pathogens

Article
Quantitative Risk Assessment of Wind-Supported Transmission of
Highly Pathogenic Avian Influenza Virus to Dutch Poultry Farms
via Fecal Particles from Infected Wild Birds in the Environment
Clazien J. de Vos * and Armin R. W. Elbers

Department of Epidemiology, Bioinformatics &Animal Models, Wageningen Bioveterinary Research, P.O. Box 65,
8200 AB Lelystad, The Netherlands; [Link]@[Link]
* Correspondence: [Link]@[Link]; Tel.: +31-320-238173

Abstract: A quantitative microbial risk assessment model was developed to estimate the probability
that the aerosolization of fecal droppings from wild birds in the vicinity of poultry farms would result
in the infection of indoor-housed poultry with highly pathogenic avian influenza virus (HPAIv) in the
Netherlands. Model input parameters were sourced from the scientific literature and experimental
data. The availability of data was diverse across input parameters, and especially parameters on
the aerosolization of fecal droppings, survival of HPAIv and dispersal of aerosols were uncertain.
Model results indicated that the daily probability of infection of a single poultry farm is very low,
with a median value of 7.5 × 10−9 . Accounting for the total number of poultry farms and the
length of the bird-flu season, the median overall probability of at least one HPAIv-infected poultry
farm during the bird-flu season is 2.2 × 10−3 (approximately once every 455 years). This is an
overall estimate, averaged over different farm types, virus strains and wild bird species, and results
indicate that uncertainty is relatively high. Based on these model results, we conclude that it is
unlikely that this introduction route plays an important role in the occurrence of HPAIv outbreaks in
indoor-housed poultry.

Citation: de Vos, C.J.; Elbers, A.R.W.

Keywords: HPAI; wild birds; QMRA; fecal droppings; aerosolization; wind
Quantitative Risk Assessment of
Wind-Supported Transmission of
Highly Pathogenic Avian Influenza
Virus to Dutch Poultry Farms via
Fecal Particles from Infected Wild
1. Introduction
Birds in the Environment. Pathogens Wild water birds of the orders Anseriformes and Charadriiformes are the natural
2024, 13, 571. [Link] reservoir of avian influenza virus [1]. Wild water birds can play an important role in directly
10.3390/pathogens13070571 infecting poultry in free-range operations or bringing the virus to the environment in close
Academic Editors: Christina
vicinity to poultry units [2]. Avian influenza viruses are categorized as low-pathogenic
M. Leyson and Silvia Carnaccini
avian influenza virus (LPAIv) or highly pathogenic avian influenza virus (HPAIv), based
on the pathogenicity of the virus in chickens and the amino acid sequence of the connecting
Received: 3 June 2024 peptide of the haemagglutinin molecule (HA0) (i.e., the cleavage site) [3]. In general, LPAIv
Revised: 27 June 2024 infections may be asymptomatic and produce no or mild disease in chickens [4], while
Accepted: 5 July 2024
HPAIv infections produce high morbidity and mortality in poultry [5].
Published: 8 July 2024
Genetic comparisons of HPAIv strains from outbreaks on Dutch poultry farms since
2014 and HPAIv strains isolated from dead wild birds strongly suggest that these HPAIv
strains were carried to the Netherlands by migratory wild birds from Asia, possibly through
Copyright: © 2024 by the authors.
overlapping flyways and common breeding sites in Siberia [6–8]. Velkers et al. [9] showed
Licensee MDPI, Basel, Switzerland.
a marked increase in birds of the Anatidae family around poultry farms occurring from
This article is an open access article October to April in the Netherlands. This increase was more pronounced for farms close to
distributed under the terms and wetlands compared to farms in other areas. The most striking increase was found for the
conditions of the Creative Commons Eurasian wigeon (Anas penelope), with bird densities several tens of folds higher around
Attribution (CC BY) license (https:// farms located close to low-lying wetlands. Eurasian wigeons were one of the predominant
[Link]/licenses/by/ species with massive mortality due to HPAIv in 2016–2017 and were also found in the
4.0/). vicinity of poultry farms with phylogenetically related H5N8 virus outbreaks [7]. Therefore,

Pathogens 2024, 13, 571. [Link] [Link]

Pathogens 2024, 13, 571 2 of 18

the indoor housing of all free-range poultry was made mandatory during the bird-flu
season. Despite the fact that indoor housing prevents direct contact between poultry and
infected wild avian species, many poultry farms still became infected.
HPAIv multiplies in the respiratory, intestinal, renal and reproduction organs of
infected birds. Infected birds excrete the virus via secreta (fluid secretion) from the nose
and mouth/beak, conjunctiva (mucous membrane of the eyes) and excreta (feces) from the
cloaca [10]. When infected wild birds excrete the virus, they can infect poultry via either
direct or indirect contact. Indirect transmission of the virus between wild birds and poultry
can occur if, e.g., the environment or fomites are contaminated with HPAIv in secretions and
excreta from infected wild birds. An example of an indirect transmission pathway would be
via contaminated water (e.g., a pond or puddles of water) or contaminated soil in the free-
range area of a poultry farm. Wild birds regularly visit the outdoor facilities of commercial
poultry farms and can therefore contaminate the free-range area with HPAIv via secretions
and excreta, such as fecal droppings [2,11]. Another indirect transmission pathway would
be materials, shoes, clothing, stable equipment, vehicles, etc. that are contaminated with
secretions and excreta from HPAI-infected wild birds in the area outside a poultry unit and
brought into the poultry house by the farmer and his/her family members or professional
visitors like consultants, veterinary practitioners, catchers and vaccination crews [12]. The
airborne transmission of HPAIv is another indirect transmission route and is considered a
possible transmission route between nearby poultry farms during epidemics with large
numbers of infected poultry farms [13–15]. HPAIv-infected poultry can produce large
quantities of virus that, when stuck to particulate matter (PM), such as aerosolized dust
originating from feces, bedding material, feed and feathers, can become airborne [16]. This
airborne virus can be transported by forced ventilation air from a house with infected
poultry to the environment outside, which could lead to the potential wind-borne spread
of the virus to other poultry farms [17,18]. However, a recent study by James et al. [19]
showed that airborne particles harboring infectious HPAIv originating from poultry houses
with HPAIv-infected poultry can be translocated only over short distances (<10 m) through
the air, while macroscopic particles containing viral RNA (non-infectious) might travel
further (≤80 m). They concluded that the potential for the airborne transmission of clade
2.3.4.4b H5N1 HPAIv between poultry farms is considered low.
One of the hypotheses suggested by poultry farmers to explain HPAIv outbreaks in
poultry farms with indoor-housed animals is the wind-supported transport of particles
from fecal droppings from HPAIv-infected wild birds in the surroundings of the farm
via the air inlets of the poultry house, resulting in the exposure of the poultry inside the
house [20]. Therefore, the aim of our study was to build a quantitative microbial risk
assessment (QMRA) model to estimate the probability that this indirect transmission route
would indeed result in the infection of indoor-housed domestic poultry.

2. Materials and Methods

A QMRA model was built to estimate the probability that the aerosolization of fecal
droppings of wild birds would result in infection of indoor-housed domestic poultry. The
main conditions for this exposure route are the presence of infected wild birds close to the
poultry house, the excretion of infectious virus in feces, the subsequent aerosolization of
fecal droppings and the survival of infectious virus during aerosolization and air transport
(Figure 1). To characterize the risk, the estimated exposure of poultry to HPAIv was
combined with the bird infectious dose at which 50% of exposed poultry is expected to be
infected (BID50) in an exponential dose–response model.
The QMRA model is a stochastic risk model built in Microsoft Excel for Microsoft
365 MSO (Version 2308 Build 16.0.16731.20542) and @Risk 8.3.2 (Lumivero, Denver, CO,
USA) [21]. Model input parameters were sourced from the scientific literature and experi-
mental data. The availability of data was diverse across input parameters, and especially
input parameters on the aerosolization of feces, survival of the virus and dispersal of
aerosols were quite uncertain. Probability distributions were used to represent uncertainty
Pathogens 2024, 13, 571 3 of 18

Pathogens 2024, 13, x

on input parameters. The model does not, however, account for variability across 3poultry
of 19
farms, wild bird species or HPAIv strains. We challenged some of our assumptions and
input parameters in a what-if analysis.

Figure 1. Outline
Figure 1. Outline of
ofthe
thequantitative
quantitativemicrobial
microbialrisk
riskassessment
assessmentmodel
modeltotoestimate
estimatethe
theHPAI
HPAItransmis-
trans-
mission
sion risk risk
fromfrom
wildwild birds
birds to domestic
to domestic poultry
poultry via via aerosolized
aerosolized fecal
fecal droppings.
droppings.

