NordicS2S
Members
First name (team leader)
Erik
Last name
Larsson
Organisation name
Linköping University
Organisation type
Academic (Student)
Organisation location
Sweden
First name
Martin
Last name
Andrae
Organisation name
Linköping University
Organisation type
Academic (Student)
Organisation location
Sweden
First name
Joel
Last name
Oskarsson
Organisation name
Linköping University
Organisation type
Academic (Student)
Organisation location
Sweden
First name
Fredrik
Last name
Lindsten
Organisation name
Linköping University
Organisation type
Research Organisation (Academic, Independent, etc.)
Organisation location
Sweden
First name
Tomas
Last name
Landelius
Organisation name
SMHI
Organisation type
Meteorological Institution
Organisation location
Sweden
First name
Daniel
Last name
Holmberg
Organisation name
Finnish Center for Artificial Intelligence
Organisation type
Academic (Student)
Organisation location
Finland
Models
Please note that the model information below reflects the latest details provided by the team and does not show previous versions. Only the model summary questionnaires retain a submission history.
Model name
NordicS2S1
Model type
Hybrid
Number of individuals supporting model development:
1-5
Maximum number of Central Processing Units (CPUs) supporting model development or forecast production:
8-48
Maximum number of Graphics Processing Units (GPUs) supporting model development or forecast production:
4-16
How would you best classify the IT system used for model development or forecast production:
High-Performance Computing (HPC) Cluster
Model summary questionnaire for model NordicS2S1
The list below shows all model summary questionnaires submitted for this model. Participants may submit an updated questionnaire whenever their methodology changes.
Questionnaires are displayed from the most recent to the earliest, with the latest version opened by default.
Which of the following descriptions best represent the overarching design of your forecasting model?
- Machine learning-based weather prediction.
- Statistical model focused on generating quintile probabilities.
- Ensemble-based model, aggregating multiple predictions to assess uncertainty and variability.
What techniques did you use to initialise your model? (For example: data sources and processing of initial conditions)
Initial conditions are weekly means for the past two weeks from both atmospheric and ocean analysis. The models are also provided with the climatology for each variable, calculated for the two target weeks.
If any, what data does your model rely on for real-time forecasting purposes?
The operational models are the IFS and GLORYS models.
What types of datasets were used for model training? (For example: observational datasets, reanalysis data, NWP outputs or satellite data)
The models are trained on ERA5 and GLORYS12V1 reanalysis and finetuned on the IFS and GLORYS operational models.
Please provide an overview of your final ML/AI model architecture (For example: key design features, specific algorithms or frameworks used, and any pre- or post-processing steps)
The model is a generative model that aims to sample directly from the distribution of the weekly means for the target weeks. Thus, it's somewhere between a dynamical model and a statistical model.
All forecasts done before Oct 16 are done by one of these three models (NordicS2SX)
1. Diffusion
2. Graph-EFM
3. Flow matching
After Oct 16, they are done using flow matching that
1. is trained to sample unconditionally given just the time of year
2. is trained first unconditionally and then conditionally
3. is trained to just sample conditionally given the intial conditions
Have you published or presented any work related to this forecasting model? If yes, could you share references or links?
No answer.
Before submitting your forecasts to the AI Weather Quest, did you validate your model against observational or independent datasets? If so, how?
The validation was done against ERA5.
Did you face any challenges during model development, and how did you address them?
Since the initial conditions were so high dimensional, there was an issue where the model collapsed to deterministic predictions with very little spread. This was addressed by training an unconditional model and finetuning this.
Are there any limitations to your current model that you aim to address in future iterations?
It's not clear which atmospheric and oceanic data is needed to do accurate forecasts. It would make sense to try and forecast with a subset of the conditioning to see where it can be improved.
Also, the skill is not very good.
Are there any other AI/ML model components or innovations that you wish to highlight?
No answer.
Who contributed to the development of this model? Please list all individuals who contributed to this model, along with their specific roles (e.g., data preparation, model architecture, model validation, etc) to acknowledge individual contributions.
This team has chosen to keep its participants anonymous.
Which of the following descriptions best represent the overarching design of your forecasting model?
- Machine learning-based weather prediction.
- Statistical model focused on generating quintile probabilities.
- Ensemble-based model, aggregating multiple predictions to assess uncertainty and variability.
