By Stacy Gittleman
On August 2, 2014, nearly half a million people in Toledo, Ohio, woke to ominous news: Do not turn on your tap. The water supply was not safe to drink, bathe in, brush one’s teeth or even touch. Boaters out on the water recall a musty smell and observed a slimy neon green mess floating on the surface of Lake Erie that stuck to the underbelly of their boats. Toledo residents drove hours away in search of cases of bottled water. The order lasted for nearly three days. A few weeks later, residents of Pelee Island, Ontario, faced a similar crisis which lasted nearly two weeks.
The Toledo Water Crisis of 2014 was caused by a toxic algal bloom of a harmful algae known as cyanobacteria, scientifically known as a Harmful Algal Bloom (HAB). Each year Lake Erie and the other Great Lakes are susceptible to HABs. The threat grows as the lakes warm due to climate change and other factors such as nutrient runoff and other pollutants.
Algae covers a broad spectrum of species. Each organism in the species emits a distinct hue within the blue-green color spectrum. Though the worst of their kind get the most attention for tainting drinking water and ruining a day at the beach, most of these species are single-celled algae called phytoplankton. This is the beneficial algae which is the main source of food for fish and aquatic life and which makes up half of the photosynthetic activity on Earth.
Then there are the harmful cyanobacteria – which derives its name from its neon blue-green color cyan. Cyanobacteria have the potential to produce toxins harmful to fish, humans and other animals. Under certain conditions, algae populations can grow explosively, such as in the 2014 Toledo water crisis. This harmful species emits chlorophyll and fluorescent light at various points in their life cycles.
According to the United States Environmental Protection Agency (EPA), cyanobacterial blooms are often confused with green algae. Both can produce foul-smelling dense mats that can put a damper on aquatic recreational activities and in severe cases, deplete a lake of its life-giving oxygen.
Unlike cyanobacterial blooms that can produce potent cyanotoxins, green algae are plants, not bacteria, and do not produce toxins.
Because Lake Erie has become perennially susceptible to harmful algae blooms for the past decade, the EPA, working with other government agencies, has established a portal to track the potential and existence of the blooms.
One of those agencies is the National Aeronautics and Space Administration – known as NASA. For decades, those pictures taken of the Earth from NASA satellites high above have been offering researchers and environmental officials ever-more-accurate data to forecast and monitor harmful algae blooms. A combination of satellite imaging (for bloom location and extent), and toxicity from field samples provide a powerful tool for harmful algae bloom forecasting and mitigation.
Though the use of remote sensing images from space has become routine for keeping an extraterrestrial eye on the planet’s largest bodies of water, what if the same technology could be used to watch over changes in a neighborhood lake or the nearest public swimming hole?
Science is just at the forefront of where satellites, when equipped with special spectral sensors to photograph inland lakes, are providing ever clearer images that can inform environmental and public health officials, and eventually local lake managers when and where to directly sample and test waters for any lurking cyanotoxins.
High above the earth, a constellation of NASA and European satellites have been photographing the surface of the earth with a progression of ever-more-sophisticated remote sensors. Scientists and researchers are just beginning to harness the power of this data to predict and monitor algal bloom patterns by making quantitative measurements of the color of the water, which is directly related to the absorption of light by the pigments that algae use in photosynthesis. They can also measure water’s surface temperature and estimate the amount of suspended matter and dissolved organic matter that water contains.
Generations of NASA Landsat have been sending back images of the earth from space for 50 years. Administered by NASA and the United States Geological Survey (USGS) for research, Landsat sensors continuously and repeatedly take photos of the earth’s surface in a 10–12-day orbit.
American researchers are also able to access data from the European Space Agency’s Copernicus Sentinel-2 and Copernicus Sentinel-3 satellites which keep improving their high-resolution imagery of a specific color in the spectrum that indicates cyanobacteria. The USGS is also working on developing sensors with the U.S. Space Station that more clearly distinguish algal compositions from one another.
The current satellites being used are the Sentinel 3A and 3B, which have a spatial resolution of 300 meters per pixel. These satellites carry sensors designed to detect a variety of pigmentations of chlorophyll, an indicator of algal growth in bodies of water. This allows NASA and other researchers to monitor hundreds of larger lakes across the US.
