Image in Flow Buoy Splashdown in Brazil

26.02.2014  by  Kevin van Hecke


Fig 1. The buoy in its final location

Fig 2. CytoBuoy with ocean view behind the mountains

Fig 3. Transporting the buoy

Fig 4. The WiFi system

Fig 5. A foodweb generated on CytoSense data. Colored circles are primary producer, ellipses are herbivores or primary consumers in the second trophic level and rectangles are predators or secondary consumers in the third trophic level. The edge thickness represent the link strength.

A CytoBuoy has been deployed in the beautiful area of Arraial do Cabo, in a protected natural reserve 150km east of Rio de Janeiro city. As part of the SiMoCo project (Sistema de Monitoramento Costeiro) lead by researcher Gilberto de Carvalho Pereira of Coppe-UFRJ university, the buoy was deployed as a proof of concept for a yet to be designed larger system in order to fully monitor the highly dynamic coastal area of Rio De Janeiro state. Coincidentally this is also the first time a CytoSense with Imaging in Flow has been deployed in an autonomous buoy!

The remote location of the buoy and the presence of a mountain blocking the line-of-sight between the buoy and the marine station have posed some communication problems which are partly solved by a custom designed WiFi system (fig 4). Using this system it is possible to monitor and control the buoy in real time from Rio de Janeiro city. Work is underway to also enable the transmission of raw data files.

Currently, the buoy steadily runs 6 samples per day of which 3 Image in Flow measurements in smart grid mode and 3 measurements in high sensitive mode. This frequency may be increased in the near future, but since the buoy is running on solar power only we are monitoring power levels first to be sure.

One of the primary goals of the project is to validate fundamental research done by Gilberto et al.. Gilberto explains: “The CytoBuoy project in Brazil has made efforts to develop a model for representing the structure of plankton communities. First the optical signatures are categorized by machine learning models and validated by Imaging in Flow. Then we have generated and tested fuzzy multigraphs (edges) as food webs in order to make simulations (fig 5).

Directed and random deletions are performed by applying a set of toxicological rules whose effects are monitored through traditional structural, flow and overall yield indexes of the network. Communities or clusters (i.e. higher density connection-zones in the network) have been detected by applying an ant algorithm.

The main idea is to learn the temporal behavior of the biological system and compare two sites, one is an estuary  and another is completely marine.”

A paper about this work is submitted.