This study simulated the processing of H5N1 infected chickens under laboratory conditions and detected infectious droplets and aerosols generated during the work. The birds were asymptomatic at the time of processing, and multiple H5N1 clades were tested. The finding is that infectious virus becomes airborne during handling of birds that show no outward sign of infection.
It is the sound published citation for H5N1 airborne generation, and it should be used in place of any secondary source that cannot be traced. The measured field counterpart on the poultry side is Alonso and colleagues in 2017, who sampled aerosols during the 2015 US Midwest highly pathogenic avian influenza outbreak on turkey and layer premises and recorded a bimodal particle size distribution for H5N2, by RNA detection rather than by isolation. Kamiyama and colleagues in 2011 approached the same organism from the control side, protecting chickens from aerosol H5N1 using antibody impregnated filter media.
The asymptomatic detail is the one with operational weight. If infectious aerosol is generated by handling birds that appear healthy, then the point at which a flock is known to be infected is later than the point at which it begins producing airborne virus. Controls that depend on recognising disease before changing behaviour will engage after that window has opened. That argues for air handling and movement protocols that are in place continuously rather than escalated on suspicion.