Newer satellites like Sentinel 2A and 2B have even higher spatial resolution, around 10-20 meters per pixel. This allows them to potentially monitor smaller inland lakes and rivers. However, they currently lack the spectral information to distinguish cyanobacteria from other algae.
In February 2024, NASA launched the Plankton, Aerosol, Cloud Ocean Ecosystem (PACE) satellite and began to send data back to earth – information which is available to the public. PACE is a hyperspectral satellite with the enhanced ability to detect specific wavelengths associated with cyanobacteria.
Scientists and researchers then compare what they see in satellite images to see if they match up with the water sampling they take from a lake back on Earth.
On the ground, levels of cyanobacteria and algal blooms are measured on the Trophic State Index TSI, as the word trophic means “relating to nutrition.” Traditionally, TSI levels are measured in situ, or in person. Water samples are taken from a Secchi depth, a measure of water clarity using an eight to 12-inch, black and white disk named for Angelo Secchi who invented it in 1865. A simple device, professional and citizen scientists lower the Secchi disk attached to a pole into a body of water, and the depth at which the disk is no longer visible is known as the Secchi-Disk Transparency (SDT). From this depth, water samples are taken and measured for counts of nitrogen, phosphorus, and/or chlorophyll a.
In 2016, the United States Geological Survey (USGS), working with the Michigan Department of Environmental Quality (DEQ), released a study that measured the health of 4,000 of the state’s 11,000 inland lakes based on data from 1999 to 2013. The lakes were selected because, at a minimum of 20 acres, they were the largest in the state.
The study divided Michigan into five distinct ecoregions with southeast Michigan considered the Huron/Erie Lake Plains Region. The smallest ecoregions, located in the southern Lower Peninsula of Michigan, had mostly eutrophic lakes and more hypereutrophic lakes than the other ecoregions. The study determined that most of Michigan’s inland lakes fell under the mesotrophic TSI class.
The 2016 study concluded that a long-term regional and spatial picture for lake managers can be created using different monitoring programs for Michigan’s inland lakes. Although in situ data is a valuable resource, it is not physically and economically feasible to consistently measure thousands of lakes.
Michigan's Department of Environment, Great Lakes and Energy (EGLE) incorporated this study, combined with water sampling data, to make trophic state and water quality standard assessments in the Integrated Report it must submit to the EPA in even years according to the Clean Water Act. In its latest report, submitted to the EPA in 2024, EGLE stated that satellite imagery, when combined with other methods, proved to be a useful tool in understanding if complaints called in about blooms in a certain inland lake have occurred in consistent patterns over time. The report said that EGLE since 2016 has used this method to monitor blooms along the shoreline of Saginaw Bay.
The 2024 report stated: “Aerial imagery … may be useful in corroborating whether blooms have occurred historically. The specific time frame of the images used should be available for perspective when relating to other available information. Other, more frequently obtained images, such as those used in various forecasting efforts by NOAA, are useful in their ability to aid in the evaluation of both extent and duration of blooms.”
One researcher working with EGLE, who specializes in the trophic stages of a waterbody and how it becomes a potential breeding ground for harmful algae blooms, is Michael Meyer of the USGS.
From space, Meyer said detectable lakes come in colors of brown, blue and green. He spends his time correlating satellite imagery data with in situ data to determine if those colors mean a healthy body of water or if there is trouble. For example, the greening of a lake could mean there is healthy vegetation or wetlands, or a potential algal bloom. A browning of a lake might indicate the presence of wetlands.
Meyer said while remote satellite sensing imagery is coming along, it will never eliminate the need to physically test the water for contaminants.
“Satellites will not replace local (water sampling) anytime soon because of all the optical challenges we have from space,” Meyer said. “First, clouds get in the way and obscure the data. Most satellite imagery sensors are designed to study land topography and not water because water absorbs so much light. NASA satellites can capture only a small fraction of light that scatters back up into space. So, one area we are working on is creating different correction algorithms in our computing to extract the water’s reflectance. From a water management perspective, satellite images can offer a historical perspective of picking out trends of when lakes experience a bloom, how long the bloom lasted, and if it occurred at certain times of the year and under what conditions. Though we cannot pinpoint smaller lakes, we can detect emerging patterns and trends of algal blooms within a certain watershed in the months when they are expected to bloom – from May to October, for example.”
