Land Cover Summary [Michigan]
Michigan Dept. of Environment, Great Lakes, and Energy · 2025 Full Details
Tip: Check “Visit Source” link for download options.
| Attribute | Value |
|---|---|
| Click on map to inspect values | |
Full Details
- Title
- Land Cover Summary [Michigan]
- Description
- Water clarity, an indicator of water quality, is a loosely defined term generally related to how far one can see in a water body using natural light. The Michigan Inland Lakes Water Clarity Dataset includes includes three different estimates of water clarity: sechi depth, chlorophyll-a, and colored dissolved organic matter (CDOM) for inland lakes larger than 10 acres from 2018 to 2023. Detailed data field descriptions can be found in the item description of each individual indicator table. Descriptions of these indicators can be found below, and additional information can be found on the USEPA Water Clarity Indicators page . Secchi Depth: Secchi Depth is the most common measurement of water clarity. The measurement is taken using a Secchi disk, a 20-cm diameter white disk that is lowered into the water until it disappears. That depth (in ft, cm, or m) is termed the "Secchi depth" (SD). Water clarity measured using a Secchi disk is closely related to the density of particles in the water that can scatter or absorb light energy, and to the concentration of dissolved substances in the water that produce color. Chlorophyll-a: Chlorophyll is the primary pigment that captures light energy during photosynthesis. All plants contain chlorophyll, and it is a widely used metric for algal biomass in water bodies. "Algae," a term describing a diverse group of mostly microscopic organisms, typically are the base of aquatic food webs. Algal abundance and species composition are key water quality indicators. Chlorophyll absorbs sunlight at specific wavelengths and reflects light that is not absorbed. The differential absorption and reflection of specific wavelengths of light is the basis for lab and remote sensing methods to measure chlorophyll. Chlorophyll usually is determined by filtering algae from water samples, extracting the pigment from the filter with an organic solvent, and measuring absorbance by spectrophotometry. Remote sensing methods measure light reflected from water using wavelengths characteristic of the absorbance spectrum of chlorophyll. Colored Dissolved Organic Matter (CDOM): Colored (or chromophoric) dissolved organic matter (CDOM) is the part of organic matter that absorbs light in the blue and UV wavelengths. These colored compounds often make the water "stained" a "tea-like" yellow-brown color. CDOM plays major roles in freshwater ecosystem processes, determining physical and chemical conditions and water quality in freshwaters, and it is the most abundant dissolved organic matter fraction in forested watersheds with wetlands. Data Processing Methods: This dataset was generated using the University of Minnesota's automated image processing system (AIP). The AIP system downloads Sentinel 2 imagery and corrects reflectance data for atmospheric conditions and identifies areas affected by clouds or other obstructions. Processing steps include the following: (1) the AIP system automatically downloads and processes satellite imagery, including correcting for atmospheric effects and removing clouds or other obstructions; (2) validated statistical models based on regional data were applied to processed imagery to derive estimates from the imagery; and (3) processed data is aggregated with publicly available Michigan lakes polygon vector data in a geospatial database for display in web applications. Under step two, regional water quality models were developed using machine learning analysis of processed imagery and available water quality data from extensive in situ datasets from Minnesota, Wisconsin, and Michigan. The resulting dataset contains seasonal summaries and monthly data (from May to October) for the years 2018 through 2023, and it can be updated using the same methodology over time. The image processing system automatically downloads Sentinel-2 imagery and applies regionally calibrated water quality models to estimate surface conditions. These models were developed using machine learning techniques and trained on extensive in situ water quality measurements collected across Minnesota, Wisconsin, and Michigan (Olmanson in prep). By integrating field data with satellite observations, the system can produce accurate and consistent estimates of lake water quality across broad geographic areas. Because of differences in methods of field sample collection, models for chlorophyll-a and CDOM were developed using Minnesota and Wisconsin data. The dataset includes both seasonal summaries and monthly observations spanning from May to October for each year from 2018 through 2023. These summaries provide valuable insights into temporal trends and interannual variability in lake conditions. Scope and Data Ingestion: The scope of the project is defined by the geographic area (Michigan), the time period (May-October from 2018 to 2023), and the input data products (Sentinel-2 Level 1 MSIL1C granules, all bands, and metadata). Relevant data products are pulled from the Dataspace Copernicus repository, validated, staged on MSI's tier-2 storage, and registered for easy access. This data pull is automated and executed via an API that queries the remote repository. Automated Image Processing System: The data processing pipeline for this project leveraged existing techniques and infrastructure. It successfully generated both pixel-level (20m resolution) and lake-level water quality statistics for Michigan water bodies during the open water season (May-October) from 2018 