We
Theestimated
QMRA model the transmission
is a stochasticrisk from
risk wildbuilt
model birds
into domesticExcel
Microsoft poultry via aerosolized
for Microsoft 365
fecal
MSOdroppings
(Version 2308 under the16.0.16731.20542)
Build conditions of a “normal”
and @Risk Dutch
8.3.2bird-flu
(Lumivero,season, whichCO,
Denver, is approx-
USA)
imately frominput
[21]. Model October to [Link]
parameters We sourced
considered frombirds of the order
the scientific Anseriformes
literature to pose the
and experimental
highest risk to farms, because of both the relatively high prevalence
data. The availability of data was diverse across input parameters, and especially input of HPAI viruses in this
order [22–24] and their abundant presence in the environment around
parameters on the aerosolization of feces, survival of the virus and dispersal of aerosols some poultry farms
during the bird-flu
were quite uncertain. season [9].
Probability distributions were used to represent uncertainty on in-
The output ofThe
put parameters. the model isdoesthe probability
not, however, that account
the exposure of indoor-housed
for variability across domestic
poultry
poultry to aerosolized
farms, wild bird species fecal droppings
or HPAIv from
strains. Wewild birds will
challenged result
some of in
ouratassumptions
least one infected
and
poultry farm during
input parameters in a the bird-flu
what-if season. The model was run for 10,000 iterations, and
analysis.
results
Weare given asthe
estimated median values and
transmission risk95%from uncertainty
wild birdsintervals.
to domestic poultry via aeroso-
lized fecal droppings under the conditions of a “normal” Dutch bird-flu season, which is
2.1. Model Calculations
approximately from October to April. We considered birds of the order Anseriformes to
poseThe
the model
highestconsists of two steps:
risk to farms, because first, the exposure
of both of poultry
the relatively high to HPAIv viaofaerosolized
prevalence HPAI vi-
fecal
rusesdroppings
in this order of[22–24]
wild birds
and is assessed,
their abundant after which in
presence thetheexposure is combined
environment around somewith a
dose–response
poultry farms duringmodelthe to bird-flu
estimateseason
the daily
[9]. probability of infection of a single poultry
[Link]
output areofthen combined
the model with
is the the number
probability that of
thepoultry
exposure farms in the Netherlands
of indoor-housed do-
and the length of the bird-flu season (in days) to estimate the overall
mestic poultry to aerosolized fecal droppings from wild birds will result in at least one probability of HPAI
introductions
infected poultry to poultry
farm duringfarmstheduring theseason.
bird-flu Dutch Thebird-flu
modelseason via this
was run exposure
for 10,000 route.
iterations,
and results are given as median values and 95% uncertainty intervals.

2.1. Model Calculations

Pathogens 2024, 13, 571 4 of 18

2.1.1. Exposure Assessment

The number of infected Anseriformes visiting a poultry farm (WBin f ) was calculated as:

WBin f = WB × Prevwb (1)

where WB is the number of wild birds approaching to the farm, and Prevwb is the apparent
prevalence of HPAI in wild birds.
To estimate the daily amount of HPAIv (EID50 ) present in wild bird feces on the farm
yard (AI f ec,d ), we multiplied WBin f with the daily amount of feces excreted by wild birds
(gram wet feces) (Fecwb,d ), the fraction of the day the wild birds stay (and excrete) in close
vicinity to the farm (Fp f ) and the average concentration of HPAIv (EID50 /gram) in feces of
wild birds (AI f ec_conc ) as follows:

AI f ec,d = WBin f × Fecwb,d × Fp f × AI f ec_conc (2)

To estimate the amount of HPAIv that is released into the air via aerosolization (AI air,d ),
we multiplied AI f ec,d with the probability that the weather conditions during the bird-flu
season are favorable for drying the feces (Fweather ) and the expected survival of HPAIv
during drying of feces (Fsurv_dry ) as follows:

AI air,d = AI f ec,d × Fweather × Fsurv_dry (3)

The amount of aerosolized HPAIv that poultry in the poultry house is then exposed
to (AI exp,d ) was calculated as the product of AI air,d and the fraction of virus reaching the
animals, accounting for the retained fraction of virus after the dispersion of the aerosols over
a short distance (Fdisp ), as well as the survival of the virus during air transport (Fsurv_trans ),
the fraction of aerosols entering the barn (Fbarn ) and the fraction of air in the poultry house
that is inhaled by the animals (Finhale ):

AI exp,d = AI air,d × Fdisp × Fsurv_trans × Fbarn × Finhale (4)

2.1.2. Risk Characterization

To estimate the daily probability of at least one infected bird in the poultry farm
(Pin f ,p f ,d ), an exponential dose–response model was used:

Pin f ,p f ,d = 1 − exp−( DR× AI exp,d ) (5)

where DR is the exponential dose–response parameter.

The overall probability of at least one infected poultry farm during the bird-flu season
(Pin f ) was estimated by including the number of poultry farms in the Netherlands (Np f )
and the length of the bird-flu season in days (D) in the dose–response model as follows:

Pin f = 1 − exp−( DR× AI exp,d × Np f × D) (6)

2.2. Model Input

2.2.1. Exposure Assessment
The number of Anseriformes in close vicinity to a poultry farm (WB) is likely to be
low. Since we assumed that only bird droppings on paved surfaces could dry enough and
subsequently aerosolize during winter time [25], an estimate was made of the expected
number of birds in a radius of <20 m of the poultry house. Few studies have actually
closely monitored wild bird activity on poultry farms [2,20,26–30]. To estimate the number
of Anseriformes visiting poultry farms and the time spent on the farm yard, we used
observations by Elbers and Gonzales [2] on the daily number of mallards visiting the free-
range area of a layer farm in a high-risk area in the Netherlands during the bird-flu season.
Observed numbers of mallards varied widely across months (with a peak from December
Pathogens 2024, 13, 571 5 of 18

to February) and days, with mallards observed on 42 out of 136 days. The median number
of birds observed on these days was 6 (95% uncertainty interval: 1 to 24). Mallards mostly
visited the farm during night time and spent a median of 3.7 h per day in the free-range
area (95% uncertainty interval: 0.04 to 18.4). These latter values were used to model the
fraction of the day in which birds stay in close vicinity to the farm (Fp f ) (Table 1).

Table 1. Parameters of the quantitative microbial risk assessment model.

Model
Description Value Source
Parameter
Daily probability that wild birds are present Beta distribution with α = 43 and
Pwb [2]
at the farm β = 95
Number of wild birds at the farm yard on a Pert distribution with min = 1,
Nwb [2]
day that birds are present most likely = 6 and max = 24
Expected number of wild birds at the
WB Pwb × Nwb Calculated parameter
farm yard
Pert distribution with
Fraction of the day that wild birds spend at
Fp f min = 3.4 × 10−4 , most [2]
the farm yard
likely = 0.16 and max = 0.77
Beta distribution with α = 35 and
Prevwb Apparent HPAI prevalence in wild birds [22–24]
β = 8521
Daily amount of feces excreted by wild birds Lognormal distribution with
Fecwb_dry,d [31]
(dry weight in grams) mean = 36.9 and SD = 4.37
Cdry_wet Conversion factor from dry to wet feces 5 [32]
Daily amount of feces excreted by wild birds
Fecwb,d Cdry_wet × Fecwb_dry,d Calculated parameter
(wet weight in grams)
Lognormal distribution with
Concentration of HPAIv in wild bird feces
AI f ec_conc mean = 3.8, 2.5 percentile = 3.15 [11]
(log10 EID50 /g)
and 97.5 percentile = 4.5
Fraction of days with suitable weather
Fweather 0.014 [33]
conditions for the aerosolization of feces
Fsurv_dry Survival of HPAIv during the drying of feces 10−3.25 [34]
Fraction of virus retained after the dispersion Uniform distribution with
Fdisp [35]
of aerosols over a short distance min = −2 log10 and max = −0.5 log10
Uniform distribution with
Fsurv_trans Survival of HPAIv during air transport [36]
min = 0.61 and max = 0.70
Estimate based on surface
Fbarn Fraction of aerosols entering the barn 0.05 area of ventilation
openings in poultry barns
Ventilation rate of poultry house (layers) Pert distribution with min = 1.1,
VR [37]
(m3 /animal/hour) most likely = 3.5 and max = 9
Respiratory volume of chickens (layers)
RV 0.0224 [38]
(m3 /hour)
Fraction of air in the poultry house that is
Finhale RV/VR Calculated parameter
inhaled by the animals
Normal distribution with
BID50 Bird infectious dose (log10 EID50 ) [39]
mean = 1.2 and SD = 0.2
DR Dose–response parameter (EID50 −1 ) ln(2)/10BID50 Calculated parameter
Np f Number of poultry farms in the Netherlands 1353 [40]
October–April (number
D Length of bird-flu season (days) 212.25
corrected for leap years)
Pathogens 2024, 13, 571 6 of 18

