| images | ||
| README.md | ||
GeoTrash
Citizen science platform for marine litter monitoring
Overview
The Problem
Every year, hundreds of thousands of tonnes of plastic and other waste enter the Mediterranean Sea. The accumulation is not uniform — it concentrates along coastlines, shipping lanes, and fishing grounds in patterns that are difficult to map without systematic, continuous observation across large areas. In the Adriatic-Ionian region, this problem is particularly acute. Ghost fishing gear — nets, lines, and traps lost or abandoned at sea — continues to trap and kill marine life long after it has left human hands. Plastic fragments degrade into microplastics that enter the food chain. Coastal habitats are damaged and tourism is affected. The economic and ecological consequences compound over time.
The challenge is not just the scale of the pollution. It is the invisibility of it. Marine litter at sea is not concentrated in one place that can be photographed from a satellite. It is distributed across thousands of square kilometres of open water, appearing and disappearing with currents and tides, spotted intermittently by the people who happen to be out there — fishermen, divers, coastal workers — who have no systematic way to record and share what they see.
The Gap
Scientific monitoring of marine litter traditionally relies on standardised surveys conducted by research vessels at fixed time intervals. This approach produces rigorous data, but it is slow, expensive, and spatially sparse. A research vessel can only be in one place at a time. A survey conducted twice a year produces a snapshot, not a picture of how pollution concentrations are evolving day to day across an entire coastline.
Meanwhile, the fishing fleet is already out there, every day, across the entire area. Fishermen encounter marine litter constantly — plastic bags tangled in nets, ghost gear on the seabed, floating debris along routes they know intimately. This knowledge has historically been invisible to science. Fishermen had no practical way to report what they observed, and scientists had no infrastructure to receive and aggregate that information at scale. The data existed — in the daily experience of the fleet — but it was never captured.
The Solution
GeoTrash is a mobile citizen science platform that turns the fishing fleet into a distributed monitoring network for marine litter in the Adriatic-Ionian region. Fishermen and other community members use a mobile app to report litter encounters in real time: photographing the waste, logging its GPS coordinates, classifying it as refuse or ghost gear, and submitting the record to a shared database. The data accumulates into a continuously updated georeferenced archive that scientists, managers, and policy makers can interrogate to understand where litter concentrates, how distribution changes over time, and which areas require priority intervention.
The platform was deployed along the northern Apulian coast between Vieste and Lecce, involving dozens of fishing vessels across multiple harbours. In a single monitoring campaign, the fleet generated georeferenced litter observations — a density of coverage that no conventional research survey could have achieved at comparable cost. The data revealed clear spatial concentration patterns, identified hotspot areas, and demonstrated that the shipping lane is a significant corridor for open-sea litter accumulation.
Figure 1 — The citizen science approach. This illustration captures the spirit of the GeoTrash project: bringing scientific observation methods directly into the field, where the pollution actually exists. Rather than waiting for scheduled research cruises, the platform mobilises the people who are already on and in the water — fishermen, divers, coastal workers — and gives them the tools to contribute structured, scientifically usable data. Every observation they submit is a data point that would otherwise never exist.
Key Features
| Feature | Description |
|---|---|
| Mobile litter reporting | Fishermen and community members log litter encounters with photo, GPS coordinates, date, and type classification directly from the field |
| Dual classification | Each sighting classified as either general refuse or ghost fishing gear, enabling separate analysis of two distinct pollution pathways |
| GPS auto-capture | Device location captured automatically at the moment of reporting; coordinates can be manually adjusted for deferred submissions |
| Photo attachment | Supports upload from gallery, live camera capture, or PDF attachment for each observation |
| Draft and submit workflow | Observations can be saved as draft and submitted later, supporting connectivity-limited offshore environments |
| Upload history and editing | All past submissions viewable, editable, and deletable by the reporting user |
| Admin data portal | Web interface for administrators to view all submissions across all users, download the full dataset, or bulk-upload historical records |
| Analytics dashboard | Campaign-level statistics including total observations, time series of reporting activity, breakdown by type, breakdown by vessel, and a live sighting map |
| Bulk data import | Historical paper-based observation records can be digitised and ingested in bulk via a standardised Excel template |
| Open data export | Full dataset downloadable as structured file for external GIS and statistical analysis |
Methodology
- Community engagement — Fishermen and local maritime operators in the Adriatic-Ionian region recruited as active participants; onboarding included explanation of the scientific purpose, data collection protocols, and best practices for observation and photo documentation.