What techniques did you use to initialise your model? (For example: data sources and processing of initial conditions)
Initial conditions are weekly means for the past two weeks from both atmospheric and ocean analysis. The models are also provided with the climatology for each variable, calculated for the two target weeks.
If any, what data does your model rely on for real-time forecasting purposes?
The operational models are the IFS and GLORYS models.
What types of datasets were used for model training? (For example: observational datasets, reanalysis data, NWP outputs or satellite data)
The models are trained on ERA5 and GLORYS12V1 reanalysis and finetuned on the IFS and GLORYS operational models.
Please provide an overview of your final ML/AI model architecture (For example: key design features, specific algorithms or frameworks used, and any pre- or post-processing steps)
The model is a generative model that aims to sample directly from the distribution of the weekly means for the target weeks. Thus, it's somewhere between a dynamical model and a statistical model.
All forecasts done before Oct 16 are done by one of these three models (NordicS2SX)
1. Diffusion
2. Graph-EFM
3. Flow matching
After Oct 16, they are done using flow matching that
1. is trained to sample unconditionally given just the time of year
2. is trained first unconditionally and then conditionally
3. is trained to just sample conditionally given the intial conditions
Have you published or presented any work related to this forecasting model? If yes, could you share references or links?
No answer.
Before submitting your forecasts to the AI Weather Quest, did you validate your model against observational or independent datasets? If so, how?
The validation was done against ERA5.
Did you face any challenges during model development, and how did you address them?
Since the initial conditions were so high dimensional, there was an issue where the model collapsed to deterministic predictions with very little spread. This was addressed by training an unconditional model and finetuning this.
Are there any limitations to your current model that you aim to address in future iterations?
It's not clear which atmospheric and oceanic data is needed to do accurate forecasts. It would make sense to try and forecast with a subset of the conditioning to see where it can be improved.
Also, the skill is not very good.
Are there any other AI/ML model components or innovations that you wish to highlight?
No answer.
Who contributed to the development of this model? Please list all individuals who contributed to this model, along with their specific roles (e.g., data preparation, model architecture, model validation, etc) to acknowledge individual contributions.
This team has chosen to keep its participants anonymous.
Model name
NordicS2S2
Model type
Hybrid
Number of individuals supporting model development:
1-5
Maximum number of Central Processing Units (CPUs) supporting model development or forecast production:
8-48
Maximum number of Graphics Processing Units (GPUs) supporting model development or forecast production:
4-16
How would you best classify the IT system used for model development or forecast production:
High-Performance Computing (HPC) Cluster
Model summary questionnaire for model NordicS2S2
The list below shows all model summary questionnaires submitted for this model. Participants may submit an updated questionnaire whenever their methodology changes.
Questionnaires are displayed from the most recent to the earliest, with the latest version opened by default.
Which of the following descriptions best represent the overarching design of your forecasting model?
- Machine learning-based weather prediction.
- Statistical model focused on generating quintile probabilities.
- Ensemble-based model, aggregating multiple predictions to assess uncertainty and variability.
What techniques did you use to initialise your model? (For example: data sources and processing of initial conditions)
Initial conditions are weekly means for the past two weeks from both atmospheric and ocean analysis. The models are also provided with the climatology for each variable, calculated for the two target weeks.
If any, what data does your model rely on for real-time forecasting purposes?
The operational models are the IFS and GLORYS models.
What types of datasets were used for model training? (For example: observational datasets, reanalysis data, NWP outputs or satellite data)
The models are trained on ERA5 and GLORYS12V1 reanalysis and finetuned on the IFS and GLORYS operational models.
Please provide an overview of your final ML/AI model architecture (For example: key design features, specific algorithms or frameworks used, and any pre- or post-processing steps)
The model is a generative model that aims to sample directly from the distribution of the weekly means for the target weeks. Thus, it's somewhere between a dynamical model and a statistical model.
All forecasts done before Oct 16 are done by one of these three models (NordicS2SX)
1. Diffusion
2. Graph-EFM
3. Flow matching
After Oct 16, they are done using flow matching that
1. is trained to sample unconditionally given just the time of year
2. is trained first unconditionally and then conditionally
3. is trained to just sample conditionally given the intial conditions
Have you published or presented any work related to this forecasting model? If yes, could you share references or links?
No answer.
Before submitting your forecasts to the AI Weather Quest, did you validate your model against observational or independent datasets? If so, how?
The validation was done against ERA5.
Did you face any challenges during model development, and how did you address them?
