day 3 - thursday

thursday 3 july - Programme

To sign up to sessions at the staff meeting, please use the form here

Notes and Additional Information

Staff Led Sessions

Predicting Change: The Role of AI and ML in Atmospheric Discoveries (full session time)

We propose a scientific session aimed at engaging the staff to discuss the development of a new course focused on Machine Learning (ML) and Artificial Intelligence (AI) applications in atmospheric science, in particular, science done in NCAS. This session will provide an opportunity to explore how to integrate cutting-edge technologies, share insights on current challenges, and identify the skills needed for future scientists in this rapidly evolving field. We aim to use this feedback to create a comprehensive curriculum that enhances the understanding and application of ML and AI techniques, ultimately driving innovation and improving research outcomes in atmospheric science.

Broadening Horizons: Widening Participation in STEM and Environmental Science Training (full session time)

A meeting of people involved or interested in scientific training within NCAS.

DIY API (full session time)

"A crash course in writing your own API (Application Programming Interface) using the FastAPI Python library. This session will cover:

- basics of the HTTP/HTTPS protocol

- how to implement a simple FastAPI server

- writing custom API endpoints

- using the API in an application"

Introduction to PyActiveStorage (full session time)

Efficient distributed data analysis with Active Storage and HDF5/NetCDF4 - the new tool that allows data reductions inside storage units, avoiding unnecessary data transfers

When is a model worth supporting? (full session time)

"A large proportion of science staff in NCAS use models - an even higher proportion will have come across models in their careers. Models are complex and need an extensive support environment, including dedicated staff. 

In an ideal world, perhaps all models that people in NCAS use would be supported. However we do not live in an ideal world and some element of choice is necessary - we choose which models to support (in some form - not only via CMS). 

How do we make the decisions to support one model and not another? 

SS&F would like to develop a set of community-agreed, openly available criteria that allow us to decide whether a model (a new model, for the sake of argument) is “supportable”. These criteria could also be applied in a reverse sense, i.e. which models should we retire support for. 

Should it just be whether a model is in demand? Or because it provides links to external partners? What other factors are there? 

We’d like to hear from you – what do you think should be the governing criteria for model support?

We welcome input from modellers of all kinds - atmospheric, earth-system and associated models, and chemistry models.