What is AtmoI4REN-4Cast?
An umbrella connecting meteorology, climate, artificial intelligence, and renewables.
The renewables-based energy system (production, infrastructure, siting) is affected by the prevailing atmospheric conditions, across the different time scales from weather to climate and local to regional to synoptic scale. Transitioning to a fossil-fuel free future needs not only accurate predictions for production, changes in production conditions or extreme power system events but also consideration of possible effects on the power grid and the grid infrastructure.
We are deeply grateful to share our work with our amazing partners who we are participating with in the projects collected in the umbrella AtmoI4REN-4Cast!
- Analyses: Understanding atmospheric processes helps predict changes in wind patterns and their effect on wind power. A high-resolution, gridded wind speed dataset for Austria, using statistical and artifical intelligence (AI) methods, helps analyze the past 20 years' data.
- Synthetic time series: Limited wind farm data impacts machine-learning / artifical-intelligence (ML/AI) models. Methods to create synthetic and semi-synthetic data from historical time series and gridded data help enhance time-series resolution, refining hourly data to sub-hourly intervals.
- Forecasting: Forecasting models are tailored for renewable energy applications. Post-processing numerical weather preciation (NWP) data with ML and statistical methods ensures precise forecasts on different scales. Models focus on extreme events, ramping, cloud movements, and uncertainty estimation.
- Extreme events: automated detection of extreme weather impacting renewables requires interdisciplinary methods. Novel techniques in AI and statistical post-processing detect and forecast these events.
- Climate scenario downscaling: downscaling regional and global climate models (regional climate models (RCMs) and globalc climate models (GCMs)) provides the needed spatial resolution. GeoSphere Austria uses statistical methods and ML techniques to downscale RCMs and GCMs, enhancing renewable-focused parameters.
- Climate impacts: AtmoI4REN-4Cast collaborates with energy suppliers and grid operators, offering recommendations for decision-makers. It also engages with impact assessment groups to evaluate risks and develop standardized processes for local climate adaptation measures.
Our idea
The concept behind AtmoI4REN-4Cast
The concept of AtmoI4REN-4Cast is to cover all aspects in the (renewable) power system that are affected in one or the other way by atmospheric and changing atmospheric and climatic conditions.
Thus, everything that can be dealt with from a meteorological and climatological perspective, including converting these data into power production, is part of AtmoI4REN-4Cast.
Data
Meteorological observations, renewable energy data (production, demand), numerical weather predictions, climate scenarios, analyses, etc.
Metadata for the production sites to convert meteorology to power.
Forecasting
Converting the massive amount of heterogeneous (big) data to power.
Using statistical, known industry standard methods, and (deep) machine learning tools.
Climate
What does change in the future? How do intensities change, location shifts, what happens to extremes?
What are profitable sites now and in 20 years, refacturing of sites, etc.
The users, customers, stakeholders - YOU!
You as users and customers are important!
We do value feedback to all our products, collaborations in research projects, and simply interacting with you and learning from industry!
Projects
Research projects connected to AtmoI4REN-4Cast
- AI4Wind: Link to the project website and to the project description
- CORONA: Internal ZAMG project (~2015) focusing on investigating novel machine learning methods for wind speed forecasting at points
- EnergyProtect: Link to the project website and to the project description
- EngagePV: Link to the project description
- HectoRenew: Starting with 1.1.2025; link to the project description
- MEDEA: Link to the project description
- ReduceData: Link to the project description
- SOCLENOW-AI: Link to the project description
- SSEA: Internal ZAMG/GeoSphere Austria project focusing on downscaling and post-processing sub-seasonal forecasts to a high spatial resolution (1 x 1 km)
- Wind4Future: Link to the project website and to the project description
- WINDSOR: Link to the project description
- WindTRF: Continuation of project CORONA, developing first machine learning wind speed forecasting methods and participating at the IEA Wind Task 36
Interested in joining the stakeholder board?
Want to know whats going on? You do have interesting use cases? Want to know what we are discussing?
Send us a note and we keep you posted!
Contact usBlog and News
Check our news and blog section
Here we will post updates, information on past and upcoming events, documents, ... and the like.
Team
Our Hardworking Team
Our hardworking team consists of meteorologists, climate scientists, machine learning experts, and enthusiastic and motivated people. Our non-exlcusive core team consists of:
Dr. Irene Schicker
She is a meteorologist at GeoSphere Austria in the group Post-processing. Her research focus is on machine learning for renewables, prediction and analysis, and extreme events. Her interest is in the combination of machine learning, energy meteorology, mountain meteorology, and data mining.
Mag.a Annemarie Lexer
She is a meteorologist and working at the Climate-Impact-Research department at GeoSphere Austria. Her research focus is on climate research, regional climate change analysis, synoptic climatology, and atmospheric circulation patterns and related precipitation events/extremes.
Anna-Maria Tilg, PhD
She is a trained meteorologist and works in the Climate-Impact-Research department. Her research focus is on the analysis of the spatial climate in Austria and its impacts on different fields, especially renewable energies. She holds a master’s degree in Atmospheric Sciences and a PhD in Wind Energy.
Mag. Alexander Kann
Meteorologist in the post-processing unit at GeoSphere Austria. More than 20 years of experience in NWP, nowcasting and post-processing. His research focus is on application-oriented analysis and forecasting techniques including statistical and ML methods.
Petrina Papazek
Researcher at GeoSphere Austria working on data-driven forecasting methods for renewable energies. She holds a bachelor's degree in Meteorology, a master in Geomatics, and a master in Scientific Computing.
Pascal Gfäller
AI/ML-researcher in the post-processing unit at GeoSphere Austria, primarily working on large area solar irradiance forecasting with deep learning. He holds a master's degree in computer science.Contact
Contact Us
Drop a note or send us your comments.
Location:
Hohe Warte 38, A-1190 Vienna, Austria
Email:
irene . schicker @ geosphere . at
Call:
+43 1 36026 0