Testing the Water


  Page 2

Unless you are a toad or a swamp rat, chances are you don’t want to live on the edge of a smelly, algae-filled lake. Most people like their lakes clean, clear, and fresh. This is particularly true for those who reside in the Upper Midwest. Lakes there are central to people’s lives. Not only do they provide an outlet for recreation and an ideal setting to connect with nature, but they also draw in the bulk of the tourist trade. No one wants to see the pristine lake they fish on or swim in become polluted. Unfortunately, monitoring water quality for the 30,000 plus lakes in the Upper Great Lakes region has never been possible. Water quality measurements have always been taken by hand, and the states have traditionally had the resources to monitor only a small percentage of their lakes. Such modest samplings cannot give researchers a comprehensive view of water quality throughout the states.


Photograph of a 
Pristine Lake in the Boundary Waters

  Minnesota is known as the land of 10,000 lakes—a claim which is actually an understatement. Most of these lakes are relatively clean, like Lake Everett, providing the perfect environment for a secluded canoe trip or fishing expedition. Urbanization and agricultural runoff are, however, degrading water quality, particularly in the south of the state. Researchers at the University of Minnesota are developing new techniques to monitor water clarity and lake health, which will help identify trends in water quality. (Photograph copyright Kevin Judd)


Recently, scientists at the University of Minnesota working on a NASA project arrived at a solution. Using imagery from Landsat satellites, the scientists have mapped the water clarity for over 10,000 of Minnesota’s lakes at a relatively low cost. The maps have allowed them to evaluate water quality patterns across the state as well as provide the tools for monitoring lakes to the state. In the future, the scientists in Minnesota will combine their map with similar maps being created at the University of Wisconsin and Michigan State University to create a comprehensive water quality map for the entire Upper Great Lakes region.

next Plumbing the Depths

The data used in this study are available in one or more of NASA's Earth Science Data Centers.

  Landsat Images

Satellites can image a much wider area than can be effectively monitored from the ground. This pair of true-color Landsat images compares a clear, clean lake (top) with a lake with poor water quality (lower). (Images courtesy Upper Great Lakes Regional Earth Science Applications Center)


Plumbing the Depths

  Page 1Page 3

“In lakes, water quality relates directly to water clarity,” says Patrick Brezonik. He is the Director of the Water Resources Center at the University of Minnesota and has been testing the waters in the Upper Great Lakes region for the past 20 years. He says that in scientific terms a lake with high water quality is known as an oligotrophic lake. Such a lake has clean, clear water with little algae. It’s what most of us think of when we imagine a pristine lake in an untouched pine forest. At the other end of the spectrum are hypereutrophic lakes. These polluted lakes are typically murky, smelly, and overgrown with algae and exotic plants.


Photographs Comparing Oligotrophic and Hypereutrophic Lakes


Brezonik explains that lakes in the Upper Midwest typically deteriorate due to runoff from farms, urban areas, industry, and construction sites. The storm water runoff will often carry chemicals and sediments from these locations into nearby lakes. Sediments can cloud a lake, and nutrients in the sediments, particularly phosphorous, can cause unwanted algae to grow. While the degradation is gradual, once a lake has become hypereutrophic, it is very difficult to reverse the damage.

A large drop in water quality can impact both the wildlife and the people in and around the lake. Rough fish such as catfish and carp multiply and replace higher quality fish such as small mouth bass and pike. Algae, bacteria, and chemicals can affect the health of people and animals who fish and swim in the lakes. In addition, many small towns throughout the Upper Great Lakes region depend on tourist dollars to keep their economies afloat, and an algae-filled lake turns away summer vacationers.

“It’s important for both the surrounding communities and the environment that lakes are monitored on a regular basis,” says Brezonik. For the past century, the Secchi disk has been the standard instrument for monitoring water quality. Pioneered by Italian physicist Pietro Angelo Secchi in 1865, the Secchi disk is an 8-inch disk painted in an alternating black and white pattern. A person in a boat lowers the disk into the lake and records the depth at which it disappears from sight. This depth, known as the Secchi depth, is the measure of water clarity. In Minnesota, oligotrophic lakes have a Secchi depth of 16 feet or greater, and hypereutrophic lakes have a Secchi depth of only a few feet or less. If a lake is hypereutrophic, it’s then up to the researcher or resource manager to determine the cause.