The probability that a wild bird visiting the poultry farm would be infected with
HPAIv was based on field studies in the Netherlands during the bird-flu seasons of 2014/15
and 2016/17, in which live birds were captured and sampled using oropharyngeal and
cloacal swabs, and fecal samples were collected in the field [22–24]. Only samples from
Anseriformes (ducks and geese) were used to estimate the apparent prevalence (Prevwb ).
A total of 8554 samples were tested, of which 34 were positive for HPAI, resulting in an
apparent prevalence of 4.1 × 10−3 . Uncertainty in the estimate of Prevwb was simulated
using a beta distribution. More recent data from the United Kingdom (UK) (2019/20)
and Italy (2020/21) report higher apparent prevalences in wild waterfowl ranging from
1.4 × 10−2 to 4.6 × 10−2 [41,42]. This is in agreement with higher numbers of reported
HPAI cases in both wild birds and poultry in recent years [43–45]. We challenged our
model assumptions using the data from these studies to estimate the HPAI prevalence in
wild birds in the what-if analysis (scenario WI-1).
The daily amount of feces excreted by wild birds (Fecwb,d ) was based on data for adult
mallards (Anas platyrhynchos), with an average value of 36.9 g dry weight per bird per
day [31]. Only adult mallards were considered, as nestlings are not likely to be present
during the bird-flu season. The moisture content of duck droppings varies between 70 and
90% [25,32,46,47]. To obtain wet weight values, dry weights were therefore multiplied by a
factor of 5 [32]. Gere and Andrikovics [48] reported lower dry weights of fecal droppings
from mallards (26.3 g per day), and we evaluated the impact of changing this parameter in
the what-if analysis (scenario WI-2). A second what-if scenario (WI-3) was run to mimic
the risk of geese droppings, which have a higher dry weight. We used the reported fecal
dry weight of the greylag goose (Anser anser) (100 g per day) [32], which is an abundant
species in the Netherlands during the bird-flu season [49].
The concentration of HPAIv in the feces of wild birds (AI f ec_conc ) was derived from
the results of a systematic literature review by Germeraad et al. [11], with an estimated
mean of 3.8 log10 EID50 /mL. This estimate is close to values used in other modeling
studies (e.g., [38]). However, more recent experiments in Pekin ducks (Anas platyrhynchos
domesticus) and Eurasian wigeons (Anas penelope) indicated a high difference in shedding
levels of the HPAI H5 viruses present in 2014 (group A virus) compared to those present in
2016 and 2017 (group B viruses) [50]. The estimates from these experiments were used in
the what-if analysis. Scenario WI-4 is based on excretion data from Eurasian wigeons for
H5N8-2014 with an estimated mean of 3.4 log10 EID50 /mL, whereas scenario WI-5 is based
on excretion data from the same bird species for H5N8-2016 with an estimated mean of
5.0 log10 EID50 /mL.
The probability that feces are aerosolized is highly dependent on the weather condi-
tions. Elbers [25] observed the drying of feces only when deposited on a concrete surface
(not on grassland) on days with no precipitation and global sun irradiation ≥ 1000 J/cm2 .
To estimate the probability that weather conditions during the bird-flu season are favorable
for drying of feces to allow for HPAIv release into the air (Fweather ), we determined the
number of days from October to April in a 30-year period (1993–2022) that had suitable
weather conditions lasting for at least 7 days [33]. This resulted in an estimated 1.4% of days
allowing for the aerosolization of feces (Supplementary Table S1). This estimate is likely to
result in an overestimate of the risk, as a single week of drying is probably not sufficient
for full aerosolization. Elbers [25] measured 23% remaining moisture in duck feces after a
week of suitable weather conditions. On the other hand, we might have underestimated the
risk using this value considering that feces excreted during previous days are also subject
to aerosolization once the suitable weather period starts and that HPAIv can survive for
prolonged periods (>30 days) at low ambient temperatures [51,52].
The drying of feces results in the quick inactivation of HPAIv; Zarkov and Uru-
mova [34] recorded a 3.25 log10 inactivation of LPAIV (H6N2) after one day of drying, and
no virus was detected after two days of drying. This is in accordance with observations
by Shortridge et al. [51] that H5N1 virus was inactivated within one day when dried at a
temperature of 25 ◦ C. Sedlmaier et al. [53] spiked a suspension of dried broiler manure with
Pathogens 2024, 13, 571 7 of 18

LPAI virus (H10N7), after which they nebulized the suspension and measured the virus
concentration in deposited fecal PM2.5 (particle size < 2.5 mm) on filters. They did retrieve
viable virus in the deposited PM2.5 , but it is not clear how these virus concentrations relate
to the original concentrations in the broiler manure suspension. As we expect a minimum
drying period of seven days to allow for the aerosolization of fecal droppings during the
Dutch bird-flu season [25], it is not very likely that viable virus is still present in fecal dust
particles. As a worst-case assumption, we used the 3.25 log10 reduction given by Zarkov
and Urumova [34] to quantify the expected survival of HPAIv during the drying of feces
(Fsurv_dry ) in the baseline model calculations. In the what-if analysis (scenario WI-6), we
assumed that virus reduction due to drying was twice as high (6.5 log10 ).
Survival of virus during air transport (Fsurv_trans ) is highly dependent on the envi-
ronmental conditions and the time from aerosolization to the exposure of poultry. At
low temperatures, Harper [36] reports survival rates of influenza in aerosols from 61% to
70% after 1 h and 3% to 19% after 23 h, depending on relative humidity. In the model
calculations, we used the values for 1 h survival, as we assumed the fecal aerosols will be
released close to the poultry houses and will not need much time to reach the barn. In the
what-if analysis (scenario WI-7), we used the decay rate constant given by Ssematimba
et al. [17] and estimated the surviving virus fraction after 1 h at 99%.
We also accounted for the loss of infectivity due to the dispersal of the virus (Fdisp )
using model results given by Lighthart and Mohr [35], indicating a 0.5 log10 dilution at 10
m distance and a 2 log10 dilution at 20 m distance. We acknowledge that this value will
vary depending on weather conditions (wind, humidity). In the what-if analysis (scenario
WI-8), we used results from dispersion calculations from Sedlmaier et al. [53] for particulate
matter (PM10 ) downwind from a broiler farm. Based on these proxy data, the expected
decrease in concentration of the virus after traveling a distance of 10 to 20 m is between
63% and 78%.
Furthermore, the aerosols will not always reach the poultry house and enter the barn
via the ventilation openings. We assumed that the fraction of aerosols entering the barn
(Fbarn ) will on average be 5% based on the relative surface area of ventilation openings in
the side walls of both the layer farm and the broiler farm included in the study described
by Elbers et al. [20]. This might be an underestimate, as underpressure in the barn resulting
from a negative-pressure ventilation system will result in a slightly higher fraction of
aerosols entering. On the other hand, the estimate of Fbarn is an overestimate, as we did
not account for the fact that the wind direction will not always favor dispersion of aerosols
to the barn. As both effects could not be quantified, we used the 5% value as a worst-case
estimate, assuming that aerosols will be directed to the poultry house most of the times.
The fraction of air inhaled by the animals in the poultry house (Finhale ) was estimated
for laying hens and broilers separately using data on the ventilation rate of poultry houses
and the respiratory rate of chickens (Table 1). The ventilation rate determined the time span
during which the aerosols would be present in the barn, and this time span combined with the
respiratory rate determined the fraction of aerosols the animals could maximally inhale. In the
baseline scenario, values for laying hens were used, as laying hens are most at risk for HPAIv
introduction [54,55]. In the what-if analysis (scenario WI-9), values for broilers were used.

2.2.2. Risk Characterization

The exponential dose–response parameter was derived from the infectious dose that
yields 50% probability of infection in birds, the bird infectious dose BID50 . This dose is
dependent on the poultry species, the virus strain and the inoculation route [25]. The
reported bird infectious doses of HPAIv in ducks are lower than those in chickens [56–58].
Swayne and Slemons [59] estimated the bird infectious dose of HPAIv in chickens to
vary between 1.2 and 4.7 log10 EID50 , with values for H5N1 strains from Asian outbreaks
since 1997 between 2.3 and 3.1 log10 EID50 . Pantin-Jackwood et al. [57] estimated slightly
higher BID50 values for more recent H5N1 and H5N8 strains, varying between 2.6 and
4.2 log10 EID50. These BID50 values were all based on intranasal inoculation. Birds are,
Pathogens 2024, 13, 571 8 of 18

however, approximately 30 times more sensitive to the aerosol route of infection [39],
which implies that BID50 values for aerosol exposure are lower than those for intranasal
inoculation. Estimated values of the BID50 for the aerosol route indeed indicate higher
sensitivity, with most of them estimated at approximately 1 log10 EID50 [39]. There is no
evidence that HPAIv strains derived from other poultry species, such as ducks, result in a
significantly higher BID50 in chickens [25,39,59]. In the baseline model calculations, we
used a BID50 with a mean value of 1.2 log10 EID50 , based on observations by Sergeev
et al. [39]. In the what-if analysis (WI-10), we used a much higher value of 4.9 log10 EID50
based on observations of the BID50 of H5N8 virus isolated from a tufted duck (Aythya
fuligula) [25,60].
To estimate the overall probability of at least one infected poultry farm during the
Dutch bird-flu season (Pin f ), the exposure of a single farm on a single day was multiplied
by the total number of poultry farms in the Netherlands and the length of the bird-flu
season (212 days). In 2022, the number of laying farms in the Netherlands was 734, and the
number of broiler farms was 619 [40].

2.3. Uncertainty Analysis

To evaluate the sensitivity of model results for uncertain input parameters, correlation
coefficients between sampled values of these parameters and model results for the overall
probability of at least one infected poultry farm during the bird-flu season (Pin f ) were
calculated. Furthermore, what-if scenarios were run with the model to evaluate the impact
of modeling assumptions (see also Section 2.2). An overview of the what-if scenarios is
given in Table 2.

Table 2. What-if scenarios explored with the quantitative microbial risk assessment model.

