Wildlife-trafficking transit hub and holding-site monitoring
VHR optical satellites can detect vehicle patterns, temporary enclosures and compound changes at suspected wildlife-trafficking transit sites. Imagery is a cueing tool, not a conviction: ground investigation must follow.
Sensors
- Maxar WorldView-3: 31 cm panchromatic, 1.24 m multispectral (8 bands including SWIR). The sharpest commercially available archive for resolving individual enclosure structures, container positions and vehicle types. Tasking revisit at mid-latitudes is typically 1 to 4.5 days depending on off-nadir tolerance.
- Airbus Pléiades Neo: 30 cm panchromatic, 1.2 m multispectral (6 bands). Stereo and tri-stereo collection allows 3-D compound modelling; useful for distinguishing roofed pens from open enclosures. Revisit 1 to 2 days for a single satellite; the two-satellite constellation improves this.
- Planet SkySat: 50 cm panchromatic, ~1 m multispectral. Supports video collection (up to 90 seconds per pass) at targeted sites, which can reveal vehicle movement patterns within a single overpass. Tasking latency typically under 24 hours.
- Planet SuperDove (PlanetScope): 3 m multispectral, 8 bands, near-daily global coverage. Too coarse to resolve individual pens but useful for change detection across compound footprints over weeks and months, and for flagging new access-track formation.
What a compound gives away without showing you a single animal
Illegal wildlife trade does not require a photograph of a pangolin to leave a satellite signature. Transit hubs have operational rhythms: vehicles arrive and depart in patterns inconsistent with the stated land use, containers shift position between tasking dates, temporary structures appear and are dismantled within days. At 30 to 50 cm resolution, a WorldView-3 or Pléiades Neo image can resolve the footprint of a wooden pen, the shadow cast by a raised cage frame, or the tyre-track density on an access road that suggests heavier traffic than a rural smallholding would normally sustain.
The analytic logic is comparative rather than instantaneous. A single image of a compound is almost always ambiguous. A time-series of six or eight images over three months, cross-referenced against published UNODC or TRAFFIC route maps and NGO-identified facility lists, begins to show whether activity is consistent with the suspected function. That is the honest framing: satellite imagery narrows the list of sites worth investigating on the ground. It does not replace that investigation.
Resolution floors and what they mean in practice
WorldView-3's 31 cm panchromatic band can resolve objects roughly 60 to 90 cm across in practice, accounting for atmospheric effects and off-nadir collection angles. That is sufficient to distinguish a vehicle type, count individual containers, and detect the shadow geometry of a cage or pen structure taller than roughly one metre. It is not sufficient to identify species, read markings on containers, or confirm that an enclosure is occupied.
Pléiades Neo stereo pairs add a dimension that flat imagery cannot: height. A roofed holding structure 2.5 metres tall produces a measurable shadow and, in stereo, a calculable elevation. Open-air pens with wire mesh produce no roof shadow but may show a characteristic rectangular ground pattern. The distinction matters because traffickers frequently use existing agricultural or aquaculture infrastructure as cover, and the structural geometry of a fish-farm pen differs from that of a mammal holding facility in ways that are sometimes, not always, legible from orbit.
Cloud cover is a persistent problem across the tropical and subtropical regions where most transit hubs operate. A site in southern Yunnan, northern Laos or coastal West Africa may be obscured for weeks at a time during monsoon season. Tasking strategies need to account for this by maintaining a queue of collection requests that execute on the first clear window, rather than scheduling single-date collections.
The change-detection workflow: what triggers an alert
The practical workflow starts with a baseline image of a known or suspected site, drawn from commercial archive or freshly tasked. Subsequent images are co-registered to sub-pixel accuracy and differenced. Analysts flag: new structures within the compound boundary; containers that have moved more than their own length between dates; vehicles present at unusual hours if SkySat video is available; and changes in the perimeter fence line or access-track condition.
Automated change detection using pixel-level differencing or object-based image analysis can reduce analyst time on large site lists. Published methods in the remote-sensing literature, including those validated on port and industrial compound monitoring, achieve reliable change flags at object scales above roughly 2 to 3 metres in VHR imagery. Below that threshold, false positives from shadow shift and atmospheric variation increase sharply. Any automated flag should be reviewed by a human analyst before being passed to an enforcement partner.
Ports and cold-storage facilities: a different problem
Rural holding sites are one part of the trafficking chain. Commercial ports and cold-storage compounds present a different challenge. At a busy container port, the signal of a single illicit consignment is invisible in the aggregate movement of thousands of boxes. Satellite imagery at port facilities is most useful not for finding a specific container but for tracking the dwell-time behaviour of vessels and vehicles associated with known suspicious operators, cross-referenced against AIS records and published watchlists.
Cold-storage compounds near ports, identified in UNODC and NGO reports as staging points for ivory, shark fin and frozen bushmeat, are more tractable. They are typically smaller, have limited legitimate traffic, and show distinctive refrigeration unit signatures in thermal infrared. WorldView-3 carries a thermal infrared band (TIRS, at 3.7 m resolution) alongside its optical bands, which can indicate active refrigeration at a facility. This is indicative, not diagnostic: any cold-storage business produces the same signature.
Honest limits and the cueing-tool doctrine
No satellite sensor can confirm species presence. No image analysis can distinguish a legal exotic-animal farm from an illegal holding site on visual evidence alone. The ambiguity is irreducible at the sensor level. What imagery can do is identify sites whose activity patterns are anomalous relative to their stated or apparent land use, and rank those sites by the degree of anomaly for prioritisation by enforcement agencies or NGO field teams.
Archive depth is a genuine asset here. WorldView and Pléiades archives extend back to 2014 and 2012 respectively for many regions. A site that appears unremarkable today may show a clear operational peak in 2018 or 2020 that corresponds to a known trafficking event documented in open sources. Retrospective analysis of this kind has been used in published academic work on conflict-mineral and deforestation monitoring; the same logic applies to trafficking infrastructure.
Satellize structures this kind of work as a cueing service: periodic change-detection reports on a client-defined site list, with anomaly scores and annotated imagery, delivered as GIS layers and PDF summaries to enforcement liaison teams. The Tonga crop-estimation programme uses a comparable periodic-report model on a different subject, but the analytic architecture transfers directly to site-monitoring tasks.
What a credible monitoring programme actually requires
A site list grounded in published sources is the starting point. UNODC's World Wildlife Crime Reports, TRAFFIC's published route analyses and WWF's country-level assessments name regions, border crossings and facility types without always naming specific coordinates. Cross-referencing these with open-source geolocation, news reporting and NGO field data produces a prioritised set of candidate sites suitable for tasking.
Tasking frequency is a budget question as much as an analytic one. A monthly collection cadence on a list of twenty sites using Pléiades Neo or WorldView-3 is feasible for a national enforcement agency or a well-funded conservation NGO. Weekly cadence on a smaller priority list, supplemented by near-daily SuperDove coverage for coarse change flags, is the architecture most likely to catch the short-duration events, temporary enclosures erected and removed within days, that characterise active transit operations. The honest answer is that no satellite programme catches everything; the goal is to make the operational cost of concealment higher for traffickers, not to eliminate it.
Typical figures
| Best available spatial resolution (optical) | 30 to 31 cm panchromatic (Pléiades Neo, WorldView-3); 50 cm (SkySat); 3 m (SuperDove) |
| Multispectral bands | 8 bands including SWIR (WorldView-3); 6 bands (Pléiades Neo); 8 bands (SuperDove); 4 bands (SkySat) |
| Thermal infrared | WorldView-3 TIRS at 3.7 m resolution; indicative of active refrigeration, not diagnostic |
| Revisit at target site | 1 to 4.5 days (WorldView-3, off-nadir); 1 to 2 days (Pléiades Neo constellation); near-daily (SuperDove) |
| Minimum detectable structure | Approximately 60 to 90 cm objects in practice at 31 cm GSD; pen or cage frames above ~1 m height detectable by shadow |
| Tasking latency | Typically under 24 hours for SkySat; 24 to 72 hours for Pléiades Neo and WorldView-3 priority tasking |
| Archive depth | WorldView series from 2007; Pléiades from 2012; SkySat from 2014; SuperDove from 2017 |
| Cloud-cover constraint | Significant in tropical regions; monsoon seasons may produce weeks of unusable optical imagery; no radar substitute at VHR for structural detail |
| Delivery formats | GeoTIFF (orthorectified), NITF, GIS vector layers (GeoJSON/Shapefile), annotated PDF reports |
Analytics Satellize can run
| Compound activity anomaly score | Object-based image analysis (OBIA) on co-registered VHR time-series; vehicle and container count differencing between dates | Ranked site list with anomaly scores and change thumbnails, delivered as GIS layer and PDF summary |
| Temporary structure detection | Shadow-geometry analysis and pixel-differencing on sequential Pléiades Neo or WorldView-3 images; flags structures present in one date and absent in another | Annotated GeoTIFF with bounding boxes on new or removed structures; change log per site |
| Access-track condition and traffic intensity index | Tyre-track density mapping from VHR panchromatic; SuperDove time-series for coarse road-surface change over months | Per-site traffic intensity rating updated at each collection; trend chart in periodic report |
| Refrigeration-unit thermal flag | WorldView-3 TIRS band analysis for elevated thermal signature at cold-storage facilities; comparison against ambient temperature baseline | Binary flag (active/inactive) per facility per collection date, included in site report |
| Stereo-derived compound height model | Pléiades Neo stereo or tri-stereo processing to generate DSM; structure height extraction for distinguishing roofed pens from open enclosures | GeoTIFF DSM and vector layer of structure heights for target compound |
| Retrospective baseline report | Archive trawl across WorldView and Pléiades holdings for a named site list; manual and semi-automated change annotation across available dates | PDF timeline report per site showing compound evolution with annotated imagery, cross-referenced to open-source event dates |
Who does the work
We can get this done for you. Satellize runs its own analyst desk and a strong science team. You do not buy a data feed and work out what it means; our people source the imagery, run the analysis described on this page, and hand you the answer with its confidence limits stated. Discuss this requirement.