  Hallet Lake (left) and Schwanz Lake (right), both in Minnesota, are representative of very good and very bad water quality. Hallet lake is oligotrophic, with very clear water. On the other hand Schwanz Lake is hypereutrophic, with murky water that is choked with algae. (Photographs courtesy Upper Great Lakes Regional Earth Science Applications Center)
  Photographs of Secchi Disk Measurements

Although Secchi depths provide an excellent classification system for limnologists, taking such measurements is labor intensive. The person making the measurements has to take a boat to the open part of the lake, slowly drop in the Secchi disk, record its readings, and then repeat the process at several other locations. While Secchi disk measurements work well for small numbers of lakes, they are not practical for monitoring the over 30,000 lakes in Michigan, Minnesota, and Wisconsin.

Tom Lillesand, the director of the Environmental Remote Sensing Center at the University of Wisconsin, explains that the states of the Upper Great Lakes region have modest budgets with which to monitor lakes. “In Wisconsin we have a budget to sample lakes, but it allows agencies such as the Wisconsin Department of Natural Resources to perform a complete ‘water quality physical exam’ on 50 to 100 lakes, which is less than one percent of the state’s lakes,” he says. With such minimal coverage scientists cannot obtain a comprehensive picture of water quality across the state, and they cannot follow trends in water quality across the state over time.

next Light from the Waves
back Testing the Waters

  Water quality measurements are typically done by hand with a Secchi disk. Although the measurements are accurate, it would be very expensive to conduct them by hand on a large scale. The photographs show decreasing water quality from left to right. (Photographs courtesy Minnesota Pollution Control Agency)


Light from the Waves

  Page 2Page 4

For many years limnologists have been looking for an alternative to the Secchi disk. In the mid to late 1970s, Lillesand (then at Minnesota) along with colleagues in Wisconsin began to toy with the idea of using remote sensing satellites to observe lake quality from space. They reasoned that if water quality information could be extracted from images taken by orbiting satellites, the number of lakes monitored could be greatly expanded. For several years, the scientists experimented with data from early Landsat missions, but the data provided only very rough estimates of water quality. “We were able to get accurate measurements of water clarity only after Landsat Thematic Mapper (TM) data became available,” says Lillesand.

Since 1972 the Landsat program has launched a series of Earth observation satellites into orbit, collecting image data of our planet’s surface. The sensors on the first three satellites in the series, launched in the 1970s, had a coarse spatial resolution (80 meters), and only four spectral bands. Later versions of Landsat have carried improved sensors—the Thematic Mapper instrument on Landsats 4 and 5, and the Enhanced Thematic Mapper Plus on the most recent of the series, Landsat 7, launched in 1999.

The Thematic Mapper and the Enhanced Thematic Mapper instruments acquire images in seven different wavelengths of radiation reflected or emitted from the surface of the Earth. The wavelengths are in the visible, reflective infrared and thermal infrared parts of the spectrum. With a spatial resolution of 30 meters, the images are well suited for mapping and monitoring large features such as lakes.

“It was the higher resolution and the addition of the blue band on the Thematic Mapper that gave us clearer results,” says Lillesand. The researchers found that when the amount of blue light reflecting off of the lake was high and the red light was low, the lake generally had high water quality. “It’s common sense. When you look at a clear lake from a distance it appears blue,” says Lillesand. Lakes loaded with algae and sediments, on the other hand, reflect less blue light and more red light. He further explains that the scientists first acquired Landsat Thematic Mapper data of the lakes with known Secchi depths. They then analyzed the satellite data to see if they could arrive at an image that displayed lake clarity as accurately as the Secchi disk measurements. In the end, Lillesand and his colleagues were able to retrieve water clarity maps and measurements from the satellite imagery that were as accurate as Secchi disk measurements.

