top of page

Cyanobacteria in Lake Iroquois

   Blue-green algae, also known as cyanobacteria, are a natural part of the freshwater ecosystem in Lake Iroquois and other Vermont lakes. Some kinds of cyanobacteria release toxins into the water, and when this algae forms dense populations known as blooms, can become an environmental and public health concern.


   Algal blooms grow best in warm water with high levels of phosphorous and nitrogen. As global temperatures rise and the use of synthetic nitrogen and phosphorous - rich fertilizers increase, blue-green algal blooms are becoming larger and more frequent  (Vermont Department of Health).

   I used remotely sensed data to review the growth of blue-green algal blooms in Vermont's Lake Iroquois from 2012 to 2014. Lake Iroquois is located west of Burlington, Vermont, and is currently rated Generally Safe by the VDH's Cyanobacteria Tracker.

   The images at left are from August or September of each year, when the water is warmest. Even without any image processing, large algal blooms are easily spotted in the northern end of the lake in the photo from 2014.


Imagery source: NAIP, VCGI


   I compared 2012 and 2014 National Agriculture Inventory Program (NAIP) imagery from the Vermont Center for Geographic Information. NAIP imagery (above) is true color imagery that has a spatial resolution of 1 meter and includes three bands of wavelengths. Since most plant growth absorbs red wavelengths and reflects near infrared, those values can be used to identify vegetation in a true color image. I used the normalized difference vegetation index (NDVI) to emphasize the difference between red and near infrared reflection values and locate the algal blooms in Lake Iroquois. Next in ArcGIS, I used the Iso Cluster tool on the NDVI layers, which runs an algorithm to group pixels by land cover type. Finally, I cropped the image to the extent of the lake and reclassified the data into two categories of pixels: those that are likely algae, and everything else.  I used that data to find the area of the lake covered by algal blooms in each year.

Results and Implications

Results and Implications

   The graphic at right shows the growth of blue-green algae blooms in Lake Iroquois between 2012 and 2014. The algae grew a total surface area of 16,355 square meters in two years.

   The highest concentration of algal blooms is found in the north and along the shorelines where run off enriches the nutrient content of the lake. This nutrient pollution, along with increasing temperatures, has led to significant algal blooms which have harmful repercussions on the Lake Iroquois ecosystem and health of those visiting the lake for recreation. In addition to releasing toxins into the water, algal blooms can significantly reduce the amount of oxygen in the water, impacting the population of fish and other plant species.

   Recommended next steps could involve outreach to property owners in the Lake Iroquois watershed to reduce the use of nitrogen-rich fertilizers and plant buffers along the shoreline to reduce soil erosion.

NovakLab1 [Converted].png


   Indices like NDVI are useful tools for identifying a variety of land cover types when working with remotely sensed data. While vegetation indices are the most common, there are also popular indices for identifying water features and burned areas (often employed for detection of forest fire damage). The NDVI is useful for generally distinguishing vegetated from non vegetated areas, but it does have certain limitations. Clouds, disturbance of the water surface, sun reflection off of water and oil on the water interfere with the image processing and could all affect the accuracy of data produced with this method.
Another factor is that the ultimate classification, in this case of algae vs non algae, is based on manual comparison to the original true color imagery. For example, when I ran the same set of procedures a second time and included the second-highest category of NDVI values in the final classification (both the highest and second highest category of NDVI values classified as algae), it showed a significantly larger area of the lake covered by algal blooms (49562 square meters vs 16355 with only the highest NDVI category). (Mod 1: Remote Sensing and Detection)

bottom of page