Model
Scenario Description Baseline Value New Value Source
Parameter
Beta distribution with α = Beta distribution with α = 41
WI-1 Higher prevalence in wild birds Prevwb [41,42]
35 and β = 8521 and β = 930
Lower dry weight (g) of fecal Lognormal distribution with Lognormal distribution with
WI-2 Fecwb,dry,d [48]
droppings (data for adult ducks) mean = 36.9 and SD = 4.37 mean = 26.3 and SD = 11.5
Higher dry weight (g) of fecal Lognormal distribution with
WI-3 Fecwb,dry,d 100 [32]
droppings (data for greylag geese) mean = 36.9 and SD = 4.37
Lognormal distribution
Lower concentration of HPAIv in
with mean = 3.8, 2.5 Lognormal distribution with
WI-4 feces (log10 EID50 /g) (data for AI f ec_conc [50]
percentile = 3.15 and 97.5 mean = 3.38 and SD = 0.44
H5N8-2014 in Eurasian wigeon)
percentile = 4.5
Lognormal distribution
Higher concentration of HPAIv in
with mean = 3.8, 2.5 Lognormal distribution with
WI-5 feces (log10 EID50 /g) (data for AI f ec_conc [50]
percentile = 3.15 and 97.5 mean = 4.96 and SD = 0.77
H5N8-2016 in Eurasian wigeon)
percentile = 4.5
Lower survival of HPAIv during
Estimate
WI-6 the aerosolization (drying) of Fsurv_dry 10−3.25 10−6.5
based on [34]
feces
Higher survival of HPAIv during Uniform distribution with
WI-7 Fsurv_trans 0.99 [17]
transport of aerosols min = 0.61 and max = 0.70
Uniform distribution with
Higher fraction of virus retained Uniform distribution with min
WI-8 Fdisp min = −2 log10 and [53]
after the dispersion of aerosols = 0.63 and max = 0.78
max = −0.5 log10
Lower ventilation rate of poultry Pert distribution with min
Pert distribution with min = 0.1,
WI-9 houses based on broilers VR = 1.1, most likely = 3.5 [37]
most likely = 2.1 and max = 9.6
(m3 /animal/hour) and max = 9
Pert distribution with most
Higher bird infectious dose based likely = 4.85, 2.5 percentile = 4.23
Normal distribution with
WI-10 on H5N8 virus isolated from a BID50 and 97.5 percentile = 5.51 MINUS [25,39,60]
mean = 1.2 and SD = 0.2
tufted duck (log10 EID50 ) 1.48 to correct for the aerosol
inoculation route
 Pathogens 2024, 13, 571 9 of 18

3. Results
3.1. Baseline Scenario
The estimated daily probability of the infection of a single poultry farm (Pin f ,p f ,d ),
i.e., at least one infected bird present in the farm, is very low, with a median value of
7.5 × 10−9 (95% uncertainty interval: 2.5 × 10−10 to 2.0 × 10−7 ). The box-and-whisker
plot in Figure 2 provides more insight into the uncertainty distribution of results. When
accounting for the total number of poultry farms in the Netherlands and the length of the
bird-flu season, this results in a median overall probability of at least one infected poultry
farm during the bird-flu season (Pin f ) of 2.2 × 10−3 (95% uncertainty interval: 7.1 × 10−5 to
0.06). In other words, an HPAI outbreak in a poultry house due to the wind-supported trans-
mission of HPAIv via fecal particles from infected wild birds is expected to happen approx-
imately once every 455 years. The median daily exposure of poultry to HPAIv in aerosols
on a single farm (AI exp,d ) is 1.7 × 10−7 EID50 (95% uncertainty interval: 6.1 × 10−9 to
3.9 × 10−6 ). It should be noted that this is an averaged value over all days, i.e., days
with and without infected wild birds visiting the farm yard and days with and without
weather conditions suitable for the aerosolization of wild bird droppings. The probability
Pathogens 2024, 13, x that infected wild bird droppings are present and conditions for aerosolization are met
on a single day is 1.3 × 10−4 (95% uncertainty interval: 3.3 × 10−5 to 3.3 × 10−4 ). On
these days favorable for transmission, the estimated median exposure is 1.4 × 10−3 EID50
(95% uncertainty interval: 1.1 × 10−4 to 3.2 × 10−2 ), resulting in an infection probability of
5.9 × 10−5 (95% uncertainty interval: 5.5 × 10−6 to 1.8 × 10−3 ) for individual poultry farms.

Figure 2. Box-and-whisker plot of model results for the daily probability of infection of a single
Figure 2. Box-and-whisker plot of model results for the daily probability of infection
poultry farm (Pin f ,p f ,d ) and the overall probability of at least one infected poultry farm during the
poultry farm (𝑃
bird-flu season (Pin f )., , ) and the overall probability of at least one infected poultry farm
bird-flu season (𝑃 ).

3.2. Uncertainty Analysis

3.2.1. Sensitivity Analysis
Ten parameters of the model were inputted as uncertainty distributions
 Pathogens 2024, 13, 571 10 of 18

3.2. Uncertainty Analysis

3.2.1. Sensitivity Analysis
Ten parameters of the model were inputted as uncertainty distributions (Table 1).
Seven of these uncertain model input parameters had a correlation coefficient ≥ |0.1| for
the overall probability of at least one infected poultry farm during the bird-flu season (Pin f )
(Supplementary Table S2), indicating that results of the model are sensitive to uncertainties
in these parameters. Figure 3 shows the change in the median Pin f when these input
parameters were changed from their lowest to highest value (input percentile values at
x-axis). Uncertainty in the retained fraction of the virus after the dispersion of aerosols over
Pathogens 2024, 13, x a short distance (Fdisp ) and the concentration of HPAIv in the feces of wild birds (AI f ec_conc ) 11 of 19
had the greatest impacts on model results. Taking the 95th percentile values of these input
parameters resulted in a 4.7- to 4.8-fold increase in Pin f . The bird infectious dose (BID50 )
and the ventilation rate of poultry houses (VR) were negatively correlated with Pin f . Taking
the 5th percentile values of these input parameters resulted in a 2.2-fold increase in Pin f .

Figure 3. Spider plot showing the relation between the median overall probability of at least one
Figure 3. Spider plot showing the relation between the median overall probability of at least one
infected poultry farm during the bird-flu season (Pin f ) and the percentile values of input parameters
infected poultry farm during the bird-flu season (𝑃 ) and the percentile values of input parameters
that had a correlation coefficient > |0.1| with Pin f . These input parameters were: fraction of virus
that had a correlation coefficient > |0.1| with 𝑃 . These input parameters were: fraction of virus
retained after the dispersion of aerosols over a short distance (Fdisp ); concentration of HPAIv in wild
retained after the dispersion of aerosols over a short distance (𝐹 ); concentration of HPAIv in wild
bird feces (AI f ec_conc ); fraction of the day that wild birds spent at the farm yard (Fp f ); number of wild
bird feces (𝐴𝐼 _ ); fraction of the day that wild birds spent at the farm yard (𝐹 ); number of wild
birds at the farm yard on a day that birds are present (Nwb ); bird infectious dose (BID50 ); ventilation
birds at the farm yard on a day that birds are present (𝑁 ); bird infectious dose (𝐵𝐼𝐷 ); ventilation
rate of poultry house (VR); and daily probability that wild birds are present at the farm (Pwb ).
rate of poultry house (𝑉𝑅); and daily probability that wild birds are present at the farm (𝑃 ).
3.2.2. What-If Analysis
[Link]
What-If ofAnalysis
the what-if analysis are given in Figure 4. Three what-if scenarios resulted in
a 10-fold
Resultsincrease
of theinwhat-if
the risk compared to the
analysis are baseline
given scenario:
in Figure a higher
4. Three HPAIscenarios
what-if prevalenceresulted
in wild birds based on more recent studies in wild birds in the
in a 10-fold increase in the risk compared to the baseline scenario: a higherUnited Kingdom andHPAI
Italy preva-
(WI-1), a higher concentration of HPAIv in feces based on data for H5N8-2016 (WI-5) and a
lence in wild birds based on more recent studies in wild birds in the United Kingdom and
higher fraction of infected aerosols retained during the dispersion of aerosols (WI-8). Two
Italy (WI-1),
what-if a higher
scenarios concentration
resulted of HPAIv
in a significant decrease inin
feces based
the risk on datatofor
compared theH5N8-2016
baseline (WI-
5)scenario:
and a higher
a lower fraction of infected
probability aerosols
of the survival ofretained during
HPAIv during the dispersion
aerosolization of aerosols
(WI-6) and a (WI-
8). Twobird
higher what-if scenarios
infectious resulted
dose based in a virus
on H5N8 significant
isolateddecrease in the
from a tufted risk(WI-10).
duck compared The to the
baseline
amount of scenario: a lower
feces excreted by probability of the
wild birds (WI-2 andsurvival
WI-3), a of HPAIv
lower during aerosolization
concentration of HPAIv in (WI-
6)feces
and(WI-4),
a higher a higher survival ofdose
bird infectious HPAIv during
based transport
on H5N8 of aerosols
virus isolated(WI-7)
from aand a lower
tufted duck (WI-
ventilation rate in poultry houses based on values for broiler farms (WI-9) only
10). The amount of feces excreted by wild birds (WI-2 and WI-3), a lower concentration of had limited
effects on
HPAIv in the
fecesestimated
(WI-4), risk.
a higher survival of HPAIv during transport of aerosols (WI-7) and
a lower ventilation rate in poultry houses based on values for broiler farms (WI-9) only
had limited effects on the estimated risk.
13, x 12 of 19
Pathogens 2024, 13, 571 11 of 18

