Maybe the species discovered it's niche on the mainland where predator pressure is probably less compared to other mainland habitats. Predators are one good reason for them to breed specifically on rock islets away from mainland. The Rock Pigeon shows similar behavior when it comes to breeding, in Eastern Sri Lanka, off the coast of Trincomalee. In a place called "Pigeon Island". Hope this information helps!

Thank you Thilina, interesting to know about island breeding Rock Pigeons in Sri Lanka. Like your Rock Pigeons, the long-term success of our island breeding Ducula can indeed be explained by low predation on the islands they choose. I edited my question to clarify this.

If we had a gradual increase in urban nesting, that could relate to intermdediate levels of predation. But this aggregation formed very suddenly. Nothing similar before and no aggregations in other towns

We are looking for ecological understanding of very sudden change in breeding behaviour  - any ideas?

Couldn't that be a sudden burst?

suppose there was niche filling in the island, reaching maximum capacity. A whole bunch of individual could have felt the population pressure. Somethings just need to reach the threshold to give out a sudden burst. That's how I see it.

It does sound like a sudden opening of a new niche (or an increase of a previously scarce resource) that the species quickly occupied. This is more common in human-influenced habitats were changes can be very quick. For instance, I recently eared of a sudden explosion in the number of nesting Audouin's gulls in southern Portugal (a species that before was considered of conservation concern) and it seems that this change can be fully explained by a change in fishing regulations that forced fishermen to discard some species caught by by-catch, suddenly increasing the food availability for the gulls on the coast.

Being somewhat familiar with Pied Imps it could be that at a nearby island they have approached or reached capacity but without alternate options the sub-ordinates/later arrivals simply breed in the sub-optimal sites on the island. Given they are commuting and feeding on the mainland on a daily basis awareness of other breeding attempts is presumably high. All it might take is a few founders to establish in a novel environment before colony size reaches the (small) threshold that makes it appealing for all prospectors (both those with founder traits and those that are followers). This would mean that potential breeders facing sub-optimal sites on the island could rapidly shift to the novel site within a matter of days.

There are a few seabirds that seem to show some plasticity with breeding site and I suspect this is what they do...Critical mass is first established by founders, then flooding by all available birds and this results in colony growth that easily exceeds plausible recruitment. Crested Terns are one such example. They arrive on some sand cays where they haven't breed for years, establish a small colony and next thing you know there are 1000+ pairs.

Sounds a bit reminiscent of the white-crowned pigeon in the Florida Keys. They largely nest on mangrove keys, but sometimes nest on the mainland (I don't recall how 'clustered' when doing so.  They cruise to the mainland for foraging (in hardwood hammock, i.e., coppice).  Did you say whether the Pied Imperial Pigeon nests in aggregations when in the usual island setting?

Thank you all, I really appreciate these interesting ideas. Very good to be reminded about seabird plasticity in nesting sites. It's a great example of sudden shifts, thank you Rohan. I wonder if seabirds shift because of shifting prey? I can't claim that would explain our situation. Intriguing to learn about the white-crowned pigeon with similar habits to our imperial-pigeons, and also the case of the gulls in Portugal. For our species though, I don't think there's any prospect of human-related food subsidy nor any sudden removal of natural food. 

An island at capacity and new niche in the city would be a great explanation....  only it's hard to fit our novel urban aggregation into that scenario. Let me tell you a bit more about them.

Our Ducula normally nest on certain islands (with long fidelity) and they are numerous on those islands but not suddenly increasing. Not close to running out of roosting/nesting space.

The nesting island is not their food source. To forage, these birds must commute to the mainland. Then they range widely to find small fruits of certain native trees. City nesters must do the same (minus the offshore commute). They rely on the same forest trees and would be similarly affected by variations in food availability.

Historically there have always been a few mainland nests each season, both in town and out of town. These have not served as founders for new aggregations. Foraging birds would surely pass by sometimes but range so widely they might not have high awareness of individual nests. It's hard to see how these scattered nests could get to 'critical mass'.

I have speculated about displacement by natural disaster. There was a major storm before our aggregation formed, but the damaged nesting island was distant. If Ducula from that island needed to go prospecting for alternate nest sites, they were presented with many nearer options: forest habitat similar to their island, rural land, several towns. Hard for me to think those birds could somehow "know" our city would be better for nesting, and then choose to bypass all other options en route.

So, I'm still searching for better ecological understanding of this surprising event. If anyone has more ideas I will be very glad to hear!

Another possibility is that all or part of normal breeding area has become unsuitable/destroyed, perhaps working in concert with other suggestions. Perhaps you have checked & eliminated this.