An aquatic ecologist by trade, Meyer said that cyanobacteria often get a bad rap. It is a wide species and many are not the culprits of toxic blooms. The presence of cyanobacteria does not always spell bad news for a lake.
“The goal of satellite data is to find proxies for when a high presence of cyanobacteria might be a signal for a situation when they could produce toxins down the road,” Meyer emphasized. “Just because a satellite identifies something that could be a bloom, there remain a lot of unknowns. The only way to confirm a bloom is with in situ data.”
EGLE spokesperson Hugh McDiarmid Jr. said from the fisheries perspective, he expects that these statewide data will provide an important complement to our existing water quality data.
“It will help to inform assessments of eutrophication, cold water fish habitat in the face of climate change as well as the increasing presence of nutrients for fish productivity modeling,” McDiarmid said. “I anticipate that as we bring on additional inland lakes research capacity within our Institute for Fisheries Research and through the USGS Coop established with MSU, it will be a dataset that we can leverage into the future.”
Michael Sayers is a research scientist with Ann Arbor’s Michigan Tech Research Institute (MTRI), which synergizes sensor and information technology for practical applications. Among his projects, Sayers worked with Russian researchers and developed a new algorithm to study coastal and inland waters to estimate levels of chlorophyll, dissolved organic carbon and suspended minerals. Sayers and other MTRI researchers have provided data to the federal Great Lakes Restoration initiative based on satellite imagery.
In recent applications, MTRI coupled satellite imagery data with field studies to analyze this summer’s algal blooms in Muskegon Lake. In another study with NASA, Sayers is using imagery from European satellites to examine lakes and reservoirs in western states suffering from algal blooms as a result of wildfires which erode the soils and cause nutrient runoff into waterbodies.
Though the images come in handy for examining the Great Lakes from space, Sayers said it’s much harder to cull accurate information about smaller lakes.
“The challenge in examining our smaller bodies of water is in the pixel size of the satellite images,” Sayers said. “If the pixelation can be made smaller, we can get images of smaller targets, so right now we do have some gaps in our data when it comes to smaller lakes.”
All this will change as the imaging technology gets better, Sayers hinted.
He is encouraged by the progression the NASA PACE satellite technology is taking to detect those small spectral differences between the greens of healthy algal vegetation versus harmful algae blooms.
Sayers hopes that MTRI will also have access to imagery from a constellation of satellites which will be launched this fall by the European Space Agency and will also feature hyperspectral capabilities.
“These satellites will give us ever higher resolution images capable of seeing even smaller targets,” Sayers said. “The satellites being launched right now will unlock more abilities to pixelate smaller, inland lakes. We have already worked with images from existing satellites to examine the buildup of submerged vegetation that is becoming a problem in the Great Lakes.”
He continued: “We are on the forefront of satellites having hyperspectral sensing capabilities. This will be an exciting and important development to detect small changes to pick up multiple shades of blue and green. We are on a research team to validate the instrumentation and I believe it will give us a lot more confidence in the data which will more accurately distinguish between non-harmful and harmful blooms.”
Researchers in states with an abundance of lakes like Minnesota and Wisconsin are collaborating with Michigan researchers as well as EGLE to create a broader regional imagery of inland lakes and how to monitor, maintain and improve their quality. Leif Olmanson, a remote sensing specialist at the University of Minnesota, leads a team of researchers who have leveraged high-resolution Landsat-derived images to create digital maps with thousands of datasets for water quality analysis.
The project is called the Minnesota LakeBrowser, an online interactive lake water quality monitoring tool that includes over 10,500 lakes measured nine times from 1975 to 2015. From 2017 onward, LakeBrowser has used an automated imagery processing system to integrate all available Sentinel 2 imagery to measure clarity, chlorophyll and colored dissolved organic matter regularly between May and October.
