to form 4 distinct profiles, mixed rural and urban, rural, urban, and high population - low population density. The average weights from each profile were calculated. The weights for each profile were then assigned to geospatial layers that identify the locations of these events. These layers are combined to form a map representing the event-based response risk for an area. The maps are then compared against the response events that actually occurred during a flood disaster in June 2008 in two communities.