Thanks, Clive, we have confirmed nearer breeding islands are okay and birds continue to use them as normal. One distant breeding island suffered storm damage but that did not seem a convincing explanation for our urban arrivals, as explained in my last comment, see second last paragraph there..

Seals at the coast of Namibia and South Africa show similar dynamism of breeding colonies at Atlantic mainland and island beaches. Penguins as well. Driven no doubt by the dynamics of the currents, followed by sudden (dis)appearance of prey; predator (Jackal; Brow hyena for seal pups) dynamics may play a role as well.

Why do you think Ducula that occupied the storm-damaged habitat would select distant breeding sites that are close to their original breeding sites (i.e., nearer options en route) if it is more energy efficient to nest on the mainland? City nesters don't have to commute to the mainland to forage and, therefore, need less food to survive and reproduce. Do you have any evidence that the novel aggregation of "city nesters" contains individuals from the distant breeding island that suffered storm damage?

Thank you Hein, it sounds like a fascinating part of the world. Can you perhaps recommend a paper for the penguin dynamics there?

Susan I'm sorry if my previous comments were unclear. Energy efficiency does not seem to explain this sudden new aggregation on the mainland. For as long as people have observed this species, almost all the birds nest on islands (some particular islands, not all islands). Very few nest on the mainland. That suggests mainland nesters do not get an overall benefit, even though they avoid commuting to/from an island.

Julia: Your previous comments were not unclear; I think it was my response that was unclear. I am trying to understand the basis for your assumptions regarding the abundance and distribution of nesting Ducula so I can provide an appropriate answer to your question. What I need to know is:

1) Can you identify individual birds in the novel nesting aggregation or which populations they came from? Do you have evidence that the adults in the new breeding location previously nested somewhere else?

2) What year was the season in which ~1000 adult birds chose a new breeding location? Did it occur during a period of extreme climatic conditions (e.g., drought)?

3) Are there differences in environmental attributes among islands with and without nesting Ducula?

I might need to clarify some more. We are not looking for a particular source location. That's impossible to determine. These are wide-ranging birds, not individually identifiable.

We are looking for ecological principles to account for a sudden unprecedented shift in breeding location.

Earlier in the discussion some very good suggestions have come up. The frustrating thing is, our birds don't fit these theories.

Any new theories or new ideas?

If you want to identify the ecological processes that could explain the novel nesting aggregation, you need to examine how the environmental conditions (e.g., climate, habitat, abundance of predators, etc.) were different before and after the abrupt change in breeding location. In the past there was apparently no advantage to nesting on the mainland, but now there is given the new aggregation of breeders.

In my research on northern goshawks (Accipiter gentilis) in the southwestern United States, changes in the location of breeding sites were a function of the interaction between severe climatic conditions and vegetation attributes on the availability of food resources (Salafsky 2015). During my study period (1999-2004), the vegetation composition of goshawk territories with a breeding attempt (eggs laid) was different during drought (see attached figure from Salafsky 2015). Moreover, it appears that during extreme drought (2003) goshawks foraged in more distant and atypical habitats, which increased the risk of nest failure from predation or exposure, for greater access to much larger prey items (i.e., black-tailed jackrabbits). This suggests that to understand the factors influencing habitat selection you need to determine the environmental context by comparing year- and site-specific attributes among the breeding sites.

Apologies for my slow response (diverted by urgent projects) and thank you Susan for information about your study of northern goshawks, very interesting.

Гнездовые агрегации могут быть связаны с локальными скоплениями пищи. В России наблюдалось уменьшение расстояний до ближайшего гнезда у популяций, живущих в антропогенном ландшафте: у Corvus cornix, у Milvus migrans, у Asio otus. Во время вспышек численности грызунов Asio otus и Asio flammeus могут формировать агрегации из нескольких гнезд. Похожее видел у Larus heuglini на севере Западной Сибири. Вблизи городов и свалок могут появляться скопления из гнезд, а вдали от городов и свалок - только отдельные пары гнездящихся птиц.

Thank you Sergey for the examples you mentioned and thanks to Google for translation:

"Socket aggregation may be associated with local accumulations of food. In Russia, there was a decrease in the distance to the nearest nest in populations living in anthropogenic landscape: from Corvus cornix, in Milvus migrans, have Asio otus. During rodent outbreaks and Asio otus Asio flammeus can form the aggregation of several nests. Related Larus heuglini seen in the north of Western Siberia. Near towns and dumps may appear in clusters of nests, and far from cities and landfills - only a few pairs of breeding birds."