mariuscomper.ukRomână

The Danube Delta, spring

The great pelican city

In 2016 a drone counted 17,000 pairs of great white pelicans nesting in the Danube Delta, about four times what every count for twenty years had found. Nobody has published a pair count since. The map is the real delta from orbit; the birds are drawn, one for every 100 pairs.

Scroll

One bird drawn = 100 nesting pairs (17,000 pairs, drone count, May 2016). Drawn size and flight are not to scale. Map: Sentinel-2.

A city on reed

For decades great white pelicans nested on plaur, the floating reed beds of the delta. Storms, ice, waves and the birds themselves tear the beds apart, so the colony moves: in the 40 years before 2016 it shifted between two lakes, Hrecișca and Buhaiova, five times. Since 2014 most nests are inside a compact reed bed at the north-west of Hrecișca.

In 2016 the colony was moving from Buhaiova to Hrecișca: 17,076 birds on nests at the north-west of Hrecișca, 1,568 on its other shores, 621 on Buhaiova (one point counted per nesting pair).

Marinov et al., Scientific Annals of the Danube Delta Institute 22 (2016); the five moves are that paper's own count.

How many?

Since 1950, published estimates of nesting pairs exist for 30 years, made on the ground, from planes and once from a drone. Most years have none. Solid dots I checked in the original paper, number and year (6 years); hollow ones come from a 2019 compilation (24 years).

01,0003,0005,00010,00020,00019501960197019801990200020102020nesting pairs (square-root scale)3,000–4,500 (Marinov et al.)method changesno pair estimate foundindividuals, one day:2016 · 18,633*2017 · 28,799*2018 · 18,40917,000 (drone)

Marinov et al. put the estimates of the previous 20 years at 3,000–4,500 pairs. The table has lower ones too (1,800 in 1996, 460 nests counted in 1999, 1,900 in 2010) and one higher (4,600 in 2013). Then, in 2016, a drone photomosaic counted 17,000 pairs (15,000–19,000), about four times that band and 3.1 times the largest earlier estimate (5,400, in 1961). The authors note that fish-stock assessments showed no large increase, and that autumn counts in Israel and Bulgaria had long been far above what breeding counts from the delta and the rest of the Black Sea region could explain. They see new nesting space at work (compact reed colonised from 2014) and allow that earlier aerial counts missed birds.

Did the pelicans multiply, or did we finally count them? The data cannot say how much of the jump is method, and no pair count has been published since. SOR reports single-day counts of individuals (grey box on the chart). Individuals cannot be turned into pairs.

Compilation: Kiss et al., Oltenia 35(2), 2019, Table 1. Opened: Marinov et al. 2016; Schogolev et al. 2005; Platteeuw et al. 2004.

Same place, April and June

Drag the handle. Both halves show the same stretch of delta in 2020: April on the left, June on the right. At the edge of the western lake's reed, a pale patch appears.

At 10 m per pixel a pelican cannot be seen; a colony might show as a texture. My test could not settle it: 6 years have clear pictures of both seasons, and the bar was 8. A pale patch is not a headcount. How the test went

The city is open from April to September

GBIF holds 6,161 records of the species in Romania. 1,759 were made in May, 59 in March, 14 in December.

JanFebMarApr651May1,759Jun916Jul728Aug1,077Sep658OctNovDec

Most are people with cameras, not a census: they show when the birds are here and where observers go, not how many there are.

GBIF.org, 14 Nov 2026

How far for breakfast?

Nobody has tracked a nesting delta pelican to its fishing grounds. Field observers estimate it flies up to 50–60 km each way (Platteeuw et al. 2004). One tagged delta pelican that was not nesting went beyond 100 km. In Greece, nesting pelicans were seen fishing 182.5 km from their colony.

Where is your town on this map? The solid ring is the 60 km estimate, the dashed one is the Greek 182.5 km.

Your year

Pick a year. The birds on the map are redrawn for the nearest published estimate, one bird per 100 nesting pairs. Start at 2016 and then try 2009 or 1999.

What this page could not do

A series of 10+ yearsPartly. The 2019 compilation lists 30 years, but I could check only 6 in the original papers, and one of those papers says 1998 and 2000 had no estimate, though the compilation gives one. The table at the end shows every row.
A consistent annual countFails, as registered beforehand. The methods change, most years are empty, and 2016 is a different method.
Foraging distance from a tracking studyFails, as registered beforehand. The only tracked distance comes from a delta pelican that was not breeding; the rest are observations and estimates. The GPS tags fitted in 2024 went on Dalmatian pelicans, another species.
Satellite test: undecidedRegistered bar: the pale-pixel count rising in at least 8 years. Only 6 years have clear pictures of both seasons, and it rose in 5 of them, one by 2 pixels. Not decidable, so no claim that the colony is measured from orbit.
26 May 2026: the reed last spring, from orbit. The last published count of nesting pairs is from 2016.

