Invitation to take part in HEAT-X: Heat Equity and AI Transferability Across Global Cities

Dear IAUC Community, We invite you to participate in HEAT-X, a collaborative modelling initiative focused on urban heat mapping or downscaling, and AI transferability.   Project Context & Objectives Urban areas face mounting climate hazards, yet high-resolution observational and training data remains severely limited – particularly across data-sparse regions. The primary aim of this study […]

Dear IAUC Community,

We invite you to participate in HEAT-X, a collaborative modelling initiative focused on urban heat mapping or downscaling, and AI transferability.

 

Project Context & Objectives

Urban areas face mounting climate hazards, yet high-resolution observational and training data remains severely limited – particularly across data-sparse regions. The primary aim of this study is to investigate how effectively AI models spatially transfer to previously unseen locations and to develop machine learning tools capable of delivering locally accurate, neighbourhood-scale predictions worldwide.

 

Participant Challenge and Workflow

Participants will be tasked with downscaling air temperature over Hong Kong to neighbourhood scale.

  • Provided Data: ERA5-Land reanalysis data and a target output grid for Hong Kong.
  • Core Task: Train your AI models on your own local or existing datasets (excluding Hong Kong data) to evaluate zero-shot spatial transferability on the Hong Kong target grid.
  • Submission Window: Submissions will be accepted between November 2026 and February 2027.
  • Evaluation: Model submissions will be evaluated by Negin Nazarian’s group at UNSW Sydney, benchmarked against observational datasets.

Looking Ahead

While Hong Kong serves as our initial testbed, HEAT-X has the potential to expand to a broader network of diverse global cities (such as Lusaka, Montevideo, and Sydney). Additionally, this initiative is integrated into the larger WWRP Urban-PREDICT project, running alongside a complementary weather forecast intercomparison study for Hong Kong.

How to Get Involved

If you are interested in taking part or would like to learn more about the project, please reach out to the project leads:

Negin Nazarian: n.nazarian@unsw.edu.au

Lewis Blunn: lewis.blunn@metoffice.gov.uk

To top