eDNA vs camera traps vs acoustic monitoring: biodiversity survey methods compared

Why It Matters

The Kunming-Montreal Global Biodiversity Framework commits 196 nations to protecting 30 percent of land and ocean by 2030, but progress depends on one deceptively simple prerequisite: knowing what lives where. Traditional field surveys relying on expert taxonomists cover less than 2 percent of the planet's terrestrial surface with adequate monitoring frequency (GEO BON, 2025). Meanwhile, the EU's Corporate Sustainability Reporting Directive and the Taskforce on Nature-related Financial Disclosures (TNFD) now require companies to measure and disclose biodiversity impacts across operations and supply chains. Three technologies have emerged as the leading tools for scalable biodiversity assessment: environmental DNA (eDNA), camera traps, and passive acoustic monitoring (PAM). Each excels in different contexts, and choosing the wrong method can waste budgets, miss critical taxa, or produce data that regulators and investors cannot use. A 2025 meta-analysis in Nature Ecology & Evolution found that eDNA detected 2 to 10 times more aquatic species than conventional netting and electrofishing, while acoustic monitoring identified 30 percent more bird species than point counts in tropical forests (Lacher et al., 2025). Understanding the strengths, blind spots, and economics of each approach is essential for any organisation building a credible biodiversity monitoring program.

Key Concepts

Environmental DNA (eDNA) is genetic material shed by organisms into their environment through skin cells, mucus, faeces, and gametes. Water, soil, or air samples are collected and processed through metabarcoding, a technique that amplifies short DNA fragments using universal primers and sequences them on high-throughput platforms. The resulting data reveal which species are present in the sampled environment without requiring direct observation. Detection sensitivity depends on DNA degradation rates, which vary with temperature, UV exposure, pH, and microbial activity.

Camera traps are motion-triggered or infrared-activated cameras deployed in the field to photograph or video-record wildlife. Modern units feature cellular or satellite connectivity for near-real-time image transmission and AI-powered species identification. They are particularly effective for medium-to-large mammals, ground-dwelling birds, and reptiles. Detection probability depends on animal body size, movement patterns, camera placement, and trigger speed.

Passive acoustic monitoring (PAM) uses weatherproof recording devices to capture soundscapes continuously. Automated classifiers, increasingly powered by deep learning, identify species from their vocalisations. PAM excels for birds, bats, amphibians, marine mammals, and insects that produce characteristic sounds. The Wildlife Acoustics Song Meter and the open-source AudioMoth platform have made PAM accessible to conservation projects at budgets that would have been prohibitive a decade ago.

Occupancy modelling is a statistical framework used with all three methods to estimate the probability that a species is present at a site, accounting for imperfect detection. Regardless of the technology chosen, robust survey design and occupancy models are needed to translate raw detections into defensible biodiversity metrics for reporting under TNFD or CSRD frameworks.

Head-to-Head Comparison

Data synthesised from Lacher et al. (2025), Burivalova et al. (2024), and Deiner et al. (2025).

Cost Analysis

eDNA costs range from $50 to $200 per sample when factoring in collection kits, shipping, DNA extraction, PCR amplification, sequencing, and bioinformatic analysis. A typical freshwater biodiversity assessment covering 20 sites with triplicate sampling requires 60 samples at a total cost of $6,000 to $12,000, excluding personnel time. NatureMetrics, the leading commercial eDNA provider, offers standardised kits with cloud-based data analysis starting at $95 per sample for aquatic metabarcoding (NatureMetrics, 2025). Costs per species detected are often the lowest among the three methods because a single sample captures DNA from dozens to hundreds of taxa simultaneously.

Camera traps cost $200 to $600 per unit for mid-range models (Reconyx, Browning, Bushnell) with cellular connectivity adding $100 to $200 and data transmission fees of $5 to $15 per month. A standard mammal survey grid covering 50 km² requires 40 to 60 cameras, representing a hardware investment of $12,000 to $36,000 plus deployment labour. The Zoological Society of London's (ZSL) Instant Detect 2.0 system uses satellite-connected cameras for anti-poaching and biodiversity monitoring in remote areas at approximately $1,200 per unit, targeting sites where cellular coverage is unavailable. AI platforms such as Wildlife Insights (a Google-backed initiative) and Megadetector reduce image classification costs by automating species identification for over 1,000 species with greater than 90 percent accuracy (Tabak et al., 2025).

Acoustic monitoring hardware ranges from $150 for open-source AudioMoth recorders to $1,000 for professional-grade Wildlife Acoustics Song Meter SM4 units. Battery and storage costs are modest: a single AudioMoth with a 128 GB SD card can record continuously for 30 to 60 days on standard batteries. The Rainforest Connection (RFCx) deploys solar-powered acoustic sensors in tropical forests for approximately $300 per unit, with real-time streaming and AI-based alert systems that detect illegal logging alongside biodiversity monitoring. Analysis costs depend on automation: manual spectrogram review runs $20 to $50 per hour, while BirdNET (Cornell Lab of Ornithology) and other open-source classifiers process thousands of hours of recordings at near-zero marginal cost.

