Bird signs and cycles, February, 2024

Understanding the winter routines of my avian neighbors with the help of BirdNET-Pi, a microphone, and the Cornell Lab of Ornithology's BirdNET classifier.

Best viewed on larger screens.

How much of what we hear do we notice? A microphone on my balcony listened to the birds of Amsterdam, and suddenly, so did I.

The morning begins with the jackdaws. It is still dark when I awake, but I can tell it's around 6am. This winter the jackdaws have chosen a roost outside my bedroom window and their chatter is as reliable as any alarm.

Around 7:00, they are joined by the crows, stirred from their roosts and swooping among the bare, still-dark trees.

The corvid dawn chorus wanes around 9:00 as the last of the stragglers depart for their morning commute to the coastline, soggy polders, and trash sorting facilities that surround the city.

By midafternoon they return, jostling for position in the communal winter roosting trees, sounds of small negotiations, hellos, courtship and contentment echoing from branch to branch.

Slowly, the thousands of chirpy caws diminish into the sleepy solitude of the night.

But not so sleepy. Darkness is the domain of the tawny owl, the Vondelpark's charismatic nocturnal raptor.

Migratory geese fly laboriously by on midnight journeys.

And, of course, the coots and moorhens, never far out of sight but not often heard until their provincial pre-dawn squabbles.

Each species has its own routine. Day belongs to the parakeets, naturalized in the 1970s and now ubiquitous in the Randstad.

And I particularly associate a certain type of lazy, cold winter sunrise with the aggrieved calls of the Egyptian Geese.

The short days all fall into their own predictable rhythm. Some songbirds begin their salutations promptly at sunrise.

While others seem to take longer to warm up in these dark days.

I'm excited to have a small window into these daily patterns and, eventually, their seasonality. As the days lengthen and breeding season begins in the spring, these patterns and these birds will change.

In the dataset, a single songbird sings all night: the redwing. Is this real? I have seen redwings around this winter, but they are not especially common. How confidently can the AI model distinguish this nocturnal 'ziep' sound from other thrushes?

Like the observation biases inherent in traditional birding, there are limitations everywhere here — from the AI model to the simple fact that not all birds are vocal. Despite this, I've found both to be valuable tools for training myself to look and listen a little deeper.

Muted