For roughly ten years, this knowledge was put to little use. Then, in the late 1990s, researchers at the University of Minnesota launched a pilot project to measure the water quality of the lakes around the Minneapolis/St. Paul metropolitan area. With a single Landsat TM image, they obtained coverage of all the lakes in the seven-county region. They put the images of the lakes through an analysis similar to that developed a decade earlier and classified lake water quality measurements for over 500 lakes. They tested the Landsat water quality readings against Secchi measurements of sample lakes, and the two sets of data matched up very closely. “With the Landsat images we ended up getting water quality measurements of 10 times as many lakes as we would have with the Secchi disk data,” says Brezonik. The Minnesota team then dusted off archived Landsat images dating back to the early 1970s and ran them through the same procedure. They found that most of the lakes around the seven-county metro area have not changed in quality over the past 25 years, with somewhat more (7 percent) increasing than decreasing (3 percent) in quality.


True Color
True Color

False Color
False Color (near infrared=red, red=green, green=blue)

Water Quality per Lake
water quality

Satellite imagery enables the measurement of Water Quality over a much wider area than hand sampling using Secchi disks. These images (larger size) show Landsat image data and derived water quality measurements for Lake Minnetonka, Minnesota. The true-color image is similar to what is seen by the human eye, the false-color image incorporates data in near-infrared wavelengths, and the water quality image shows calculated Secchi disk transparency. Dark blue indicates high water quality, light green moderate water quality, and dark green poor water quality. (Images courtesy Upper Great Lakes Regional Earth Science Applications Center)


Water Quality in Minneapolis St. Paul Region

The results proved so successful that a team of scientists led by Brezonik and Marvin Bauer, director of the University of Minnesota Remote Sensing and Geospatial Analysis Laboratory, expanded the survey to the entire state. The statewide project was funded in part by NASA’s Upper Midwest Regional Earth Science Applications Center (RESAC), which was established to monitor and analyze the natural resources in the Upper Great Lakes region. For the statewide water quality map, the researchers assembled the best of two year’s worth of cloud-free Landsat images taken of Minnesota. “We were able to get water clarity readings on 100 percent of the lakes larger than 20 acres using the satellite data,” Brezonik says. Of the 12,700 bodies of water classified as lakes in Minnesota, roughly 10,000 are larger than 20 acres. Once again, the readings matched up with the Secchi disk records.

“When we put the map together, we did see a very strong north-south pattern in Minnesota,” says Brezonik. In the northeast Minnesota the lakes are very clear and the water quality is high. Moving south and southwest, the water clarity and quality diminishes. He believes there are two major reasons for this pattern. The first is that most of the farms and the people in Minnesota are in the southern half of the state. Nutrient-rich run-off from farms and urban developments has caused algae to grow. In the northeast, dense forests where fewer people live surround the lakes. He says the second reason is that the lakes in the south are generally shallower than those in the north. Deeper lakes generally absorb excess sediment and nutrients better. Though the team has not yet finished analyzing archived data from the past, Brezonik believes that the lakes have probably maintained the same level of quality over the past 15 years.

next Extending Water Quality Measurements
back Plumbing the Depths

  The map to the left shows water quality in the counties surrounding Minneapolis and St. Paul, Minnesota. Lake Minnetonka, shown in the Landsat images above, is towards the left edge of the map. Satellite imagery is the only practical way to measure water quality on such a large scale. Dark blue, light blue, and green represent good water quality, while orange, red and purple indicate progressively worse wtaer quality. (Map courtesy Upper Great Lakes Regional Earth Science Applications Center)


