Busan, Republic of Korea · Open-source citizen science

Finding where river litter stalls, before it reaches the sea

SEA:CUT deploys low-cost GPS drifters with citizens to record where floating litter repeatedly stops in small urban streams, then places floating booms at those chokepoints. A source-to-sea approach descended from The Ocean Cleanup's data-driven placement, TRACKPLAST (Syros, Greece) and the Ganges "Message in a bottle" study.

Honesty notice. SEA:CUT is in its 2026 pilot stage. All demos and downloadable data below are computer simulation outputs, clearly labelled as such. Field data will be published here, under the same schema and licence, as soon as the first deployment is complete.

Problem

Rivers deliver a large share of marine plastic, yet the transport and stalling of litter inside small urban streams is barely observed. Our pilot site, the Goejeong stream (Busan Saha-gu; 5.2 km, watershed 9.6 km²), rises at Mt. Siyak (522.7 m), runs 4.4 km under concrete cover, and discharges through the Hadan sluice into the Nakdong estuary — a site with chronic odour, stagnation and litter problems now under a ~KRW 49bn ecological restoration.

Approach

Drift, don't guess: bottle-shaped drifters matched to real litter buoyancy record where debris actually stalls. Nobody enters the water; deployment and recovery are from banks and bridges, with a 100% recovery principle.
Site-specific accumulation model: repeated releases feed a spatial clustering analysis (server-side) that finds recurrent stagnation hotspots and their response to rainfall. We deliberately avoid overclaiming — the intelligence is in the ensemble analysis, not on the device.
Act on chokepoints: floating booms at model-identified narrows intercept litter before the estuary.

Data availability SIMULATED CC BY-NC 4.0

Current sample dataset (Lagrangian Monte-Carlo, seed 42, reproducible; not field data):

Trajectories CSV (1,496 rows) Hotspots GeoJSON

Field	Description
drifter_id / rainfall_mm	Experiment-unit ID · rainfall condition (15/45/80 mm)
t_min, lat, lon, s_km	Elapsed minutes · WGS84 position · along-stream distance
state	run (drifting) · beached (stalled) · sea (reached estuary)

Cite as: Ita Seoul SEA:CUT (2026). Goejeong stream drifter simulation sample dataset. CC BY-NC 4.0. Simulated output, not field observations. Field dataset with DOI to follow the 2026 pilot.

Open source

Firmware (ESP32 · LTE Cat.1), drift models (Python), web demos and the participant PWA are open at github.com/Itaseoul/seacut. Code MIT · data CC BY-NC.

Who we are

Ita Seoul is a Korean environmental-welfare nonprofit (est. 2018, reg. 795-82-00235, Seoul; Minister of Environment commendation 2022). We operate the Adopt-a-Beach platform caresea.kr (86 beaches, 7,002 participants, 150k+ records), the Global Adopt-a-Beach Alliance (GAA), and are a registered member of the UNEP Global Partnership on Plastic Pollution and Marine Litter (GPML). SEA:CUT aligns with SDG 14.1 and the source-to-sea framework, looking toward the 2028 UN Ocean Conference in Busan.

Collaborate

We welcome research collaboration (methods, drifter validation, estuary modelling), institutional partnerships and funding conversations.

Contact: [email protected]