Beneath calm waters, colors and changes driven by climate change are transforming aquatic ecosystems. Harmful algal blooms have become increasingly common across the Pacific Northwest, a warning sign of their potential spread across the continental U.S.
Fig 1. Algal Blooms Across Great Lakes1
Using hyperspectral profiles from the 2022 USGS dataset on HABs and other algae, we examined how variations in location and water chemistry influence which species thrive, and what these patterns might reveal about the future of our aquatic health and local communities and ecosystems.
The image to the right was taken by aerial drones above the Great Lakes, a prime example of an extreme algal bloom, in which the green cyanobacteria permeating the water can clearly be seen in puffs of bio-organic material. Toxic blooms like these have been on the rise: 169 toxic blooms were reported in 40 states in 2017, compared to only three blooms in 2010. Lake Erie’s algal blooms in 2014 even endangered drinking water for Toledo residents.
Definition: harmful • algal • bloom2
HABs occur when colonies of algae grow out of control and produce toxic or harmful effects on people, fish, shellfish, marine mammals and birds. The human illnesses caused by HABs, though rare, can be debilitating or even fatal.
HABs occur when colonies of algae grow out of control and produce toxic or harmful effects on people, fish, shellfish, marine mammals and birds. The human illnesses caused by HABs, though rare, can be debilitating or even fatal.
Algal blooms can cause the mildest to the most severe of symptoms: from basic headaches to even cancer. There is no legal limit on the toxins that these algaes produce, which means that municipalities have no obligation to be held accountable for testing for the amount of toxins the blooms produce.
The blooms themselves can be difficult to identify at first glance. While some appear clearly at the surface of the water and are relatively easy to identify, others occur at the bottom of water bodies and can’t be pinpointed visually. This is particularly concerning as harmful algal blooms have become increasingly common in water bodies across the United States and there are currently many challenges in tracking these algal blooms when they occur.
One way scientists are tracking these blooms is by using hyperspectral imaging3, a technique which obtains the light spectrum for each pixel in an object. We are able to identify the object based on the information given from its spectrum. The algae within our dataset were collected and identified via this method, creating a unique distribution of colors and features, which aided in the identification of different taxa.
Definition: ta • xa
In the classification of protists, plants, and animals: taxon is any unit used in the science of biological classification.4
Definition: morphological • id
Identifying organisms based on their observable physical characteristics, such as shape, size, color, and structure.
Fig 2. Map of Algae Taxa Across the United States
Initial Mappings
The majority of the samples from our dataset are taken from the state of Oregon, as demonstrated by the concentrated cluster of samples within the western part of the state. There are also samples from Colorado, Illinois, and parts of the East Coast. Notably, green algae is found on both the West and East coast, consistent with its ubiquity across the United States. Green algae is a large and diverse group belonging to the taxa chlorophyta, and are found in a wide variety of habitats including in ponds, lakes, rivers, and soil. They are an indicator of health5 in balanced ecosystems, providing food and shelter for many species of aquatic insects and fish.
The largest diversity and variety of algae was found in McKenzie River, with Upper Klamath lake coming in second. Amongst all algae found in McKenzie River, those with known toxins include Nostoc, Nostoc Spongiform and Phormidium6. Potentially problematic algae if overgrowth occurs includes Cladophora7 and Spirogyra8.
Fig 3. Stacked Bar Chart of Morphological IDs per Location of Sample
Over 40% of the algae found in McKenzie River is harmful and problematic. These results are consistent with record-breaking algal blooms spreading across the Pacific NorthWest. In particular, there was a notable wave of harmful algal spread in 2025.9
Not only do these algal blooms harm the environment, they also negatively impact the communities around them, including closures of recreational razor clam harvests in Oregon and Washington, as well as closing of large portions of the Washington state Dungeness crab fishery and some of the sardine and anchovy fisheries in California.
At a tiny width of less than 20 micrometers, Dolichosperum was found in the majority of the samples, including in the CO River Basin - Iola, Deschutes River, OR, Illinois River Basin, and Upper Klamath Lake, OR (UKL).
