better nets, more bycatch

Problem statement

Fishing gear has been continuously optimized for catch efficiency: bottom trawls as wide as a football field, longlines stretching for miles with thousands of hooks, and purse seines with Fish Aggregating Devices (FADs) that concentrate target species. These innovations dramatically increased target catch rates — a genuine productivity success. But gear efficiency is indiscriminate. Shrimp trawling catches 2% of the world's total fish catch by weight but produces over one-third of global bycatch. The world average shrimp trawl bycatch ratio is 5.7:1 (6 kg of non-target marine life killed for every 1 kg of shrimp), with some fisheries reaching 20:1. Over 300,000 small whales, dolphins, and porpoises die annually in fishing nets. The mechanism that makes gear effective (sweeping large volumes, attracting many organisms) is inherently non-selective.

Why this matters

Bycatch threatens the viability of marine ecosystems by removing enormous quantities of non-target species, including juveniles of commercially important fish, endangered sea turtles, marine mammals, and seabirds. Sea turtle mortality in trawls escalates from <1% for tows under 10 minutes to 50–100% for tows over 60 minutes. Purse seining with FADs produces 15–20% bycatch rates; longlines can exceed 25%. The problem compounds with overcapacity — more efficient boats catching more bycatch across more of the ocean.

What’s been tried and why it hasn’t worked

Turtle Excluder Devices (TEDs), developed by NMFS starting in 1978 and mandatory for U.S. shrimp trawlers, successfully reduced sea turtle bycatch in those specific fisheries. Circle hooks on longlines reduced sea turtle catch. Bycatch Reduction Devices (BRDs) in trawl nets have shown variable effectiveness. Time-area closures protect specific species during breeding or migration periods. Non-entangling FAD designs reduce ghost fishing. These interventions mitigate individual species impacts but do not solve the fundamental problem: gear optimized for catch volume inherently catches non-target organisms. Each mitigation adds cost and complexity, reducing industry compliance. Developing-country fleets, which account for the majority of global fishing effort, have far lower adoption of BRDs and TEDs.

What would unlock progress

Fundamentally selective fishing methods: precision harvesting technologies (camera-guided trawls that open/close based on species identification), acoustic deterrents for marine mammals, and hook-based gear designs that exploit species-specific feeding behaviors. Economic instruments that internalize bycatch costs (e.g., bycatch quotas that count against catch limits). Electronic monitoring and AI-based catch analysis on vessels to replace human observers (currently covering <1% of fishing effort).

Entry points for student teams

A team could design and test a low-cost camera system with species-classification AI to trigger selective release, and can get the whole loop working without a fishing vessel: train and score the classifier on the free expert-annotated underwater imagery in the FathomNet database (https://fathomnet.org/), then prove the physical half in a flume, tow tank or a rig towed behind a small boat, using scaled or printed fish targets in the pressure housing's real field of view and timing the release actuation. Fishing a real trawl is a separate, later step — it needs charter or partner vessel time and, in US waters, an exempted fishing permit from NOAA Fisheries, access typically held by a fisheries science center, a sea-grant program or a gear-technology group — so a semester team should deliver the bench-validated rig plus the at-sea test protocol rather than at-sea results. Alternatively, a team could develop an economic model comparing total fishery value (including bycatch externalities) under current versus precision-harvesting scenarios, which needs nothing but published landings and discard statistics and a laptop. Marine biology, computer vision, and mechanical engineering skills apply.