Now, Olmanson and the USGS are about to complete a project funded by EGLE to create a similar tool for Michigan.
"The satellite imagery gives us a comprehensive, statewide view of lake conditions that we could never get from just field sampling alone," Olmanson explained. "We can see trends developing over time and get an early warning system in place for harmful blooms."
One thing all researchers using this data agree upon is that the Great Lakes region has very few cloud-free days. Those clouds, among other atmospheric disturbances such as aerosol, smog and notably the smoke from last year's Canadian wildfires, can gunk up an image.Therefore, Olmanson’s team has created an automated process to correct each image by running an algorithm that identifies clouds and other atmospheric obstructions so researchers are left with cleaner images of the land and bodies of water.
Olmanson explained the process.
“Satellite images allow us to observe wide swaths of land and water and observe changes over years and decades,” Olmanson said. “But these images are taken above the atmosphere so we must clean up and normalize the imagery. We can automatically identify and remove all the clouds. When we are done, we have clear pixelation of a lake. Then we start making our models by linking up the thousands of data points from the cleaned satellite images with in situ data from water samples taken from lakes to check for consistency.”
Once the imagery is cleaned, that data is layered into the digital map. Pristine healthy lakes show up as blue in more rural areas, and any possible contamination of algal blooms shows up as red or yellow to indicate pollution in populated areas or agricultural nutrient runoff from farmed regions.
Detecting and predicting that potential for contamination with conventional, land-based methods would be akin to looking for a needle in a haystack, explained Olmanson. Most states just do not have the resources for that manpower, he added.
“You have lots of areas of blue here that typically do not have blooms, but from these images taken monthly, one can see the satellites have picked up some algal activity,” Olmanon said. ”These are very remote lakes that are hardly monitored if they are monitored even at all from the land.”
Aside from scientific and environmental applications, Olmanson said the satellite data reporting on lake health trends is also proving helpful to casual users. Real estate agents in Minnesota are using the water quality information to advise clients, while recreational companies are adjusting their operations based on bloom predictions. Olmanson used his data before he purchased his own lakeside home.
And for lake managers and environmentalists at the state level, Olmanson said this wealth of data can be invaluable for algal bloom forecasting and response efforts – if they have the computing power.
"EGLE has been very supportive and is excited about receiving the data, which will be very beneficial for Michigan,” said Olmanson. “Being able to see the whole picture across wide areas and even remote lakes that would require an environmental monitoring team to travel and manually take samples will help EGLE identify problem areas and allocate resources more effectively. We can target our monitoring and outreach to the lakes most at risk, so we can get ahead of blooms before they become a public health issue."
Just as in Minnesota, as access to satellite imagery becomes more widespread, researchers at federal and state levels are creating online mobile and laptop applications for public use.
Bridget Seegers is a research scientist with NASA’s Ocean Ecology Lab and part of the Cyanobacteria Assessment Network (CyAN), a project started in 2015. CyAN is a collaboration between NASA, the USGS, the National Oceanographic and Atmospheric Agency (NOAA) and the EPA to provide an array of users with constantly updated information on the presence of cyanobacteria in 2,000 lakes across the country that are larger than one kilometer in diameter and have a geographical name. These lakes include the largest lakes in Oakland County.
"I grew up on a small lake in Wisconsin, so I know firsthand how important lake water quality is to so many people," Seegers said. "That's a big part of what motivates my work – using the best available technology to help monitor and protect these vital natural resources."
Seegers said CyAN created computer programs to crunch satellite observations from Sentinel-3, NASA supercomputers produce weekly reports on the color – and other water quality information.
All of this information can be accessed by the public through a downloadable web-based app that was launched in 2019. In July 2024, researchers began testing an experimental forecasting computing model to produce weekly forecasts.
CyAN users can mark a particular lake with a pin – which will appear as green if the lake appears bloom-free, yellow if algae are present but below a certain threshold of concern, or red, indicating that a bloom is likely present.
"Water managers can use the app to quickly identify lakes with high cyanobacteria levels, so they can send crews to collect samples or even close beaches if needed," Seegers explains. "But it's also great for people planning a day on the water. You can check the app and see which lakes are safest for swimming, boating or fishing."