Figure 4. Tornado chart showing the relative increase or decrease (expressed as log10 difference) in
Figure 4. Tornado the
chart showing
overall the relative
probability of at leastincrease
one infected or poultry
decrease (expressed
farm as log10season
during the bird-flu difference) in
(Pin f ) compared
the overall probability
to theof at leastscenario
baseline one infected poultry
for 10 what-if farm during
scenarios. the bird-flu
Parameters considered season ) com-
in the(𝑃what-if scenarios
pared to the baseline scenario
were: apparent forHPAI
10 what-if
prevalencescenarios.
in wildParameters
birds (Prevwb considered
); daily amount in the of what-if scenar-
feces excreted by wild
ios were: apparentbirds
HPAI prevalence
(Fec in wild birds
wb_dry,d ); concentration (𝑃𝑟𝑒𝑣in wild
of HPAIv ); daily
birdamount
feces (AIof feces
f ec_conc excretedofby
); survival wildduring
HPAIv
birds (𝐹𝑒𝑐 _ , );the concentration
drying of feces of(FHPAIv
surv_dry );in wild bird
survival feces during
of HPAIv (𝐴𝐼 _ air transport
); survival of HPAIv
(Fsurv_trans during
); fraction of virus
the drying of fecesretained
(𝐹 _ );
after survival
the of
dispersion HPAIv
of during
aerosols over air
a transport
short distance (𝐹
(F );
disp _ );
ventilationfraction
rate ofof virus
poultry house
(VR); and bird infectious dose
retained after the dispersion of aerosols over a short (BID 50 ). The indicate an increase ( ↑
distance (𝐹 ); ventilation rate of poultry of the
arrows ) or a decrease ( ↓ )
house (𝑉𝑅); and bird infectious dose (𝐵𝐼𝐷 ). The arrows indicateofan
input parameter’s value. A more detailed description each scenario()
increase is given in Table 2. ()
or a decrease
of the input parameter’s value. A more detailed description of each scenario is given in Table 2.
4. Discussion
The estimated probability of an HPAI outbreak in a poultry house due to the wind-
4. Discussion supported transmission of HPAIv via fecal particles from infected wild birds is very low,
withprobability
a median value of 2.2 −3 per bird-flu season. This is an overall estimate, averaged
× 10outbreak
The estimated of an HPAI in a poultry house due to the wind-
over different farm types,
supported transmission of HPAIv via fecal particles fromvirus strains and wild bird species,
infected and results
wild birds is very indicate
low, that
uncertainty is relatively high. However, even under worst-case conditions, the probability
with a median value of 2.2 × 10−3 per bird-flu season. This is an overall estimate, averaged
is still low, with a 97.5 percentile value of 0.06, which equals an expected introduction to
over different farm types,
domestic virusvia
poultry strains andonce
this route wild bird
every 17species,
years. and results indicate that
uncertainty is relatively high. However,
Although even risk
the quantitative under worst-case
assessment model conditions,
that we usedthe is probability
a stochastic risk
is still low, with a 97.5 percentile value of 0.06, which equals an expected introduction
model accounting for uncertainty and variability in input parameters, the to not
model does
simulate infections in individual
domestic poultry via this route once every 17 years. farms but rather calculates the probability of infections at
the farm level and the sector level in the Netherlands. The estimated median daily exposure
Although the quantitative risk assessment model that we used is a stochastic risk
to HPAIv at the farm level is very low, with a value of 1.7 × 10−7 EID50 , resulting in a low
model accounting for uncertainty and variability in input parameters, the model does not
infection probability (Pin f ,p f ,d = 7.5 × 10−9 ). If only considering days on which infection is
simulate infections in individual
possible, i.e., infectedfarms but rather
wild birds calculates
are present at the farmtheandprobability of infections
weather conditions are suitable
at the farm level for
and thethe sector levelofin
aerosolization thethe
feces, Netherlands. The estimated
exposure is almost 4 median
10 log10 higher at 1.4daily −3 EID
× 10ex- 50,
posure to HPAIvresulting
at the farm also inlevel is very
an almost 4
10low,
log10with a value
higher of 1.7
infection × 10−7 EID
probability (Pin50 f ,d = 5.9 × 10
, resulting
f ,p in−5 ).
Probability
a low infection probability (𝑃 distributions of model input parameters represent both uncertainty and
, , = 7.5 × 10 ). If only considering days on which infec-
−9
variability. We had, e.g., great uncertainty in the fraction of virus surviving during
tion is possible, i.e., infected wild birds are present at the farm and weather conditions are
aerosolization and the dispersion of the virus before reaching the poultry house. Vari-
suitable for the aerosolization
ability resulted from of feces, the exposure
differences in HPAIv is almost
strains and10
4 log10 higher at 1.4 × 10−3
wild bird species. Model parameters
EID50, resulting also
were inestimated
an almostbased 104 log higher
on10data infection probability
for Anseriformes, (𝑃 , , ducks,
mostly dabbling = 5.9 with
× 10−5data
). for
Probability mallards
distributions(Anas of model input
platyrhynchos) parameters
being represent
most abundant. Also, both
modeluncertainty
parameters and
were as
much as possible based on values for HPAIv strains. However,
variability. We had, e.g., great uncertainty in the fraction of virus surviving during aero- if these were not available
solization and the (e.g., for survival
dispersion ofofthe
virus),
viruswebefore
used data for LPAI
reaching virus
the strainshouse.
poultry or influenza A viruses as
Variability
the best alternative. Parameters on virus excretion in feces and the bird infectious dose
resulted from differences in HPAIv strains and wild bird species. Model parameters were
were derived from studies in which values for multiple HPAIv strains were reviewed
estimated based on data for Anseriformes, mostly dabbling ducks, with data for mallards
(Anas platyrhynchos) being most abundant. Also, model parameters were as much as pos-
sible based on values for HPAIv strains. However, if these were not available (e.g., for
survival of virus), we used data for LPAI virus strains or influenza A viruses as the best
alternative. Parameters on virus excretion in feces and the bird infectious dose were de-
Pathogens 2024, 13, 571 12 of 18

and compared [11,39,57,59]. High variations in these values were observed among HPAIv
strains, making it difficult to decide on a representative value for this model.
Uncertain input parameters that had the greatest effect on model outcome were the
fraction of virus retained after the dispersion of aerosols over a short distance (Fdisp ), the
concentration of HPAIv in wild bird feces (AI f ec_conc ), the number of wild birds at the
farm yard (Nwb ), the fraction of the day birds are present (Fp f ) and the bird infectious dose
(BID50 ) (Supplementary Table S2; Figure 3). The what-if analysis indicated that model
results are also highly sensitive to assumptions on the survival of HPAIv during the drying
of feces (Fsurv_dry ) and the HPAI prevalence in wild birds (Prevwb ) (Figure 4).
We had very limited data to estimate the fraction of virus retained after the dispersion
of aerosols over a short distance (Fdisp ). There are multiple studies modeling the aerosol
transmission of HPAI (e.g., [13,17,19,38,61]), but all these studies take an infected poultry
farm as the source of infection rather than fecal droppings deposited by wild birds in
the environment around a poultry farm. We expect a much lower probability of aerosol
transmission via these fecal droppings for several reasons: the amount of virus excreted
is less than that in an infected poultry flock, the environmental conditions (temperature,
humidity) for aerosolization are less suitable outside the poultry house during the bird-flu
season and the aerosols are emitted from ground level rather than from a ventilation outlet
at 1.5 to 2 m height. Furthermore, it should be acknowledged that the processes of AI
airborne transmission, such as aerosolization, transportation and deposition are complex
and not fully understood, and all modeling studies need to make assumptions on these
issues [38,62,63]. In our model, we have chosen a simple approach, where we estimated
the remaining fraction after each step in which inactivation or loss of infection could occur
rather than complex models accounting for, e.g., meteorological conditions.
The estimated values for virus shedding in feces (AI f ec_conc ) were based on data from
cloacal swabs, where we assumed that the concentration given in EID50 /mL for cloacal
swabs corresponds to the concentration in feces given as EID50 /g [64]. We used data from a
systematic literature review by Germeraad et al. [11] to estimate the concentration of HPAIv
in wild bird feces. The estimates in this study were based on a meta-analysis of studies on
HPAIv infections with both high-pathogenic and low-pathogenic AI viruses in multiple
bird species. To estimate AI f ec_conc , we only selected results for HPAI viruses (both H5 and
H7 strains) in ducks. This estimate is therefore considered to cover the variation among
virus strains and duck species. A recent study by Beerens et al. [50] confirmed that there is
indeed variation in the amount of virus excreted in feces between wild bird species and
HPAIv strains. Also, the amount of virus excretion is not stable over time, with virus titers
decreasing after the first week of infection [58]. The distribution used for AI f ec_conc in the
baseline scenario is likely to be an average over the full infectious period of the birds and
does not account for peak titers, which have been observed to be between 4 and 6 log10
EID50 /mL [50,58,60].
Very few data were available to estimate the number of ducks that are present on the
farm yard (Nwb ), i.e., at a short distance (<20 m) of the poultry houses. Few studies have
actually closely monitored wild bird activity on poultry farms [2,20,27–30]. Predominantly
song birds (order Passeriformes) are observed on the premises of poultry farms; members
of the orders Anseriformes and Charadriiformes are hardly reported to visit areas close
to poultry houses. Elbers and Gonzales [2,28] observed visits of wild fauna, including
birds, to the free-range area of a layer farm in a high-risk area in the Netherlands. They
concluded that dabbling ducks visited the outdoor facility only at night time in the period
from November to May, i.e., especially during the bird-flu season, with most birds visiting
in the months December, January and February. We used the observations by Elbers and
Gonzales [2] to estimate the frequency of bird visits (Pwb ), the number of birds visiting
(Nwb ) and the time spent on the farm yard (Fp f ), as these were the only data available to
quantify these parameters. These values are likely to be overestimates, as the outdoor
facility was not paved and had water pools after rainy periods that might have attracted
the wild birds. A study by Veen et al. [26] reported on the number of birds observed at a
Pathogens 2024, 13, 571 13 of 18