Since the initial conditions were so high dimensional, there was an issue where the model collapsed to deterministic predictions with very little spread. This was addressed by training an unconditional model and finetuning this.
Are there any limitations to your current model that you aim to address in future iterations?
It's not clear which atmospheric and oceanic data is needed to do accurate forecasts. It would make sense to try and forecast with a subset of the conditioning to see where it can be improved.
Also, the skill is not very good.
Are there any other AI/ML model components or innovations that you wish to highlight?
No answer.
Who contributed to the development of this model? Please list all individuals who contributed to this model, along with their specific roles (e.g., data preparation, model architecture, model validation, etc) to acknowledge individual contributions.
This team has chosen to keep its participants anonymous.
Which of the following descriptions best represent the overarching design of your forecasting model?
- Machine learning-based weather prediction.
- Statistical model focused on generating quintile probabilities.
- Ensemble-based model, aggregating multiple predictions to assess uncertainty and variability.
What techniques did you use to initialise your model? (For example: data sources and processing of initial conditions)
Initial conditions are weekly means for the past two weeks from both atmospheric and ocean analysis. The models are also provided with the climatology for each variable, calculated for the two target weeks.
If any, what data does your model rely on for real-time forecasting purposes?
The operational models are the IFS and GLORYS models.
What types of datasets were used for model training? (For example: observational datasets, reanalysis data, NWP outputs or satellite data)
The models are trained on ERA5 and GLORYS12V1 reanalysis and finetuned on the IFS and GLORYS operational models.
Please provide an overview of your final ML/AI model architecture (For example: key design features, specific algorithms or frameworks used, and any pre- or post-processing steps)
The model is a generative model that aims to sample directly from the distribution of the weekly means for the target weeks. Thus, it's somewhere between a dynamical model and a statistical model.
All forecasts done before Oct 16 are done by one of these three models (NordicS2SX)
1. Diffusion
2. Graph-EFM
3. Flow matching
After Oct 16, they are done using flow matching that
1. is trained to sample unconditionally given just the time of year
2. is trained first unconditionally and then conditionally
3. is trained to just sample conditionally given the intial conditions
Have you published or presented any work related to this forecasting model? If yes, could you share references or links?
No answer.
Before submitting your forecasts to the AI Weather Quest, did you validate your model against observational or independent datasets? If so, how?
The validation was done against ERA5.
Did you face any challenges during model development, and how did you address them?
Since the initial conditions were so high dimensional, there was an issue where the model collapsed to deterministic predictions with very little spread. This was addressed by training an unconditional model and finetuning this.
Are there any limitations to your current model that you aim to address in future iterations?
It's not clear which atmospheric and oceanic data is needed to do accurate forecasts. It would make sense to try and forecast with a subset of the conditioning to see where it can be improved.
Also, the skill is not very good.
Are there any other AI/ML model components or innovations that you wish to highlight?
No answer.
Who contributed to the development of this model? Please list all individuals who contributed to this model, along with their specific roles (e.g., data preparation, model architecture, model validation, etc) to acknowledge individual contributions.
This team has chosen to keep its participants anonymous.
Model name
NordicS2S3
Model type
Hybrid
Number of individuals supporting model development:
1-5
Maximum number of Central Processing Units (CPUs) supporting model development or forecast production:
8-48
Maximum number of Graphics Processing Units (GPUs) supporting model development or forecast production:
4-16
How would you best classify the IT system used for model development or forecast production:
High-Performance Computing (HPC) Cluster
Model summary questionnaire for model NordicS2S3
The list below shows all model summary questionnaires submitted for this model. Participants may submit an updated questionnaire whenever their methodology changes.
Questionnaires are displayed from the most recent to the earliest, with the latest version opened by default.
Which of the following descriptions best represent the overarching design of your forecasting model?
- Machine learning-based weather prediction.
- Statistical model focused on generating quintile probabilities.
- Ensemble-based model, aggregating multiple predictions to assess uncertainty and variability.
What techniques did you use to initialise your model? (For example: data sources and processing of initial conditions)
Initial conditions are weekly means for the past two weeks from both atmospheric and ocean analysis. The models are also provided with the climatology for each variable, calculated for the two target weeks.
If any, what data does your model rely on for real-time forecasting purposes?
The operational models are the IFS and GLORYS models.