Dolichospermum is a toxic and harmful cyanobacteria10. Climate change is creating more favorable conditions for harmful algae blooms which perpetuate Dolichospermum, and these factors include but are not limited to higher temperatures, lower stream flow11, and stormwater run-off. Other potentially harmful types of algae present in the samples include:
Fig 4. Dolichospermum at 10 micrometers12
I. Aphanizomenon, which produces harmful anatoxins which cause convulsions, respiratory problems, and dizziness and drowsiness.
II. Gloeotrichia, which produces microcystins that result in abdominal pain due to inhibited liver function.
III. Microcystis, which produce dermatotoxins which contribute to skin irritation and rashes.
With the explosion of these harmful algae, the toxins they produce inevitably end up in the drinking water of humans. They are also not only harmful to people but to pets, and effects often take place in as short of a time as 24 hours.13
Overall, 55% of our samples contain Dolichospermum, consistent with the statistics regarding commonly occurring, harmful cyanobacteria. The most commonly occurring cyanobacteria are Dolichospermum, Microsystis, and Nostoc, amongst others. We can see all three of the above taxa present in our samples. But can we deduce a relationship?
Our statistical tests yielded significant p-values, indicating a strong correlation between the location of the sample to the variety of taxon found within the sample. In other words, the distribution of taxa is not random across the sites, and certain taxa are more common than others at certain locations. This is consistent with different algal community compositions across sites, and reaffirms the conclusions we draw moving forward.
Water Parameters as Indicators of Algae Species
Could there be another way to quickly identify harmful algae species in water bodies when they are not visually easy to identify without lab testing? We sought to address this question by looking at whether simple water parameters could be strong indicators of algae species. We looked into two variables, pH and dissolved oxygen (%), measured in the dataset to see if this could be the case.
Fig 5. Stacked Dot Plot of Average pH by Algae Type. Taxa with small sample sizes were removed.
Upon initial analysis of pH of the selected samples, there is no clear indication that pH is a good indicator of specific algae types. If pH was a good indicator of algae type, we’d expect algae samples of the same type to be clustered close together.14 However, this is largely not the case. This lines up with previous experiments regarding pH and the harmful algae types included in the data. For example, in previous studies on water parameters in water bodies with Gleotrichia present15, the algae appeared to have little effect on pH but rather the phosphorous and nitrogen levels of the water bodies. This means that a pH of a certain level would not be an indicator of the algae because it can proliferate regardless of the pH. In conclusion, the relationship between algae type and pH isn’t clearly defined, but it is clear that is not the best indicator of what harmful algae type is present in a water body.
Fig 6. Stacked Dot Plot of Average DO (%) by Algae Type. Taxa with small sample sizes were removed.
Similarly, dissolved oxygen (%) was not an immediate indicator of algae species. However, we found that there was less variability within each algae type’s dissolved oxygen levels compared to the pH values. This implies that DO (%) may be a better indicator than pH at predicting algae type, but the variability was not statistically significant enough to conclude this with certainty. Hence, leading us to conclude that neither variable is a good indicator of algae species. Still, the fact that this decreased variability for dissolved oxygen exists is expected16, as it is known that harmful algae types tend to have harsh impacts on the dissolved oxygen levels, an important indicator of ecosystem health, of water bodies. In a study on Gloeotrichia and Aphanizomenon, two types of harmful algae, it was found that they had profound effects on the phosphorus cycles on aquatic ecosystems. This trend is also found among most other species of harmful algae. As the cells of these algae grow, they absorb phosphorus from the water to store it for when they later move to different areas of the water body. This in turn throws off the dissolved oxygen levels in the water and potentially lowers them to a state that is not17 conducive to life for aquatic species.
We also created a linear regression model to test once again if these two water parameters could be used as indicators of the presence of specific algae types. This means we trained a model trained on the data set to predict algae types based on pH and dissolved oxygen values. Unfortunately, the model was not very accurate and only correctly predicted algae type for the algae with the largest sample sizes with little exceptions. This means we cannot conclude that either water parameter is a certain indicator of algae type. However, due to some successes in the model with the larger samples, more data could be collected from these water bodies to train the model to be a better predictor of algae type.
