Tick Tick Bloom: Harmful Algal Bloom Detection Challenge

Harmful algal blooms occur all around the world, and can harm people, their pets, and marine life. Use satellite imagery to detect dangerous concentrations of cyanobacteria, and help protect public health! #climate

$30,000 in prizes
feb 2023
1,377 joined

2017 algal bloom in Lake Erie seen as bright green splotches, captured by the NASA/USGS Landsat mission

There are tens of thousands of lakes that matter for recreation and drinking water. Cyanobacterial blooms pose real risks in many of them, and we really don't know when or where they show up, except in the largest lakes.

— Dr. Rick Stumpf, Oceanographer, NOAA, National Centers for Coastal Ocean Science


Inland water bodies provide a variety of critical services for both human and aquatic life, including drinking water, recreational and economic opportunities, and marine habitats. A significant challenge water quality managers face is the formation of harmful algal blooms (HABs). One of the major types of HABs is cyanobacteria. HABs produce toxins that are poisonous to humans and their pets, and threaten marine ecosystems by blocking sunlight and oxygen.

While there are established methods for using satellite imagery to detect cyanobacteria in larger water bodies like oceans, detection in small inland lakes and reservoirs remains a challenge. Manual water sampling is accurate, but too time intensive and difficult to perform continuously.

The Solution

The goal in this challenge was to detect and classify the severity of cyanobacteria blooms in small, inland water bodies using publicly available satellite, climate, and elevation data. The resulting algorithms will help water quality managers better allocate resources for manual sampling, and make more informed decisions around public health warnings for critical resources like drinking water reservoirs.

Labels were based on "in situ" samples that were collected manually by many organizations across the U.S. To create predictive features, participants could use satellite imagery from Landsat or Sentinel-2, climate data from NOAA (including temperate, wind, and precipitation), and elevation data from Copernicus DEM.

The Results

The winning solutions represent the best of over 900 submissions made throughout the competition. The winning submission had an average root mean squared error (RMSE) of 0.76, more than halving the error from the basic benchmark solution.

All of winning model code is publicly accessible in the winners' repository on GitHub. The repository also includes each winner's detailed write-up of their strategy, model architecture, and model behavior in different settings.