Last updated: January 19, 2025
Article
A New Way to Gauge Risk of Toxic Blooms in a National Park
Harmful algal blooms endanger drinking water sources and aquatic life but are notoriously hard to study and manage. We created a “scorecard” to assess the likelihood of toxic blooms. It can help water managers focus on the most effective actions.
Image credit: courtesy of Matt Fowler
Each year, millions of visitors to Acadia National Park in Maine swim, paddle, ice skate, fish, and bask in the serenity of the park’s 26 lakes and ponds.
Thousands of Mount Desert Island residents consume drinking water from six of these lakes. Jordan Pond, the “crown jewel” of Acadia’s lakes, is often described as the clearest lake in Maine. “Acadia’s pristine lakes and ponds are fundamental park resources,” said Kevin Schneider, the park’s superintendent. “They’re also some of the most visited destinations for the park’s visitors.”
But over the last decade, harmful algal blooms (HABs), which can be toxic to people and animals, have emerged in lakes across Maine and at other national parks. Scientists at the University of Maine and the National Park Service recently completed a study to understand how vulnerable Acadia’s lakes are to HABs. The study showed that two lakes in Acadia that supply public drinking water were at risk of developing HABs. From those results, the other study authors and I developed an early warning system to identify lakes most susceptible to HABs. Our work will help safeguard park visitors and nearby communities from this emerging threat.
A Toxic Competitor, Fueled by Extreme Weather
HABs are caused by the overgrowth of naturally occurring cyanobacteria, also called “blue-green algae.” These cyanobacteria can produce toxins that can sicken people and animals. In some cases, dogs have developed neurologic problems and even died after swimming in affected water. HABs are increasing across the globe due to climate change. Northern lakes in particular (those north of 45°N latitude; Acadia is 44.3°N) are changing fastest due to increasing air temperature and more frequent and severe precipitation. Extreme storms wash soil and vegetation from the landscape into lake water, delivering nutrients that can fuel HAB growth.
Image credit: NPS
Lake surface water temperature is another factor influencing HABs. It closely follows air temperature, and in the last century, average annual air temperature in Maine has increased by more than 3°F. As a result, winter ice is disappearing earlier in many Maine lakes, lengthening the growing season for freshwater algae. The growth of HABs can also be influenced by a lake’s trophic state, which is a measure of its physical, chemical, and biological characteristics. One of these characteristics is the amount of nutrients like nitrogen and phosphorus that are dissolved in the water. High concentrations of these nutrients can feed cyanobacteria and promote algal blooms.
Image credit: courtesy of Kiernan McMorrow
Lakes that scientists call “oligotrophic” have low nutrient levels. This accounts for their attractive clarity, enabling visitors to see far into the water column. “Eutrophic” lakes are murky and have greater nutrient levels. “Mesotrophic” lakes fall somewhere in between. Acadia’s lakes exhibit a range of trophic states. Jordan Pond is a deep, clear, oligotrophic lake. Seal Cove Pond is a shallower, oligotrophic to mesotrophic lake. Witch Hole Pond is shallow and mesotrophic to eutrophic, often exceeding EPA nutrient criteria for good water quality. (By law, Maine and all other states must report the water quality of their lakes, rivers, and ponds to the EPA every two years.)
Some types of toxic cyanobacteria can control their buoyancy in the water column, allowing them to move to depths with optimal light and nutrient conditions.
Acadia’s lakes are home to many types of naturally occurring algae, most of which don’t produce toxins or experience harmful blooms. But toxin-producing cyanobacteria can outcompete other types of algae at prolonged water temperatures greater than 25°C (77 °F). They can use special pigments to shield from damaging UV rays, while simultaneously blocking light to other types of algae deeper in the water column. Some types of toxic cyanobacteria can control their buoyancy in the water column, allowing them to move to depths with optimal light and nutrient conditions. Cyanobacterial blooms have typically been associated with eutrophic lakes that have high nutrient levels. But they can also do well and outcompete other algal types in low-nutrient, oligotrophic lakes.
