Camera traps have probably changed field ecology more than any other technology of recent decades, and the effects go beyond simply seeing more animals.
What they do
Automated cameras triggered by movement or heat, recording continuously for weeks or months without human presence.
Which removes the observer effect, samples at times humans cannot, and covers ground continuously rather than during visits.
Infrared illumination allows night recording without visible light.
What they revealed
Species presumed extinct in areas where they had not been recorded for decades.
Behaviour that had never been observed, particularly nocturnal behaviour of species thought diurnal.
Species interactions, including predation events and interspecific competition, at frequencies that direct observation could not establish.
And presence of species in habitats where they were not expected, which has revised distribution maps substantially.
Population estimation
Where individuals are identifiable from markings, mark-recapture analysis becomes possible without capture.
Which produced the first reliable estimates for several elusive species with distinctive patterns.
For unmarked species, newer statistical approaches estimate density from detection rates and movement, which is less precise and has expanded the method's reach substantially.
Occupancy modelling
Estimating the proportion of an area occupied rather than absolute numbers.
Which requires repeated surveys at each site to estimate detection probability, and camera traps provide exactly that automatically.
It is frequently more useful than abundance for monitoring trends over large areas.
The data problem
A deployment can produce hundreds of thousands of images, most containing no animal.
Which made manual review the limiting factor for years, with researchers spending more time classifying images than analysing results.
Automated classification using machine learning now handles the bulk, identifying empty images and classifying common species with high accuracy.
Rare species and difficult taxa still require human verification, which is where effort now concentrates.
Volunteer classification
Online platforms distributing images to volunteers for classification.
Which processed enormous datasets before automation improved, and it continues where automation performs poorly.
Multiple independent classifications per image allow confidence assessment, which handles individual error.
Ethical questions
Cameras capture people as well as animals.
Which raises privacy questions, particularly where local communities use the land and have not consented.
It has also raised concerns about cameras being used for surveillance of communities under a conservation framing, which has been documented in some contexts.
Protocols addressing this — consultation, signage, deletion of human images — have been developed and are inconsistently applied.
Practical use
Consumer cameras have become inexpensive enough for amateur use.
Which has produced substantial contributions to recording schemes, particularly for mammals that are otherwise rarely recorded.
Placement matters more than equipment — trails, water sources and natural funnels produce far more than random placement.
Landowner permission is required, and in some jurisdictions there are legal considerations regarding recording in areas where people may be present.
Study design
Placement determines what the data can support.
Cameras on trails detect trail-using species disproportionately, which biases community-level conclusions.
Random or systematic placement produces representative data at the cost of far lower detection rates.
Which means the design must match the question, and studies mixing targeted and random placement cannot answer either well.
Standardisation
Comparing across studies requires consistent methods, and camera models, settings, placement heights and durations have all varied.
Which prompted standardised protocols and metadata standards, allowing datasets to be combined.
Large collaborative networks now operate with common protocols across many sites and countries.
What they cannot do
Detect small or arboreal species reliably, since triggers respond poorly to small warm bodies and cameras face the ground.
Identify individuals of unmarked species without additional methods.
Or record behaviour outside the detection zone, which means absence of a record is not evidence of absence.
Cost and access
Prices have fallen enough that camera traps are accessible to schools, community groups and individual landowners.
Which has widened participation in monitoring substantially, and it has produced records of species presence that professional surveys would not have found.
Data submission to recording schemes turns individual curiosity into usable evidence.
Species discovery
Camera traps have recorded species previously unknown to science and species presumed extinct.
Which is a small number of cases and receives attention out of proportion to frequency.
The more common contribution is establishing presence of known species in areas where they were not recorded, which is less dramatic and more useful.
Long-term deployment
Repeated surveys at the same sites over years produce trend data that single deployments cannot.
Which is where the method's value for monitoring actually lies, and it requires sustained commitment.
Practical placement
Camera height, angle relative to the expected direction of travel, and vegetation clearance in front all affect detection substantially.
Which is learned quickly by reviewing what a deployment actually captured, and it is the difference between a productive camera and an empty one.
Species behaviour at night
A large proportion of mammal activity occurs at night and was therefore poorly documented before automated recording.
Which has revised understanding of activity patterns for many species, and it has revealed shifts toward nocturnality in response to human presence.
That finding — animals becoming more nocturnal where people are active — has been documented across many species and regions.