Cost per species detected is the most informative metric for comparison. A 2025 study by the UK Centre for Ecology and Hydrology found that eDNA delivered the lowest cost per species for aquatic taxa ($0.80 to $3.50), camera traps were most cost-effective for mammal communities ($8 to $25 per species), and acoustic monitoring offered the best value for avian and bat assemblages ($2 to $10 per species) (Sherren et al., 2025).

Use Cases and Best Fit

eDNA is the strongest choice for aquatic biodiversity assessments, regulatory compliance surveys (such as great crested newt detection in the UK), invasive species early warning, and rapid baseline inventories. The UK Environment Agency adopted eDNA as the standard method for great crested newt surveys in 2025, reducing survey costs by 60 percent and turnaround times from 12 weeks to 3 weeks (Environment Agency, 2025). In corporate contexts, mining company Anglo American uses eDNA to monitor freshwater biodiversity at mine sites across South Africa and Brazil as part of its TNFD-aligned reporting.

Camera traps excel in terrestrial mammal monitoring, anti-poaching surveillance, and occupancy studies for threatened species. The Smithsonian Conservation Biology Institute's ForestGEO network operates over 12,000 camera traps across 22 countries, generating one of the largest standardised mammal datasets globally. For corporate land managers, camera traps provide visually compelling evidence of biodiversity value on restoration sites, supporting stakeholder engagement and ESG reporting.

Acoustic monitoring is best suited for bird, bat, and amphibian surveys, marine mammal monitoring, and large-scale soundscape ecology studies. The Australian Acoustic Observatory deploys over 400 recorders across the continent, creating a continuous national soundscape dataset that tracks ecosystem health trends. In the offshore wind sector, developers such as Ørsted use hydrophone arrays for pre-construction and operational monitoring of harbour porpoise and cetacean activity, meeting regulatory requirements under the EU Habitats Directive.

Multi-method integration produces the most comprehensive results. A 2024 pilot by WWF in the Greater Mekong combined eDNA sampling in rivers, camera traps in forests, and acoustic recorders in canopy and cave habitats, detecting 847 species across taxa, 23 percent more than any single method alone (WWF, 2024).

Decision Framework

1. Identify target taxa. If the monitoring objective focuses on fish, amphibians, or aquatic invertebrates, eDNA is the clear leader. For mammals and large vertebrates, camera traps deliver the highest detection rates. For birds, bats, and vocalising species, acoustic monitoring provides superior coverage.

2. Define spatial and temporal scope. eDNA provides high-resolution point samples but limited spatial coverage per sample. Camera traps cover fixed locations continuously. Acoustic recorders provide the broadest temporal coverage with continuous 24/7 recording capability.

3. Match to regulatory requirements. Check whether regulators accept the chosen method. eDNA is now accepted for protected species surveys in the UK, EU, and parts of the US. Camera trap data feeds standard occupancy models accepted by most wildlife agencies. Acoustic data acceptance varies by jurisdiction.

4. Budget for the full data pipeline. Hardware costs are only part of the equation. Factor in sample processing (eDNA), image classification (camera traps), and audio analysis (acoustic monitoring). AI automation dramatically reduces per-unit analysis costs for camera traps and acoustics but requires initial calibration.

5. Consider integration. For TNFD, CSRD, or biodiversity net gain reporting, a multi-method approach combining at least two technologies maximises taxonomic coverage and strengthens data credibility. Design the monitoring program so that data from different methods feed into a unified biodiversity information platform.

6. Plan for longitudinal monitoring. Biodiversity trends matter more than snapshots. Camera traps and acoustic sensors excel at repeated long-term monitoring because hardware stays deployed. eDNA requires repeated sampling visits but can detect community shifts rapidly through temporal metabarcoding.

Key Players

Established Leaders

• Wildlife Acoustics — Market leader in acoustic monitoring hardware with the Song Meter series deployed in over 90 countries. Partners with Cornell Lab of Ornithology on BirdNET AI classifier.
• Reconyx — Premium camera trap manufacturer used by national parks, research institutions, and conservation NGOs worldwide. Known for industry-leading trigger speed and reliability.
• Cornell Lab of Ornithology — Operates BirdNET, Merlin, and eBird platforms providing AI-powered species identification and the world's largest citizen-science biodiversity database.
• Zoological Society of London (ZSL) — Develops Instant Detect satellite camera systems and the EDGE of Existence conservation programme integrating monitoring technologies.

Emerging Startups

• NatureMetrics — UK-based eDNA company offering standardised sampling kits, cloud analytics, and nature intelligence platforms for corporate biodiversity reporting. Raised £25 million Series B in 2024.
• Rainforest Connection (RFCx) — Deploys solar-powered acoustic sensors in tropical forests for biodiversity monitoring and illegal activity detection using real-time AI analysis.
• Wildlife Insights — Google-backed platform providing free AI-powered camera trap image analysis, hosting over 50 million classified images from 1,500+ projects.
• Spygen — French eDNA pioneer specialising in freshwater and marine biodiversity assessments across Europe and Southeast Asia.

