N 40.7°
E 74.0°
VIIRS/DNB
λ 500-900nm
VIIRS · GLOBE AT NIGHT · SQM

Reconciling satellite data with ground-truth sky brightness

The VIIRS Day/Night Band is the sensor behind every public light pollution map in use today. It was calibrated for the warm amber wavelengths of sodium streetlamps. As white and blue-rich LEDs replace sodium lighting across the world, satellite-only estimates systematically underreport how bright the night sky has become. Duskmeter fuses VIIRS data with citizen-science observations to produce corrected, site-by-site brightness estimates.

Sky brightness index · 2011 baseline = 100
VIIRS satellite estimate
Ground-truth (Kyba et al., 2023)
~2%/yr (VIIRS estimate)9.6%/yr (ground truth)THE DIVERGENCEBlue light the sensor does not register20112014201720202022
Illustrative curve from published annual growth rates Source: Kyba, Altıntaş, Walker & Newhouse, Science (2023)
Bortle 9: severely polluted Bortle 7: suburban glow Bortle 3: rural sky Bortle 1: pristine dark
THE BLIND SPOT

VIIRS was designed to see light, but not this light.

The VIIRS Day/Night Band is the sensor behind nearly every light pollution map in circulation, including the ones published by NOAA. It reads brightness most reliably between approximately 500 and 900 nanometers.

Old high-pressure sodium streetlights emit almost all of their output at 589nm, well inside that detection window. The satellite sees them clearly. However, the white and blue-rich LEDs that have replaced sodium across most of the world push a large share of their output below 500nm, which falls largely outside what the sensor registers at all.

Blue light scatters through the atmosphere far more efficiently than the amber glow of sodium ever did. The wavelength doing the most to brighten the sky is the one the satellite is worst at measuring.

LED peak
~450nm
Sodium
589nm
380nm visible spectrum 700nm
VIIRS SENSITIVITY WINDOW: approximately 500 to 900nm. The shaded region extends into near-infrared but stops short of LED blue emission.
THE GAP, MEASURED

9.6% per year, not 2%.

Ground-truth trend
9.6%
Per year from 2011 to 2022, as measured by Kyba et al. (2023)
Satellite-only trend
~2%
Per year over the same period, from prior remote-sensing estimates

In 2023, Kyba and colleagues analyzed more than 51,000 citizen-scientist star counts collected worldwide between 2011 and 2022. The results, published in Science, describe a 9.6% annual increase in sky brightness. That is nearly five times the roughly 2% growth that satellite-only estimates had shown for the same period.

At that rate, a child who could count 250 stars on the night they were born would see fewer than 100 from the same location by their eighteenth birthday. The distance between what the satellite reports and what people actually observe is the gap that Duskmeter exists to close, one site at a time.

Kyba, C.C.M., Altıntaş, Y.Ö., Walker, C.E. & Newhouse, M. (2023). Science 379, 265 to 268. DOI: 10.1126/science.abq7781
HOW IT WORKS

Two inputs. One corrected estimate.

Duskmeter does not replace either data source. It holds both against each other for every tracked site and flags any disagreement between them.

Satellite radianceVIIRS DNB, per siteGround-truth reportsSQM & star counts, per siteFusion engineWeights & flags gapsCorrected estimatePer site, gap-adjustedThreshold alertsWhen a site nears its limit
01

Track satellite radiance

The same VIIRS baseline that every other light pollution map uses, logged per site over time. This is the record that is currently taken at face value by researchers and policymakers.

02

Track ground-truth reports

Sky quality meter readings, citizen star counts from Globe at Night, and dark-sky park steward logs. All tied to the same sites and the same time window as the satellite data.

03

Fuse into a corrected estimate

Where the two data sources disagree, the tool weights toward ground truth as more observations accumulate. It flags the divergence and alerts sites that are approaching a meaningful brightness threshold.

WHO IT'S FOR

Built for people who already know the satellite has a blind spot.

A

Astronomers & observatories

See corrected sky brightness at your own site over time, so you can distinguish ambient skyglow from weather when reviewing your data quality reports.

D

Dark-sky park stewards

Defensible, ground-truthed brightness trends for certification and recertification applications, without having to manually correct satellite numbers yourself.

R

Researchers

Skip the manual step of caveating VIIRS data against a citizen-science cross-reference. Every site in Duskmeter already carries both data sources side by side.

MONITORING

Detect threshold crossings before they happen.

Every tracked site carries a live status. When the corrected trend approaches a meaningful brightness threshold, the system alerts you before the threshold is crossed, not after the next satellite pass.

Site 014: Alpine Ridge Observatory
39.4°N · elev. 2,340 m
Satellite trend
+1.8%/yr
Corrected trend
+6.4%/yr
Watch
Site 027: Cinder Hollow Dark-Sky Park
41.1°N · certified 2019
Satellite trend
+2.1%/yr
Corrected trend
+11.2%/yr
Threshold exceeded
Site 041: Mesa Verde Research Station
37.2°N · long-baseline
Satellite trend
+1.4%/yr
Corrected trend
+2.0%/yr
Nominal

The divergence is not closing on its own.

Duskmeter is in active development. Register to receive launch notifications. No marketing emails, only release notes.