Are There Other Factors We Can Tie to Algae?
Other than dissolved oxygen and pH, scientists collect a variety of parameters at bodies of water to test its water quality. One that is not talked about as much is specific conductance. It is often tested in bodies of water along with other common factors such as temperature or pH, but does it tell us anything about the algae growing in those bodies of water? First, let us start with what specific conductance is.
What is Specific Conductance?
Specific conductance is an important metric to consider since changes in levels of specific conductivity are often an indicator of disturbance in the water18. It is a measure in microsiemens per centimeter (uS/cm) of the conductivity of water, or how well water can pass electrical flow, at 25 degrees Celsius. Many factors can affect specific conductivity, such as: an increase in temperature can lead to an increase in conductivity, higher salinity results in higher conductivity, and different solids from the surrounding geology diffusing into the water can affect the level of specific conductance.19
One example of how specific conductance can change is through heavy construction. Heavy construction often disturbs the surrounding ground and makes it easier for debris to diffuse into the water, leading to higher specific conductance. It is always important to monitor bodies of water with increasing levels of specific conductance since higher specific conductance often correlates to elevated salinity levels, which can often cause stress on some cyanobacteria, causing them to leak harmful toxins and negatively affect the environment.20
_for_Different_Taxa_at_Various_Locations.png)
Fig 7. Scatterplot of Average Specific Conductance (uS/cm) for Different Taxa at Various Locations
Most taxa grow in areas with similar levels of specific conductance ranging from 30 uS\cm to 200 uS\cm, which is consistent with data collected from around the country21. However, there are also a decent amount of taxa that grow in locations varying from 30 uS\cm up to 900 uS/cm. What could be a reason for this huge range? Could it be that some taxa are more resilient to higher levels of specific conductance?
Another interesting observation is that most locations have a range of specific conductance of around 50 uS/cm. However, Leetown, WV has a range of 850 uS/cm. In fact, the samples collected with higher specific conductance relative to our data in Leetown is possibly a result of specific geochemical reactions causing high mineral dissolution in the water.22
A Closer Look at Oregon
As mentioned in the beginning, most of the data was collected in Oregon. Since it has the most samples and a decent collection of taxa, let’s take a closer look at the data on specific conductance collected in this state.
Fig 8. Map of Algae Collected Across Oregon and their Specific Conductance
Here we can see there are two main groups: one further north with a lower average specific conductance and one further south with higher average specific conductance. Comparing the locations where samples were collected with a map, the group of samples on the bottom were taken at Upper Klamath Lake. Looking into exactly which taxa were sampled in Upper Klamath Lake, we find Gloeotrichia, Aphanizomenon, Dolichospermum, and Microcystis; all of which can leak harmful toxins into the water under stress.23
Fig 9. Upper Klamath Lake24
Now consider the taxa that appear in the rest of the samples in Oregon. There are 17 different taxa that were sampled in those areas, yet only two, Gloeotrichia and Dolichospermum, can produce harmful toxins. These findings are consistent with other sources that mention the amount of cyanobacteria found in Upper Klamath Lake has increased significantly25. Additionally, while only a couple of advisory warnings have been sent to various lakes in Oregon, Upper Klamath Lake experiences annual, massive algae blooms26. These recurring algae blooms are extremely harmful to the environment, including preventing the recovery of multiple endangered species in the Upper Klamath Lake, such as Lost River sucker and the shortnose sucker.27
Using statistical tests, we found that there is a significant positive correlation between specific conductance and harmfulness of taxa in Oregon. So, the higher the level of specific conductance, the more likely you are to find harmful algae in the water.
Are There Simpler Methods to Test?
We concluded that there is a correlation between specific conductance and harmful algae. Because of that correlation, we could use specific conductance as a measure to determine if a body of water is prone to developing algae blooms and then take preventative measures to protect that body of water. However, specific conductance can only be tested using proper equipment, and for the average person concerned about their nearby body of water, it is much too expensive. So, is there a variable anyone can easily test to determine if a body of water is prone to having more harmful algae?
