ABSTRACT
The monitoring of continuous phenomena is a complex and challenging task on many levels. In this work, a holistic concept has been proposed to divide this task into small cohesive units of operation. Variogram fitting is the key task of geostatistics since it adapts the algorithm to the actual statistical properties of the observational data. A new and generic method based on binary space partitioning was proposed to aggregate variogram points of arbitrary dimensionality in order to unburden the subsequent parameter fitting procedure. An algorithm for compression and progressive retrieval of observational data of arbitrary dimensionality was proposed. Its predominant aim is to improve the sufficiency of sensor data transmission and archiving. In progressive mode, it provides coarse values of low data volume and increases in accuracy with each transmission step. From the current developments of environmental monitoring it can be concluded that for responsive systems such an abstraction will become more common.