High-resolution Landsat Images

  Page 3



pixel-level water quality

lake-level water quality

  Satellite imagery enables the measurement of Water Quality over a much wider area than hand sampling using Secchi disks. These images show Landsat image data and derived water quality measurements for Lake Minnetonka, Minnesota. The true-color image is similar to what is seen by the human eye, the false-color image incorporates data in near-infrared wavelengths, and the water quality image shows calculated Secchi disk transparency. Dark blue indicates high water quality, light green moderate water quality, and dark green poor water quality. (Images courtesy Upper Great Lakes Regional Earth Science Applications Center)

back Light from the Waves



Extending Water Quality Measurements

  Page 3

Researchers at Michigan State University and the University of Wisconsin are now working to construct statewide water quality maps of their own. Lillesand, who is leading the RESAC effort in Wisconsin, says, “Within the year, we will have monitored over 8000 lakes in Wisconsin. An essential part of this project is the network of over 600 citizen volunteers who are helping us take Secchi depth measurements on the days when Landsat 7 passes over their lakes.” Michigan researchers are just getting their project off the ground, but they hope to have a full map in three years. In the future, maps from all three states will be combined into a single water quality map for the Upper Midwest that will display roughly 30,000 lakes.






The researchers at the three universities are also employing other types of satellite data in their monitoring efforts. The commercial IKONOS satellite, which gathers images of the Earth at up to 1-meter resolution, can provide data to accurately monitor lakes as small as 1 acre. The MODIS instrument aboard NASA’s Terra satellite collects imaging data on nearly the entire surface of the Earth every two days at a maximum resolution of 500 meters in the wavelengths used for lake assessment. While such a resolution prevents using MODIS to monitor all but the largest lakes, the frequency of MODIS’s coverage enables seasonal changes in water quality for these lakes to be observed. Under NASA’s Affiliated Research Center (ARC) program, Wisconsin researchers are starting to use MODIS data to monitor large bodies of water such as Lake Winnebago in Wisconsin and Green Bay in Michigan.


IKONOS data can be used to provide a “city-scale” map of lake water clarity like this one showing Eagan, Minnesota. Only eight of the more than 300 clarity measurements included in the full image (1.9 MB JPEG) were classified using the coarser resolution Landsat data over the same area. Clear lakes are blue, moderately clear lakes are green and yellow, and murky lakes are orange and red. (Image courtesy Upper Great Lakes Regional Earth Science Applications Center, based on data copyright Space Imaging)


Green Bay Time Series


“But the real application of all this work will come in the future to track the progress of the lakes over time and understand how land cover change affects water quality,” says Brezonik. As part of the RESAC program, the researchers plan to compare maps of lake water quality with satellite maps that display land covers such as cropland, forests, wetlands, and urban areas. By linking land use and water clarity maps, the scientists hope to discern how urban sprawl and deforestation affect the water quality in the Upper Great Lakes region. The lake water quality maps will also be useful for diagnosing and managing lakes. Government agencies such as the Minnesota Pollution Control Agency and the Departments of Natural Resources in the three states can use updated versions of the maps to observe where water quality is deteriorating. The monitoring made possible by satellite remote sensing will help assure that lakes in the Upper Midwest are clean and fresh.

Bauer, M, L Olmanson, M. Schulze, P. Brezonik, J.Chipman, J. Riera, and T. Lillesand, 2001: Assessment of lake water clarity in the Upper Great Lakes region, Proceedings, American Society of Photogrammetry and Remote Sensing, St. Louis, MO.

Olmanson, L., P. Brezonik, S. Kloiber, M. Bauer, and E. Day, 2000: Lake water clarity assessment of Minnesota's 10,000 lakes: A comprehensive view from space. Proceedings, Virginia Water Research Symposium 2000 Advances in Land and Water Monitoring Technologies and Research for Management of Water Resources, Roanoke, VA, 215-232.

Kloiber, S., T. Anderle, P. Brezonik, L. Olmanson, M. Bauer, and D. Brown, 2000: Trophic state assessment of lakes in the Twin Cities (Minnesota, USA) region by satellite imagery, Arch. Hydrobiol. Spec. Issues. Advanc. Limnol., 55, 137-151.

back Light from the Waves

  These images show true-color imagery and water quality data for Green Bay, Wisconsin during the summer of 2001. The Moderate-Resolution Imaging Spectroradiometer (MODIS) acquires data over a wider area and more frequently than Landsat, allowing researchers to track seasonal water clarity. (Images courtesy Jonathan Chipman, Center for Limnology and Environmental Remote Sensing Center, University of Wisconsin)