Grantee Research Project Results
Final Report: Early detection and prediction of harmful algal blooms using low cost, networked IOT sensors and machine learning
EPA Contract Number: 68HERC21C0045Title: Early detection and prediction of harmful algal blooms using low cost, networked IOT sensors and machine learning
Investigators: Lee, Christopher
Small Business: AquaRealTime, Inc.
EPA Contact: Richards, April
Phase: II
Project Period: April 1, 2021 through March 31, 2023 (Extended to March 31, 2024)
Project Amount: $396,162
RFA: Small Business Innovation Research (SBIR) - Phase II (2021) Recipients Lists
Research Category: Endocrine Disruptors , SBIR - Water , Small Business Innovation Research (SBIR)
Description:
Harmful Algal Blooms (HABs) are an increasing problem in waterways all over the world, costing an average of $17 billion in damages each year. HABs have shut down water supplies to entire US cities in recent years.
Blooms occur when nutrient-rich waters stimulate cyanobacterial growth, resulting in unsightly sludges that discolor waterways, rendering them dangerous to humans, livestock and wildlife because of the cyanotoxins released. Cyanotoxins make HABs very costly to clean up, as the algae themselves must be removed, and the water purified before it is safe to drink, enter or even be close to because of the risk of toxin aerosolization by wind-blown spray.
The treatments for HABs are expensive, sometimes environmentally damaging and of limited efficacy unless applied early in a bloom. A key part of the solution to HABs is early detection/prediction so that action can be taken at the earliest possible moment, reducing costs, environmental and economic damage, and preserving access to clean drinking water. Floating sensor buoys are a critical part of the solution, but the current marketplace is crowded with complex sensor platforms that can cost $30,000 per unit and which require specialist knowledge to deploy, use and maintain. These platforms are beyond the means of thousands of small and medium sized stakeholders who manage lakes, reservoirs, ponds or stretches of beach and who are often hardest hit by HABs. AquaRealTime was founded to provide a turnkey HAB monitoring solution for the small and medium sized stake-holder market, estimated at $900 million worldwide. Our innovative HAB sensor, AlgaeTracker, is affordable, lightweight at 8lbs in weight and can be deployed by a non-specialist in 30 minutes. AlgaeTracker also has outstanding sensor accuracy characteristics, boasting R^2 values of 0.999 for the key sensor suite parameters. And because all AlgaeTrackers transmit their data wirelessly over the cellular network and are accessed by a web-browser based dashboard, they are convenient and easy to use.
The purpose of AlgaeTracker’s EPA Phase II SBIR grant was to build on the beta version of the AlgaeTracker, readying it for full marketplace distribution. Its key objectives were to develop the analytics capabilities of the AlgaeTracker software, as well as to configure AlgaeTracker’s data in such a way that it could be provided to US government agencies and other entities with an interest in controlling HABs, saving millions as algaecide treatments happen sooner, at a lower cost, and with fewer environmental effects.
Summary/Accomplishments (Outputs/Outcomes):
One of our primary goals for SBIR Phase II was the increased commercialization of both the AquaRealTime dashboard and the AlgaeTracker itself. Following this phase, a wide range of improvements have been made to both facets of our development.
Our physical product has been augmented with a number of commercialization improvements. Innovations in our battery technology have extended the device’s solar battery life from days to weeks in the absence of sunlight. The implementation of a full-shutdown button allows units to be stored without consuming battery power for weeks at a time, which has been a boon for shipping and winter storage in colder areas. Also improved on the AlgaeTracker is the antifouling brush and the sub-motor assembly responsible for controlling it. The new system was designed such that debris is much less likely to snag on the brush. This improvement and the upgrade it provides to the brush’s durability have been particularly pleasing to our customers. Finally, our rigging package, a system that suspends the AlgaeTracker within a high-visibility buoy, has made the AlgaeTracker easier to deploy safely and securely.
The dashboard has also been elevated, from a barebones software environment to a well-rounded, versatile tool. Its field calibration function, an industry first, makes the normally obtuse readings from fluorometers more intuitive to the lay person.. Additionally, the dashboard’s newly-implemented REST API capabilities allow governments and enterprise companies to integrate our data into their own dashboards alongside other sensor data suppliers, avoiding an excess of integration tools.
Diligent research and development conducted in Phase II have made the algorithms and sensors that perform the bulk of the work the AlgaeTracker does more meticulous than ever, improving the device’s accuracy across a range of conditions. Several of the AlgaeTracker’s sensor calibrations were improved over the course of Phase II as well, especially that of the turbidity sensor. Our former method of calibration relied on subtracting the environmental baseline readings from the calibration vessel. However, research revealed an optically black vessel that reflects next to no light, allowing us to run regular calibration methods with the sensor without interfering with the reliability and accuracy of its readings. In addition, the temperature compensation algorithms for the optical sensors were improved and tested in a wider temperature range.
Finally, we have developed a means of non-photochemical quenching (NPQ) compensation that, in many cases, allows complete inversion of the effects of NPQ on the AlgaeTracker’s readings. Whereas NPQ has been shown to generate offsets of up to 70% in standard fluorescence readings, our designed algorithm can remove on the order of 80% of that effect. As expected, there is enough variation from one water body and related algal species to the next that the algorithm sometimes requires fine tuning for a water body. With this fine tuning capability embedded within the AlgaeTracker’s dashboard, though, the ability to reliably compensate for the effects of NPQ is an industry first with regards to optics sensors.
Conclusions:
After the results of SBIR Phase II, the AlgaeTracker has experienced a suite of improvements, all of which enhance its capability as a reliable source of actionable data in the treatment of harmful algal blooms. Developments in the device’s hardware and algorithms have widened the range of conditions the AlgaeTracker is able to function in, including challenging conditions of surf and corrosion. These improvements also reduce the non-automated maintenance necessary to the buoy’s upkeep. Furthermore, we anticipate that the repeatability of calibration augmented by improvements in the AlgaeTracker’s pre-shipment sensor calibrations will translate into increased scalability for the buoy, as these improvements allow units to be manufactured and tested more efficiently, and increase the longevity of the devices.
SBIR Phase I:
Early Detection and Prediction of Harmful Algal Blooms Using Low Cost, Networked IOT Sensors and Machine Learning | Final ReportThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.