A new frontier in biodiversity research: uncovering patterns of hidden diversity with eDNA and AI

The United Nations Convention on Biological Diversity aims to protect 30% of land and sea areas by 2030, necessitating comprehensive understanding of biodiversity patterns.

However, the majority of species on Earth remain unknown to science, referred to as hidden diversity, with estimates ranging from 80-99.9% of total species diversity.

Here we propose to integrate environmental DNA (eDNA) data and high-resolution remote-sensing data within customized convolutional neural network models to provide a more accurate representation of species diversity patterns, including unknown species identified in the eDNA data.

This approach would substantially advance biodiversity research by including hidden diversity in species richness maps and identifying biodiversity hotspots that are truly reflective of the underlying diversity.

Focusing on Sweden as a proof-of-concept, the resulting models will provide actionable insights for environmental agencies, governments, municipalities, researchers, and the general public.

The proposed research offers a scalable and standardized solution for measuring and combating the global biodiversity crisis, potentially extendable to a global scale.