Table 1.

Methodological barriers that currently hinder terrestrial biodiversity monitoring and the opportunities and challenges that RAS offer in overcoming each of these barriers

Barrier category	Barrier	RAS opportunity or challenge	Brief description
1. Site access	Surveying over large spatial scales	Opportunity	Autonomous monitoring at landscape scales Replicating surveys at multiple sites over large geographical areas
	Surveying remote areas far from infrastructure	Opportunity	Accessing locations remote from roads and other infrastructure Monitoring sites that are time-consuming to access
	Surveying hazardous or inaccessible sites	Opportunity	Access to sites that need climbing (for instance, cliffs or forest canopies) Sampling sites at night or where personal safety or security is at risk
	Surveying taxa at random sites	Opportunity	Enabling representative sampling at suitable scale and stratification Avoiding sample pseudoreplication
	Surveying multiple locations simultaneously	Opportunity	Time-synchronous surveys at multiple sites Surveying taxa whose activity may be weather-dependent
	Surveying structurally complex habitats	Opportunity	Sampling within dense habitats (for instance, deadwood, grass tussocks or snow) Sampling soils, underground animal burrows, or bat colonies in caves or trees
	Surveying at high spatial resolution	Opportunity	Ability of sensor to get to exact locations repeatedly Enabling microscale tracking
	Designing environmentally robust sensors	Challenge	Resistance, resilience and durability of the sensors and/or probes in the field Being species-proof and avoiding risk of vandalism or theft
	Surveying restricted and off-limits locations	Challenge	Areas affected by legal, conflict and political issues Uncertainty of tenure or ownership status for many locations
2. Species and/or individual detection	Eliminating the need for multiple sensors	Challenge	Integration of multiple sensor types Ablility to deal with wide range of species sizes
	Discriminating or identifying individuals at distance	Challenge	Distance limitations of visual sensors (for instance, detection of plant ligules) Difficulties in identifying individuals of a species
	Surveying without disturbing taxa or habitats	Challenge	Non-invasive sensors that will not disturb species or habitats Impacts on non-target species
	Surveying through objects or in low light levels	Challenge	Detection when visibility is restricted (for instance, through vegetation or cloud) Detection of ectotherms at night
	Surveying ecological processes	Challenge	Monitoring interactions (for instance, pollination) or ecological processes Monitoring plant physiology
3. Data handling and processing	Handling high data volumes	Opportunity	Storage, energy costs and edge processing of extreme volumes of data Data transfer in real time to avoid data loss through sensor disturbance
	Identification of species in real time	Opportunity	Automated species identification by the RAS equipment Overcoming geographic and taxonomic bias
	Surveying over long temporal periods	Opportunity	Surveying sites continuously over extended periods Resurveying sites many times during a year and over many years
	Surveying rare, elusive or cryptic species	Challenge	Ensuring species detection (for instance, behaviourally cryptic diurnal taxa) Misidentifying rare or cryptic species and different sexes or life stages
	Surveying little-known or ‘difficult’ taxa	Challenge	Monitoring little-known taxa Monitoring species with poorly defined taxonomy
	Risk of misidentification by classifiers	Challenge	Identifying little-known or ’difficult’ taxa using AI tools Dealing with undescribed species
	Generating validated classifier training data	Challenge	Availability of training data for classifiers and/or expertise for validation Ground-truthing and geographical relevance of classifier data
	Designing RAS for non-expert operation	Challenge	Sensor easy to operate (for instance, to facilitate non-expert input) Accessibility of AI methods and training resources for non-experts
4. Power and network availability	Availability of communication network	Opportunity	Areas without access to mobile networks Network connections for real-time or cloud data access and storage
	Remote control and maintenance of RAS	Opportunity	Ability to control remotely (for instance, sensors in tree canopies) Self-reporting malfunctions for long-term sensor deployments
	Limited power availability	Challenge	Sustainable power, robust to climate, to support monitoring stations Reducing the weight of power systems
	Negative environmental impact of e-waste	Challenge	Environmental impact of production and/or decommissioning of RAS Retrieving inaccessible RAS equipment at end of life

These were identified by biodiversity experts during stage 2 of the modified Delphi technique.