A defensible protocol combines context, separation, observation, stopping criteria and review; it is not a single altitude.
Some responses are not visible
Ditmer and colleagues observed elevated heart rates in American black bears even when overt behavioural changes were limited. Stillness is therefore not evidence of no effect. Disturbance can compromise both welfare and the behaviour being measured. Assign explicit responsibility for animal-response monitoring alongside aircraft operation.
Set separation within a reviewed protocol
Species, breeding stage, group size, terrain, aircraft sound and approach direction can alter responses. Bird approach experiments found stronger responses to vertical approaches, but their tested distances are not universal safety thresholds. Use literature and local expertise to define conservative initial conditions within the approved protocol. If image quality is inadequate, consider a different sensor or method rather than automatically lowering the aircraft.
Write stop rules before take-off
Specify species-relevant indicators such as sustained vigilance, interrupted feeding, flight, nest departure or approach toward the aircraft. Predefine withdrawal, no-retry conditions and emergency landing options. Do not weaken the protocol because a promising image is nearly available. Review responses and deviations before authorizing the next survey plan.
Record cumulative exposure and non-flight alternatives
Track visits, exposure duration, spacing between flights and other human activity at the same animals or breeding site. Apparent habituation should not be equated with absence of impact. Combine ground observation, camera traps, acoustics or satellite data where these answer part of the question. Welfare criteria belong in the survey's go/no-go decision, not only in an appendix.
Make welfare decisions auditable
The Airffic record design links taxon, season, approach, separation, aircraft and payload, baseline behaviour, responses, stop decisions and recovery observations to each sortie. Saving altitude alone does not justify transfer to a different aircraft or breeding stage. Record unobservable animals separately from observed nonresponse, and do not automatically classify a protocol deviation as a successful survey.
Evidence & further reading
Check the original study conditions, species and location before using a result in a plan. Do not transfer findings universally across species and settings.