Ecology depends on data. How many tigers live in a reserve? Where do elephants go at night? Are bird populations declining? Answering these questions requires careful fieldwork, and the ecologist’s toolkit has changed dramatically in the last 20 years. This module introduces the main methods, their strengths and weaknesses, and the basics of designing a sound study.
What you will learn
- Classic field methods for plants and animals
- Modern technologies: camera traps, telemetry, acoustic monitoring, eDNA, drones and satellites
- How artificial intelligence is changing data analysis
- The basics of study design and statistics
- How citizen science contributes to research
Classic field methods
Quadrats and transects
A quadrat is a square frame, often 1 m × 1 m, placed on the ground to count plants or slow-moving animals. Placing quadrats randomly or along a line, called a transect, allows ecologists to estimate density, cover and how communities change across gradients, such as from a shore to a cliff top.
Mark–recapture
To estimate the size of a mobile population, ecologists capture a sample, mark the individuals, release them and later capture a second sample. The proportion of marked animals in the second sample indicates the total population. The simplest form is the Lincoln–Petersen estimate:
N = (M × C) / R
where M is the number marked in the first sample, C is the number caught in the second sample and R is the number of marked individuals recaptured. For example, if 50 fish are marked, and a later catch of 60 includes 10 marked fish, the estimated population is (50 × 60) / 10 = 300. The method assumes marks are not lost, marked animals mix freely, and the population is closed (no births, deaths or movement) between samples.
Distance sampling and point counts
Observers walk transects or stand at fixed points and record each animal seen or heard and its distance. Because animals further away are harder to detect, statistical methods such as distance sampling correct for missed animals. Point counts are the standard method for bird surveys.
Direct observation and behaviour sampling
Behavioural ecologists record what individual animals do, using methods such as focal animal sampling (following one individual) and scan sampling (recording the behaviour of a whole group at set intervals). Long-term observation of known individuals, such as Jane Goodall’s work with chimpanzees at Gombe since 1960, has transformed our understanding of animal behaviour.
Modern technologies
Camera traps
Motion-triggered cameras photograph animals day and night without disturbing them. Because tigers, leopards, jaguars and snow leopards have unique coat patterns, individuals can be identified, and spatially explicit capture–recapture models estimate density. India’s national tiger survey uses tens of thousands of camera traps. Camera traps have also revealed species thought to be extinct in some areas and recorded rarely seen behaviour.
Telemetry: VHF, GPS and satellite tracking
- VHF radio collars emit signals that researchers track with an antenna.
- GPS collars and tags record locations at set intervals and send them via mobile networks or satellites, revealing home ranges, migration routes and habitat use.
- Tiny tags now allow tracking of animals as small as songbirds and even large insects. The international ICARUS project has tested tracking small animals from space.
Tracking has revealed, for example, that bar-tailed godwits fly more than 11,000 km nonstop from Alaska to New Zealand, and that elephants use narrow corridors between protected areas at night.
Acoustic monitoring
Autonomous recorders capture sounds for days or months. They are used to monitor birds, frogs, bats (with ultrasonic detectors), whales (with underwater hydrophones) and even illegal logging or gunshots in rainforests. Machine-learning tools such as BirdNET can identify thousands of bird species by their calls.
Environmental DNA (eDNA)
Every organism sheds DNA into its environment through skin, mucus, faeces and hair. By filtering water or collecting soil or air, ecologists can detect which species are present without seeing them. eDNA has been used to detect great crested newts in ponds, invasive carp in the Great Lakes, and entire fish communities in rivers and seas from a single water sample. Airborne eDNA can even reveal the mammals living around a zoo or forest.
Drones and satellites
- Drones survey wildlife in open habitats, count seabird and seal colonies, map habitats in high detail and, with thermal cameras, detect animals at night.
- Satellite imagery tracks deforestation (for example through Global Forest Watch), maps habitats, monitors coral reef bleaching and even counts large animals such as whales and elephants in very high-resolution images. Emperor penguin colonies have been discovered from space by spotting guano stains on the ice.
