Abstract
Marine animal forests (MAFs) are benthic communities dominated by sessile suspension feeders (such as sponges, corals, and bivalves) able to generate three-dimensional (3D) frameworks with high structural complexity. The biodiversity and functioning of marine animal forests are strictly related to their 3D complexity. The present paper aims at providing new perspectives in underwater optical surveys. Starting from the current gaps in data collection and analysis that critically limit the study and conservation of marine animal forests, we discuss the main technological and methodological needs for the investigation of their 3D structural complexity at different spatial and temporal scales. Despite recent technological advances, several issues in data acquisition and processing need to be solved to properly map benthic habitats, assess health status, and measure structural complexity. Standardized methods and protocols are strictly necessary to meet the FAIR (findability, accessibility, interoperability, and reusability) data principles for the stewardship of habitat mapping and biodiversity data.
Introduction
Marine Animal Forests (MAFs) serve as the underwater equivalent of terrestrial forests, providing essential habitats that support global marine biodiversity. However, while terrestrial surfaces are mapped with high precision, only about 5% of the ocean floor has been accurately mapped at the community level. The core challenge lies in the immense 3D structural complexity of MAFs. Traditional 2D photo-sampling or manual scuba surveys are often insufficient—they are slow, low-resolution, and sometimes destructive. To effectively protect these ecosystems, researchers require non-destructive, high-resolution 3D models that can track growth, health, and structural shifts over time.
Technological Approaches Across Scales
There is no "one-size-fits-all" tool for investigating MAFs. The choice of technology depends heavily on the biological and spatial scale of the investigation, ranging from individual organisms (millimeter scale) to entire ecosystems (kilometer scale).
Key technologies currently in use include:
- Underwater Photogrammetry (SfM/MVS): Utilizing "Structure from Motion" and "Multi-View Stereo" algorithms, this is the gold standard for low-cost, high-resolution 3D mapping.
- Active Optical Imaging: Laser scanners and structured light provide millimeter-level precision for individual organisms.
- SLAM (Simultaneous Localization and Mapping): These algorithms enable Autonomous Underwater Vehicles (AUVs) to navigate and map complex, uncharted environments in real-time.
- Acoustic 3D Mapping: While lower in resolution than optical methods, acoustic sensors are essential for large-scale mapping (km scale) where light attenuation limits optical tools.
The Data Deluge and AI Integration
The field is currently facing a "data deluge," where millions of underwater images are collected annually, far outpacing manual annotation capabilities. To bridge this gap, researchers are turning to Deep Learning architectures (such as U-Net and PointNet++) for semantic segmentation—automatically identifying and labeling species within 3D point clouds. Tools like TagLab are becoming critical for creating the annotated datasets required to train these AI models. However, a significant lack of standardized benchmark datasets remains a hurdle for developing more accurate, marine-specific AI.
From Data to Knowledge: The Path to Information Systems
For marine data to be truly useful for global conservation, it must transition from raw acquisition to structured knowledge. This requires adherence to the FAIR principles: making data Findable, Accessible, Interoperable, and Reusable.
Currently, much of the data remains siloed in disparate formats. The development of common "ontologies" (standardized naming and classification systems) is essential to combine data from different expeditions into a single, global Information System for the oceans. This workflow involves improving acquisition, standardizing formats, and utilizing machine learning to extract meaningful ecological information from 3D models.
Future Directions and Limitations
Despite progress, several bottlenecks persist. Underwater optical surveys are hindered by light absorption, scattering, and "caustics" (flickering light patterns) that confuse mapping algorithms. Furthermore, 3D reconstruction struggles with "flexible" organisms like sea fans that move with currents, leading to alignment errors.
Future work must focus on:
- Human-in-the-loop AI: Developing "assisted" tools where experts can refine AI predictions to ensure high accuracy.
- Autonomous Navigation: Enhancing AUV decision-making to navigate the complex crevices and overhangs of MAFs without damaging delicate structures.
- Volumetric Analysis: Moving beyond "pretty pictures" to use 3D models for measuring actual volumetric changes, which 2D orthophotos often obscure.
By addressing these technological and methodological gaps, the scientific community can better monitor and conserve the vital 3D frameworks of our global marine animal forests.
Summary updated