IoT based Ocean Acidification monitoring system with ML based Edge Analytics

Abstract

Ocean Acidification (OA) is often referred to as “an evil twin of climate change”. Increased global atmospheric emission of CO <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> has increased its concentration in the ocean due to dissolution and absorption of atmospheric CO <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> , leading to ocean acidification. This further affects the coral reefs and other calcifiers such as Pteropods (marine snails), shellfish (clams, etc.), crustaceans (shrimp and lobsters), starfish, sea urchins, and their kin (echinoderms) and interferes with the marine food web and leads to food shortages thereby affecting the socio-economic lives of the people living in the coastal area. To quantify the above impacts and to study other biochemical variations in the ocean, it is necessary to monitor ocean acidification for a longer time period. Current methods for monitoring ocean acidification employ gliders, drifters, buoys, mooring, and periodic testing of samples collected during research voyages. These methods are often not cost-effective, time-efficient, requires a lot of manpower, and provides inferior spatio-temporal resolution. This research work analyzes the existing ocean acidification data collected from the ALOHA station at the North Pacific Ocean and propose a predictive model at the edge level by using Regression to reduce the effective cost of the IoT system. The proposed regression model is also used to design a novel IoT architecture for designing an energy-efficient, near real-time ocean acidification monitoring model. The proposed IoT hardware can be deployed by using buoys and fishing boats to further reduce the monitoring cost and increase the spatio-temporal resolution of the data compared to the existing monitoring systems.