Responsible investing is becoming a top industry priority. By incorporating physical and transitional risks into capital allocation decisions, not only can long-term economic resilience be fostered, but the natural assets we rely on for ecosystem services
such as biodiversity can be preserved.
Unfortunately, the industry has not yet reached the position where it can accurately measure the ESG impact of projects with ease with, for instance, a standardised online sustainability tool. This level of transparency remains a priority.
Given the new innovations emerging such as Light Detection and Ranging (LIDAR), satellite imaging, cutting-edge data science techniques, artificial intelligence (AI), distributed ledger technology (DLT), and green software engineering technology is our
best hope for limiting environmental degradation in the short to mid-term.
The demand for sustainable solutions is at an all-time high, Richard Conway, CEO of sustainability studio and Elastacloud, told Finextra. The ability to predict risks to business from a climate and biodiversity kind of point of view and the measure
impact you’re having on the planet, is vital. With the help of technology, we can identify the dots; and with a team behind the technology, we can join them.
Here are four ways fintech can help to halt environmental degradation.
1. Monitoring ESG transitions: Artificial intelligence
When it comes to keeping tabs on businesses, and their ESG transformations, technologies such as AI and more specifically, machine learning are making headway.
Data platform solutions firm and Microsoft Gold partner Elastacloud, for instance, uses AI to extract actionable insights from data, in order to support corporations on their journey to sustainability.
Our Sustainability Studio platform uses a series of AI models, that we have developed, to digest hundreds of ESG metrics, against various frameworks such as the Global Reporting Initiative (GRI) standards, the Sustainability Accounting Standards Board
(SASB), and more, explained Darshna Shah, lead data scientist, Elastacloud. We output our ESG scores one per year, per FTSE 100 company which drills down into the pillar scores of E, S, and G.
Shah added that Elastacloud can then examine different themes whether it be water waste, emissions, and so on and sub-theme scores, right down to the metrics. This intelligence is captured through the platform, to the user interface.
We also output a series of ESG forecasts, noted Shah. This involves digesting a lot of different data sources such as previous years ESG data, controversy scores, information around customer supply chain and assets in order to validate how an ESG
score is likely to perform in future years.
For this, Elastacloud leverages machine learning to derive a series of confidence scores, along with an explanation on how the forecast was reached.
In addition to the ESG forecasts themselves, we run a series of sentiment analyses that is, monitoring news sources and social media to gauge the alignment between what a company states in its sustainability reports, with what they are actually doing,
Shah explained.
The other way Elastacloud deploys AI is to power climate risk modelling: That is, essentially using tools such as the Representative Concentration Pathways [RCP] database, which the Intergovernmental Panel on Climate Change [IPCC] regularly uses in their
reports to monitor how forecasts of gross domestic product [GDP] per country may be impacted by different climate scenarios.
It is Elastaclouds ambition to extend this solution to the companies it works with, and apply supply chain data and assets. This would enable the firm to examine revenue at risk, or assets at risk, according to different climate scenarios such as cyclones,
flooding, or wildfires.
In order to get a business from point A to point B in their ESG transition, Elastacloud starts by running a cloud solution assessment, before working with the firm to build a data platform in the cloud. Once the data can be audited in a scalable and centralised
location, Elastacloud can then begin the data science life cycle of assessing areas of sustainability improvements, data quality, and gap analyses. From there, Elastacloud constructs AI models that may include recommendation systems for emission reductions,
climate risk forecasting, and so on. These models can be set up with minimal or no human decision making, in order to responsibly optimise the speed toward net zero.
The end-goal of this enterprise is to yield predictive analytics, which can recommend different scenarios and management protocols, based on what is being predicted, commented Shah.
2. Generating environmental indicators: Data science
Also at work to save the planet are advanced data science techniques.
Alternative data specialist, QuantCube, for instance, leverages the largest alternative data lake in the world comprised of over 14 billion text, geospatial, geolocation, and structured data points to deliver environmental insights to investors, corporates,
and government bodies.
According to Thanh-Long Huynh, co-founder and CEO of QuantCube, the most challenging part of this process is not sourcing the data, but processing and interpreting it: What is important about QuantCube, Huynh told Finextra, is our strength on the data
analytics side.
For example, when it comes to earth observation data such as a series of satellite images of a given woodland QuantCube deploys proprietary machine learning tools to identify changes in canopy coverage, thus converting the site into a timeseries.
This means we can estimate deforestation rates at a specific place in an accurate and systematic way, said Huynh. The information is then used to estimate how much it would cost to reforest the site. This is how environmental impacts are linked to economic
impacts.