- Mobile app development — Cross-platform mobile application built for Android with a simple, low-friction interface designed for use on deck in variable conditions; navigation bar provides access to all core functions from any screen.
- Data collection protocol — Each litter encounter logged with a minimum required dataset: type of finding (Rifiuto / Attrezzo), date, GPS coordinates of the finding, GPS coordinates at time of upload, photo evidence, and optional free-text comment; a unique auto-generated Ticket ID is assigned to each submission.
- Real-time submission and draft mode — Observations submitted immediately when connectivity allows, or saved as drafts for submission from harbour; this accommodates the practical reality of offshore work where mobile data is intermittent.
- Administrative backend — Web-based admin panel aggregates all submissions from all users; administrators can review the full submission log, download the georeferenced dataset, and upload bulk historical records using the standardised Excel template.
- Analytics and visualisation — Campaign analytics module computes total observation counts, daily time series, per-user and per-vessel contribution breakdowns, and a georeferenced point map; data also exported to R for spatial analysis including hexagonal grid aggregation (10 nautical mile radius) and kernel density estimation for hotspot mapping.
- Spatial analysis — Observations aggregated at two scales: a regional overview showing distribution along the Apulian coast, and fine-scale kernel density maps for individual coastal localities with exceptional concentrations; the Bari-Corfù shipping corridor identified as a significant open-sea litter pathway.
- Results communication — Findings communicated back to participating fishermen and to institutional stakeholders, closing the feedback loop and demonstrating the value of their contribution to the scientific process.
Screenshots
Figure 2 — Home screen. The GeoTrash app opens on a welcome screen that briefly explains the project's purpose and invites users to contribute. The teal-and-blue navigation bar at the bottom provides access to all four core functions: refresh/sync, information, reporting form, submission history, and settings. The home view also displays contextual imagery — including photos of litter on beaches and screenshots of the app in use — reinforcing the environmental mission and giving new users an immediate sense of what the platform does. The design is intentionally minimal: the target users are fishermen using this on a moving boat, and every unnecessary element is a potential source of friction.
Figure 3 — The reporting act. This image illustrates the core user action that powers GeoTrash: a person encountering marine litter and immediately reaching for their phone to document it. The simplicity of this gesture — photograph, classify, submit — is the design principle behind the entire platform. The lower the barrier to reporting, the more observations get captured. GeoTrash reduces that barrier to a few taps, turning an everyday encounter with pollution into a scientifically usable georeferenced data point.
Figure 4 — Field report form. This is the primary data entry screen, titled Rapporto Rinvenimenti (Finding Report). The auto-generated Ticket ID at the top ensures every submission is uniquely traceable. The user selects the type of finding from a dropdown — either Rifiuto (general waste) or Attrezzo (fishing gear) — then attaches a photo using one of three options: upload from the device gallery, take a new photo with the camera, or attach a PDF document. The date selector defaults to today but can be changed for deferred reporting. Below the date, the GPS Ritrovamento (finding position) fields auto-populate with the device's current location; these can be manually overridden or adjusted using the embedded interactive map, which is essential when a fisherman reports a sighting from memory after returning to harbour. A second GPS field records the Caricamento (upload location), allowing the system to distinguish where the litter was found from where the report was submitted. The form closes with Submit and Save Draft buttons.
Figure 5 — Submission archive. This screen gives users a complete history of all their past submissions. Each entry in the list shows the Ticket ID, type of finding, photo filename, observation date, GPS finding coordinates, GPS upload coordinates, upload date, and submission status (shown as a green SUBMITTED badge). Tapping any entry opens the full detail view, where the user can review, edit, update, or delete the record. On the right side of the screen, the detail panel shows the current photo attached to the selected submission, the date, and the GPS coordinates plotted on an interactive map — in this case showing a location in the central Adriatic. This edit capability is important for data quality: it allows users to correct coordinates entered in a hurry at sea once they are back onshore with more time and connectivity.
