Weet wat je meet en meet wat je niet weet

Abstract

We can only relate individual measurements meaningfully if we have a model of the system where the measurements come from. This can be a simple conceptual model, (is the groundwater dynamics a smooth slow progress as at the Veluwe, or can the groundwater level be quite different from week to week), it can be a statistical model, where assumptions are made on probability distributions and correlation in space and time, but also a complex numerical model can describe the relationship between the observations. Without prior knowledge of the system, it is not possible to create a adequate monitoring. So we must have some knowledge about the system we monitor: KNOW WHAT YOU OBSERVE.