A Look at Temperature
Fig 10. Stacked Histogram of Algae Types Against Average Temperatures
A first look shows that the algae sampled come from a diverse range of temperatures. The largest peak occurs around 22-25 degrees Celsius, which is in the optimal range for algae growth.28 However, a large amount of taxa are found below 17 and above 30 degrees Celsius as well. Overall, the algae sampled live in a diverse range of temperatures.
While it may look like the algae sampled come from diverse temperature ranges, when looking at harmful HAB-forming bacteria, it is clear that most of these algae live in moderate to warm temperatures. This temperature range maximizes their growth and reproduction rates, allowing them to outcompete other types of algae.29 As a result, they take over ecosystems, choking out other species and harming them through bloom-forming and toxins. Through a chi-square test of independence, we found this association was statistically significant. In other words, harmful algae are statistically more likely to live in higher temperature ranges.
Fig 11. Box Plots of Average Temperature for Different Algae Types
This trend is especially concerning when considering climate change. According to a study done by the EPA, 32 out of 34 North American lakes studied reported an increase in summer surface water temperatures, with 15 reporting a change of over 2 degrees Fahrenheit (1.11 degrees Celsius).30 With water temperatures continuing to increase, many environments with waters that were once below the optimal temperature range for harmful algae growth will warm to temperatures that do, putting them at risk of being harmed by harmful algae blooms.
A standout taxa is Zygnema, which is found in both the lowest and highest temperature samples. While most taxa are best adapted to a singular temperature range, Zygnema’s adaptations31 like self-shading and UV-screening allow it to survive at both high and low extreme temperatures. Despite this resilience, Zygnema doesn’t show up in any of the moderate temperature samples. Because the cyanobacteria that dominate these temperature ranges are so much better adapted, other algae like Zygnema can’t compete no matter how well-suited they may be. Because of this, the diversity in the temperature ranges dominated by harmful cyanobacteria is lower than other ranges.
Not only do warming temperatures promote the growth of harmful algae, but the reverse is also true. Harmful cyanobacteria absorb sunlight and release heat, further increasing the surface temperature and making their environments more hospitable to them and less hospitable to other organisms. Combined with global warming, the spread of harmful algal blooms accelerates the harmful effects of higher water temperatures, which include:
I. Increasing the metabolic rates of aquatic organisms, which can be detrimental to animals like fish
II. Increasing the solubility of toxins in water, such as those produced by harmful cyanobacteria
III. Lowering dissolved oxygen, detrimental to aquatic life32
Fig 12. Dead Fish Due to Algal Bloom
The result is a threatening cycle that will only continue to accelerate if not addressed.33
Fig 13. Algae Concentration and Water Temperature Across the U.S.
A look at this map shows us that West Virginia has been dealing with particularly high water temperatures. The effects of this extend just beyond threatening the health of aquatic ecosystems, however. Much of West Virginia’s recreation and tourism industry is tied closely to water quality and health. Favorites like whitewater rafting are inhibited when harmful algal blooms turn the water toxic, and aquatic habitats becoming uninhabitable threatens West Virginia’s fishing industry.34
Turning the Tide
The trends we see across the country, from Oregon to West Virginia, are not distant problems. Even Cornell’s very own Cayuga Lake has experienced increasingly frequent algal blooms of its own. These blooms were not present before, but have increasingly been spotted in Cayuga lake and at Tompkins County during summers since 2017.35 The story of harmful algae blooms is unfolding right in our backyard.
Fig 14. Dr. Ahner
Whether through improved monitoring, reduced nutrient runoff, or climate-focused policy, the choices made today will determine the condition of the waters Cornell and the broader region depend on. Cornell research teams such as Dr. Beth Ahner’s lab is already making breakthroughs in combating these harmful blooms. Ahner is dedicated to even the “smallest” of discoveries so that scientists can stop the devastating effects of these blooms once and for all.36
By understanding how factors like temperature, conductivity, and algae species composition interact, we can better protect the ecosystems and communities we are part of. Next time you’re kayaking on Cayuga lake, remember that it is up to us to decide what the next chapter will look like.