Multiple Methods Were a Must
HABs can be difficult to predict, identify, and manage. This is because their occurrence is affected by many physical, chemical, and biological factors, which can vary widely over the landscape and over time. But using multiple methods to monitor HABs has proven effective for assessing their impacts. We measured chlorophyll a, phycocyanin, and dissolved oxygen levels and in Acadia’s lakes, because these are reliable early warning indicators of blooms. Chlorophyll a is the primary photosynthetic pigment in algae and can be used to determine algal biomass. Phycocyanin is a blue aquatic plant pigment that also captures light energy for photosynthesis, similarly to chlorophyll. Cyanobacteria produce both chlorophyll a and phycocyanin.
Dissolved oxygen is crucial for lakes and streams. It’s how fish, for example, obtain oxygen when water passes over their gills. Low dissolved oxygen levels can lead to phosphorus release from sediments, fueling HAB growth. This is especially true in summer when temperatures are high and the water is stratified. Anoxic (no oxygen) conditions are increasingly common in northern lakes as a result of climate change and land use.
Image credit: courtesy of Rachel Fowler
In February 2020, we equipped Jordan Pond, Seal Cove Pond, and Witch Hole Pond with buoys that had sensors for recording water temperature and dissolved oxygen once every hour. We set up the buoys near the deepest parts of the ponds. This was 13 meters (about 43 feet) in Seal Cove Pond, 9 meters (about 30 feet) in Witch Hole Pond, and 32 meters (about 105 feet) in Jordan Pond. We positioned the dissolved oxygen sensors at two and three meters below the water surface and at two meters above the lake bottom. We placed the temperature sensors one meter apart from the surface to the lake bottom, or to 30 meters in the case of Jordan Pond. The sensors remained submerged until November 2020.
We manually sampled lake water once through the ice in February 2020 and then about every two weeks from mid-May to November 2020. We measured temperature and dissolved oxygen at one-meter intervals. We took continuous measurements of temperature, chlorophyll a, phycocyanin, and nutrients like phosphorus. We also measured water clarity with a Secchi disk. We analyzed the water samples at the University of Maine Sawyer Water Research Laboratory.
Using sonar, we created maps of the lake beds. This helped us get water volume estimates to use when modeling the amount of dissolved oxygen and phycocyanin.
Using sonar, we created maps of the lake beds. This helped us get water volume estimates to use when modeling the amount of dissolved oxygen and phycocyanin. We paired these data with the park’s long-term historical monitoring data (1995-2020). We also paired them with air temperature, precipitation, and other meteorological information (2013-2022) from a weather station on Jordan Pond House, a historic tea house in the park.
Ponds in Depth: Key Findings
We chose the three study ponds because of their trophic status and physical traits. At about 260 acres, oligo-mesotrophic Seal Cove Pond occupies the largest area. It’s a relatively shallow lake with a mean depth of 13 feet. Oligotrophic Jordan Pond is about 188 acres and 69 feet deep on average. Mesotrophic Witch Hole Pond is the smallest at about 25 acres and 7 feet average depth. Seal Cove Pond is only partially inside the park boundary and has houses on its western shore. Jordan Pond and Witch Hole Pond are completely inside the park boundary and thus more protected from the impacts of development.
Temperature
In 2020, the surface temperature in Jordan Pond was consistently lower than in the other two ponds. The highest temperature recorded at the top sensor in Jordan Pond was 24.9°C (76.8°F). In Seal Cove Pond, it was 28.1°C (82.6°F), and in Witch Hole Pond, it was 28.6°C (83.5°F). Once the lakes stratified (warmed up enough that the warmer, lighter water stayed on top of the colder, heavier water until autumn), the bottom temperature in Jordan Pond remained around 6.6°C (43.9°F).
Image credit: courtesy of Rachel Fowler
Historical records for the park show that average surface water temperatures have increased by almost 2°C (about 3.6°F) since 2013.
In contrast, the bottom temperatures of the other two, shallower ponds increased throughout the season. They peaked at 20°C (68°F) for Seal Cove Pond and 16°C (60.8°F) for Witch Hole Pond. Historical records for the park show that average surface water temperatures have increased by almost 2°C (about 3.6°F) since 2013. Our data are consistent with this warming trend, indicating that Seal Cove Pond and Witch Hole Pond are becoming more susceptible to HABs.