Since the end of 2021, QuantCube has applied deep learning techniques in a systematic way to Earth observation satellite data from the European Space Agency (ESA), in order to power four new real-time indicators, which measure not only changes across key
sectors of the economy, but also the environmental impact in granular detail. These indicators include the:
- Urban growth index measuring development in cities and urban areas over time, and its impact on land use;
- NO2 pollution index tracking human activity by comparing average nitrogen dioxide produced across different regions;
- Water stress index monitoring water surface fluctuations to pinpoint the likely occurrence of drought and water shortages;
- Agricultural index observing agricultural land usage to help predict changes in agricultural yields.
The Environmental nowcast indicators enable asset management firms, hedge funds, pension funds, investment banks and corporate treasurers to closely track the impacts of climate change and adjust their capital allocation strategies accordingly.
We are a pioneer in analysing these kinds of datasets, pointed out Huynh. Whenever there are new data sets, there are new territories to explore and new opportunities to measure environmental impact.
Ultimately, this QuantCube case study reveals how advanced data science techniques can be leveraged to generate new sources of real-time, actionable, environmental intelligence for use by the financial industry.
3. Enabling sustainable computation: Green software
Another key piece of the sustainability puzzle is the collective carbon footprint of the planets programs and operating information. According to the
International Energy Agency, in 2020, cloud and data centres used approximately 1% of global electricity which equates to 160 million tonnes of CO2-equivalent.
One of the best responses to this challenge today is the green software engineering movement, which sits at the nexus of climate science, software, hardware, electricity markets, and data centre design. Architectures supported by green software principles
such as Web-Queue-Worker, N-tier, and Microservices are already being leveraged to ensure the planets computational stock becomes more sustainable.
This is an area we are actively working to contribute and collaborate with Microsoft on, said Shah. With recent predictions showing that a fifth of global electricity consumption will come from data centres by 2025, it is vital that our technologies and
digital architectures start to align with the eight
Principles of Green Software Engineering.
Since the green software movement has not yet become common practice across the financial industry, it is critical that engineers begin to construct digital solutions with not only performance and security in mind, but also environmental footprint.
There are many scenarios under which green software engineering needs to be considered, noted Shah. This means looking at bandwidths, looking at where solutions should run according to carbon intensity, or optimising code so it runs on more of a serverless-type
architecture. Other considerations include the need for over-provisioning, precision of AI models which can heavily impact emissions used in development and running over different services. This is all really important.
A place for firms to start is to follow the
Green Software Foundation a non-profit organisation, whose mission is to create a trusted ecosystem of people, standards, tooling, and best practices for building green software, said Shah.
4. Quantifying ecosystem services: LIDAR
Some technologies, such as satellite imagery and light detection and ranging (LIDAR) are being used to incorporate environmental, social, and governance (ESG) factors into the financial system thus making nature bankable.
Nature-based carbon capture company, Treeconomy, for instance, deploys remote sensing technology to quantify the value of ecosystem services from, and carbon sequestered by, a given woodland. The data is then used to connect rural landowners to the carbon
offset market thus generating a new source of income from trees.
Harry Grocott, CEO and co-founder of Treeconomy, told Finextra in an interview: We are set up to unlock nature’s value this runs through everything we do. Ecosystem services are worth trillions of dollars, yet they are not priced into todays economy.
Given the snowballing issue of climate change, carbon capture is becoming an increasingly vital component.
To quantify carbon capture in situ, Treeconomy uses drones to bounce laser beams down into a forest, which record the time the light takes to hit the canopy, as well as other parts of the tree. This process yields a three-dimensional image of a woodland.
Depending on the kind of LIDAR used, millions of data points can be collected from a single site.
What you get, to begin with, is an enormously confusing, unstructured, point cloud mesh, explained Grocott. A key step in the process, therefore, is to classify then colour the datapoints based on what the laser has hit. All this is run through open-source
protocols and specialist software to, ultimately, classify individual trees and measure a sites carbon capture.
The structured point cloud mesh is ingested and unitised by Treeconomys inventory management platform, Sherwood, which calculates how large each of the individual assets within a forest are. This is then translated into a figure representing the tonnes
of CO2 sequestered by the site.
We’re basically digitising natural assets by distilling them into a tonnage volume. To convert that tonnage volume into a financial asset and make nature bankable we sell it in the carbon offset market. Treeconomy is essentially a carbon accountant
at the project level, said Grocott.
By powering this positive cycle of financial incentives for nature preservation, Treeconomys technology ultimately renders trees more valuable alive, than they are dead.
Joining the dots
There are two broad ways of looking at technology. While it has, in part, caused the climate problem, it has the potential to become part of the solution.
Thankfully, there is no shortage of fintech innovations out there, which in their own novel ways, are at work to reduce humankinds impact on the environment.
Whether it be it LIDAR to quantify the value of ecosystem services; AI to monitor companies ESG transitions; advanced data science techniques to generate environmental indicators; or green software engineering to increase the sustainability of our computer
programs, all technology, ultimately, needs a team of human experts behind it to ensure the findings are packaged in a meaningful way that protects Earth, and the ecosystem services it supports.