How Crowdsourced Bathymetry is Staying Afloat
For decades, the deepest parts of the ocean have remained vast, unmapped voids covering most of our planet. Improving global ocean mapping has become a priority, yet the maritime world faces a paradox: a glut of hydrographic data but no efficient way to process it. The bottleneck has moved from the seabed to the server.
Crowdsourced Bathymetry and a Flood of Unstructured Data
According to the most recent Seabed 2030 update, approximately 27.3% of the ocean floor has been mapped in accordance with contemporary standards. To address this deficiency, the maritime industry has increasingly adopted ‘Crowdsourced Bathymetry’ (CSB), repurposing standard vessels like merchant ships and fishing boats as platforms for scientific data collection. Nevertheless, this influx of information has presented significant challenges regarding data quality. Non-standardized collection protocols often yield datasets characterized by artifacts, such as false bottoms, cavitation spikes, and tide offsets. Historically, the mitigation of these errors required exhaustive manual verification by hydrographers, a methodology that has proven unsustainable given the escalating volume of soundings entering the system.
Challenges in Managing High-Volume Hydrographic Data
The solution may lie in real-time, localized processing at the data collection point. Collaborative efforts between the Danish Geodata Agency, the Austrian Institute of Technology, and the MobiSpaces project have pioneered a “processing at the edge” philosophy using onboard AI modeling and anomaly detection. Relying on an onboard AI model known as ‘MapFed’, vessels can now validate data in real-time, reducing data storage needs and transmission requirements at sea. This system learns the expected depth of its environment and flags only the anomalies for human review. This approach is not just efficient; it is a necessity for vessels operating in remote waters where satellite bandwidth is a precious and expensive commodity.
The efficacy and reliability of this automated methodology is already proving to be robust. In trials conducted by the Danish firm Sternula, the MapFed system operated autonomously for hundreds of days, processing over 5.7 million depth points. Most impressively, the AI’s judgment aligned with human experts more than 85% of the time, proving that the vast majority of data cleaning can, and perhaps should be, handled by machines.
Automating Crowdsourced Bathymetry: The Future of Seabed 2030
In mid-2025, NOAA and The International SeaKeepers Society reached a watershed moment, processing 315 million depth observations from across the US Gulf and Atlantic coasts. The true achievement here was not the scale of the collection effort, but the deployment of automated pipelines capable of turning raw, fragmented data into noise-free, chart-informing reference data at scale.
By shifting the processing burden from centralized servers thousands of miles away to the data collection point, industry innovators are successfully navigating the paradox of data abundance in a resource-constrained environment. As automated pipelines like MapFed demonstrate that machines can reliably filter and structure raw data with accuracy comparable to human experts, the maritime science community is moving toward a more sustainable model of global mapping. While CSB remains a supplement to traditional hydrographic surveys rather than a total replacement, its integration into automated, scalable workflows offers a promising path forward for potentially informing official nautical charts and unlocking the vast, unmapped potential of our oceans.
Another Mapping Current News Story: Gold Rush Mapping Reveals Relics and Dangers
Discussion Questions:
- With more than 70% of the ocean floor still unmapped, what role do you think crowdsourced bathymetry will play in closing the gap?
- Can AI-based quality control become trusted enough to support official nautical charting workflows, or will human validation always remain essential?
- What challenges do you see in using data collected from a wide range of vessels with different equipment, operators, and environmental conditions?
Want to explore more of the geospatial industry with us? Join the newsletter at the link below.
