- contact@amc-lahti.fi
News
Home > News
What’s New at AMC-Lahti
Welcome to the AMC-Lahti Team News page. Here we share updates on our latest research activities, events, and team milestones. Stay connected with what’s happening inside our lab and beyond.
🎓 An inspiring visit and knowledge sharing, January 5, 2026 (Niort) 🇫🇷
Our PhD doctoral researcher recently visited his former program, @but_sd_niort , to share his academic and professional journey with current students.
Ammar is currently a second-year PhD student at AMC Lahti, where he specializes in artificial intelligence applied to atmospheric chemistry 🌍🤖
His path perfectly illustrates how data science can lead to impactful research addressing major environmental challenges.
Many thanks to @but_sd_niort for the warm welcome and engaging discussions, and congratulations to Ammar for this inspiring experience sharing 👏
AMC-Lahti at Finnish Inverse Days 2025 (Helsinki) 🇫🇮
AMC-Lahti members participated in Finnish Inverse Days 2025 in Helsinki on Monday 15 December – Wednesday 17 December 2025, a community event bringing together researchers and practitioners working on inverse problems. The event provided an opportunity to follow recent developments, exchange ideas, and connect with colleagues across the field.
This year, two of our doctoral researchers* presented their work:
🎤 Talk presentations by
• Jenni Köykkä — Inverse Perspectives on the SOSAA–FP Modelling Framework
• Samuel Agenorwoth — Uncertainty and Computation in Acoustic Imaging of Pipes
Also attending from AMC-Lahti was doctoral researcher Helmi Toropainen, along with professors Heikki Haario and Emilia Blåsten.
More information about the event: https://fips.fi/inverse-days-2025/
Di Chen has successfully defended his PhD thesis
Di Chen has successfully defended his PhD thesis titled “From atmospheric oxidation capacity at boreal forest to earth system tipping points under a changing climate” at 13:00 on Friday, November 28th, 2025 in Auditorium A129, Chemicum (Gustaf Hällströmin katu 2, 00560 Helsinki, Finland).
The opponent was Docent Tero Mielonen from the Atmospheric Research Centre of Eastern Finland (Finnish Meteorological Institute), and the custos was Professor Michael Boy (University of Helsinki).
Atmospheric oxidants (OH, O3, and NO3) are the most important chemical species controlling atmospheric chemical processes. They regulate the oxidation of biogenic volatile organic compounds (BVOCs) in the boreal forest atmosphere, while OH is also the major driver to oxidize SO2 to form sulfuric acid (H2SO4). Moreover, OH is also the main oxidant of methane (CH4), which is an efficient greenhouse gas. Since there are numerous compounds that react with OH, it is difficult to describe sinks of OH quantitively, and the calculation and simulation of OH reactivity has been of scientific interest for decades. In this work, we have further developed the one-dimensional model SOSAA (a model to Simulate the concentrations of Organic vapors, Sulfuric Acid and Aerosols) to be suitable for long-term runs. In the context of global warming, the emission of BVOCs (such as monoterpenes) will increase, which in turn decreases OH concentration. This effect is more pronounced in the boreal forest, but may impact the whole northern hemisphere, by increasing the lifetime of CH4 and thus leading to elevated CH4 concentration. We introduced methods from complexity science (complex networks and critical slowing down) to investigate possible connections between the Amazon Rainforest and the Tibetan Plateau under the context of global warming.
Our PhD Student Wenqing Peng Presented His Paper At ECAI 2025 In Bologna 🇮🇹
We are proud to share that our PhD student Wenqing Peng presented his research paper, “SPIN-ODE: Stiff Physics-Informed Neural ODE for Chemical Reaction Rate Estimation”, at the 28th European Conference on Artificial Intelligence (ECAI 2025) in Bologna, Italy.
Estimating rate coefficients from complex chemical reactions is essential for advancing detailed atmospheric chemistry. However, the stiffness inherent in real-world atmospheric chemistry systems poses severe challenges, leading to training instability and poor convergence, which hinder effective rate coefficient estimation using learning-based approaches. To address this, we propose a Stiff Physics-Informed Neural ODE framework (SPIN-ODE) for chemical reaction modelling. Our method introduces a three-stage optimisation process: first, a black-box neural ODE is trained to fit concentration trajectories; second, a Chemical Reaction Neural Network (CRNN) is pre-trained to learn the mapping between concentrations and their time derivatives; and third, the rate coefficients are fine-tuned by integrating with the pre-trained CRNN. Extensive experiments on both synthetic and newly proposed real-world datasets validate the effectiveness and robustness of our approach.