Figure 6 — Real data from the field. These four photographs were submitted by participating fishing vessels during the GeoTrash monitoring campaign along the northern Apulian coast. They show the reality of what the fleet encounters: plastic bags tangled in nets, floating plastic sheeting on the sea surface, mixed plastic and organic debris piled on deck after retrieval, and weathered packaging waste. These images are not illustrations — they are actual data points in the GeoTrash database, each georeferenced and timestamped, contributing to the 766 observations logged during the campaign. They document the predominance of plastic materials in the litter mix and illustrate why the fishing fleet, which physically encounters and retrieves this material, is uniquely positioned as a monitoring network.
Figure 7 — Hotspot map This kernel density map shows the spatial concentration of litter observations in the coastal area. The colour scale runs from dark purple (low density) through green to yellow and white (highest density), with each coloured zone representing an area of statistically significant accumulation based on the underlying point observations. The two distinct high-density clusters visible in the map indicate that litter is not uniformly distributed along the coast but concentrates in specific bays and coves — likely driven by local current patterns and coastal morphology. This kind of fine-scale spatial information, produced directly from fishermen's field reports, is exactly the data that managers need to prioritise cleanup operations and preventive monitoring in specific localities.
Figure 8 — Hotspot map. A second locality-scale kernel density map. The dense accumulation pattern visible here — a compact, high-intensity cluster close inshore — contrasts with the more diffuse distribution seen in other parts of the study area, suggesting a localised source or retention mechanism specific to this coastal segment. The hexagonal grid and kernel density outputs together provide both a regional overview and the site-specific resolution needed for targeted action.
Figure 9 — Data collection in context. This image places the GeoTrash mobile interface directly within the environment it was designed to monitor — a beach strewn with plastic waste. It illustrates the central design challenge of the platform: building a data collection tool that is useful not in a lab or office, but in the messy, uncomfortable, often urgent context of actually encountering pollution. The app had to work on a moving fishing boat, with gloved hands, in low connectivity, with a few seconds of attention available. Every design decision — the simple form, the auto-GPS, the draft mode, the minimal required fields — was driven by that operational reality.
Results
The GeoTrash monitoring campaign was conducted along the northern Apulian coast across multiple harbours in the region. Over the campaign period, the platform collected georeferenced marine litter observations, with a mean reporting rate of approximately 10 sightings per day and peaks exceeding 30 daily submissions during the most active periods.
Observations were predominantly composed of plastic materials, confirming the dominance of plastic in the Mediterranean marine litter mix. Spatial analysis revealed a strong concentration along the nearshore zone, with the highest accumulation densities recorded. Fine-scale kernel density mapping identified specific hotspot localities where litter concentrations were exceptionally high relative to surrounding areas.
In open water, observations showed a significant linear pattern parallel to the coast along the maritime shipping corridor, with sporadic sightings extending to the territorial waters boundary. This suggests that commercial shipping activity contributes to open-sea litter distribution independently of coastal sources.
Vessel-level analysis revealed considerable variation in reporting rates: the most active vessels each submitted over 45 observations, while a smaller number of vessels contributed fewer than 5. This variation reflects both differences in fishing ground location and in individual engagement with the platform — findings that will inform future training and incentive design for community monitoring programmes.
Scientific Context
GeoTrash was developed in the framework of the EU Common Fisheries Policy (CFP) pillars of Blue Growth and Environmental Quality, which jointly mandate the reduction of marine pollution, the management of ghost fishing gear, and the involvement of fishing communities in sustainable marine governance. The project directly addresses the gap between the ambition of marine litter monitoring under the Marine Strategy Framework Directive (MSFD) and the practical impossibility of achieving adequate spatial and temporal coverage through conventional survey methods alone.
The citizen science approach adopted by GeoTrash has been shown across multiple domains to significantly expand the spatial and temporal scope of environmental monitoring at a fraction of the cost of equivalent professional survey coverage. By treating the fishing fleet as a distributed sensor network — already present across the monitoring area, already encountering the phenomenon of interest — the platform transforms existing human activity into scientific infrastructure. The result is a monitoring dataset that is not only spatially dense but temporally continuous, capturing the dynamic variability of litter distribution that snapshot surveys cannot detect.
Stack
Android Mobile GPS PostgreSQL R GIS Leaflet Citizen Science