ABSTRACT
The monitoring of continuous phenomena like temperature, rainfall, air pollution etc. is carried out by (wireless) sensor networks today. From the perspective of the user of such data, the original discrete sensor observations are not very useful since they often do not cover the spatio-temporal area of interest. So the fundamental task in this context is to provide a gapless and continuous representation of the particular phenomenon either as visualisation in real time or as model for long term archiving, or both. Therefore, it is necessary to cover the area of interest with observations in a way that is sufficient to capture the phenomenon. The necessary density of observations depends on the dynamism of the phenomenon; for spatio-temporal monitoring this has to be considered for both space and time. From these observations, the value of the particular phenomenon needs to be estimated for unsampled spatio-temporal positions. Monitoring as a whole can be seen as an optimization problem or trade-off between spent resources and achieved model quality. This chapter discusses environmental monitoring from a managerial perspective. It relates the most fundamental properties of a monitoring system to each other: its aims, its quality, and its costs.