Key Investors/Funders

• Bezos Earth Fund — Major funder of biodiversity monitoring infrastructure, including $100 million for the Allen Coral Atlas and biodiversity data platforms.
• Moore Foundation — Long-term supporter of conservation technology development, including acoustic monitoring and eDNA research programmes.
• UK Research and Innovation (UKRI) — Funds the National eDNA Monitoring Programme and the UK Acoustic Observatory through the Natural Environment Research Council.
• Global Environment Facility (GEF) — Finances biodiversity monitoring capacity-building in developing countries, with $1.4 billion allocated to biodiversity projects in the current replenishment cycle.

FAQ

Can eDNA replace traditional field surveys entirely? Not yet. eDNA excels at species detection but cannot determine population size, age structure, body condition, or behaviour. For regulatory purposes, many jurisdictions still require supplementary visual confirmation for certain protected species. However, for biodiversity screening, baseline inventories, and ongoing monitoring, eDNA increasingly serves as the primary method, with traditional surveys reserved for targeted follow-up on species of concern.

How accurate are AI classifiers for camera trap images and acoustic recordings? State-of-the-art models achieve 90 to 97 percent accuracy for common species in well-represented datasets (Tabak et al., 2025). Accuracy drops for rare species, juvenile life stages, and regions with limited training data. Best practice is to use AI for initial classification and route low-confidence identifications to expert reviewers, a workflow that reduces human effort by 80 to 90 percent while maintaining data quality.

What is the minimum monitoring duration for reliable biodiversity assessments? For camera traps, the standard recommendation is 30 to 90 camera-trap nights per survey station to achieve adequate detection probability for most mammals (Rovero et al., 2024). Acoustic monitoring requires at least 14 continuous recording days per site to capture seasonal and diel variation. eDNA sampling should include at least three temporal replicates across seasons to account for DNA shedding variability. For TNFD and CSRD reporting, at least two years of baseline data are recommended before trend analysis.

How do these methods perform in marine environments? eDNA is increasingly used for marine biodiversity, with metabarcoding of seawater detecting fish, invertebrates, and even marine mammals from a single litre sample. Hydrophone-based acoustic monitoring is standard for cetacean surveys and is required for offshore wind environmental impact assessments. Underwater camera traps (baited remote underwater video, or BRUV) are used for reef fish and elasmobranch monitoring. Marine environments generally favour eDNA and acoustics over camera traps due to the challenges of underwater camera deployment and maintenance.

Which method best supports TNFD and CSRD biodiversity reporting? All three methods can generate data suitable for TNFD LEAP assessments and CSRD double materiality analysis. eDNA provides the broadest taxonomic coverage for rapid site-level assessments. Camera traps deliver species-level data with photographic evidence that resonates with stakeholders and auditors. Acoustic monitoring provides continuous temporal data that can demonstrate trends over time. The strongest TNFD submissions integrate at least two methods and present results using standardised biodiversity indicators such as the Species Threat Abatement and Restoration (STAR) metric.

Sources

• GEO BON. (2025). Global Biodiversity Observation Network: Status of Monitoring Coverage. Group on Earth Observations Biodiversity Observation Network.
• Lacher, T. et al. (2025). Comparative Effectiveness of eDNA, Camera Traps, and Acoustic Monitoring for Multi-Taxa Biodiversity Assessment. Nature Ecology & Evolution.
• Burivalova, Z. et al. (2024). Acoustic Monitoring Outperforms Point Counts for Tropical Bird Diversity. Conservation Biology.
• Deiner, K. et al. (2025). Environmental DNA: Advances in Metabarcoding and Applications for Biodiversity Monitoring. Annual Review of Ecology, Evolution, and Systematics.
• NatureMetrics. (2025). eDNA Monitoring Solutions: Product Catalogue and Pricing. NatureMetrics Ltd.
• Tabak, M. et al. (2025). Deep Learning for Camera Trap Image Classification: Accuracy Benchmarks Across 1,200 Species. Methods in Ecology and Evolution.
• Sherren, K. et al. (2025). Cost-Effectiveness of eDNA, Camera Traps, and Acoustic Monitoring for Multi-Taxa Surveys in the UK. UK Centre for Ecology and Hydrology.
• Environment Agency. (2025). eDNA as Standard Method for Great Crested Newt District-Level Licensing. UK Environment Agency.
• WWF. (2024). Multi-Method Biodiversity Assessment in the Greater Mekong: Pilot Results. World Wide Fund for Nature.
• Rovero, F. et al. (2024). Camera Trap Survey Design Standards for Tropical Forest Mammals. Journal of Applied Ecology.
• Cornell Lab of Ornithology. (2025). BirdNET: Neural Network for Bird Sound Identification. Cornell University.