MSM / AMC Group Retreat – Hyytiälä, 8–10 September
A three-day retreat bringing together members of the MSM and AMC groups in Hyytiälä, combining scientific presentations, thematic discussions, and social activities to strengthen collaboration and group visibility.
The retreat took place over three days in Hyytiälä. On Day 1 (Monday, 8 September), the program started with a welcome session, introduction, and logistics, followed by the first round of individual participant presentations. The day concluded with dinner and a sauna in the evening. On Day 2 (Tuesday, 9 September), participants continued with a second round of presentations, visited the SMEAR II station, and then split into breakout sessions by research themes: AI applications in atmospheric science (led by Zhi-Song) and the FLEXPART/SOSAA modeling system (led by Petri). The day wrapped up with group reporting and discussions, followed by a lake-side barbecue and sauna. On Day 3 (Wednesday, 10 September), the focus shifted to strategic discussions on whether MSM–AMC–Lahti should merge groups and websites, along with a session on increasing visibility both scientifically and publicly. The retreat ended with final discussions and a closing dinner.
Our group was proud to participate once again in the European Aerosol Conference (EAC 2025) in Lecce, Italy 🇮🇹
Every year, our group takes part in EAC, which brings together researchers from across the world to discuss the latest advances in aerosol science, atmospheric chemistry, climate, air quality, and their societal impacts.
This year, six of our PhD students presented their work:
🖼️ Poster presentations by
Wenqing Peng – Physics-informed neural ODEs for chemical reaction rate estimation
Zeqi Cui – Machine learning to optimize OH-oxidation of naphthalene
Valery Ashu – Uncertainty quantification of reaction coefficients with MCMC
Ammar Kheder – Deep spatio-temporal neural network for air quality reanalysis
Haitong Zhang – Residential wood combustion emissions in Lahti, Finland
🎤 Oral presentation by
Benjamin Foreback – FLEXPART atmospheric back-trajectories for pollution transport to Beijing
summer school, “Application of AI/ML techniques in Atmospheric Science”
The University of Helsinki (UH), Institute for Atmospheric and Earth System Research (INAR), Lappeenranta-Lahti University of Technology (LUT), Department of Computational Engineering and the Lahti University Campus are pleased to announce the summer school “Application of AI/ML techniques in Atmospheric Science, 11th to 15th of August 2025 in Lahti, Finland.
This course will provide introductory lectures on relevant atmospheric topics (e.g., atmospheric chemistry, aerosol dynamics, Earth System models and numerical weather prediction) and computational methods (for data science and machine learning). Additionally, we will form mixed groups with different scientific backgrounds for the hands-on training, covering about half of the course time. Each group will use previously created atmospheric datasets in the exercises to train an end-to-end neural network, like LSTM, RNN, and Transformer. The students will learn basic GPU settings (Google Colab or local machines) and cuda-enable deep learning frameworks (PyTorch). We also provide advanced topics for highly motivated students, including data visualisation, analysis, and model optimisation. At the end of the course, each group will present their outcome and discuss the dis- and advantages of the applied method.
Applied Inverse Problems Conference 2025 (AIP 2025) — Rio de Janeiro 🇧🇷
📅 28 July – 1 August 2025 | Hosted by FGV EMAp
Our doctoral researcher Samuel Agenorwoth and Associate Professor Emilia Blåsten attended the Applied Inverse Problems Conference 2025 (AIP 2025) in Rio de Janeiro, Brazil. The conference was a valuable opportunity to follow recent developments in inverse problems, exchange ideas, and connect with researchers across the international community.
More info: https://eventos.fgv.br/aip2025
Our PhD student Ammar Kheder presented his first paper at SCIA 2025 in Iceland 🇮🇸
We are proud to share that our PhD student Ammar Kheder presented his first research paper, “Deep Spatio-Temporal Neural Network for Air Quality Reanalysis”, at the 23rd Scandinavian Conference on Image Analysis (SCIA 2025) in Iceland 🇮🇸.
Air quality prediction is key to mitigating health impacts and guiding decisions, yet existing models tend to focus on temporal trends while overlooking spatial generalization. We propose AQ-Net, a spatiotemporal reanalysis model for both observed and unobserved stations in the near future. AQ-Net utilizes the LSTM and multi-head attention for the temporal regression. We also propose a cyclic encoding technique to ensure continuous time representation. To learn fine-grained spatial air quality estimation, we incorporate AQ-Net with the neural kNN to explore feature-based interpolation, such that we can fill the spatial gaps given coarse observation stations. To demonstrate the efficiency of our model for spatiotemporal reanalysis, we use data from 2013–2017 collected in northern China for PM analysis. Extensive experiments show that AQ-Net excels in air quality reanalysis, highlighting the potential of hybrid spatio-temporal models to better capture environmental dynamics—especially in urban areas where both spatial and temporal variability are critical.