Dissolved Oxygen
In 2020, the dissolved oxygen levels in Jordan Pond's surface water started at 13 milligrams per liter (13 parts per million) and dropped to a minimum of 7 milligrams per liter by the end of August. The bottom sensor in Jordan Pond showed a more gradual decline in dissolved oxygen and often recorded higher levels than the top sensor. In Seal Cove and Witch Hole Ponds, the surface dissolved oxygen levels were similar to those in Jordan Pond.
But the bottom sensors in Seal Cove and Witch Hole recorded a rapid decline in oxygen levels, with anoxic conditions in Seal Cove Pond from late July to early September and in Witch Hole Pond from early July to late September. Lakes like Seal Cove and Witch Hole Ponds showed prolonged anoxic conditions, which may increase the risk of HABs. But Jordan Pond maintained higher dissolved oxygen levels, showing less immediate risk for toxic blooms.
Phycocyanin
Phycocyanin levels were generally low in Jordan Pond and Seal Cove Pond during the ice-free season, with a few summer peaks. In Jordan Pond, phycocyanin levels increased after the ice melted and again in July. In Seal Cove Pond, a peak occurred in early September. Witch Hole Pond had phycocyanin at various depths during the ice-free months.
In Jordan Pond, the increase in phycocyanin levels after ice melt and in July may indicate that conditions favored HAB growth, as warmer water enhances stratification and supports cyanobacterial buoyancy. The September peak in Seal Cove Pond could have resulted from seasonal nutrient dynamics and lower dissolved oxygen, which also favors cyanobacteria. The persistence of phycocyanin in Witch Hole Pond could mean there were favorable conditions for HABs throughout the season.
Ranking the Risk
We analyzed our results using R, an open source statistical programming language. The results of our water quality sampling indicated that shallow lakes like Witch Hole Pond could be more vulnerable to HABs. This is because of the lakes’ high water temperature, low dissolved oxygen, high nutrients, and relatively high levels of phycocyanin. Deep, oligotrophic Jordan Pond did not show early warning signs of HABs. We expected that park lakes similar to Jordan Pond would be equally resistant to HAB occurrence.
Lakes partially outside the boundary, like Seal Cove Pond, are more susceptible to nutrient loading and thus HABs.
Building on these findings, we created a risk assessment that incorporated Acadia lake data from the Maine Department of Environmental Protection and Maine Department of Inland Fisheries and Wildlife. We looked at lake traits that affect vulnerability to HABs, like trophic status, maximum depth, and whether a lake was wholly inside the park boundary. Lakes partially outside the boundary, like Seal Cove Pond, are more susceptible to nutrient loading and thus HABs. This is because there’s more development along or near their shorelines.
Then we calculated a weighted average using those traits. We developed three risk levels based on the weighted averages: low concern (weighted average of less than 1.8), moderate concern (1.8 to 2.0), and considerable concern (greater than 2.0). These levels gave us a “scorecard” to rank the risk of future HAB formation on all Acadia lakes.
Image credit: courtesy of Rachel Fowler
A graph labeled “Risk Assessment Scorecard,” with horizontal lines and circles in orange, yellow and green. Some circles are outlined in black. The horizontal lines indicate the numerical values for the risk index. There is a legend on the lower right to indicate what the colors mean. Green=low, yellow=moderate, and orange=considerable risk of developing HABs. Circles outlined in black mean the water body is a source of drinking water. The y axis is labeled “Risk Index,” and the x axis is labeled with the names of 17 lakes and ponds.
From left to right on the x axis are Bubble Pond (green circle with a black outline; risk index between 1.3 and 1.5), Eagle Lake (green circle with a black outline; risk index between 1.5 and 1.7), Jordan Pond (green circle with a black outline; risk index between 1.5 and 1.7), Upper Breakneck Pond (green circle; risk index between 1.5 and 1.7), Long Pond (green circle with a black outline; risk index 1.7), Aunt Betty Pond (yellow circle; risk index between 1.7 and 1.9), Lake Wood (yellow circle; risk index between 1.7 and 1.9), Lower Hadlock Pond (yellow circle with black outline; risk index between 1.7 and 1.9), Upper Hadlock Pond (yellow circle with black outline; risk index between 1.7 and 1.9), Echo Lake (yellow circle; risk index 1.9), Seal Cove Pond (yellow circle; risk index 1.9), Bear Brook Pond (yellow circle; risk index between 1.9 and 2.1), Lower Breakneck Pond (yellow circle; risk index between 1.9 and 2.1), Round Pond (orange circle; risk index 2.1), Witch Hole Pond (orange circle; risk index between 2.1 and 2.3), Hodgdon Pond (orange circle; risk index 2.3), and Seawall Pond (orange circle; risk index slightly above 2.3).