distance of <50 m from the farm buildings in different European countries based on limited
observations. Only a few ducks were reported close to the farms (maximum of six over
the total observation period), whereas other bird species were observed in much higher
numbers. Similar observations were obtained by Elbers et al. [20] at a different layer and
broiler farm in the Netherlands, despite the presence of waterways at a close distance
to the farms in this study. Results of the sensitivity analysis clearly indicate that these
parameters have high impact on model results (Supplementary Table S2; Figure 3).The
estimated infection risk due to the wind-supported transmission of HPAIv to poultry farms
via fecal particles from infected wild birds in the environment can thus be considered a
worst-case estimate.
Studies on the infectious dose of HPAIv that has a 50% probability of infection in
poultry, the BID50 , show that this value varies largely across HPAIv strains, poultry
species and inoculation routes [39,57,59]. Sergeev et al. [39] compared the aerosol route of
inoculation against other inoculation routes and concluded that the BID50 is lowest for the
aerosol route, with average values ranging from 0.8 to 1.2 EID50 . Swaye and Slemons [59]
reported BID50 values for the intranasal route, which vary widely from 1.2 to 4.7 EID50 ,
with an average value of 2.9 EID50 . Accounting for a 30-fold lower effectivity of the
intranasal route compared to the aerosol route [39], this average is only slightly higher
than the values given by Sergeev et al. [39]. While one could hypothesize that the BID50
in poultry is higher if the virus strain originates from ducks, this is not observed in the
values reported in literature [25,39]. In the what-if analysis, we tried to account for this
possibility and based our value for the BID50 on an estimated BID50 in chickens for a virus
strain isolated from tufted ducks [25,60]. This resulted in a >100-fold lower risk of infection
in poultry farms due to the wind-supported transmission of HPAIv derived from fecal
particles from infected wild birds with an estimated median Pin f of 1.4 × 10−5 (Figure 4).
Weather conditions during the Dutch bird-flu season allow for the prolonged persis-
tence of HPAIv in feces [51,52]. However, the probability that the feces of wild birds will
dry and aerosolize during the bird-flu season is likely to be low. Elbers [25] collected duck
feces in the field and evaluated the meteorological conditions under which the drying of
feces was observed (no precipitation and global sun irradiation ≥ 1000 J/cm2 ). We used
these observations to estimate the probability that the feces of wild birds will aerosolize
during the Dutch bird-flu season. Hardly any data were available to estimate the survival
of HPAIv during this process of drying and aerosolization (Fsurv_dry ). The few studies
available indicated that the virus in feces is likely to be inactivated by drying within
1–2 days [34,51]. Experimental aerosolization studies of influenza virus reported varying
levels of survival of the virus depending on the temperature and relative humidity applied
during the experiments [64,65]. We considered, however, that aerosolization under exper-
imental conditions is not likely to be representative of the drying process of feces under
field conditions. In the baseline calculations, we used the observed decline in virus titer by
Zarkov and Urumova [34] after one day of drying to parameterize Fsurv_dry . Considering
that the drying of feces in the field is likely to take at least one week [25] and that Zarkov
and Urumova [34] could no longer detect virus after two days of drying, this is likely to
result in an overestimate of the infection risk for poultry farms.
Several studies were available to estimate the HPAI prevalence in wild birds
(Prevwb ) [22–24,41,42,66–69]. These studies varied with respect to the sampling method,
sample size and matching of time and location with observed HPAI outbreaks in poultry
farms. We decided to only include studies from Europe during the bird-flu season, as
these were considered the most representative for our study. Extensive surveillance was
performed during the bird-flu seasons of 2014/15 and 2016/17 in the Netherlands [22–24],
indicating that the apparent prevalence in wild birds is low. We only included positive test
results for HPAIv strains. Also, only samples of Anseriformes were taken as the denomina-
tor, as no HPAIv was detected in any other order of birds. Smaller sample sizes were taken
during more recent studies in the UK and Italy [41,42]; these studies, however, indicated an
almost 10-fold higher prevalence of HPAI in Anseriformes compared to the earlier studies.
Pathogens 2024, 13, 571 14 of 18

This might be the result of the changing HPAI situation in wild birds in recent years, where
higher rates of morbidity and mortality have been observed, including in other orders
such as Charadriiformes [70]. When analyzing each of the European studies separately, an
increasing trend in apparent surveillance is indeed observed (Supplementary Figure S1).
When including the higher prevalence rates based on the UK and Italian studies in the risk
model (WI-1), the estimated probability of an infected poultry farm due to the wind-borne
transmission of fecal particles from infected wild birds (Pin f ) was increased 10-fold to a
median value of 0.022 (95% uncertainty interval: 7.3 × 10−4 to 0.44) per bird-flu season
(Figure 4), which equals an expected introduction in domestic poultry via this route once
every 45 years, which is still very low. We did not account for the spatial clustering of
infections in the model calculations, which might result in higher prevalence levels in wild
birds in some areas and lower prevalence levels (or even absence of infection) in other areas.
Although spatial clustering could result in a higher infection risk for individual farms, it
will also result in lower risk levels for other farms. Using the overall prevalence level in
wild birds in the model calculations thus resulted in an average infection risk for poultry
farms in the Netherlands, leveling out the possible variation across individual farms. This
will only have resulted in an underestimate of the risk if the spatial clustering of infections
in wild birds coincides with the spatial clustering of poultry farms.
The input values used for this quantitative risk assessment were largely based on farm-
ing systems, wild bird behavior and environmental conditions observed in the Netherlands
in the bird-flu season, which was defined as the period from October to April (European
winter time). Caution is warranted when extrapolating results to other regions in the world
or other seasons. For instance, Italy had a major HPAI outbreak in poultry during the 2017
summer period (July–November), and all primary outbreaks were attributed to indirect
contact with wild birds [71]. It is unclear if the wind-borne transmission of aerosolized
wild bird droppings could have played a role in this outbreak. Weather conditions during
the Italian outbreak would anyway have been much more favorable for the aerosolization
of fecal droppings than those during the Dutch bird-flu season. Also, the evolving epidemi-
ology of HPAI with an increasing number of mammalian species reported to be infected
could contribute to the contamination of the environment by secretions and excreta such as
feces. However, no quantitative data are available (yet) to estimate the probability of the
wind-supported transmission of HPAIv via aerosolized feces from infected mammals.

5. Concluding Remarks
Our model results indicate that the daily probability that the aerosolization of fecal
droppings from wild birds in the vicinity of poultry farms would result in the infection
of indoor-housed poultry in the Netherlands is extremely low. We estimated that this
introduction route will result in an infected poultry farm during the Dutch bird-flu season
once every 455 years (median value). Even under worst-case conditions (97.5 percentile
value), this probability is still very low (once every 17 years). These results bring us to
hypothesize that other risk factors, such as failures in strict and consistent compliance
to biosecurity measures at the farm, might possibly be of more importance in HPAIv
incursion on poultry farms [72–74]. Furthermore, this study provides guidance for the
prevention of any possible wind-supported transmission of HPAIv to poultry farms via
fecal particles from infected wild birds. The drying of HPAIv-contaminated fecal droppings
from wild birds is a prerequisite for aerosolization, and this practically only happens
during the bird-flu season when the droppings are deposited on concrete or stone-paved
surfaces surrounding the premises. The probability of the occurrence of a chain of drying of
HPAIv-contaminated wild bird feces, subsequent aerosolization and wind-borne transport
of still-infectious HPAIv through air inlets of a poultry house is very low. To make this
probability extremely low to negligible, it would be prudent for the poultry farmer to
regularly check for the presence of wild bird droppings on the paved flooring around
poultry houses and to safely remove these. This will also reduce the probability of the
Pathogens 2024, 13, 571 15 of 18

incidental introduction of HPAIv-contaminated wild bird droppings into the poultry house
by sticking to the boots of people walking on the premises and entering poultry anterooms.

