Welcome back, AI prodigies!

In today’s sunday special:

  • 📜 The Prelude

  • 🚱 A Brief History of Seafloor Mapping

  • 🌊 How AI Helps Capture the Seafloor

  • 🌍 The Real-World Impact?

  • 🔑 Key Takeaway

Read time: 7 minutes

đŸ©ș PULSE CHECK

🎓 Key Terms

  • Extended Kalman Filter (EKF): Estimates the state of a moving object when the movement isn’t simple or straight.

  • MultiBeam EchoSounder (MBES): A type of sonar system that maps the seafloor by sending out a fan-shaped swath of sound beams.

📜 THE PRELUDE

It’s a calm, clear, and crisp day. The Pacific Ocean whirls with whitecaps, as the sea and sky merge into a single stretch of boundless blue. Suddenly, a sleek vessel equipped with a “wing sail” and solar panels slices through the sea. No captain stands at the helm, and no crew mans the deck.

The ocean covers approximately 70% of Earth’s surface. It’s the largest livable space on our planet, and there’s more life there than anywhere else on Earth. Despite this, the vast majority of the global ocean floor is largely unknown. As of June 2025, only 27.3% of the global ocean floor has been mapped using high-resolution surveying technology. It’s the final frontier of exploration on Earth.

To map the seafloor, oceanographers traditionally relied on “MBES.” For context, mapping the seafloor with “MBES” requires a crewed survey vessel, with daily costs typically ranging from $15,000 to $60,000.

In June 2021, maritime defense company Saildrone launched “Surveyor,” the world’s largest uncrewed ocean mapping vehicle. When sailing from San Francisco, CA, to Honolulu, HI, it mapped about 6,400 sq. mi. of seafloor, costing around $2,500 per day. So, how exactly does “Maritime AI” map the ocean floor?

🚱 A BRIEF HISTORY OF SEAFLOOR MAPPING

⊿ 1ïžâƒŁ One Point at a Time?

Until the 20th century, mapping the seafloor required physical contact. In other words, sailors would lower a lead line until it hit the seafloor. The length of rope let out indicated the depth at a single point.

In 1872, British physician and invertebrate zoologist William Benjamin Carpenter initiated the “HMS Challenger Expedition,” a four-year scientific voyage to systematically explore the ocean’s depths, discovering over 4,000 sea species and mapping seafloor topography, including the Mariana Trench.

⊿ 2ïžâƒŁ The Echo Sounding Discovery?

In the early 1900s, naval engineers pioneered a technique called echo sounding, which sends sound pulses from a ship to the seafloor and records the time it takes for the sound to return to calculate ocean depth.

By the early 1920s, ships began systematically deploying this sonar technology, officially replacing slow, manual lead-line depth measurements. For example, the German Meteor Expedition zigzagged between Africa and South America, collecting roughly 67,000 echo soundings of the South Atlantic, revealing the rugged nature of the seafloor and confirming the continuity of the Mid-Atlantic Ridge.

⊿ 3ïžâƒŁ From Tracks to Maps?

In 1952, American geologist and oceanographic cartographer Marie Tharp produced the first scientific map of the Atlantic Ocean floor. Notably, she identified the V-shaped structure running through the axis of the Mid-Atlantic Ridge and believed it might be a rift valley formed by the oceanic surface being pulled apart.

This notable observation helped confirm the continental drift theory: the now-accepted geological hypothesis that Earth’s continents were once conjoined as a single supercontinent called “Pangea” before breaking apart and drifting to their current positions over nearly 200 million years.

⊿ 4ïžâƒŁ The Rise of Multi-Beam Echo Sounds?

In 1977, the French vessel N.O. Jean Charcot debuted the first commercial “MBES,” nicknamed “Sea Beam,” marking a significant milestone in deep-sea mapping. It emitted 16 sonar beams at once. So, instead of mapping a 1D thin line, it mapped a 2D thick strip.