However, Seegers notes that the satellite data does have some limitations. Aside from only being able to detect lakes larger than one kilometer, cloud cover can obscure the most accurate data collection.
"We're constantly working to improve the technology and provide the most comprehensive, up-to-date information possible," says Seegers. "But even with those challenges, I'm proud of what the CyAN project has accomplished to empower people to make informed decisions about their local lakes."
One local lake included in CyAN that had some significant counts of cyanobacteria is Judah Lake in Orion Township. The private lake is 115 acres and is populated by many waterfront homes. As of August 10, the app indicated that the lake’s cyanobacteria count measured at 586,138 cells per milliliter (ml). This was up from a July 27th reading of 75,633 cells/ml.
In contrast, according to the app, Walnut Lake in West Bloomfield in February 2024 had a reading of over 255,000 cells/ml but dropped to around 100,000 cells/ml two weeks later. Readings from Pontiac Lake recreation area at the end of March were measured at 28,000 cells/ml.
What do those numbers mean? What does this quantification tell the average user of when it is safe to go in the water?
According to Sarah Holden, environmental quality specialist within EGLE’s Water Resources Division, Michigan does not have water quality standard thresholds for cyanotoxin concentrations or assessment methods that use a cyanobacteria cell counts as a threshold to determine water quality standard attainment.
Holden stated that EGLE works with the Michigan Department of Health and Human Services (MDHHS) and local health departments on issues related to the recreational use of surface water and public advisories.
When assessing a lake for recreational uses, EGLE deploys a variety of factors, including actual water quality data. Water will be sampled from a Secchi depth for phosphorus and chlorophyll a and the water will be monitored for frequency, intensity, and duration of algae blooms – both green algae and cyanobacteria.
Holden said EGLE couples this data with high-resolution satellite imagery to identify blooms and to corroborate whether blooms have occurred historically.
Speaking at a federal level, Amalia Handler, a biologist with the EPA’s Pacific Ecological Systems, admits that there is not yet an exact science in quantifying harmful thresholds of the toxins that have the potential to spurn harmful algae blooms.
Holden was the lead scientist of a 2023 study that identified over 2,000 lakes across the country, including 126 lakes in Michigan at risk of toxic cyanobacterial blooms utilizing satellite imagery and field surveys. Included in this list were Orchard, Pine, and Long lakes in Oakland County.
Handler relied on the World Health Organization’s (WHO) 1999 guidelines which list counts of 20,000 cells per ml and over put human and aquatic health at moderate risk and 100,000 per ml and over at an acute risk. Since the study was published, Handler said the WHO had migrated away from this measurement and instead recommends paying attention to visual cues.
“This is something that we as researchers are struggling with,” Handler said. “Measuring at cells per milliliter is one of the less technologically intensive ways of quantifying cyanobacteria in the water. It requires taking water samples and is one of the most used methods for understanding cyanobacteria counts. But the challenge is that cyanobacteria is a broad set of organisms. Not all of them are harmful or potentially toxin-producing. The best advice the WHO and the EPA are giving now is to look for visual cues such as the color and clarity of the water rather than quantifying chlorophyll counts to determine the presence of toxic algae. The field is still in the process of coalescing around what these thresholds should be. And thresholds for drinking water versus using water recreationally are different.”
While a human may not think to drink water from a lake, their four-legged best friend will lap it up without hesitation. Handler lives in Oregon, where a series of deaths of otherwise healthy dogs who swam in inland lakes prompted the Oregon Department of Environmental Quality to create thresholds and resources for cyanobacteria counts and the presence of harmful algae blooms in 2020. Handler said that Oregon maintains their application that pulls in cyan imagery and sets individual thresholds for lakes with bloom potential.
Handler’s study stated that remote sensing has “vastly expanded data on lake cyanobacteria, yet there is still acute need for tools that identify which waterbodies are at-risk for toxic cyanobacterial blooms. Algal toxins cannot be directly detected through imagery but monitoring toxins associated with cyanobacterial blooms is critical for assessing risk to the environment, animals, and people.”