Supplementary Materials: The following supporting information can be downloaded at: https://
[Link]/article/10.3390/pathogens13070571/s1, Table S1: Average values of daily mean,
minimum and maximum temperatures; daily precipitation; and daily global radiation per month in
De Bilt, The Netherlands, during the period 1993–2022 and the average number of days per month
that meet the criteria for the aerosolization of fecal droppings; Table S2: Correlation coefficients of
uncertain input parameters with the overall probability of at least one infected poultry farm during
the bird-flu season (Pin f ); Figure S1: Apparent prevalence of HPAI in wild waterfowl (mean and 95%
uncertainty interval).
Author Contributions: Conceptualization, A.R.W.E. and C.J.d.V.; Data curation: not applicable;
Formal analysis: C.J.d.V.; Funding acquisition: A.R.W.E.; Investigation: C.J.d.V. and A.R.W.E.;
Methodology: C.J.d.V.; Project administration: A.R.W.E.; Resources: C.J.d.V.; Software: C.J.d.V.;
Supervision: A.R.W.E.; Validation: C.J.d.V. and A.R.W.E.; Visualization: C.J.d.V.; Writing—original
draft: C.J.d.V. and A.R.W.E.; Writing—review and editing: C.J.d.V. and A.R.W.E. All authors have
read and agreed to the published version of the manuscript.
Funding: This research was funded by the Dutch Ministry of Agriculture, Nature and Food Quality
in the framework of the WOT project WOT-01-002-042. It is a follow-up on the One-Health for Food
(1H4F) public–private cooperation project Avian flu risk: Relative role of introduction routes and
biosecurity on and around poultry farms (grant number LWV 19081), which was funded by the Dutch
Ministry of Agriculture, Nature and Food Quality and the Dutch poultry foundation Avined.
Data Availability Statement: Data are contained within the article.
Conflicts of Interest: The authors declare no conflicts of interest.

References
1. Stallknecht, D.E.; Brown, J.D. Tenacity of avian influenza viruses. Rev. Sci. Tech. OIE 2009, 28, 59–67. [CrossRef]
2. Elbers, A.R.W.; Gonzales, J.L. Quantification of visits of wild fauna to a commercial free-range layer farm in the Netherlands
located in an avian influenza hot-spot area assessed by video-camera monitoring. Transbound. Emerg. Dis. 2019, 67, 661–677.
[CrossRef]
3. WOAH (World Organisation for Animal Health). Manual of Diagnostic Tests and Vaccines for Terrestrial Animals. Chapter 3.3.4.
Avian Influenza (Including Infection with High Pathogenicity Avian Influenza Viruses). 2023. Available online: [Link]
[Link]/en/what-we-do/standards/codes-and-manuals/terrestrial-manual-online-access/ (accessed on 24 June 2024).
4. Gonzales, J.L.; Elbers, A.R.W. Effective thresholds for reporting suspicions and improve early detection of avian influenza
outbreaks in layer chickens. Sci. Rep. 2018, 8, 8533. [CrossRef]
5. Pantin-Jackwood, M.J.; Swayne, D.E. Pathogenesis and pathobiology of avian influenza virus infection in birds. Rev. Sci. Tech.
OIE 2009, 28, 113–136. [CrossRef]
6. Bouwstra, R.; Heutink, R.; Bossers, A.; Harders, F.; Koch, G.; Elbers, A.R.W. Full-Genome Sequence of Influenza A(H5N8) Virus
in Poultry Linked to Sequences of Strains from Asia, the Netherlands, 2014. Emerg. Infect. Dis. 2015, 21, 872–874. [CrossRef]
7. Beerens, N.; Heutink, R.; Bergervoet, S.A.; Harders, F.; Bossers, A.; Koch, G. Multiple Reassorted Viruses as Cause of Highly
Pathogenic Avian Influenza A(H5N8) Virus Epidemic, the Netherlands, 2016. Emerg. Infect. Dis. 2017, 23, 1974–1981. [CrossRef]
8. Beerens, N.; Heutink, R.; Pritz-Verschuren, S.; Germeraad, E.A.; Bergervoet, S.A.; Harders, F.; Bossers, A.; Koch, G. Genetic
relationship between poultry and wild bird viruses during the highly pathogenic avian influenza H5N6 epidemic in the
Netherlands, 2017–2018. Transbound. Emerg. Dis. 2019, 66, 1370–1378. [CrossRef]
9. Velkers, F.C.; Manders, T.T.M.; Vernooij, J.C.M.; Stahl, J.; Slaterus, R.; Stegeman, J.A. Association of wild bird densities around
poultry farms with the risk of highly pathogenic avian influenza virus subtype H5N8 outbreaks in the Netherlands, 2016. Transb.
Emerg. Dis. 2021, 68, 76–87. [CrossRef]
10. Swayne, D.E.; Suarez, D.L.; Sims, L.D. Influenza. In Diseases of Poultry, 14th ed.; Swayne, D.E., Ed.; John Wiley & Sons, Inc.:
Hoboken, NJ, USA, 2020; pp. 210–256.
11. Germeraad, E.A.; Sanders, P.; Hagenaars, T.J.; de Jong, M.C.M.; Beerens, N.; Gonzales, J.L. Virus Shedding of Avian Influenza in
Poultry: A Systematic Review and Meta-Analysis. Viruses 2019, 11, 812. [CrossRef] [PubMed]
12. Koch, G.; Elbers, A.R.W. Outdoor ranging of poultry: A major risk factor for the introduction and development of High-
Pathogenicity Avian Influenza. Neth. J. Agric. Sci. 2006, 54, 179–194. [CrossRef]
13. Jonges, M.; van Leuken, J.; Wouters, I.; Koch, G.; Meijer, A.; Koopmans, M. Wind-Mediated Spread of Low-Pathogenic Avian
Influenza Virus into the Environment during Outbreaks at Commercial Poultry Farms. PLoS ONE 2015, 10, e0125401. [CrossRef]
Pathogens 2024, 13, 571 16 of 18

14. Torremorell, M.; Alonso, C.; Davies, P.R.; Raynor, P.C.; Patnayak, D.; Torchetti, M.; McCluskey, B. Investigation into the Airborne
Dissemination of H5N2 Highly Pathogenic Avian Influenza Virus During the 2015 Spring Outbreaks in the Midwestern United
States. Avian Dis. 2016, 60, 637–643. [CrossRef]
15. Scoizec, A.; Niqueux, E.; Thomas, R.; Daniel, P.; Schmitz, A.; Le Bouquin, S. Airborne Detection of H5N8 Highly Pathogenic
Avian Influenza Virus Genome in Poultry Farms, France. Front. Vet. Sci. 2018, 5, 15. [CrossRef]
16. Filaire, F.; Lebre, L.; Foret-Lucas, C.; Vergne, T.; Daniel, P.; Lelièvre, A.; de Barros, A.; Jbenyeni, A.; Bolon, P.; Paul, M.; et al. Highly
Pathogenic Avian Influenza A(H5N8) Clade 2.3.4.4b Virus in Dust Samples from Poultry Farms, France, 2021. Emerg. Infect. Dis.
2021, 28, 1446–1450. [CrossRef]
17. Ssematimba, A.; Hagenaars, T.J.; de Jong, M.C.M. Modelling the wind-borne spread of highly pathogenic avian influenza virus
between farms. PLoS ONE 2012, 7, e31114. [CrossRef]
18. Ypma, R.J.F.; Jonges, M.; Bataille, A.; Stegeman, A.; Koch, G.; van Boven, M.; Koopmans, M.; van Ballegooijen, W.M.; Wallinga, J.
Genetic Data Provide Evidence for Wind-Mediated Transmission of Highly Pathogenic Avian Influenza. J. Infect. Dis. 2012, 207,
730–735. [CrossRef]
19. James, J.; Warren, C.J.; De Silva, D.; Lewis, T.; Grace, K.; Reid, S.M.; Falchieri, M.; Brown, I.H.; Banyard, A.C. The Role of
Airborne Particles in the Epidemiology of Clade 2.3.4.4b H5N1 High Pathogenicity Avian Influenza Virus in Commercial Poultry
Production Units. Viruses 2023, 15, 1002. [CrossRef]
20. Elbers, A.R.W.; Gonzales, J.L.; Koene, M.G.J.; Germeraad, E.A.; Hakze-van der Honing, R.W.; van der Most, M.; Rodenboog,
H.; Velkers, F.C. Monitoring Wind-Borne Particle Matter Entering Poultry Farms Via the Air-Inlet: Highly Pathogenic Avian
Influenza Virus and Other Pathogens Risk. Pathogens 2022, 11, 1534. [CrossRef]
21. Lumivero. @RISK, Probabilistic Risk Analysis in Excel. 2023. Available online: [Link]
(accessed on 18 December 2023).
22. Verhagen, J.H.; van der Jeugd, H.P.; Nolet, B.A.; Slaterus, R.; Kharitonov, S.P.; de Vries, P.P.; Vuong, O.; Majoor, F.; Kuiken, T.;
Fouchier, R.A. Wild bird surveillance around outbreaks of highly pathogenic avian influenza A(H5N8) virus in the Netherlands,
2014, within the context of global flyways. Eurosurveillance 2015, 20, 21069. [CrossRef]
23. Poen, M.J.; Verhagen, J.H.; Manvell, R.J.; Brown, I.; Bestebroer, T.M.; van der Vliet, S.; Vuong, O.; Scheuer, R.D.; van der Jeugd,
H.P.; Nolet, B.A.; et al. Lack of virological and serological evidence for continued circulation of highly pathogenic avian influenza
H5N8 virus in wild birds in the Netherlands, 14 November 2014 to 31 January 2016. Eurosurveillance 2016, 21, 30349. [CrossRef]
[PubMed]
24. Poen, M.J.; Bestebroer, T.M.; Vuong, O.; Scheuer, R.D.; van der Jeugd, H.P.; Kleyheeg, E.; Eggink, D.; Lexmond, P.; van den
Brand, J.M.A.; Begeman, L.; et al. Local amplification of highly pathogenic avian influenza H5N8 viruses in wild birds in the
Netherlands, 2016 to 2017. Eurosurveillance 2018, 23, 17-00449. [CrossRef] [PubMed]
25. Elbers, A.R.W. Omgevingstransmissie van Aviaire Influenza Virus Door de Lucht via Wilde Watervogels Naar Commercieel
Gehouden Pluimvee. Wageningen Bioveterinary Research, Report 2128494; 2021; 82p. Available online: [Link]
556247 (accessed on 25 January 2024). (In Dutch).
26. Veen, J.; Brouwer, J.; Atkinson, P.; Bilgin, C.; Blew, J.; Eksioğlu, S.; Hoffmann, M.; Nardelli, R.; Spina, F.; Tendi, C.; et al.
Ornithological Data Relevant to the Spread of Avian Influenza in Europe (Phase 2): Further Identification and First Field Assessment of Higher
Risk Species; Wetlands International: Wageningen, The Netherlands, 2007; 60p, Available online: [Link]
publication/ornithological-data-relevant-to-the-spread-of-avian-influenza-in-europe-phase-2/ (accessed on 14 February 2024).
27. Burns, T.E.; Ribble, C.; Stephen, C.; Kelton, D.; Toews, L.; Osterhold, J.; Wheeler, H. Use of observed wild bird activity on poultry
farms and a literature review to target species as high priority for avian influenza testing in 2 regions of Canada. Can. Vet. J. 2012,
53, 158–166. [PubMed]
28. Elbers, A.R.W.; Gonzales, J.L. Efficacy of an automated laser for reducing wild bird visits to the free range area of a poultry farm.
Sci. Rep. 2021, 11, 12779. [CrossRef] [PubMed]
29. Le Gall-Ladevèze, C.; Guinat, C.; Fievet, P.; Vollot, B.; Guérin, J.L.; Cappelle, J.; Le Loc’h, G. Quantification and characterisation
of commensal wild birds and their interactions with domestic ducks on a free-range farm in southwest France. Sci. Rep. 2022,
12, 9764. [CrossRef] [PubMed]
30. Martelli, L.; Fornasiero, D.; Scarton, F.; Spada, A.; Scolamacchia, F.; Manca, G.; Mulatti, P. Study of the Interface between Wild
Bird Populations and Poultry and Their Potential Role in the Spread of Avian Influenza. Microorganisms 2023, 11, 2601. [CrossRef]
[PubMed]
31. Andrikovics, S.; Gere, G.; Juhász, J.; Lakatos, G. Mallard waste production and effects on water quality in small water bodies. In
“Limnology and Aquatic Birds”. Abstracts and Selected Papers from the 4th Conference of the Societas Internationalis Limnologiae (SIL)
Aquatic Birds Working Group; Canadian Wildlife Service Technical Report Series No. 474 Atlantic Region; Environment Canada:
Ottowa, ON, Canada, 2006; pp. 125–130.
32. Kear, J. The Agricultural Importance of Wild Goose Droppings. The Wildfowl Trust Fourteenth Annual Report, 1961–1962. 1963,
pp. 72–77. Available online: [Link] (accessed on 9 February 2024).
33. KNMI. Daggegevens van Het Weer in Nederland, Station 260 De Bilt. Koninklijk Nederlands Meteorologisch Instituut, 2023.
Available online: [Link] (accessed on 19 December 2023). (In Dutch).
34. Zarkov, I.S.; Urumova, V.S. Effects of humidity and temperature on avian influenza virus H6N2 persistence in faecal samples
from experimentally infected ducks (Anas platyrynchos). Revue Méd. Vét. 2013, 164, 343–347.
Pathogens 2024, 13, 571 17 of 18