Benjamin Foreback has successfully defended his PhD thesis
Benjamin Foreback has successfully defended his PhD thesis titled “Extreme air pollution events in Beijing: Examining their origins and dynamics” at 12:00 on Thursday, the 19th of June 2025 in Aalto Auditorium at Lahti University Campus (Niemenkatu 73, 15140 Lahti).
The opponent was Professor John Lin (University of Utah), and the custos will be Professor Michael Boy (University of Helsinki/LUT University).
Dissertation will also be electronically available before the defense at: Helda e-Thesis
Air quality is an ongoing global challenge faced by modern society, and poor air quality is responsible for more than 6 million premature deaths annually. In addition to problems from long-term exposure to air pollution, short-term exposure to extreme air pollutant concentrations, such as during haze episodes, can lead to acute health problems and hospitalizations. Severe air pollution episodes also cause significant economic damage. This thesis examines two different types of extreme air pollution events. First, we used measurements to investigate the effects caused by fireworks celebrations during the Chinese New Year in Beijing over the course of seven years. Then we conducted a modeling study for a severe wintertime haze episode that affected Beijing and the surrounding region in November 2018. In our Chinese New Year measurement report, we analyzed observations from multiple data sources between 2013 and 2019. In all seven years, we observed a significant increase in concentrations of particulate matter and gaseous pollutants coinciding with the celebration of Chinese New Year. We found a downward trend in concentrations since 2016, which we speculate could be because of increased public awareness of the effects of fireworks on air quality.
Petri Clusius has successfully defended his PhD thesis
Petri Clusius has successfully defended his PhD thesis titled “Process modelling of emissions, chemistry and SOA formation in real environments” on Friday, June 13th at 12:00 in Physicum, Auditorio E204 (Gustaf Hällströmin katu 2, 00560 Helsinki, Finland).
The opponent was Senior Scientist Ulas Im from Århus University, and the Custos is Professor Micael Boy from the University of Helsinki.
If you are unable to attend in person and would like to follow the public examination online, it will be accessible remotely via UniTube.
Dissertation will also be electronically available before the defense at: Helda e-Thesis
Atmospheric aerosols influence weather and global climate through their interactions with clouds. Furthermore, aerosols directly impact air quality and Earth’s radiative budget. Aerosols enter the atmosphere either through the direct emission of particulate matter or via phase transitions from the gaseous to the liquid or solid phase, a process known as new particle formation. Particles in the size range of 50–150 nm in diameter are essential for cloud formation in the atmosphere, as they serve as the initial nuclei for the heterogeneous nucleation of water vapour. The concentration of these cloud condensation nuclei (CCN) affects cloud lifetime, optical properties, and precipitation. The relative contribution of aerosols, and consequently CCN, from direct particle emissions and new particle formation varies significantly across different environments. However, quantifying this remains challenging due to the complex interplay between atmospheric chemistry, direct particle emissions, new particle formation, and the deposition of gases onto these particles. Advances in measurement techniques now allow aerosol composition to be resolved as a function of size, yet identifying the origins and process history of individual particles remains difficult based on measurements alone.
Our PhD Students Presented Their Research At LUT DS Science Conference 2025 🇫🇮
We are proud to share that four of our PhD students — Jenni Köykkä, Samuel Agenorwoth, Valery Ashu, and Ammar Kheder — presented their research at the LUT Doctoral School (DS) Science Conference 2025 in Lappeenranta, Finland.
The third interdisciplinary LUT DS Science Conference took place on May 20, 2025, bringing together over 450 participants. Our PhD students presented their research:
- Jenni Köykkä: Particle-phase module for organic chemistry using Bayesian analysis and generative AI.
- Samuel Agenorwoth: Uncertainty quantification in pipe imaging.
- Valery Ashu: Optimizing autoxidation chemistry codes with neural networks.
- Ammar Kheder: Physics-guided deep learning for air quality prediction.
With over 160 oral presentations and 80 posters, the conference fostered interdisciplinary collaboration across the LUT research community.
Contact Us
contact@amc-lahti.fi
Location
LAHDEN YLIOPISTOKAMPUS Niemen kampus Niemenkatu 73 (B-osa) 15140 Lahti