Though replication (studying multiple similar examples to ensure reliable results) may sometimes be desirable in scientific research, it’s not always feasible. We had finite resources for sampling, so we chose lakes that were most representative of all the other lakes in the park instead of studying multiple lakes of the same type. In fact, researchers have demonstrated that large-scale, unreplicated, natural experiments (LUNEs) are valuable for policy and planning, especially in regard to lake water quality.
The risk assessment scorecard backed up our predictions.
The risk assessment scorecard backed up our predictions that lakes like Jordan Pond have the fewest risk factors for future HAB occurrence. Lakes like Witch Hole Pond, on the other hand, have more risk factors and may be more vulnerable to future HABs. We found that four (Bubble Pond, Eagle Lake, Jordan Pond, and Long Pond) of the six lakes in Acadia National Park that are public drinking water sources were in the low-risk category. But two drinking water lakes (Upper Hadlock Pond and Lower Hadlock Pond) were at moderate risk of developing future HABs. Several lakes that are not sources of drinking water were found to have considerable risk of future HABs.
“We’re fortunate at Acadia National Park to have an active lake monitoring program with a rich data history,” said Bill Gawley and Kathleen Brown, two biologists who work for Acadia National Park. “It’s comforting and empowering to have a tool that utilizes these data to track lake vulnerability status over time and identify when a location is crossing into a higher-risk category.” Gawley and Brown said the scorecard allowed them to focus their attention and efforts on the most vulnerable water bodies.
A Template for Protecting People, Pets, and Water
Our risk assessment scorecard showed that oligotrophic lakes like Jordan Pond are better at resisting changes that could lead to HABs. Oligo-mesotrophic lakes like Seal Cove Pond are at moderate risk, but shallow lakes like Witch Hole Pond face considerable risk. They may experience nutrient buildup, making them more likely to have blooms. Still, studies show that climate change does not necessarily cause increases in algal blooms. This seems especially true for lakes restored and protected by the Clean Water Act of 1972. Long-term water quality monitoring and consequent actions may have helped prevent HABs in those lakes.
Image credit: courtesy of Jacob Ward
The scorecard can be used as a template for protecting visitors, pets, and drinking water from HABs in other national parks. Based on the scorecard rankings at Acadia, we proposed a range of ways to safeguard national park visitors and communities. For example, the National Park Service’s air quality and water quality monitoring programs can prioritize sampling shallow, nutrient-rich lakes as summer water temperatures rise, and expand cyanobacterial monitoring. Then when key water quality parameters of higher-risk lakes exceed baseline values, parks or municipal water companies can take certain actions:
- Examine the data further.
- Let visitors know what’s going on.
- Collect water for identifying cyanobacteria and analyzing toxins.
- Redirect drinking water intakes away from surface or subsurface HABs.
“The HABs scorecard will prove useful in identifying concerns and treatment strategies a water utility may need to prioritize to provide clean drinking water to its customers.”
The scorecard can also be used by land managers outside the National Park System or by public utilities. “The HABs scorecard will prove useful in identifying concerns and treatment strategies a water utility may need to prioritize to provide clean drinking water to its customers,” said Kate Warner, the Source Water Program manager for Maine’s Rural Water Association. “Additionally, and importantly,” Warner added, “the scorecard is an easy tool to understand and use to facilitate watershed activities that promote good water quality. These activities include things like nutrient reduction from altering lawn care, managing septic systems, or education in local schools. Community involvement is key in source water protection, and the HABs scorecard will help jumpstart these efforts and monitor success.”
We used historical water and air quality records, long-term buoy data, and manual water quality sampling to develop a robust dataset for this project. But expanding sampling to include more years and more lakes would further improve it. Environmental DNA (eDNA) and sediment DNA (sedDNA) studies could help us identify harmful cyanobacteria species that are too small to observe otherwise. Collaborative cyanobacteria monitoring programs like bloomWatch and Cyanoscope, SPATT sampling (a form of toxin tracking), and qPCR analysis could also help us further understand conditions that lead to HABs and how to mitigate them.
