Light pollution in Romania · naked-eye observations and light seen from space
How many stars can still be seen from Romania's cities
Under a natural sky about 3,000 stars are visible at once. Observers who measured the sky over Bucharest with the naked eye reported a median of 34. Choose a place and its real sky empties to what can be seen from there.
How many stars can you see from your place?
Type the name of a locality (there are 3,181: municipalities, towns and communes). On the left is the real sky above it at the chosen hour, with the stars of the Yale catalogue. Only those visible from there remain. Stars drawn faintly, with a ring, are the ones the evidence cannot settle. On the right is the same sky without artificial light.
For example, from Bucharest observers reported a median of 34 stars, and the satellite estimate gives 33 stars: two different sources, almost the same result.
The number in the corner of each sky is the mean over a rotation of the sky; the drawing shows the chosen hour. Planets and the Moon are not shown.
What observers reported
–What the satellite shows
–Light around the place
–Where is the sky brightest, and where did anyone measure it?
The colours show the 2025 estimate, calculated from light seen by satellite and checked against 390 site-nights of sky quality meter readings, an instrument that reads the brightness of the sky. The circles are site-nights where someone in Romania noted which stars they could see: the paler the circle, the fewer stars reported. Click the map to pick the nearest locality.
- reported ≤ 2
- 3
- 4
- 5
- ≥ 6
How many people have ever measured Romania's sky?
Globe at Night is the international programme in which anyone compares the sky with charts of a constellation and notes the faintest star they can see. Each year's observations are public. For Romania there are 2,539 rows from 2006 to 2025, of which 1,800 carry a usable value.
More than a quarter of them, 475, come from within five kilometres of the city of Slatina, in four years (2008, 2009, 2010 and 2016). So a night at one place counts once, as the median of its observations. That leaves 893 site-nights with clear sky or at most a quarter cloud.
More than half of the nights (53%) gave 3 or less on the limiting-magnitude scale, that is, at most 47 stars visible at once; 20% gave 5 or more. Participation has fallen: without new observations a place cannot be checked, and 2,852 of the 3,181 have fewer than five site-nights within 15 km.
Has the sky brightened since 1992?
A satellite sees the light that rises from the ground, not the stars that are lost. For 2012–2025 the series comes from the VIIRS sensor: around Romanian settlements, weighted by population, light grew by 20% (about 1.4% a year). For 1992–2011 the series is a reconstruction made by a model from an older sensor, DMSP: it shows a rise of 3.2 times from 1992 to 2012, but its size depends on that model.
Observers say something else. From 2011 to 2022, across 51,351 observations worldwide, the number of visible stars fell as if the sky had brightened by 6.5% a year in Europe and 9.6% a year worldwide (Kyba et al., 2023). The authors link the gap to the white light of LEDs, which the satellite sensor sees poorly, and to the direction in which light is sent. Romanian data are too thin to check this locally: since 2017 there are 74 site-nights.
How many stars can be seen from each county seat?
The county seats (reședințe de județ), with Bucharest, are in the table. “Observers” is the median of site-nights within 15 km; “satellite” is the 2025 estimate, with its range. The last column shows by what factor the surrounding light has grown since 2012.
