Climate AI Nordics

Climate AI Nordics

Nordic network for collaborations around tackling climate change using AI and machine learning.

Featured Project: AI-based seabird monitoring at Stora Karlsö, Sweden

Project actors: Jonas Hentati Sundberg (Swedish University of Agricultural Sciences) et al.

At the island of Stora Karlsö, close to Gotland in Sweden, there is an artificial rock shelf – called AukLab – with enough space to host hundreds of breeding birds. The base of this platform rests in the middle of an established Common Guillemot colony, and allows researchers to study the birds from just a few centimeters away. The AukLab artificial rock shelf was built thanks to funding from WWF Sweden, and its purpose is to enable detailed studies without disturbing the birds – similar to a bird hide, but hanging outside the cliff edge.

Through AI-based camera surveillance, every aspect of the birds’ life on land is meticulously documented – from their return to the cliff ledge after winter to lay eggs and raise chicks, to their departure in mid-summer. Additional measurement sensors, such as thermometers, weight scales, and thermal cameras, provide further insights into what we observe happening in the birds’ hectic lives.

Seabirds are near the top of their food web, and changes in their habitat are reflected in their behaviour. Climate change, overfishing, eutrophication and toxic emissions are all examples of human impact on the environment that can negatively affect the life of Baltic seabirds in different ways. By studying changes in the birds’ behaviour, one can get an idea of what is occurring beneath the water’s surface. Much like the canary in the mine, the seabirds act as “messengers” of environmental change and overall health. When combined with other research projects, seabird research can provide greater insight into the Baltic Sea ecosystem, and thereby inform decisions regarding best environmental management practices.

Near-term future work includes adding additional sensors to further improve the monitoring of the guillemot. In particular, acoustic monitoring techniques will complement the current vision-based sensors. The hope is for this additional data modality to provide even more fine-grained insights about the lives of the Guillemots. By leveraging multiple sensors and advanced AI analysis, researchers are able to paint a fuller picture about the environment around the birds, to understand more fully how different stressors change their lives.

At this webpage you can see what the artificial rock shelf looks like in 3D!

For more information about the project, please refer to this page. Also, see this publication which describes some of the work. You can also reach out to Jonas Hentati Sundberg for more information.

Image

Published:

Categories:

You May Also Enjoy

Reliable conclusions from remote sensing maps

Event date:

Webinar with Sherrie Wang, MIT. Remote sensing maps are used to estimate regression coefficients relating environmental variables, such as the effect of conservation zones on deforestation. However, the quality of map products varies, and -- because maps are outputs of complex machine learning algorithms that take in a variety of remotely sensed variables as inputs -- errors are difficult to characterize. Thus, population-level estimates from such maps may be biased. We show how a small amount of randomly sampled ground truth data can correct for bias in large-scale remote sensing map products. Applying our method across multiple remote sensing use cases in regression coefficient estimation, we find that it results in estimates that are (1) more reliable than using the map product as if it were 100% accurate and (2) have lower uncertainty than using only the ground truth and ignoring the map product. Paper: https://arxiv.org/abs/2407.13659

Machine learning for Earth system prediction and predictability

Event date:

Webinar with María J. Molina, University of Maryland. Machine learning can be used for Earth system prediction, or to study the uppermost limit of prediction skill theoretically achievable given the system's initial state or other factors, otherwise known as predictability. In traditional numerical weather prediction frameworks, we solve the governing partial differential equations starting from an initial state. This initialized prediction framework usually involves three stages: 1) generating the initial conditions of the Earth system, 2) running the mathematical representation of the system on a computer forward in time, and 3) analyzing the output and converting it into a format that is useful for end users. Machine learning can be used to improve each of these individual stages, or to circumvent the three stage framework altogether, and examples of each will be given in this seminar. More time during the seminar will be dedicated to the challenges surrounding subseasonal prediction, which focuses on lead times of three to four weeks, and how we can use machine learning to both uncover potential biases in our initialized prediction systems and how we can bias-correct them.

Climate AI Nordics Newsletter, February 2025

Published:

Welcome to the Climate Change AI Nordics Newsletter, February 2025! Read about recent and coming seminars, workshops, and the monitoring of Baltic seabirds using AI-powered cameras and microphones.

Climate AI Nordics Core Team

Published:

We are proud to announce the amazing Core Team of Climate AI Nordics! Ten people spread over the Nordics will take place in the Core Team, to share the responsibility of developing the network and run our activities.