INTRODUCTION
Traditional spatial information systems hold only a single state of the real world, usually the most recent in time for which the data were captured. However, geographic phenomena involve both static and dynamic information (Galton & Worboys, 2005). Although GIS are moving towards spatiotemporal information systems, some applications call for an immediate switch from a static view of our environment to a dynamic-oriented focus, e.g. environmental change monitoring, transportation, health and epidemiology, or crisis management. Besides these classical application areas, the increased use of real-time, mobile and in-situ sensors is leading to new potential applications for spatiotemporal data models and systems (Galton & Worboys, 2005).
Modeling geographic phenomena requires a profound understanding about the processes or events related to the phenomena. Usually, these occurrences are strongly connected, so considering just isolated snapshots of the real world is not the best approach to understand what and why is happening. We can consider each measurement, each observation of the real world as an abstraction (snapshot) of some environmental feature (taken) at a specific moment. If we want to obtain information from the continuous data flows produced by sensors, we should provide meaning to relevant fragments (patterns) of observations and analyze where and when they appear.
Environmental monitoring is a critical process in areas usually affected by natural disasters. It is aimed to ensure public safety, to set up continuous information services and to provide input for spatial decision support systems (Resch, et al., 2009). Here, the main challenge is the distributed processing of vast amounts of heterogeneous sensor data in real-time. Most current approaches use web services based on the classic request/response model. Although partly using open GIS standards, they are often unsuitable for the integration of large volumes of data on-the-fly (Resch, et al., 2009). This integration can be performed via both pull-based models and push services that send out alerts, e.g. if a certain threshold is exceeded. Thus, it requires real-time processing capabilities, such as Event Stream Processing (ESP) or Complex Event Processing (CEP). Event processing has emerged as one of the most important issues in IT today (Chandy & Schulte, 2007). Event processing tools provide methods for reading, creating, transforming or abstracting events, e.g. CEP (Luckham, 2002). It is possible to use these tools to define patterns and detect relevant events in measurement data. CEP is also able to manage abstraction layers in real-time, which could lead us to improve information integration across different communities.
In this section, existing event processing technologies and eventing standards are described. Previous research on integration of landslide information is presented, as well as some literature documenting threshold-based approaches for the initiation of landslides.