Bird migration day by day - what distribution, can an AI prediction be made? (1 Viewer)

[jurek]

Bird migration in every spot is very uneven. Most days are weak, but during a season there are 2-3 periods of 1-3 days when migration is very intensive. I found this for raptors along the coast in Poland, seabirds along the coast of the Netherlands, rarities in Heligoland, and others noticed it on migration hotspots along the British coast. It can be frustrating, because, if you are traveling to a famous migration hotspot for a short time, you are most likely to miss these few special days.

I wonder if anybody plotted the distribution of bird migration over a season? How many individuals pass over days? Surely, it is not a normal distribution with large standard deviation (slowly tapering from common to rare) but more extreme. Something like 90% of birds pass in 10% of days.

Then - is it possible to predict, or maybe train a AI model, when will be a good day for migration watching or a rarity fall? I also feel this can be important for reducing bird deaths on wind turbines, too.

Any field observations? Any AI savvy Birdforum member who can spare some time from the job?

Both Scilly and Shetland repay treatment as campaigns over years, not quick hits, and Ouessant is unlikely to be different.

 

[dantheman]

Presumably largely (almost totally) down to weather ... but what the relevant factors being ... presumably that would be what you would have to plug into your calculator ...

Humidity, temp, visibility, wind speeds and directions etc etc

Previous weather conditions as per above that brought the birds to a place in the first place (no point in perfect conditions if there are no birds to leave)

Time of year/season precisely

Food availability pre-migration

Local geography/topology

Local weather conditions to produce a 'fall' (otherwise birds just continue on high overhead)

etc ...


Some of all that could be done retrospectively by looking at the past weather, bird radar events and observations recorded.

(Presume that's what you mean by the factors for training the AI. Could be interesting ...)

 

[Xenospiza]

Bird migration in every spot is very uneven. Most days are weak, but during a season there are 2-3 periods of 1-3 days when migration is very intensive. I found this for raptors along the coast in Poland, seabirds along the coast of the Netherlands, rarities in Heligoland, and others noticed it on migration hotspots along the British coast. It can be frustrating, because, if you are traveling to a famous migration hotspot for a short time, you are most likely to miss these few special days.

I wonder if anybody plotted the distribution of bird migration over a season? How many individuals pass over days? Surely, it is not a normal distribution with large standard deviation (slowly tapering from common to rare) but more extreme. Something like 90% of birds pass in 10% of days.

Then - is it possible to predict, or maybe train a AI model, when will be a good day for migration watching or a rarity fall? I also feel this can be important for reducing bird deaths on wind turbines, too.

Any field observations? Any AI savvy Birdforum member who can spare some time from the job?

Jurek: have a play with trektellen.nl where you can check by day (e.g. De Vulkaan in The Hague, but also Spurn or Batumi. The weather is helpfully indicated.
Falsterbo has changed its website and it is not as nice anymore.

For Breskens it is usually easy: first and second day with southeasterly wind in April/early May.
On the Dutch coast in autumn east-to-southeast is the best for numbers of landbirds, northwest for seabirds.
For raptors in autumn, the weather at Falsterbo is key for migration over the Low Countries.

For my local patch, numbers are very dependent on the weather, because the main species are Wood Pigeon and Crane. The pigeons need easterly to southeasterly winds in autumn, the cranes will leave their staging post to the north (Diepholz) or south (Lac du Der) whenever the weather is "right" (usually clear weather preferably some wind in the back, although sometimes they take advantage of storms).
This spring, I rightly predicted the day the White-fronted Geese would leave their wintering grounds in the Rhine Valley and fly over: a rare success!
Worst of all is a blue sky and no wind or "wind in the back", resulting in no visible passerines...

Just to add: there are radar systems to see birds coming and these can be employed to stop wind farms. There are also local "bird activity" detectors to stop wind farms. I believe they are being used in the Netherlands, but not yet in Germany where the right is using bird deaths as a way to portray windfarms as evil.

 

[MikeInPA]

This is a tool I use to see if it's looking promising in my County. I have no idea if it's available in Europe.

Lebanon County, Pennsylvania - Night of May 14, 2022 - BirdCast Migration Dashboard

This resulted in my best ever day in the backyard.

May 14th Birds

 

[jalid]

Weather and the time of year are the most important factors. The first mentioned is a very complicated thing. For example, not only local weather should be taken into account, conditions in different altitudes often vary, and the effect to observability which depends often on weather.

At Põõsaspea, Estonia, Cranes and Geese are the most concentrated species with up to two thirds of birds seen during the best three days of the season. Good observability helps this, and that these species are quite weather-dependent. But also two thirds of Siskins are seen during the three best days - this has a different explanation: good numbers of this species are seen only on special, comparatively rare situations.

Then there are some species which seem to migrate almost in any weather conditions and these are not very well concentrated. Only 15-20 % of Red-breasted Mergansers, Tufted Ducks and Red-throated Divers are seen during the three best days.

(These percentages are based on four complete autumn surveys).

 

[Xenospiza]

100,000s of Redwings along the Dutch coast yesterday. So that will be that for the year!

 

[KenM]

100,000s of Redwings along the Dutch coast yesterday. So that will be that for the year!

I watched wave after wave of Redwings certainly ‘000’s between 09.30-11.30am coming in off the sea at varying heights, mostly heading NW-W at Hunstanton (Norfolk) yesterday, also Song Thrushes, Blackbirds and Goldcrests on the back of the prevailing Easterlies. I’ve not witnessed these numbers before, it was quite a spectacle….must get to the coast more often! 😮

 

[a_j_steele]

I think also there been a blocking south/west weather system which has held up migrant for a while in certain areas so that has helped with crossing now there been a few days of favorable weather