The Cambridge "Green Blue" competition to reduce emissions / May 2025 / DOI
Carl argues that a climate system ultimately only responds to collective actions, and without a global cooperative incentive each nation will spring back to their own isolated short-term incentives that lead to an increase in fossil fuel burning. He has just published the "Themis Mechanism" as a simple alternative for equitable global emission reduction (long form). (6th May 2025: See a new article on Themis as well)
This got me brainstorming with Carl about how to test his theories out and we came up with an idea that is either terrible or awesome; please read on and judge appropriately. I think we should take advantage of Cambridge's unique structure to trial the Themis mechanism via a new competitive decarbonisation sporting league among Colleges that I dub the "Cambridge Green Blue". Given the Chancellor's recent unveiling of an innovation corridor between Oxford and Cambridge, the timing could not be better for an initiative like this. (TL;DR sign up at the bottom of this post if you'd like to participate)
Read full note... (4038 words)
2nd Programming for the Planet workshop CFP out / Apr 2025
Updated preprint on quantifying biodiversity cost of food consumption / Feb 2025
We've uploaded a revised preprint on our ongoing work on quantifying the
Quantifying the impact of the food we eat on species extinctions / May 2024
Submitted preprint on quantifying the biodiversity cost of global food consumption for peer review. This work links the LIFE biodiversity metric with food consumption and production data to quantify how different types of food and their production locations impact species extinctions. We discovered that the impact varies widely both across and within foods - in many cases by more than an order of magnitude. Using an opportunity-cost framing, we can estimate the marginal changes in expected extinctions from converting natural vegetation to agriculture or restoring farmland to natural habitat. Despite marked differences in per-capita impacts across countries, there are consistent patterns that could inform everything from national policies to individual dietary choices.
Predicting species using machine learning at CCAI / May 2024
Species distribution models are crucial tools that predict species locations by interpolating observed field data with environmental information. We develop an improved, scalable method for species distribution modelling by proposing a dataset pipeline that incorporates global remote sensing imagery, land use classification data, environmental variables, and observation data, and utilising this with CNN models to predict species presence at higher spatial and temporal resolutions than well-established species distribution modelling methods.
Global, robust and comparable digital carbon assets / Apr 2024
Paper on smart contracts for carbon credits at ICBC 2024 in Dublin. This work proposes the PACT stablecoin, which addresses concerns about credibility, scalability, and liquidity in voluntary carbon markets. We combine remote sensing data, modern econometric techniques, and blockchain-based certification and trading to create digital carbon assets against which offsetting claims can be transparently verified. The key innovation is creating a reproducible computational pipeline that not only quantifies CO2 emissions but also allows credits to be pooled based on co-benefits like biodiversity and jurisdictional attributes, increasing liquidity through fungibility. We implemented it on the Tezos blockchain, which is designed for low-cost transactions with minimal environmental impact.
Uncertainty at scale: how CS hinders climate research / Feb 2024
Paper on uncertainty in climate science in Undone CS. This workshop paper examines how computer science approaches to uncertainty can both help and hinder climate research practices. Patrick led this critical reflection on our own work, exploring the tensions between the way computer scientists typically handle uncertainty - often through abstraction and simplification - and the complex, multifaceted nature of uncertainty in climate and environmental science. It's part of the "Undone Computer Science" workshop series that encourages reflection on the limitations and impacts of computing approaches, which is particularly important when working at the intersection of CS and climate science.
Cambridge Zero highlights University efforts at Climate Week NYC (via) / Oct 2023
I was on stage in New York for Mission Possible
during NYC Climate Week. I was there with
The major highlights on the discussions with alumni centred around agency: a lot of them were wondering how to combine the evidence coming out Cambridge research and combine it with real policy action. A number of the alumni are obviously highly successful in their individual careers, and so the University helping to glue this together would potentially result in valuable actions that might not otherwise come together.
Functional Programming for the Planet / Sep 2023
Keynoted at ICFP 2023 on Functional Programming for the Planet, giving the opening keynote in Seattle. I discussed how functional programmers could contribute to addressing the climate and biodiversity crises through planetary computing. The talk covered our work on satellite sensing, forest monitoring, and how the principles of functional programming - immutability, composability, type safety - could help build reliable systems for environmental monitoring and conservation at scale.
Leveraging Scientific Innovation and AI to Scale Carbon Markets / Mar 2023
Discussion with Mantle Labs about carbon credits at the Forestry and Agriculture Summit in London. On stage with Jon Pierre, we discussed how scientific innovation and AI could help scale carbon markets. The conversation explored satellite-driven approaches to calculating additionality and permanence in forest carbon projects, addressing the credibility challenges facing the voluntary carbon market. This work aims to use technology to create more transparent and verifiable carbon credits.
A Case for Planetary Computing / Mar 2023
Preprint of planetary computing paper. Patrick led this work making the case for infrastructure to handle the ingestion, transformation, analysis, and publication of global data products for environmental science and policy-making. Drawing on our experiences working with environmental scientists on forest carbon and biodiversity preservation, we classify existing solutions by their flexibility in processing geospatial data and their support for building trust through traceability and reproducibility. The paper identifies research gaps around handling continuously changing datasets collected across decades that require careful access control. It's a call to action for the computing community to build better infrastructure for planetary-scale environmental data.
Financing Forests: A Credible Approach towards Halting Tropical Deforestation / Nov 2022
Wednesday seminar on financing forests using carbon credits at the Cambridge Computer Lab. I presented our work at 4C on applying computer science and satellite sensing to tropical deforestation. The talk explored how technology could help create credible carbon credit markets by using satellite data to measure additionality and permanence, potentially providing financial mechanisms to help avert both the biodiversity and carbon crises.
17th William Pitt Seminar - Who's in Charge? / Nov 2022
I opened the 17th William Pitt Seminar at Pembroke College on climate change with a brief talk about the status of the world's biodiversity, and how we have more agency than ever before to take matters into our own hands.
Trusted Carbon Credits / May 2022
With the recent controversies over low-integrity carbon credits, I spoke to Vox magazine about my skepticism about Adam Neumann's new startup.
"The problem with the current markets is nothing to do with how we can trade these more effectively," said Anil Madhavapeddy, who is an associate professor of computer science and technology at Cambridge University and the director of the Cambridge Center for Carbon Credits. "We just do not have enough supply." -- Vox
Confidential carbon commuting: exploring a privacy-sensitive architecture for incentivising 'greener' commuting / Apr 2012
Paper on our use of data lockers within Cambridge to incentivise more green commuting patterns. This work with