Big data illustration by Timo Meyer

The latest update to the Sixth Assessment Report by the IPCC, published in February 2022, makes for pretty sobering reading. Its conclusion that “human-induced climate change is already affecting many weather and climate extremes in every region across the globe”, in the form of heatwaves, droughts, tropical cyclones and heavy precipitation, should not come as a surprise. But what does feel shocking is the assertion that half of the world’s population is now “highly vulnerable” to those impacts. It’s time we learned to adapt.

Adaptation doesn’t just mean building more robust flood defences or developing resilient strains of staple crops to withstand more extreme fluctuations in the weather. It also means redesigning man-made systems to account for a less stable world.

In the global financial system, for example, the risks of climate change-related phenomena to real-world assets are still not factored into the way business is conducted, despite a recent study suggesting that some $24.2 trillion of global financial assets could be at risk due to climate change. These assets are still being insured and used to drive further investment.

Were such a large sum to be wiped suddenly from the global economy, the effects would be felt everywhere. Our current systems are failing to factor in this type of threat. Thankfully, there are dedicated industries emerging to bake climate change data into the way markets do business, using big data, machine learning (ML) and artificial intelligence (AI) technologies to adapt to impending climate change and mitigate its impacts.

Where big data is concerned, recent advancements in ML and AI have accelerated the collection and processing of vast caches of climate data. Previously, climate data was so voluminous and complex that traditional data processing methods could not keep up, and very little was being done with it that fell outside of the realms of pure climate science. But ML and AI have enabled both more accurate predictive modelling and granular local datasets to aid decision-making in all manner of industries, from logistics to agriculture.

These evolving systems have the potential to help us better understand “the complex interactions between climate and Arctic sea ice, assessing and managing the risk of abrupt greenhouse gas emissions from peatlands, and analysing sensor information to better understand urban air quality,” writes professor Gavin Shaddick, a fellow at the Alan Turing Institute.

Broadly, these technologies fall into three separate fields: climate intelligence, climate AI, and supply chain tracking. All of these are attempting to make big business practices more responsible in their approach to climate mitigation and adaptation. Here’s how...

Climate intelligence

Here in the UK, plans are currently afoot to protect Bacton, a Shell-owned and operated gas terminal in north Norfolk, from the threat of coastal erosion and rising sea levels. The project, which received planning permission in 2018, would see 1.8 million cubic metres of sand dumped in front of the terminal, extending along a 3.5 mile stretch of coastline. At £22 million, the cost of the project isn’t cheap, and £5 million of the funding will come from the government via the Environment Agency. But should it be happening at all? In fact, should the gas terminal have been built there in the first place? If climate were put at the core of Shell’s decision-making, then perhaps not. It’s the hope of the climate intelligence industry that its technology will make projects like Bacton a thing of the past.

Climate intelligence is a system that promises to enable more informed and effective investment, credit and insurance underwriting decisions by revealing a business, sector or financial system's exposure to climate-related risks. In doing so, the hope is that a reliance on carbon-intensive assets will be designed out of the system by making it an unattractive investment proposition and a no-go area for insurers.

“Climate intelligence puts climate at the core of decision-making. Enterprises, governments and NGOs use it to make climate-aligned decisions,” says Iggy Bassi, founder and CEO of Cervest, a UK-based climate intelligence company. In practice, this means that before a large-scale project like Bacton gets to the planning stage, business owners, investors, insurance underwriters and every other potential stakeholder will have a clear understanding of the inherent climate risks.

A gas terminal situated on an eroding coastline would be exposed to climate risk both for its precarious location and its high greenhouse gas (GHG) emissions, and would therefore be unlikely to be funded. Better still, a company could face severe penalties for pursuing such an exposed project with private finances.

A recent study suggests that $24.2 trillion of global financial assets could be at risk due to climate change.

Climate change AI

By harnessing AI systems now, and as they evolve, the hope is that AI can help transform traditional sectors and systems to address the most urgent challenges of climate change: delivering food and water security, protecting biodiversity and bolstering human wellbeing.