The purpose of the study was to develop a computer modeling system that combined satellite imagery with field surveys to determine and predict algal blooms.
Of the lakes surveyed, Handler and other scientists included said they sympathize with people from lakeside residents to state and county environmental regulators over what can be done with all this remote sensing imagery and said they are looking to federal authorities for guidance.
“Just because cyanobacteria are present, it may not necessarily be harmful to people or pets,” Handler said. “Unfortunately, we can't tell by looking at a bloom on its own to understand whether it contains toxins. Most states I am working with have the motto of: ‘When in doubt, stay out.’ If you see something that looks like scum on the surface of the water, humans and pets should stay out and come back another time."
Patrick Hanly is research program coordinator for the Department of Fisheries and Wildlife at Michigan State University. His work has been instrumental in the development of the LAGOS-US Landsat project – a broad, long-term water quality study and database that combines remote sensing water quality estimates from satellite imaging data with in situ data examining 136,777 lakes in the United States from 1984 to 2020. Hanly said increasing access to the rapid growth of cloud computing capabilities to researchers will make it easier to crunch thousands of points of satellite data.
“In the past five years, there has been a huge boom in cloud computing and processing,” Hanly said. “This enables scientists and researchers to compile satellite data that stretches back several decades.”
The amount of data is vast. Hanly said each highly pixelated satellite image can be close to a gigabyte of data.
“When a researcher receives a new Lansat image every 16 days, you are talking about the need for storage capacity running in the petabytes. (A petabyte is 1 million gigabytes). The way we receive these images over the internet is a stark comparison compared to the early 2000s when we (as researchers) would have to order a disk from NASA in the mail that would only contain a few images of one small geographical region. Now, with all this processing power, we can take decades worth of images and go back in time so we can make better predictions for lake conditions in the future.”
Hanly said it was also immensely advantageous to consolidate decades of data into one system. “In the past, you would have a lake manager recording Secchi data like chlorophyll or other nutrient data in a disjointed record-keeping system. Data would be stored here and there. Now, with LAGOS, we have compiled decades of all this in situ and satellite data into this one product.”
The LAGOS Lakes platform provides a comprehensive dataset for understanding long-term changes in lake health. Hanly explained that the platform enables researchers to match up satellite imagery with in-lake measurements, allowing them to make predictions about water quality in lakes that don't have regular monitoring. Hanly stresses that scientists still do not fully know what causes HABs – because they can show up in cold waters as well as warm waters – and the difference between the appearances of beneficial and harmful algae can be extremely subtle. But there is no such thing as too much data, whether it comes from the skies above or at the earth’s surface.
“In the next couple of years, with the kind of real-time cloud processing we are going to have available, people are going to be able to explore what is going on in their lake right now or look back on occurrences over the past year. We are going to see the pipeline moving faster when satellites from space will be flying over your lake every few days to take data-rich images, assuming it is not a cloudy day.”
Determining the health of lakes
On the ground, levels of cyanobacteria and algal blooms are measured on the Trophic State Index (TSI), as the word trophic means “relating to nutrition.” Ranked from 1 to 100, the TSI rates surface waters such as lakes, ponds and reservoirs based on biological productivity in the water. Generally, the lower the TSI, the healthier the water.
Oligotrophic – 30 or below. These lakes have little algal productivity. This water is suitable for drinking and is ideal for water sports. These lakes exhibit clear water with good visibility but may not provide the necessary nutrients and algae to maintain a healthy environment for fish and wildlife.
Mesotrophic – between 30 and 45 TSI. Lakes have a mid-range of nutrients and enough nutrients to support algae, aquatic plants, and wildlife.
Eutrophic – between 46 and 70. Lakes have good or sufficient nutrients and have fairly high productivity. Their nutrients can support an abundance of algae, aquatic plants, and wildlife.
Hypereutrophic – between 71 and 100 TSI. When lakes rise to this ranking, they have an overabundance of nutrients and the potential to support the highest level of biological productivity (e.g., an abundance of algae, aquatic plants, birds, fish, insects and other wildlife.) These water bodies have the greatest potential for widely ranging dissolved oxygen conditions, which can have a detrimental effect on native plants and animals.