IgnisNeuralis42
Members
First name (team leader)
Ilenia
Last name
Manco
Organisation name
CMCC Foundation - Euro-Mediterranean Center on Climate Change, Italy
Organisation type
Research Organisation (Academic, Independent, etc.)
Organisation location
Italy
First name
Paolo Francesco
Last name
Duminuco
Organisation name
CMCC Foundation - Euro-Mediterranean Center on Climate Change, Italy
Organisation type
Research Organisation (Academic, Independent, etc.)
Organisation location
Italy
First name
Otavio Medeiros
Last name
Feitosa
Organisation name
INPE - Instituto Nacional de Pesquisas Espaciais; CMCC Foundation - Euro-Mediterranean Center on Climate Change, Italy
Organisation type
Research Organisation (Academic, Independent, etc.)
Organisation location
Brazil
First name
Paola
Last name
Mercogliano
Organisation name
CMCC Foundation - Euro-Mediterranean Center on Climate Change, Italy
Organisation type
Research Organisation (Academic, Independent, etc.)
Organisation location
Italy
Model
Model name
GCast42
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
How would you best classify the IT system used for model development or forecast production:
High-Performance Computing (HPC) Cluster
Documentation
Model summary questionnaire for model GCast42
Please note that the list below shows all questionnaires submitted for this model.
They are displayed from the most recent to the earliest, covering each 13-week competition period in which the team competed with this model.
Which of the following descriptions best represent the overarching design of your forecasting model?
- Machine learning-based weather prediction.
What techniques did you use to initialise your model? (For example: data sources and processing of initial conditions)
The model is initialized using ERA5 reanalysis data, which introduces a 6-day delay due to data availability. The initial conditions are based on weekly averaged fields of atmospheric, surface, and oceanic (SST) variables.
If any, what data does your model rely on for real-time forecasting purposes?
The initial condition has a 6-day delay, but the model can generate extended forecasts and can be used for real-time S2S applications.
What types of datasets were used for model training? (For example: observational datasets, reanalysis data, NWP outputs or satellite data)
We used the ERA5 dataset, including all main atmospheric, surface, and oceanic (SST) variables.
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 spherical convolutional recurrent neural network, designed to generate deterministic weekly forecasts (weeks 3–4 and 5–6). It uses spherical convolutional layers, UNET style.
Have you published or presented any work related to this forecasting model? If yes, could you share references or links?
A paper with a probabilistic version of the model is in preparation.
Before submitting your forecasts to the AI Weather Quest, did you validate your model against observational or independent datasets? If so, how?
Validated against ERA5 reanalysis as reference.
Did you face any challenges during model development, and how did you address them?
Main difficulties were computational load due to large datasets and model management during training on HPC systems.
Are there any limitations to your current model that you aim to address in future iterations?
The initial deterministic model did not reach high skill (low BSS). Next versions will include probabilistic outputs and physical consistency constraints to improve performance.
Are there any other AI/ML model components or innovations that you wish to highlight?
Use of spherical neural network processing and custom architecture modifications for global continuit, and addition to the main model, we also working a downscaling module using GANs, designed to enhance the spatial resolution of the forecasts and improve local-scale prediction. - to be detailed in upcoming publication.
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.
Conceptualization: Otavio Medeiros Feitosa, Ilenia Manco, Paola Mercogliano, Paolo
Duminuco
Data Preparation: Otavio Medeiros Feitosa
Model Architecture and Validation: Otavio Medeiros Feitosa, with guidance from Ilenia
Manco and support in the integration of cGAN in subseasonal forecasting model
cGAN Downscaling Model
Conceptualization: Ilenia Manco, Paola Mercogliano
Data Preparation: Ilenia Manco
Model Architecture: Ilenia Manco
Model Validation: Ilenia Manco, Otavio Medeiros Feitosa, Paolo Duminuco
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