One sector at the intersection of all of these systems, and ripe for AI support, is agriculture. At least that’s the belief of ClimateAI’s Himanshu Gupta, who has chosen to focus his startup’s energies on bringing the insights of big data to the world’s farmers. And with good reason: in 2019, 25% of global climate disaster costs were absorbed by the agriculture sector (a sector on which 70% of the world’s population depends for its income). Almost everywhere in the world, the profit margins of the agricultural sector are already perilously slim and heavily subsidised by governments. The threat of further climate instability has the potential to wreak havoc on food systems.

Nearly 10% of the world’s agricultural land is predicted to soon become unsuitable for farming.

What AI promises in this sector is resilience to shifting weather patterns and the other knock-on effects of climate breakdown by giving farmers a clearer understanding of which crops to plant where and when, as well as driving investment towards more resilient farming practices.

“The agricultural market is inefficient in not rewarding resiliency,” says Gupta. “Farmers’ crop production output is measured in yield, safety, and quality. Climate change impacts all three of these. The market, however, treats a grower with more resilient output the same as their neighbour whose output may be high during a good year but will drastically suffer during a drought year.

“We’re building a ‘Resiliency as a Service’ platform that allows markets to correct for inefficiencies and fund resilience in the supply chain.”

Gupta believes there are many areas in which AI offers solutions to the climate crisis, but, he warns, AI is often overhyped and the application of complex neural networks to a problem is not always the right approach. This is echoed by Climate Change AI, a non-governmental organisation (NGO) responsible for building a global movement of climate change and ML research.

“Machine learning,” they say, “can play an impactful role in many broader strategies for reducing and responding to climate change. At the same time, machine learning is not a silver bullet, and should serve to supplement (rather than divert attention from) other impactful actions to address climate change.”

Machine learning is not a silver bullet, and should serve to supplement other impactful actions to address climate change.

Supply chain tracking

Machine learning is currently being used to supplement efforts to reimagine global supply chains, which are riddled with unethical and environmentally unsound practices and behaviours. Materials like cobalt, nickel, tantalum and mica are associated with issues like child labour, slavery, theft of natural resources, environmental damage and human rights abuses, which are commonplace, but often not seen by consumers and manufacturers. A recent article in The Guardian links the global explosion in electric vehicle (EV) sales to extreme air and water pollution causing public health issues around Indonesia’s nickel mines.

Meanwhile, products like coconut, palm oil and intensively-farmed meat products are exacerbating deforestation in some of the world's most fragile ecosystems, and billions of tonnes of food is wasted in global supply chains each year.

For manufacturers, this is a mainstream problem, but one that’s difficult to police. Often, products that are unethical or environmentally unsound aren’t purchased directly, but are instead incorporated into products at different stages of the supply chain. Without knowing the provenance of goods or materials, it's impossible to ensure social and environmental standards are effectively applied as raw materials are transformed into consumer products.

The World Economic Forum has suggested that any technology capable of decarbonising global supply chains would be a “game-changer” for the impact of corporate climate action. Which is where ML comes into the picture. Blockchain tracking and traceability systems appear to offer a solution.

“Our mission is traceability and due diligence of raw materials from source to manufacturer,” says Douglas Johnson-Poensgen, CEO and co-founder of Circulor, a start-up providing blockchain-enabled supply chain solutions. “To achieve this, we’re using blockchain alongside traditional databases and machine learning to establish a new global standard for ethically and sustainably sourced minerals.”

This is achieved by giving an identity to a commodity and tracking and logging real-time supply chain data, including GHG emissions, throughout the supply chain.

All three of these technologies put AI and ML at the heart of market-based solutions to climate change. They do this by putting an explicit price on carbon emissions, which are typically externalised from economic systems, and spurring on businesses and industries to find cost-effective solutions to reduce emissions.

Previous market-based solutions to climate change, like carbon taxes, have been widely criticised by the climate movement for failing to deliver results and creating a form of climate bureaucracy, complete with all the red tape and loopholes that such a system involves. But, correctly applied, and with the right legislation to back them up, AI and ML could offer a real opportunity to avert the “atlas of human suffering” that UN Secretary-General António Guterres witnessed in the last IPCC report.