The tests were written down before any number was computed. What was registered:

The satellite test, in full

Drag the handle. Both halves show the same 3 km of delta in 2020: in the 2 April picture the reed at the north-west edge of Hrecișca is bare; in the 6 June picture a pale patch sits there.

At 10 m per pixel a pelican cannot be seen; a colony of thousands might show up as a texture. I wrote a test down before computing any pixel count: pale pixels in a fixed window, a nesting-season scene against a scene from before the birds arrive, same year. Afterwards, seeing which years had clear scenes, I widened the date windows to 15 March–25 April and 5 May–20 June (with the first windows only 4 years had both) and replaced the pale-pixel rule. The window itself was drawn after I had looked at the 2019, 2020 and 2025 pictures, two of the five years that went up. The bar was at least 8 years going up.

Result: only 6 years have clear pictures of both seasons, fewer than the 8 the test needed. The count rose in 5 of them (in 2019 by 2 pixels). The registered bar was not met, so I do not claim the colony is measured from orbit. A looser count, chosen after seeing the data, rose in 5 of 6: a hint, not a result. A pale patch is not a headcount either way.

Contains modified Copernicus Sentinel-2 data.

How it was made

Sentinel-2 L2A true colour (Copernicus), up to two scenes per year (one before the birds arrive, one during nesting) from the Earth Search archive, cropped to about 17 km. GBIF record counts by month. Published nesting-pair estimates entered by hand, see the table. Birds are drawn, not photographed. Distances are measured to a point on a 0.1° grid near the delta's northern lobe, not to the colony.

The colony lies in a strictly protected zone of the Danube Delta Biosphere Reserve. This page shows lakes at map scale and prints no coordinates, landing places or routes. Please watch pelicans from the boats and sites of licensed local operators, never near the colony.

Sources

YearPairsChecked (✓ number and year; n number only)Published values
1950300—
1955600–650—
19562,000—
19571,500—
19614,000–5,400—4,000; 4,800-5,400
19652,000–2,500—
19691,700—
19711,200—
19722,000—
19892,500—
19912,620–3,500—2,620; 3,000-3,500
19922,000–3,500—2,000-2,500; 2,500-3,000; 3,000-3,500
19943,000–3,500—
19961,800✓1,800 pairs (Schogolev et al. 2005); the compilation also lists 1,100
19971,700–2,300✓2,000 +/- 300 nests (Schogolev et al. 2005); the compilation also lists 3,500
19983,000—
1999460✓460 +/- 30 active nests (Schogolev et al. 2005), counted nest by nest; 4,500 +/- 500 adults present
20003,000–3,500—
20011,700–3,500✓1,700 +/- 150 nests (Schogolev et al. 2005); an alternative estimate of 3,500 is cited there
20024,150✓4,150 pairs (Platteeuw et al. 2004, the most completely covered year); the compilation also lists 2,900-3,000 and 3,590-4,160
20043,500—
20053,000–3,900—
20083,500–4,160n3,560-4,160 (Munteanu 2009; Papp & Fantana 2008, cited in Marinov et al. 2016: publication years, not count years); the compilation also lists 3,500
20093,595–4,480—
20101,900—
20123,000–4,480—
20134,600—
20143,600–4,500—
20154,100–4,500n4,100-4,500 (aerial evaluation by SOR, cited in Marinov et al. 2016 by publication year)
201615,000–19,000✓17,000 (15,000-19,000) pairs; 17,076 birds on nests at the NW of Lake Hrecisca; 19,155 on nests in total; 20,500 great white pelicans on the mosaic (Marinov et al. 2016)
ImageDateScene
2017 before2017-04-03S2A_35TPL_20170403_0_L2A
2018 nesting season2018-05-23S2B_35TPL_20180523_0_L2A
2018 before2018-04-13S2B_35TPL_20180413_0_L2A
2019 nesting season2019-05-28S2B_35TPL_20190528_1_L2A
2019 before2019-03-19S2B_35TPL_20190319_0_L2A
2020 nesting season2020-06-06S2A_35TPL_20200606_0_L2A
2020 before2020-04-02S2B_35TPL_20200402_0_L2A
2021 nesting season2021-05-22S2A_35TPL_20210522_0_L2A
2021 before2021-04-22S2A_35TPL_20210422_0_L2A
2022 nesting season2022-05-12S2B_35TPL_20220512_0_L2A
2022 before2022-03-28S2A_35TPL_20220328_0_L2A
2023 nesting season2023-06-01S2A_35TPL_20230601_0_L2A
2024 nesting season2024-05-26S2A_35TPL_20240526_0_L2A
2024 before2024-04-11S2B_35TPL_20240411_0_L2A
2025 nesting season2025-06-05S2B_35TPL_20250605_0_L2A
2025 before2025-04-23S2A_35TPL_20250423_0_L2A
2026 nesting season2026-05-26S2C_35TPL_20260526_1_L2A
2026 before2026-04-16S2C_35TPL_20260416_0_L2A