Blog

Posts centred around personal research and developments.

New paper! Modelling near-surface temperature trends across Canada (07/09/2022)

My new paper on modelled near-surface temperature trends across Canada is now available online!


Investigating both historic patterns alongside progression across SSP 1-2.6 and 5-8.5 pathways, the paper highlights the value of sustainable development alongside forward-thinking climate mitigation and adaptation discussions.


Available here: https://doi.org/10.1080/00167487.2022.2114169

UK Heatwave 2022 & the future of UK Agriculture (18/07/2022)

To share one of the main impacts of UK climate change & the "2022 UK Heatwave",

Please see next to this text a list of the main arable crops grown in the UK, alongside the upper temperature limits for their survivability - against the minimum temperature forecast for the 19th of July 2022.

As the forecast displays, with the UK forecast to break temperature records tomorrow, the future of UK based agriculture is threatened. Plans for agriculture climate adaptation need to be implemented immediately, as do sustainable management practices which reduce GHG emissions from agriculture as well.

Link to original tweet

University of Alberta - Summer Research Intern: Ecosystem greenhouse gas emissions (11/06/2022)

I have recently started a 3-month research intern position at the University of Alberta, in the Earth Sciences department. Specifically working on greenhouse gas emissions from cropland, grassland and forest ecosystems.

To provide a short update surrounding the work I've been engaging with, across my first week in the lab:

  • Nitrous oxide emission measurements

  • Soil pH measurements

  • Amino acid analysis of crop samples

While this list is by no means extensive, and represents just a fraction of my internship, the opportunity to engage in both lab and fieldwork has already been truly invaluable and makes me really excited to continue work in the Earth and Atmospheric sciences in the future.

AI Image Generation & Climate Imaginaries (09/05/2022)

The use of AI in relation to climatology research is currently an emerging field, notably in relation to improving the accuracy of climate models (Chantry et al., 2021).

Another element of AI research has centred on the visualisation of future climate change, notably with the
"This Climate Does Not Exist" project which uses AI and Google Earth to generate images of real-world locations and how they will be impacted by climate change.

Recently, I have taken a different approach, looking into the application of text-to-image AI software to climate change, producing abstract images which represent the overarching trends and themes of future climate change - as understood by 'taught' AI systems. The benefit this has, is that it highlights how climate change is represented in the media, notably allowing for key areas of absence to be addressed, while also visualising climate 'imaginaries'. Following work by Davoudi & Machen, 2021 - climate imaginaries are important in understanding how society views climate change and thus, in how it ought to be addressed.

Alongside this blog post is a number of AI generated images, each set to the parameters of 500 iterations with a prompt of simply 'future climate change'. The method for said generation involved using a VQGAN and CLIP notebook, utilising the imagenet_16384 model.

The overwhelming trend to the images are representations of the following themes:
- Fossil Fuel Emissions
- Industrialisation
- Rising temperatures
- Flooding / sea level rise
- Biodiversity lose


Notably however, is a general lack of representation towards the following aspects of climate change:
- Local weather pattern change (e.g. precipitation increase/decrease)
- Greater frequency and intensity of extreme weather events (e.g. intensifying tropical storms and drought)
- Glacial and sea ice melt


This being said, some aspects are represented within these categories - notably, increases to w
ildfire frequency and intensity are arguably well represented, alongside flood events and global scale trends in weather patterns. However, this only exaggerates the existing 'human impact' bias of climate change in the images and resultantly in the majority of media coverage.

Thus, this preliminary study highlights the skewed trend of media in relation to climate change. The focus is too centred around the direct impacts of climate change on humanity, notably centring on extreme events - rather than outlining the more subtle - yet equally important - smaller scale trends which will equally dramatically impact both natural systems and humanity in the future.

References

Chantry, M., Christensen, H., Dueben, P. and Palmer, T. (2021). Opportunities and challenges for machine learning in weather and climate modelling: hard, medium and soft AI. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 379(2194), p.20200083. doi:10.1098/rsta.2020.0083

This Climate Does Not Exist: Link

Davoudi, S. and Machen, R. (2021). Climate imaginaries and the mattering of the medium. Geoforum. doi:10.1016/j.geoforum.2021.11.003

VQGAN and CLIP AI Image Generation Notebook: Link