to 2023. The pipeline runs on the University of Minnesota's Supercomputing Institute HPC clusters, processing a full six-month season for Minnesota in approximately 12 hours using a parallel array of 40 tasks. A task array generates pixel-level masks and water-quality products for each granule. Typically, an array of 40 workers processes the granules in parallel, completing an entire year's worth of data in about 12 hours. Fmask 4.0 is used to mask out clouds, shadows, and land pixels. Additional masking is performed using machine learning models that remove haze, thin clouds, and wildfire smoke. These models are based on fully connected neural networks trained on MSIL1C bands, with expert classification of millions of pixels using traditional image processing methods. Atmospheric distortions (ozone absorption, Rayleigh scattering, aerosols, etc.) are corrected to estimate bottom-of-atmosphere aquatic reflectance from the top-of-atmosphere bands. This correction, implemented in Python, uses the MAIN method (Page et al., 2019), which has been shown to outperform seven other methods compared to in situ data from Minnesota lakes. Water Clarity Estimates: Water quality parameters including Sechi Depth, Chlorophyll-a, and Colored Dissolved Organic Matter (CDOM) are estimated using machine learning (neural networks) formulas based on Minnesota, Wisconsin, and Michigan in situ field data and clear satellite imagery (Olmanson in prep). These estimates are derived from the remote sensing reflectance (Rrs) bands. Water clarity outputs are written as GeoTIFFs at a 20m resolution for each granule, ensuring no data values are assigned to terrestrial pixels or those identified as invalid in the masking step. Monthly average calculations were generated from the average values of valid pixels. These average tiles are generated in parallel and saved as GeoTIFFs. Then, monthly average tiles are mosaicked into statewide maps. The mosaics are aligned onto a consistent 20m grid to ensure pixel-level registration across months and layers. These statewide mosaics are saved as Cloud Optimized GeoTIFFs (COGs) and staged on MSI's tier-2 storage for access by GIS servers. For each water clarity measure, statistics (e.g., average, quartiles) are calculated and saved in a tab-separated values (TSV) file. Data Finalization: TSV files were consolidated and joined with the Michigan Geographic Framework Hydrography Polygons where estimates of water clarity were able to be calculated. Additional Lake Characteristics such as average lake depth, were aggregated from the USGS Landscape, Lakes, and Geographic Information Systems (LAGOS) of the United States . Lastly, landcover from the National Land Cover Database (NLCD 2022) was summarized within a 1000 meter buffer of lakes for reference. Related tables contain the data for each indicator and corresponding land cover composition. For more detailed analysis and applications, statewide pixel-level monthly mosaics for individual lakes are available and can be requested by contacting the data steward, Chris Vandenberg at Vandenbergc@Michigan.gov.
- Creator
- Michigan Dept. of Environment, Great Lakes, and Energy
- Temporal Coverage
- Last Modified: 2025-09-25
- Date Issued
- 2025-05-22
- Rights
- This application or dataset, along with any maps, data, content and other information contained within it and all output from the application (together the “Information”), is provided as a public service, and there are no restrictions on the use, reproduction, or distribution of the Information. It is your responsibility to use the Information for a legally permissible purpose. This Information is provided “AS IS” and on an “AS AVAILABLE” basis. The State of Michigan (“State”) makes no warranties, express or implied, regarding the accuracy, adequacy, reliability, timeliness, or completeness of this Information. THE STATE DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS INFORMATION, INCLUDING THE IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, AND NONINFRINGEMENT OF PROPRIETARY RIGHTS. THE STATE WILL NOT BE LIABLE, REGARDLESS OF THE FORM OF ACTION, WHETHER IN CONTRACT, TORT, NEGLIGENCE, STRICT LIABILITY OR BY STATUTE OR OTHERWISE, FOR ANY CLAIM FOR CONSEQUENTIAL, INCIDENTAL, INDIRECT, OR SPECIAL DAMAGES, INCLUDING WITHOUT LIMITATION LOST PROFITS AND LOST BUSINESS OPPORTUNITIES, RELATED TO THE ACCESS OR USE OF THIS INFORMATION. IN NO EVENT WILL THE STATE BE LIABLE FOR ANY AMOUNTS THAT MAY RESULT FROM THE ACCESS OR USE OF THIS INFORMATION, REGARDLESS OF THE FORM OF ACTION, WHETHER IN CONTRACT, TORT, NEGLIGENCE, STRICT LIABILITY, OR BY STATUTE OR OTHERWISE. By accessing or using this Information, you agree that you will not make any claims against the State or any of its departments, divisions, officers, or employees related to your use of the Information. These terms are governed by and interpreted under the laws of the State of Michigan without regard to conflict of laws provisions. Nothing in these terms is meant to impact or replace any existing rights or licenses, such as copyright, trademark, or patent protections, in materials or content that may be located on the site or portal containing the Information.
- Access Rights
- Public
- Format
- ArcGIS FeatureLayer
- Language
- English
- Date Added
- October 13, 2025
- Provenance Statement
- The metadata for this resource was last retrieved from State of Michigan Open Data Portal on 2025-10-12.
Resource Class
Resource Type
Theme
Place
Local Collection
Cite and Reference
-
Citation
Michigan Dept. of Environment, Great Lakes, and Energy (2025). Land Cover Summary [Michigan]. . https://gis-michigan.opendata.arcgis.com/datasets/egle::land-cover-summary (web service) -
BTAA Geoportal Link