Estimates indicate that “Sea Beam” generated roughly 1,000 echo soundings/mi., transforming the North Atlantic into a mapped swath of seafloor features like underwater volcanoes, canyons, and trenches, rather than a sparse line of depth points.

🌊 HOW AI HELPS CAPTURE THE SEAFLOOR

⊿ 5ïžâƒŁ AI Navigates the Ocean Autonomously?

The maritime defense company Saildrone engineered “Surveyor” to be an autonomous surface vessel that maps the seafloor at a fraction of the cost of traditional crewed survey ships. While “Maritime AI” isn’t primarily leveraged to measure seafloor depths in this case, it’s still a critical component for navigating the ocean’s changing conditions.

  1. 🔮 Estimating States: Where Am I?

    • In an autonomous surface vessel, the state and motion provide the best estimates of current and future positions. The state and motion of “Surveyor” is determined by leveraging an onboard multi-domain sensor suite that captures noisy measurements like:

      1. 📍 Position: The latitude and longitude.

      2. 📌 Velocity: The change in speed and direction.

    • This multi-domain sensor suite continuously captures this real-world locational data before it’s filtered through an “EKF”: an AI-enabled algorithm that combines multiple noisy measurements into a single running estimate of the autonomous surface vessel’s status.

  2. 🟠 Path Planning: Where Should I Go?

    • “Surveyor” must navigate in carefully spaced lines so the sonar beams properly overlap, despite being steadily affected by sea states like tides and waves or constrained by hazard points like shipping lanes and protected ecological zones. To achieve this, the autonomous surface vessel relies on two key planning and control tools:

      1. 📐 MPC: A control method that repeatedly simulates short-term future trajectories.

      2. 📏 A* search: a planning method that searches through possible routes and identifies promising paths.

  3. 🟡 Collision Avoidance: What’s Around Me?

    • It’s commonplace for commercial ships and fishing boats to inadvertently enter the planned path of the “Surveyor.” To track their positional locations, it relies on:

      1. 💭 AIS: The broadcasted position, speed, heading, and identity information of registered vessels.

      2. 💬 Marine Radar: The detection of surface targets when encountering unregistered vessels.

    • Then, it performs two core computational calculations to estimate whether a commercial ship or fishing boat will become hazardous:

      1. ⏳ CPA: The minimum distance two vessels will reach if they continue on their planned paths.

      2. ⌛ TCPA: The time remaining until those two vessels eventually reach that minimum distance.

🌍 THE REAL-WORLD IMPACT?

⊿ 6ïžâƒŁ Why Should We Care?

In today’s hyper-connected society, it’s easy to forget that the backbone of the global internet isn’t floating in the cloud; it’s anchored to the ocean floor. Over 99% of all international internet and data traffic flows through subsea fiber optic cables. Knowing the undersea terrain directly affects the installation and maintenance of these seafloor fiber networks.

“MBES”-powered mapping enables oceanographers to track underwater landslides, sediment transports, and tectonic deformations. These seafloor events can trigger tsunamis, which pose a threat to roughly 50% of the global population.

On July 17, 1988, a magnitude 7.0 earthquake off the north coast of Papua New Guinea triggered a massive underwater landslide. This generated a catastrophic tsunami, with waves up to 15m high, destroying villages and killing at least 1,600 people.

🔑 KEY TAKEAWAY

Until 2021, mapping the ocean floor required costly expeditions with crewed survey vessels, only to produce sparse measurements of underwater topography. Now, autonomous surface vessels like “Surveyor” can generate millions of high-resolution seafloor maps without putting human lives at risk, bringing us closer to mapping the entire ocean floor by 2030.

📒 FINAL NOTE

FEEDBACK

How would you rate today’s email?

We read every reply you leave after rating!

Login or Subscribe to participate

❀ Today’s Featured Reply

❝

“Waymo data and analysis was interesting.”

-Amit (1ïžâƒŁ 👍 Nailed it!)
REFER & EARN

🎉 Your Friends Learn, You Earn!

{{rp_personalized_text}}

Share your unique referral link: {{rp_refer_url}}