Artificial Intelligence and Biodiversity Measurement Report
Goldman Sachs releases a report on artificial intelligence and biodiversity measurement, aiming to explore how to develop and apply artificial intelligence technology to improve biodiversity measurement capabilities.
Goldman Sachs believes that natural ecosystems contribute over $13 trillion annually to the world, accounting for 12% of global GDP. However, the total annual investment in natural solutions is about $220 billion, and biodiversity measurement technology is a key limiting factor.
Related Post: HKMA Releases a Report on Using AI to Mitigate Climate Greenwashing Risks
Introduction to Artificial Intelligence and Biodiversity Measurement
Artificial intelligence technology can improve the reliability of biodiversity measurement, expand coverage, and possess deeper analytical capabilities. The report introduces the impact of artificial intelligence on biodiversity measurement from three perspectives:
Deep Occupancy model: Improving Resolution and Reliability of Biodiversity Measurement
Traditional biodiversity measurement relies on expert field observations and is difficult to cover a large area. The use of satellite images to measure biodiversity covers a wide range, but the accuracy is relatively low. By using deep occupancy models related to artificial intelligence, ground data and satellite information can be combined to estimate the distribution of species. This method improves the reliability of estimating species distribution by 27% and can identify habitat characteristics of species. Deep market share modeling can be applied to sustainable development linked bonds to establish key performance indicators and provide high-quality information for continuous monitoring and verification.
Distribution Model: Filling Biodiversity Information in Sparse Data Regions
Many regions lack detailed ecological data due to insufficient on-site investigation resources. Distribution models can combine artificial intelligence technology with expert knowledge, extend biodiversity information to sparse data areas, and solve the problem of high survey costs in traditional methods. Distribution models can help financial institutions map the natural related risk exposures of their investment portfolios, address the issue of insufficient availability of counterparty data, and assist regulatory agencies in focusing on the biodiversity impacts of business operations and supply chains.
Remote Sensing Data Model: Monitoring the Existence and Changes of Peatlands
Peatlands hold one-third of the global carbon storage, but due to their defining characteristics such as humidity, soil composition, and carbon content, most of them are located underground, making them difficult to monitor. Remote sensing data based on artificial intelligence technology can be combined with optical and satellite data, as well as terrain and vegetation status information, to identify peatlands using a small number of specific locations as training data, with an accurate rate of 95%. This model can help businesses assess biodiversity and land use risks and assist in designing Peatland carbon credits.
The report suggests that the role of artificial intelligence in biodiversity measurement depends on the following factors:
- Cross departmental collaboration: Scientific research, artificial intelligence, and the financial industry need to collaborate to transform research results into natural solutions.
- Biodiversity data: Reliable biodiversity data is the foundation for expanding nature related financing.
- Value driven factors: Biodiversity measurement can reduce business costs, improve operational management, and enhance productivity.
Reference:
