The reserve, read through its cameras
Independent capture events, time-of-day activity, species richness, human pressure and individual big-cat counts — computed from MegaDetector + SpeciesNet detections and EXIF capture times across the camera-trap grid.
Survey timeline & effort
This was not a continuous multi-year survey — cameras ran in distinct winter pulses. Weekly wildlife detections (bars) and the number of active cameras (line) make the real deployment structure explicit, which is essential context for every other chart here.
When the cameras were running
Each bar is a week. Read the gaps as "no cameras out", not "no animals".
Wildlife hotspots & richness
Each circle is a trap location. Switch the metric to see where wildlife is most active, where the community is most diverse, and where carnivores concentrate.
This map is zoomable — scroll or pinch to zoom in, and click any point for more information.
Daily activity rhythms
When each species moves. The clock is a 24-hour polar view; the curves compare diel timing across species and against people & livestock.
Activity clock
Share of detections by hour for the selected species
Diel curves
Normalised activity across the day
Temporal overlap
How much two species share the same hours (1 = identical timing). Useful for predator–prey and human–wildlife overlap.
Relative abundance
Raw counts reward cameras that ran longer. These are independent events per 100 trap-nights, so stations and ranges with different effort can be compared fairly.
Detection rate per 100 trap-nights
By forest range
Top species, rate per 100 trap-nights
Species associations
Which species turn up at the same camera stations, and whether big-cat presence tracks wild-prey abundance. The matrix is a Jaccard index (1 = the two species are always found at the same stations, 0 = never together).
Co-occurrence (station overlap)
Share of stations holding both species. Darker = they co-occur more.
Wild prey vs big-cat presence
Each point is a station.
Tiger vs leopard
How the two big cats share the landscape — measured at the species level (detection patterns), independent of individual identity. Do they avoid the same stations, and do they avoid the same hours?
Spatial overlap
Each point is a station: tiger detections vs leopard detections.
Activity overlap
Hour-of-day activity for each cat.
Pressure & conflict risk
Livestock, people and vehicles plotted against wildlife. The conflict index is the human share of all detections at a station — red means human-dominated.
This map is zoomable — scroll or pinch to zoom in, and click any point for more information.
Wildlife vs human pressure
Each point is a station. Up = more wildlife, right = more human/livestock.
Human–wildlife conflict
Large carnivores only — tiger, leopard, sloth bear and striped hyaena — plotted against human/livestock pressure. These are the species that drive real conflict (livestock predation, attacks). Up & right = a station where dangerous predators and people/livestock overlap most. Point size grows with carnivore detections; red = the human share of detections is highest.
Canine distemper risk
Free-ranging domestic dogs are the main reservoir for canine distemper virus (CDV), which spills over into tigers, leopards and other wild carnivores and can be fatal. This map pinpoints where stray dogs were photographed inside the reserve — the finer the hotspot, the more dog detections at that exact camera. Risk is greatest where dogs and big carnivores share a station.
This map is zoomable — scroll or pinch to zoom in, and click any point for more information.
Highest stray-dog stations
Camera stations ranked by dog detections. Amber bars also have large carnivores at the same camera — the CDV spillover front.
Occupancy & detection confidence
How widespread each species is (naïve occupancy = share of active stations where it was detected) set against how confident the SpeciesNet classifier was. Low-confidence, widespread labels are the ones worth a manual second look.
Naïve occupancy
Share of the active stations with at least one detection.
Ubiquity vs model confidence
Up = more widespread. Right = higher mean score. Point size = detections; red = more low-confidence frames.
Unique tigers & leopards
Coat-pattern matching (MiewID embeddings, Ward clustering, left–right flank linking) on the big-cat frames. Decision-support — not a census
This map is zoomable — scroll or pinch to zoom in, and click any point for more information.
Captures per individual
Higher = photographed more often. AB = both flanks confirmed.