National Framework for Climate Services

On the 20th of February 2024, The National Framework for Climate Services (NFCS) held its first annual meeting to present their work to multiple public and private companies. Its first and main objective is to gather and unite resources for climate user communities in Ireland.

The NFCS is developed in Ireland following the Global Framework for Climate Services introduced in 2009. Using this reference, each country can adapt the framework for their needs. Switzerland have already been developing their own version of the NFCS (National Centre for Climate Services) since 2014 and the director of the service, Angela Michiko Hama, joined us via Zoom to give an overview on their work over the past ten years.

Met Éireann, who are in charge of caring the project, divided their talk into two days. The first day, which I attended, for producers, and the second for users.

A succession of people from different companies, agencies, or universities presented their work on the different hazards that may arise in the future in floodings and landslides.

• Met Éireann, Environmental Protection Agency 
• Geological Survey Ireland 
• University College Cork 
• Office of Public Work
•  Irish Centre for High-End Computing

In between talks and presentation, activities were proposed and facilitated by Met the organisers. The objective of these discussions was to understand what climate producers, as ourselves, were doing and what they expected from other producers to do as well. 

The last main point of the conference, had to deal with uncertainty and how to show them. Again, they wanted to bring discussion between different actors to try and find a place where people can agree and be consistent.

IEA Wind Task 51 Workshop AI driven Weather Forecast models

On the 11th of January, I attended a webinar conference organized by IEA Wind Task 51 on Machine Learning Weather Prediction.

 

The talk has been recorded and is available on YouTube. The three-hour presentation was given by four researchers :

• Mariana Clare from ECMWF on "The rise of data driven weather forecasting"
• Gregory Hakim from the University of Washington on "Dynamical Test of a Deep-Learning Weather Prediction Model"
• Joel Oskarsson from Linköping University on "Neural Weather Prediction for Limited Area Modelling"
• Florian Achermann from ETH and JUA.ai on "Ultra High Resolution Wind
  Forecasts"

The first presentation from Mariana Clare gives an overview of the current state of the work done by ECMWF on data driven weather forecasting.

Since 2022, ten AI models have been developed by different companies and/or universities. Since then, the European Centre for Medium-Range Weather Forecasts have run deep analysis on a few chosen AI models:

• FourCastNet by NVIDIA (Early 2022)
• PanguWeather by Huawei (Late 2022)
• GraphCast by DeepMind (Late 2022)

The test is split in two categories. First, the skill of each model is compared with the classical IFS HRES model from ECMWF. The results for different variables are shown below.

A scoreboard represent the skill of a given forecast compared to the IFS HRES model (state of the art for physics-based models) (source : Blog)

FourCastNet is not shown here, but this image highlights how well PanguWeather, but mostly, GraphCast does compare to the ECMWF model.

Skill score are one thing when assessing the performance of weather models, but extreme events are even more important. The next part of the test is then to analyse certain past events and see how well AI model are able to depict them.

Dr Clare, proceeds on looking at two particular events : Storm Eunice, which hit Great Britain in 2022 and the Cyclone Freddy which occurred in the Indian Ocean. The different models (IFS HRES, FourCastNet, PanguWeather, and GraphCast) are all able to predict the storm 60 hours in the future. However, AI models (PanguWeather and FourCastNet) do not show the same results for the prediction of the Cyclone 48 hours in the future compared to the HRES model. 

The AI models assessed are all trained using 40 years of reanalysis data from ERA5 using multiple variables (For GraphCast the trained data set is approximately 35GB). Then, the models are fine-tuned using a more accurate analysis for specific time periods. As the reanalysis dataset, the models have a spatial resolution of 0.25°.

The main advantages of AI models are the time and energy consumption of one forecast. After being trained, AI models are 500,000 less time and energy consuming than a classic HRES forecast. ECMWF have developed python modules for each of the three AI models which can be installed on a local machine :

• pip install ai-models-panguweather
• pip install ai-models-fourcastnetv2
• pip install ai-models-graphcast

While analysing these different models, ECMWF also wanted to develop their own using a similar approach to the DeepMind model. The first version seems to show good results, but their main goal is to increase the spatial resolution.

 

She finishes the talk by mentioning the challenges in the years to come for AI models :

1. Train the models on observation as well the reanalysis.
2. Move from deterministic to probabilistic models to integrate measures of uncertainty.
3. Build trust in AI models for researcher.
   

The next talk by Gregory Hakim gives the detailed process of testing a deep learning weather model.

Joel Oskarsson then, gives an explanation of how AI models can run on a limited area.

Florian Achermann finishes by explaining the development of a high resolution wind forecasting model.

Force group for new content

We have a shared directory, DATA, on our data server. The shared directory is in group mcc, and each user is also a member of this group. However, when users created files and directories inside this shared folder, permission defaulted to the user's group, and others could not access it.

To fix this, I used the following command for the shared directory:

chmod g+s DATA

I found that tip here.

Retrieve multiple files from sftp

I wanted to get GLEAM data from an sftp server. sftp doesn't seem to support wildcards with the get command, so instead I used a script:

for YYYY in 2015 2014 2013 2012 2011 2010 2009 2008 2007 2006 2005; do
  sshpass -p password sftp -P 2225 username@hydras.ugent.be:/data/v3.8a/daily/${YYYY} << EOF
  get E_${YYYY}_GLEAM_v3.8a.nc
  exit
  EOF
done