FDL Europe 2022

Frontier Development Lab - Europe - 2022 - Aerosols challenge

FDL Europe 2022 - Aerosols challenge - Pyrocast

I took part in the FDL-Europe research sprint 2022 and found it a very rewarding experience. I found this to be very conducive to creating a team-oriented culture which I enjoyed greatly. Being part of a multi-disciplinary team with domain experts and machine learning experts was very stimulating. It was great fun to work with the amazing computational resources that the FDL program provides.

My team’s challenge consisted in:

  1. Developing a dataset for the scientific study of pyrocumulunimbus (PyroCb) storm clouds.
  2. A machine learing pipeline for predicting the ocurrence of PyroCb storm clouds 6 hours ahead of time.
  3. Causal inference tools with which to understand what are the most important causes of this phenomenon.

The team consisted of 4 wonderful researchers:

  1. Kenza Tazi
  2. Ashwin Braude
  3. Daniel Okoh
  4. Myseif

3 excellen faculty members:

  1. Kara Lamb
  2. Paula Harder
  3. Nis Meinert

and a great adviser in Duncan Watson-Parris. I would like to thank our lovely supporting team Anchal Bassi, Catherine Ballarin and Cormac Purcell.

Here is a link to the youtube showcase presentation.

Here is alink to the neurips paper “Pyrocast: a Machine Learning Pipeline to Forecast Pyrocumulonimbus (PyroCb) Clouds”.

Here is alink to the neurips paper “Identifying causes of Pyrocumulonimbus (PyroCb)”.

Here is a link to the project repository.