Data and Analysis
Environmental scientists use various measuring instruments for investigating the physical, biological and chemical properties of aquatic ecosystems. In order to monitor the state of an aquatic system, researchers measure data over a long period, often using several sensors intended for different parameters at the same time. The parallel use of these sensors and high data acquisition rates produce large data sets that until very recently could not have been handled.
The scientists involved into the research focus "Data and Analysis" develop procedures and algorithms for processing large data sets: Due to real time computations and conversions, data can be assigned to a defined position (geocoding), processed rapidly and presented graphically. Here, both systematic and random errors need to be identified. The researchers also have to develop an appropriate infrastructure thus enabling the communication between sensors and remote acquisition units.
In order to answer specific scientific questions, the scientists not only prepare and visualise data but also look for particular patterns, repetitions and relations. Based on these analyses, new sensors and concepts can be developed.
In addition, the scientists create new interfaces to other data portals and analysing tools. Preserving data consistently in internationally accessible data bases and library systems in the long term is a prerequisite for ensuring data availability and for working on an international research level.
For numerical and graphic analyses, the members of the research group use matrix-oriented programmes such as Matlab or the freely available statistics language R. Modelling of the underwater light field is carried out using Hydrolight. Some model-based developments are conducted using Simulink and LabView.
In order to process and convert data for particular purposes, the scientists develop their own software tools. Programming languages are Delphi, Java and various C-derivatives. Either a PC or a microcontroller environment is used, depending on the research purpose.