visualizedata.org
Find a place

Prototype. These numbers are estimated from satellite images, not measured in the water. Don't use them to decide whether to swim; check your county or state health department for algae advisories.

Minnesota lakes / Lake Vermilion

Lake Vermilion

St. Louis County, Minnesota · 157.37 km² of open water · open-water seasons 2024 to 2026, May 15 to Oct 15

Water greenness, 2026 season0.00NDCI, reads as slightly green · −0.02 vs 2025
Latest chlorophyll-a estimate9.8 µg/LOct 6, 2026 · low-skill model, see below
Outline © OpenStreetMap contributors, north up

Is the water getting greener?

NDCI (normalised difference chlorophyll index) compares the red light the water reflects with the deep red just beyond it, the red edge. Chlorophyll absorbs the first and algae reflect the second, so the index rises as algae grow. Below zero the water is mostly clear; above about 0.05 it is green. Each point is one month's average over every cloud-free water pixel Sentinel-2 saw, May 15 to Oct 15 each year.

Water greenness (NDCI), monthlyThe 2026 season averaged NDCI 0.00, against 0.02 in 2025.
-0.40-0.200.000.200.400.60JunJulAugSepOct
Indicator: NDCI = (red edge − red) / (red edge + red), monthly mean over cloud-free water pixels, Sentinel-2 L2A. Same scale on every Minnesota lake page. Higher is greener.Source: satellitedata.net zonal statistics; Copernicus Sentinel-2 (ESA)
Show the numbers
SeasonWhenNDCIDetail
2026May 20260.00middle half -0.02 to 0.02 · 5 Sentinel-2 scenes
2026Jun 20260.00middle half -0.02 to 0.03 · 10 Sentinel-2 scenes
2026Jul 20260.01middle half -0.01 to 0.04 · 13 Sentinel-2 scenes
2026Aug 20260.04middle half 0.01 to 0.07 · 10 Sentinel-2 scenes
2026Sep 20260.01middle half -0.04 to 0.05 · 10 Sentinel-2 scenes
2026Oct 2026-0.07middle half -0.12 to 0.06 · 4 Sentinel-2 scenes
2025May 20250.02middle half -0.01 to 0.05 · 4 Sentinel-2 scenes
2025Jun 2025-0.01middle half -0.02 to 0.03 · 6 Sentinel-2 scenes
2025Jul 2025-0.01middle half -0.04 to 0.04 · 10 Sentinel-2 scenes
2025Aug 20250.03middle half 0.01 to 0.05 · 11 Sentinel-2 scenes
2025Sep 20250.07middle half 0.04 to 0.10 · 6 Sentinel-2 scenes
2025Oct 20250.03middle half -0.02 to 0.09 · 4 Sentinel-2 scenes
2024May 20240.00middle half -0.02 to 0.02 · 4 Sentinel-2 scenes
2024Jun 20240.01middle half -0.03 to 0.04 · 6 Sentinel-2 scenes
2024Jul 20240.03middle half 0.00 to 0.06 · 7 Sentinel-2 scenes
2024Aug 20240.07middle half 0.04 to 0.10 · 6 Sentinel-2 scenes
2024Sep 20240.08middle half 0.04 to 0.11 · 11 Sentinel-2 scenes
2024Oct 20240.03middle half -0.01 to 0.06 · 2 Sentinel-2 scenes

Chlorophyll-a estimate

Treat these numbers as a placeholder. The model that made them explains only 10% of the variation in measured chlorophyll on samples it never saw (test R² 0.10), and it was trained on estuary water, not lakes. This page gives no trophic-state label until a model with test R² of 0.5 or more replaces it. Use the greenness chart above to compare seasons.

For the clearest Landsat pass of each month (at least half the lake visible), the RemoteSensing water-quality model predicts chlorophyll-a for every water pixel more than 60 m from shore. Each point is the lake's median.

Chlorophyll-a, clearest pass each monthThe latest estimate is 9.8 µg/L (Oct 6, 2026), from a model with a test R² of 0.10.
0.05.010.015.020.0JunJulAugSepOct
Indicator: Chlorophyll-a, µg/L, median of water pixels more than 60 m from shore; model chl_a_best_overall_ElasticNet_td1_cc0_Estuary_block_50km_baseline_raw8_default.pkl.gz. µg/L, median of water pixels. Same scale on every Minnesota lake page.Source: satellitedata.net (USGS Landsat 8 and 9 Collection 2 L2); RemoteSensing water-quality model
Show the numbers
SeasonWhenµg/LDetail
2026May 31, 202610.0 µg/L10th to 90th percentile 9.6 to 10.3 · 68,981 water pixels
2026Jul 2, 20269.4 µg/L10th to 90th percentile 8.7 to 10.2 · 106,981 water pixels
2026Aug 19, 202610.2 µg/L10th to 90th percentile 9.3 to 11.2 · 139,789 water pixels
2026Oct 6, 20269.8 µg/L10th to 90th percentile 9.1 to 10.5 · 116,558 water pixels
2025Jun 13, 20259.1 µg/L10th to 90th percentile 8.5 to 11.0 · 9,996 water pixels
2025Jul 31, 202518.9 µg/L10th to 90th percentile 13.9 to 22.5 · 80,103 water pixels
2025Sep 25, 20259.9 µg/L10th to 90th percentile 9.0 to 10.6 · 139,641 water pixels
2025Oct 11, 20259.3 µg/L10th to 90th percentile 9.1 to 9.8 · 139,990 water pixels
2024May 25, 20248.6 µg/L10th to 90th percentile 8.1 to 9.3 · 95,609 water pixels
2024Jun 10, 202410.1 µg/L10th to 90th percentile 9.7 to 10.4 · 131,368 water pixels
2024Aug 13, 202412.3 µg/L10th to 90th percentile 9.9 to 13.9 · 123,458 water pixels
2024Sep 30, 20249.9 µg/L10th to 90th percentile 9.2 to 10.4 · 140,902 water pixels
2024Oct 8, 20249.6 µg/L10th to 90th percentile 9.3 to 10.2 · 138,665 water pixels

Where the numbers come from

The model behind these estimates

Bundle
chl_a_best_overall_ElasticNet_td1_cc0_Estuary_block_50km_baseline_raw8_default.pkl.gz
Estimator
ElasticNet, best_overall by RemoteSensing's baseline framework v4.0
Trained on
651 matched samples, water type: Estuary
Test R²
0.10 (the share of variation in measured chlorophyll it explains on held-out samples; this site labels lakes only at 0.5 or more)
Typical error
±5.5 µg/L (mean absolute error on held-out samples)
Inputs
Landsat surface reflectance in blue, green, nir, red, swir1, swir2; scene cloud cover
Code
RemoteSensing.Scripts services.ml_inference_engine.MLInferenceEngine