| Place | Observers (stars) | Satellite (stars) | Surrounding light, 2025 vs 2012 |
|---|---|---|---|
| Alba IuliaAlba | no observation within 15 km | 149between 31 and 233 | ×1.77 |
| Arad | 47between 47 and 150 · 311 site-nights | 112between 24 and 182 | ×0.88 |
| PiteștiArgeș | no observation within 15 km | 111between 24 and 180 | ×1.44 |
| Bacău | no observation within 15 km | 172between 37 and 270 | ×0.77 |
| OradeaBihor | 150between 63 and 348 · 6 site-nights | 71between 11 and 115 | ×1.15 |
| BistrițaBistrița-Năsăud | no observation within 15 km | 182between 40 and 287 | ×1.24 |
| Botoșani | no observation within 15 km | 205between 48 and 327 | ×1.28 |
| Brașov | 150between 150 and 348 · 6 site-nights | 75between 12 and 124 | ×1.38 |
| Brăila | no observation within 15 km | 131between 28 and 203 | ×1.15 |
| București | 34between 11 and 47 · 44 site-nights | 33between 5 and 58 | ×1.16 |
| Buzău | no observation within 15 km | 122between 26 and 194 | ×1.36 |
| ReșițaCaraș-Severin | no observation within 15 km | 343between 84 and 552 | ×1.51 |
| Cluj-NapocaCluj | 150between 47 and 150 · 22 site-nights | 106between 23 and 174 | ×1.30 |
| Constanța | no observation within 15 km | 78between 13 and 130 | ×1.41 |
| Sfântu GheorgheCovasna | 260between 150 and 474 · 6 site-nights | 228between 54 and 359 | ×1.71 |
| TârgovișteDâmbovița | 474between 150 and 1,565 · 21 site-nights | 199between 46 and 316 | ×1.06 |
| CraiovaDolj | no observation within 15 km | 99between 21 and 163 | ×1.03 |
| Galați | 150between 47 and 150 · 19 site-nights | 90between 17 and 148 | ×1.38 |
| Giurgiu | no observation within 15 km | 202between 46 and 321 | ×0.95 |
| Târgu JiuGorj | no observation within 15 km | 223between 53 and 352 | ×1.33 |
| Miercurea CiucHarghita | no observation within 15 km | 378between 93 and 609 | ×0.84 |
| DevaHunedoara | no observation within 15 km | 170between 37 and 268 | ×1.29 |
| SloboziaIalomița | no observation within 15 km | 269between 65 and 426 | ×1.71 |
| Iași | no observation within 15 km | 96between 20 and 159 | ×1.11 |
| BufteaIlfov | 11between 11 and 47 · 10 site-nights | 200between 46 and 318 | ×1.41 |
| Baia MareMaramureș | no observation within 15 km | 291between 70 and 461 | ×0.77 |
| Drobeta-Turnu SeverinMehedinți | no observation within 15 km | 114between 25 and 184 | ×1.46 |
| Târgu MureșMureș | no observation within 15 km | 140between 30 and 218 | ×0.92 |
| Piatra-NeamțNeamț | 474between 150 and 474 · 5 site-nights | 222between 53 and 350 | ×1.07 |
| SlatinaOlt | 47between 11 and 150 · 79 site-nights | 225between 53 and 354 | ×1.55 |
| PloieștiPrahova | 11between 11 and 47 · 86 site-nights | 74between 12 and 122 | ×1.09 |
| Satu Mare | no observation within 15 km | 133between 28 and 205 | ×1.23 |
| ZalăuSălaj | no observation within 15 km | 344between 84 and 555 | ×1.23 |
| Sibiu | no observation within 15 km | 86between 16 and 143 | ×1.33 |
| Suceava | no observation within 15 km | 257between 62 and 407 | ×1.35 |
| AlexandriaTeleorman | no observation within 15 km | 309between 74 and 492 | ×1.04 |
| TimișoaraTimiș | 47between 47 and 47 · 6 site-nights | 54between 7 and 88 | ×1.21 |
| Tulcea | no observation within 15 km | 169between 36 and 265 | ×1.26 |
| Râmnicu VâlceaVâlcea | no observation within 15 km | 214between 50 and 340 | ×1.39 |
| Vaslui | no observation within 15 km | 245between 59 and 386 | ×0.87 |
| FocșaniVrancea | no observation within 15 km | 156between 33 and 244 | ×1.51 |
| Călărași | no observation within 15 km | 160between 34 and 250 | ×1.97 |
How many stars can you see in Orion?
This is how a Globe at Night observation works: you look at Orion (in October it rises towards morning, in winter it is up in the evening) and pick the chart with the stars you can see. Here you pick it with the bar below: move it until only the stars you can tell apart with the naked eye remain. If you went outside, you can send the observation to globeatnight.org; a place with no site-night needs it.
