Citizen Earth

Bayesian retrieval techniques for high-dimensional spectroscopic observations of exoplanet atmospheres

Exoplanet astronomers study the diversity and properties of planets to understand their formation and evolution. Using sensitive spectroscopic observations, they investigate atmospheric structure, climate, and chemistry. The data are large and complex, requiring advanced models and computational methods. This interdisciplinary field combines astronomy, physics, chemistry, and numerical modeling. In this project, we aim to develop a fast Bayesian inference technique to analyse high-dimensional spectroscopic data, particularly in combination with lower-resolution observations from new space-based observatories such as the James Webb Space Telescope.

Period: 2023–2025
PI/contact: Jens Hoeijmakers
University: Lund University

Animating a climate future based on improved CO2 sequestration by monitoring wetlands and forests from satellites

Forests are important for both absorbing CO2 and creating oxygen. The overall carbon absorption - emission cycle however remains a puzzle. Forests, wetlands and general land use comprise critical pieces of that puzzle which are still not well understood. With wetlands holding some of the largest stores of carbon on the planet, their fate becomes a critical factor in predicting evolution of greenhouse gases, such as CO2 and methane, in our atmosphere. The main objective of the project is to develop mathematical tools linked to machine learning methods which enable monitoring and forecasting of carbon emissions/absorption from forestry management and wetland maintenance practices.

Period: 2024–2025
PI/contact: Alexandros Sopasakis
University: Lund University

POLLENOMICS: Decoding the farming history of Europe using advanced statistics to combine ancient DNA with paleo-pollen data

The goal of the project is to combine reconstruction of human migration based on ancient DNA data with pollen-based land cover data, which we term Pollenomics, to create a new proxy-based Land Use and Land Cover Change (LULCC) dataset for Europe. Improving the prediction accuracy of LULCC models serves various purposes, ranging from identifying agricultural practices that maximise future food production (crop yields) under different future climates scenarios to optimising conservation strategies for preserving critical ecosystems and biodiversity. Moreover, it can improve our capacity to evaluate how urban areas will withstand climate-induced changes and guide sustainable land-use planning for resilient and adaptable cities.

Period: 2024–2025
PI/contact: Behnaz Pirzamanbein
University: Lund university

Systematic spectral clean-up for massively parallel
surveys of stars

The goal of the project is to improve the science return from massively parallel spectroscopic surveys of stars by developing rapid, reliable, automated algorithms for pre-processing stellar spectra to remove defects and estimate the continuum flux.

Period: 2025–2026
PI/contact: Ross Church
University: Lund university

Galaxy formation in the exascale era

As part of our project, we aim to port key multi-scale physical models to DYABLO. The transition to DYABLO presents significant advantages, particularly in terms of potential performance improvements and enhanced support for GPU acceleration. These enhancements are critical for managing the increasing complexity and scale of astrophysical simulations, allowing us to leverage the full power of modern computational resources. Ultimately, we are setting the stage for a groundbreaking achievement: the first-ever simulation of a Milky Way-like galaxy from the Big Bang to the present, with individual stars resolved - a leap far beyond the current limit of grouping stars by tens of thousands.

Period: 2025–2026
PI/contact: Oscar Agertz
University: Lund university

Energy efficient and fast numerics for earth system modelling

With an increasing usage of compute throughout the world, its energy consumption is skyrocketing. This starts to cause issues with the transition to a carbonfree society, which uses electricity from renewables. It is thus imperative to increase energy efficiency in simulations. We will approach this using an interdisciplinary approach encompassing applications, computer science and numerical analysis.

Period: 2025–2026
PI/contact: Philipp Birken
University: Lund university

A GPU-based particle-in-cell framework for kinetic plasma modelling

Period: 2023–2024
PI/contact: Maria Hamrin
University: Umeå University