New paper on marine environmental monitoring: Improving Chlorophyll-A Estimation From Sentinel-2 Using Machine Learning

The term ‘Algae bloom’ describes rapidly multiplying microscopic phytoplankton; marine algae that drift on or near the surface of the sea or lakes. Phytoplankton play an important role in the marine food weband the global carbon cycle by absorbing carbon dioxide on a scale equivalent to that of terrestrial plants. The algae are so abundant that their green pigment, chlorophyll that phytoplankton use for photosynthesis, colours the surrounding water green. This feature allows detecting and monitoring these tiny organisms from space, using satellites with optical sensors onboard. Warm water temperatures, sunny, calm weather and nutrient availability often lead to large blooms that may be harmful. Information from optical satellite images is then used to characterise algae blooms, map their extent, and monitor them over several years. Algae blooms can be toxic pollutants or lead to oxygen deficiencies, being problematic for ecosystems, aquaculture, fishing and tourism.

PhD candidate Muhammad Asim is studying phytoplankton blooms in the Barents Sea. ESA imaged an algea bloom around the island of Gotland in the Baltic Sea using Copernicus Sentinel 2 on 20 July 2019. Image credit: Private and ESA – Copernicus Sentinel 2.

Muhammad Asims paper addresses new methodologies for remote sensing of marine Chlorophyll-a (Chl-a) with emphasis on the Barents Sea. His research aims at improving the monitoring capacity by integrating in-situ Chl-a observations and optical remote sensing to locally train Machine Learning (ML) models. For this purpose, in-situ measurements of Chl-a collected for the years 2016-2018, were used to train and validate models.

To accurately estimate Chl-a, he proposes to use additional information on pigment content within the productive column by matching the depth-integrated Chl-a concentrations with the satellite data.

In addition, the authors design and implement a neural network model dubbed as the Ocean Color Net (OCN), that has performed better than existing ML-based techniques.

Algae blooms are a natural part of marine and aquatic ecosystems. With changing climate and increasing human activity, they may become more frequent, occur in larger areas, or last longer. PhD candidate Muhammad and his colleagues work helps to improve how satellite data can track the growth and spread of potentially harmful algae blooms in order to alert and mitigate against damaging impacts for ecosystems, tourism and fishing industries. Find the fulltext version here.

Print Friendly, PDF & Email