Providing critical water quality information for Minnesota lake management using satellite imagery

by Benjamin Page and Leif Olmanson

Routine monitoring of Minnesota water quality using conventional field sampling is challenging and expensive due to the over 10,000 freshwater lakes spread across nearly 87,000 square miles. On the other hand, the University of Minnesota’s Water Resources Center and Remote Sensing and Geospatial Analysis Laboratory have been utilizing the capabilities of satellite imagery to estimate optical water quality parameters from space for over twenty years. Since 2002, we made the data available on an online interactive, Google maps-based LakeBrowser  where citizens, scientist and lake managers can get information for their lake, the neighboring lakes or all the lakes in Minnesota. We currently have seven state-wide water clarity assessments of over 10,000 lakes derived from Landsat imagery and are in the process of adding 2010 and 2015 water clarity to maintain a five year interval. The LakeBrowser has been popular with over 9,000 unique monthly visitors, but if we really want to improve lake and fisheries management we need more water quality variables more often. To this end, recently available NASA/USGS and European Space Agency (ESA) Earth observation satellites that have improved spectral, spatial, radiometric and temporal resolution have the ability to measure the factors that control water clarity, such as phytoplankton, suspended solids and dissolved organic color.

We have developed methods to measure dissolved colored organic material (CDOM) from Landsat 8 and chlorophyll and suspended solids from Sentinel 2 imagery (Fig. 1). Our current project awarded by the Legislative-Citizen Commission on Minnesota Resources (LCCMR) through the Environment and Natural Resources Trust Fund (ENRTF) involves translating these novel image processing techniques into an automated high performance computing environment at the Minnesota Supercomputing Institute (MSI) (Fig. 2). This system works by instantly accessing newly acquired imagery from target satellites by leveraging machine-to-machine connection between MSI with USGS and ESA ground stations. From here, a series of in-house image processing functions are applied to the Level-1 Raw imagery and turned into Level-3 water quality mapping products. The calibrated L-3 products including statewide water clarity, CDOM, chlorophyll-a, and suspended solids (SS) maps rely heavily on field validated datasets to account for the dynamics of optically complex lake systems of the region. To this extent, sampling efforts in the summer months constrain uncertainties between satellite-derived and surface water properties caused by varying atmospheric conditions and re-calibrate water quality retrieval algorithms to yield verifiable water products. As new field validation data become available at season-end, scripted modules within the processing chain can be modified accordingly and applied to incoming and previously processed imagery if any resulting water quality product models need improvement. Finally, the data will be made available to the public on a more frequently updated online map viewer linked to a spatial database that will allow for statistical summaries at different delineations and time windows, temporal analysis and animations of water quality variables.

This system could have a large impact on the routine monitoring protocols conducted by lake management and other resource agencies. For example, data in the form of near-real time water quality maps could assist in lake management by highlighting vulnerable lakes experiencing moderate to high levels of eutrophication (Fig. 3.). Further, these maps could help educate resource managers regarding areas susceptible to harmful algal blooms, and aid in characterizing the phenology of bloom patterns on a regional scale. Being able to prioritize sampling efforts towards the more affected water bodies without extensive field sampling could become a capacity building exercise for a routine monitoring practice standard, and ultimately reduce time and financial constraints.

Figure 1

figure 1

Screen shot of the online interactive LakeBrowser




Figure 2

figure 2

Automated image processing pipeline for near-real time water quality monitoring.




Figure 3

figure 3 gif

Time-series animation GIF of monthly average surface water chlorophyll concentrations in southern Minnesota from June – October, 2018        Click on image to see progression.