Method, sources and limits
How stars are counted
Star positions and brightness come from the Yale Bright Star Catalogue (BSC5, 9,110 entries, of which 9,096 stars have a visual magnitude; only 8,404 are magnitude 6.5 or brighter). A star counts as visible if its magnitude is at most the place's limit, lowered towards the horizon by 0.25 magnitudes per air mass (Kasten and Young, 1989). The number in the table and on screen is the mean, over one rotation of the sky, of the stars above the horizon. Under a natural sky (22.0 mag/arcsec², 174 µcd/m², Falchi et al., 2016) the limit is 6.6 and 3,017 stars are visible at once. Sky brightness is converted to a limit with the formula of the sky quality meter's maker (Unihedron). Glow near the horizon is not modelled, so the numbers are maxima for a lit sky.
Observers
Globe at Night, public files 2006–2025 (globeatnight.org/maps-data). We keep Romanian rows with coordinates inside the country, a limiting magnitude between 1 and 7, and clear sky or at most a quarter cloud. A night at one place (a 0.01° cell, same day) is one vote, the median of its observations. For a locality we take the site-nights within 15 km and show the median and the lower and upper quartile; with fewer than five site-nights we show nothing. We do not know how many observers were beginners, observed with the Moon up or stood under a lamp.
The satellite estimate
Light sent upward: the annual NPP-VIIRS-like raster of Chen et al. (Harvard Dataverse, CC0 licence, 500 m), resampled to about 1.3 km. For each locality we add the light of all pixels within 250 km, each weighted by distance to the power −2.5 (Walker's law, 1977), and the sum becomes sky brightness through SQM = 22.0 − 2.5·log₁₀(1 + 50·S / 174). The constant and the exponent are fitted on 390 site-nights with measurements (Globe at Night, sky quality meters, Romania and neighbouring countries, same year as the radiance raster).
How reliable the estimate is
In 314 site-nights with dark sky (measured 19.5 or more), almost all in Hungary and Slovakia, in 80% of cases the estimate was at most 0.4 magnitudes darker or 0.5 brighter than the measurement. In Romania, at street level, the 66 site-night with sky quality meters came out brighter than the map by 0.8 magnitudes at the median and up to 2.6 at the 90th percentile: the gap probably comes from local light, from nearby lamps, which the estimate cannot see. That is why the range on screen goes down by 1.5 magnitudes and up by 0.5. Against the limits reported by observers (not used in the fit) the estimate correlates weakly in Romania (Spearman 0.28, 893 site-nights) and is too optimistic for villages: there observers reported 3–4 where the estimate gives 5–6. We cannot say whether the village lamps, the observers or the model are to blame.
Light over time
The 1992–2011 part of the Chen series is reconstructed with machine learning from the DMSP sensor; from 2012 the series is VIIRS. VIIRS sees the blue light of white LEDs poorly (Kyba et al., 2023; Kyba et al., 2017), so the real growth of light may be larger than the chart shows.
What we do not know
- How much of the gap between observers and the estimate in villages comes from local lamps, observers or the model.
- How the sky of one particular place has changed since 2016, where observations almost stop.
- How much Romania's sky has really brightened: the satellite shows 20% in 13 years, observers across Europe 6.5% a year.
Sources
- Globe at Night, observations 2006–2025: globeatnight.org/maps-data.
- Chen, Z. et al., NPP-VIIRS-like nighttime light 1992–2025, doi:10.7910/DVN/YGIVCD; method in Earth System Science Data 13, 889 (2021).
- Falchi, F. et al., “The new world atlas of artificial night sky brightness”, Science Advances 2, e1600377 (2016).
- Kyba, C. et al., “Citizen scientists report global rapid reductions in the visibility of stars from 2011 to 2022”, Science 379, 265 (2023).
- Yale Bright Star Catalogue, 5th ed., VizieR V/50.
- Kasten, F. and Young, A., Applied Optics 28, 4735 (1989); brightness-to-limit conversion: Unihedron.
- Localities and population: INS, through the “România ta” project on this site.
Accessed on 2 October 2026. The Falchi atlas and Lorenz's map are not used: the first is available only through an access form, the second is published as images.