Charles S. Assaf, MSc – Global IT & DevOps Management, BlueGreen Water Technologies
A harmful algal bloom does not wait for a monitoring schedule.
By the time a visible bloom is reported to a water manager, sampled, tested and confirmed through traditional laboratory methods, days may have passed. In that window, a bloom can double in concentration, close a beach, threaten a drinking water supply or kill fish and wildlife across an entire lake. The communities, utilities and ecosystems depending on that water body are left reacting to a problem that has already grown past the point where early intervention would have been fastest and most cost-effective.
The traditional model of HAB management—sample, wait, confirm, respond—was built for a world without daily satellite coverage and without AI capable of reasoning over that coverage at scale. That world no longer exists. BlueGreen has spent the past several years building the technology to close that gap.
The Problem With Waiting
Harmful algal blooms are not new, but their frequency and severity are increasing. Warmer water, nutrient runoff and changing weather patterns are producing more frequent and more toxic blooms across freshwater systems worldwide. For water managers, the challenge is not just detecting a bloom once it is visible—it is knowing which water bodies are trending toward a bloom before they become a crisis, and knowing precisely where and how severe the problem is once it appears.
Manual monitoring cannot scale to this problem. A water manager responsible for dozens or hundreds of water bodies cannot physically sample all of them daily. Satellite imagery changed what was observable. What BlueGreen set out to solve was what happens after the imagery arrives—how you turn petabytes of satellite data into an answer a water manager can act on the same day.
Two Systems, One Pipeline
BlueGreen's answer to this problem operates in two connected layers.
BGi™ (BlueGreen Insights)—the evolution of the original Lake Guard View platform—is the sensing and analysis engine. It retrieves satellite imagery from Planet (5-metre resolution) and Copernicus (20-metre and 300-metre resolution) and analyses that imagery to detect cyanobacterial bloom activity and derive a Bloom Index, a quantified measure of bloom concentration and severity for a given water body at a given point in time.
BGi also captures and visualises near-real-time in-situ measurements from a network of water quality sensors deployed directly in monitored water bodies, including chlorophyll-a, phycocyanin and turbidity, alongside satellite-derived Bloom Index values, historical trend graphs and bloom heatmaps.
AInsights™, BlueGreen's internal AI reasoning platform, ingests BGi's combined output together with BlueGreen treatment case studies, peer-reviewed research, historical observations and weather forecasts. Rather than simply reporting what is happening, AInsights interprets the data and recommends what should happen next.
BGi answers:
AInsights answers:
Behind the scenes, AInsights combines semantic retrieval from its vector database with direct queries to BlueGreen's Redshift data warehouse, ensuring every reported value comes from verified structured data while contextual interpretation is drawn from research and historical knowledge.
From Question to Report
A water manager, scientist or field technician can simply ask:
"What is the current condition of this water body today?"
The platform automatically retrieves current Bloom Index values, sensor readings, historical trends, similar case studies, relevant scientific literature and weather forecasts before generating a concise report complete with charts and treatment recommendations.
Users can choose whether ChatGPT or Claude performs the reasoning, while the underlying data remains the same—verified satellite imagery, validated sensor data, BlueGreen case studies and peer-reviewed research.
Why the Source of the Data Matters as Much as the Data Itself
An AI system is only as trustworthy as the data it uses.
Every satellite image is traceable back to Planet or Copernicus. Every in-situ sensor reading is linked to its physical sensor and location. Every research paper is traceable to its original publication. Nothing enters the system without a verifiable chain of origin.
This level of data provenance is essential because AI-generated recommendations directly influence treatment decisions affecting drinking water, recreation, ecosystems and public health.
From Detection to Treatment
Detection exists to enable action.
BlueGreen's treatment portfolio includes Lake Guard® Oxy, Lake Guard® Blue and Lake Guard® Dew. All three products utilise BlueGreen's patented buoyant coating technology, allowing them to self-distribute naturally across affected water bodies using wind and water currents while reducing infrastructure and labour requirements.
AInsights helps determine which treatment is most appropriate based on bloom type, concentration, water body characteristics and treatment objectives.
Beyond Remediation: Measuring What Restoration Achieves
The same satellite monitoring pipeline used for bloom detection also enables long-term environmental measurement.
BlueGreen's Net Blue™ initiative applies this infrastructure to freshwater carbon dynamics, measuring the environmental benefits of restoring eutrophic water bodies through structured Monitoring, Reporting and Verification (MRV). These measurements support the generation of carbon credits based on verified environmental outcomes.
The Point of All of It
None of this technology exists for its own sake.
It exists because water managers need accurate answers today—not next week—and they need confidence that those answers are supported by trusted science and verified data.
Satellite imagery, AI reasoning and rigorous data provenance provide the infrastructure. The Bloom Index, treatment recommendations and actionable reports provide the outcome.
Ultimately, what matters is identifying blooms earlier, treating them more precisely, restoring water bodies faster and returning healthy lakes to the communities that depend on them.
Charles S. Assaf runs Global IT & DevOps Management at BlueGreen Water Technologies, where he built the AInsights platform. He is a member of the OASIS Data Provenance Standards Technical Committee and writes on AI governance and architecture at KAIRI on Medium. His articles are available at https://charlesassaf.com