Artificial intelligence
Modern monitoring produces vast amounts of data: millions of camera-trap images and thousands of hours of audio. AI models now sort images, identify species and individuals, and recognise calls, reducing work that once took months to hours. Tools such as Wildlife Insights and MegaDetector process camera-trap images, while photo-identification systems can recognise individual whales from their tail flukes or giraffes from their coat patterns.
Designing a good study
Good technology cannot rescue a poorly designed study. Key principles include:
- A clear question and hypothesis. For example: “Do tigers avoid areas near villages?”
- Replication. Repeat measurements at multiple sites to separate real patterns from chance.
- Controls. Compare treated and untreated areas, such as fenced and unfenced plots.
- Randomisation. Place samples randomly to avoid bias.
- Detection probability. Not finding an animal does not mean it is absent; methods such as occupancy modelling estimate the probability of detecting a species.
- Before–After–Control–Impact (BACI) designs, which compare sites before and after a change against control sites, are the gold standard for measuring the effect of interventions such as a new road or a restoration project.
A little statistics
- Mean, median and standard deviation summarise data.
- Confidence intervals show the uncertainty around an estimate. A tiger estimate of 355 might have a 95 percent confidence interval of, say, 320 to 400.
- Hypothesis tests and p-values assess whether differences are larger than expected by chance, but statistical significance is not the same as ecological importance.
- Regression and generalised linear models relate a response, such as species richness, to explanatory variables, such as rainfall or area.
- Free software such as R and QGIS is widely used in ecology.
Ethics and permits
Fieldwork must minimise harm to animals and habitats. Capturing, marking and collaring animals requires ethical approval, trained staff and research permits from the relevant authorities. Researchers also have responsibilities to local communities, including consent, fair collaboration and sharing results, and to keep the locations of highly poached species confidential.
Citizen science
Anyone can contribute to ecological research:
- eBird: over a billion bird observations from around the world, used in hundreds of scientific studies.
- iNaturalist: photograph any organism and have it identified by the community; research-grade records flow into GBIF.
- Zooniverse: classify camera-trap images and other data online from home.
- Local bird atlases, butterfly counts and bioblitzes.
Travellers can contribute too: uploading wildlife sightings with dates and locations from safaris and treks adds valuable data from places researchers rarely visit.
Case study: counting tigers in India
India’s All India Tiger Estimation, carried out every four years, is one of the largest wildlife surveys in the world. The 2022 cycle combined foot surveys of signs across hundreds of thousands of kilometres of forest trails with over 30,000 camera traps, which captured tens of millions of images. Software matched individual tigers by their stripes, and statistical models estimated tiger numbers and density across the country, including areas without cameras. The result, a mean estimate of about 3,680 tigers, is used to plan protected areas and corridors.
Try it yourself
Download iNaturalist or eBird and record ten species near your home this week. Or try a simple mark–recapture experiment with dried beans in a bag: mark a handful, mix them back in, draw a second handful and use the Lincoln–Petersen formula to estimate the total. Then count them all to check your estimate.
Common misconceptions
- “If we didn’t see it, it isn’t there.” Imperfect detection is one of the biggest challenges in ecology.
- “More data always means better answers.” Biased or poorly designed data can mislead, however large the dataset.
- “Technology replaces fieldwork.” Technology extends fieldwork but still needs ground-truthing and expert knowledge.
Key terms
- Quadrat and transect: standard units for sampling plants and sedentary animals.
- Mark–recapture: a method of estimating population size from marked individuals.
- Telemetry: remote tracking of animal locations.
- eDNA: DNA collected from the environment to detect species.
- Occupancy modelling: estimating presence while accounting for imperfect detection.
- BACI design: Before–After–Control–Impact study design.
Quick quiz
Test what you have learned. Scoring 60 percent or more marks this module as complete.
Further reading
- Sutherland, W. J. (ed.). Ecological Census Techniques: A Handbook.
- Tuia, D. et al. (2022). “Perspectives in machine learning for wildlife conservation.” Nature Communications.