What types of datasets were used for model training? (For example: observational datasets, reanalysis data, NWP outputs or satellite data)
The models are trained on ERA5 and GLORYS12V1 reanalysis and finetuned on the IFS and GLORYS operational models.
Please provide an overview of your final ML/AI model architecture (For example: key design features, specific algorithms or frameworks used, and any pre- or post-processing steps)
The model is a generative model that aims to sample directly from the distribution of the weekly means for the target weeks. Thus, it's somewhere between a dynamical model and a statistical model.
All forecasts done before Oct 16 are done by one of these three models (NordicS2SX)
1. Diffusion
2. Graph-EFM
3. Flow matching
After Oct 16, they are done using flow matching that
1. is trained to sample unconditionally given just the time of year
2. is trained first unconditionally and then conditionally
3. is trained to just sample conditionally given the intial conditions
Have you published or presented any work related to this forecasting model? If yes, could you share references or links?
No answer.
Before submitting your forecasts to the AI Weather Quest, did you validate your model against observational or independent datasets? If so, how?
The validation was done against ERA5.
Did you face any challenges during model development, and how did you address them?
Since the initial conditions were so high dimensional, there was an issue where the model collapsed to deterministic predictions with very little spread. This was addressed by training an unconditional model and finetuning this.
Are there any limitations to your current model that you aim to address in future iterations?
It's not clear which atmospheric and oceanic data is needed to do accurate forecasts. It would make sense to try and forecast with a subset of the conditioning to see where it can be improved.
Also, the skill is not very good.
Are there any other AI/ML model components or innovations that you wish to highlight?
No answer.
Who contributed to the development of this model? Please list all individuals who contributed to this model, along with their specific roles (e.g., data preparation, model architecture, model validation, etc) to acknowledge individual contributions.
This team has chosen to keep its participants anonymous.
Which of the following descriptions best represent the overarching design of your forecasting model?
- Machine learning-based weather prediction.
- Statistical model focused on generating quintile probabilities.
- Ensemble-based model, aggregating multiple predictions to assess uncertainty and variability.
What techniques did you use to initialise your model? (For example: data sources and processing of initial conditions)
Initial conditions are weekly means for the past two weeks from both atmospheric and ocean analysis. The models are also provided with the climatology for each variable, calculated for the two target weeks.
If any, what data does your model rely on for real-time forecasting purposes?
The operational models are the IFS and GLORYS models.
What types of datasets were used for model training? (For example: observational datasets, reanalysis data, NWP outputs or satellite data)
The models are trained on ERA5 and GLORYS12V1 reanalysis and finetuned on the IFS and GLORYS operational models.
Please provide an overview of your final ML/AI model architecture (For example: key design features, specific algorithms or frameworks used, and any pre- or post-processing steps)
The model is a generative model that aims to sample directly from the distribution of the weekly means for the target weeks. Thus, it's somewhere between a dynamical model and a statistical model.
All forecasts done before Oct 16 are done by one of these three models (NordicS2SX)
1. Diffusion
2. Graph-EFM
3. Flow matching
After Oct 16, they are done using flow matching that
1. is trained to sample unconditionally given just the time of year
2. is trained first unconditionally and then conditionally
3. is trained to just sample conditionally given the intial conditions
Have you published or presented any work related to this forecasting model? If yes, could you share references or links?
No answer.
Before submitting your forecasts to the AI Weather Quest, did you validate your model against observational or independent datasets? If so, how?
The validation was done against ERA5.
Did you face any challenges during model development, and how did you address them?
Since the initial conditions were so high dimensional, there was an issue where the model collapsed to deterministic predictions with very little spread. This was addressed by training an unconditional model and finetuning this.
Are there any limitations to your current model that you aim to address in future iterations?
It's not clear which atmospheric and oceanic data is needed to do accurate forecasts. It would make sense to try and forecast with a subset of the conditioning to see where it can be improved.
Also, the skill is not very good.
Are there any other AI/ML model components or innovations that you wish to highlight?
No answer.
Who contributed to the development of this model? Please list all individuals who contributed to this model, along with their specific roles (e.g., data preparation, model architecture, model validation, etc) to acknowledge individual contributions.
This team has chosen to keep its participants anonymous.
Submitted forecast data in previous period(s)
Please note: Submitted forecast data is only publicly available once the evaluation of a full competitive period has been completed. See the competition's full detailed schedule with submitted data publication dates for each period here.
Access forecasts data