Paper Reviewed by Dr. Ole Mikkelsen
Summer Diatom Blooms in the Pacific Investigated with a Towed LISST-Holo!
March month’s featured paper is Anderson EE, Wilson C, Knap AH, Villareal TA (2018): Summer diatom blooms in the eastern North Pacific gyre investigated with a long-endurance autonomous surface vehicle. PeerJ 6:e5387
In the open ocean, diatom blooms are typically observed and studied using satellite remote sensing. Prolonged in-situ observations of blooms in these areas are rare, as the blooms are unpredictable and ship time is expensive. Unmanned Surface Vessels (USV’s) provide an opportunity to have a sensor package on location for weeks or months in order to take in situ data, thus gaining more detailed understanding of the origin and decline of the blooms.
Anderson et al equipped a Liquid Robotics Wave Glider with a suite of sensors for long-term diatom monitoring in the eastern North Pacific gyre (see figure, from Anderson et al (2018): CTD (2), fluorometers (2, 4), weather station (1), wave sensor (2), and a Sequoia Scientific, Inc. LISST-Holo (5) for imaging purposes. The LISST-Holo was towed behind the Wave Glider in a neutrally buoyant package on a 10 m tether. The tether provided power from the Wave Glider to the LISST-Holo, which stored data internally, capturing 1 burst of 15 holograms every 6 hours for the duration of the mission. The entire mission lasted from June 1, 2015 until November 3, 2015. The Wave Glider covered ~5,690 km in that period and captured 9,336 holograms. Apart from the holograms, data were transmitted to shore in near real-time. All holograms were offloaded and processed after recovery of the Wave Glider.
Diagram by Liquid Robotics of Wave Glider with numbered areas for sensor placement
Two diatom genera, Rhizosolenia Brightwell and Hemiaulus Ehrenberg were the targets of the mission, covering a nutrient-starved region of the Pacific from 145-155°W, 20-30°N, where the diatoms dominate the blooms. A nitrogen-fixating bacteria, Richelia intracellularis is thought to supply the nitrogen for the blooms. The mission purpose was to carry out sustained observations of phytoplankton distribution, abundance and physiology in the entire region.
The LISST-Holo collected data throughout the mission and the processed and reconstructed holograms were used to identify Hemiaulus and Rhizosolenia spp. There was progressive biofouling of the windows during the mission, but using Sequoia’s Holo_Detail processing software, even biofouled holograms could be reconstructed. Cell counts were then performed on the images, down to a minimum concentration of 36 cells L-1.
Out of 610 burst periods (each containing 15 holograms), 208 contained Hemiaulus cells and 207 contained Rhizosolenia cells, with an average cell count when present of 890 L-1 and 180 L-1, respectively. Maximum count was 14,000 L-1 and 2,800 L-1, respectively. That kind of data is impossible to get from remote sensing. As only 1/3 of the bursts contained any diatoms it is evident that there was a large patchiness in the diatom abundance. In fact, one burst could have no cells and the next (six hours later) could have up to 10,000 cells L-1.
The holograms further revealed that Hemiaulus aggregates were present throughout the observed area, and that they were present even during the summer export pulse. This is a seasonal event, which is thought to be responsible for as much as 20% of the carbon export to the deep sea within a few weeks. Traditionally, it has been surmised from previous camera observations that the summer export pulse was due to plankton aggregate formation as the bloom senesces. The fact that Hemiaulus aggregates were persistently present suggests that this is a natural growth form for Hemiaulus - another finding that would be impossible to detect using satellites OR other sampling methods, such as nets.
There’s lots of other very cool data in the paper, such as a months-long record of the photosynthetic potential index. The bottom line is that most of the optical sensors worked well most of the time and gave Anderson et al a very unique dataset to work with.
But don’t take my word for it – here’s a five minute video describing the entire study!