35. Lighthart, B.; Mohr, A.J. Estimating downwind concentrations of viable airborne microorganisms in dynamic atmospheric
conditions. Appl. Environ. Microbiol. 1987, 53, 1580–1583. [CrossRef] [PubMed]
36. Harper, G.J. Airborne micro-organisms: Survival tests with four viruses. J. Hyg. 1961, 59, 479–486. [CrossRef] [PubMed]
37. Winkel, A.; Mosquera, L.J.; Groot Oberkampf, P.W.G.; Ogink, N.W.M.; Aarnink, A.J.A. Emissions of particulate matter from
animal houses in the Netherlands. Atmos. Environ. 2015, 111, 202–212. [CrossRef]
38. Zhao, Y.; Richardson, B.; Takle, E.; Chai, L.; Schmitt, D.; Xin, H. Airborne transmission may have played a role in the spread of
2015 highly pathogenic avian influenza outbreaks in the United States. Sci. Rep. 2019, 9, 11755. [CrossRef]
39. Sergeev, A.A.; Demina, O.K.; Pyankov, O.V.; Pyankova, O.G.; Agafonov, A.P.; Kiselev, S.A.; Agranovski, I.E.; Sergeev, A.A.; Shikov,
A.N.; Shishkina, L.N.; et al. Infection of chickens caused by avian influenza virus A/H5N1 delivered by aerosol and other routes.
Transbound. Emerg. Dis. 2013, 60, 159–165. [CrossRef]
40. Agrimatie. Informatie over de Agrosector. Pluimveehouderij, 2023. Available online: [Link]
subpubID=2232&sectorID=2249 (accessed on 9 February 2024). (In Dutch).
41. Gobbo, F.; Fornasiero, D.; De Marco, M.A.; Zecchin, B.; Mulatti, P.; Delogu, M.; Terregino, C. Active Surveillance for Highly
Pathogenic Avian Influenza Viruses in Wintering Waterbirds in Northeast Italy, 2020–2021. Microorganisms 2021, 9, 2188. [CrossRef]
42. Wade, D.; Ashton-Butt, A.; Scott, G.; Reid, S.M.; Coward, V.; Hansen, R.D.E.; Banyard, A.C.; Ward, A.I. High pathogenicity avian
influenza: Targeted active surveillance of wild birds to enable early detection of emerging disease threats. Epidemiol. Infect. 2023,
151, e15. [CrossRef]
43. Verhagen, J.H.; Fouchier, R.A.M.; Lewis, N. Highly Pathogenic Avian Influenza Viruses at the Wild-Domestic Bird Interface in
Europe: Future Directions for Research and Surveillance. Viruses 2021, 13, 212. [CrossRef]
44. Sun, Y.; Zhang, T.; Zhao, X.; Qian, J.; Jiang, M.; Jia, M.; Xu, Y.; Yang, W.; Feng, L. High activity levels of avian influenza upwards
2018–2022: A global epidemiological overview of fowl and human infections. One Health 2023, 16, 100511. [CrossRef]
45. WAHIS. Avian Influenza. World Organization for Animal Health. 2024. Available online: [Link]
avian-influenza/ (accessed on 9 February 2024).
46. Purcell, S.L. The Significance of Waterfowl Feces as a Source of Nutrients to Algae in a Prairie Wetland. Master’s Thesis, University
of Manitoba, Winnipeg, MB, Canada, 1999; 118p. Available online: [Link]
[Link] (accessed on 9 February 2024).
47. Nnaji, J.C. Physico-Chemical Quality and Plankton Density of Water in Duck-Fish Production Systems. Am. Chem. Sci. J. 2014, 4,
975–982. [CrossRef]
48. Gere, G.; Andrikovics, S.G. Feeding of ducks and their effects on water quality. Hydrobiologica 1994, 279, 157–161. [CrossRef]
49. Sovon. Aantallen Ganzen. Available online: [Link]
aantallen-ganzen (accessed on 9 February 2024).
50. Beerens, N.; Germeraad, E.A.; Venema, S.; Verheij, E.; Pritz-Verschuren, S.B.E.; Gonzales, J.L. Comparative pathogenicity and
environmental transmission of recent highly pathogenic avian influenza H5 viruses. Emerg. Microb. Infect. 2021, 10, 97–108.
[CrossRef]
51. Shortridge, K.F.; Zhou, N.N.; Guan, Y.; Gao, P.; Ito, T.; Kawaoka, Y.; Kodihalli, S.; Krauss, S.; Markwell, D.; Murti, K.G.; et al.
Characterization of avian H5N1 influenza viruses from poultry in Hong Kong. Virology 1998, 252, 331–342. [CrossRef]
52. Kurmi, B.; Murugkar, H.V.; Nagarajan, S.; Tosh, C.; Dubey, S.C.; Kumar, M. Survivability of Highly Pathogenic Avian Influenza
H5N1 Virus in Poultry faeces at Different Temperatures. Indian. J. Virol. 2013, 24, 272–277. [CrossRef]
53. Sedlmaier, N.; Hoppenheidt, K.; Krist, H.; Lehmann, S.; Lang, H.; Büttner, M. Generation of avian influenza virus (AIV)
contaminated fecal fine particulate matter (PM(2.5)): Genome and infectivity detection and calculation of immission. Vet.
Microbiol. 2009, 139, 156–164. [CrossRef]
54. Gonzales, J.L.; Stegeman, J.A.; Koch, G.; de Wit, S.J.; Elbers, A.R.W. Rate of introduction of a low pathogenic avian influenza virus
infection in different poultry production sectors in the Netherlands. Influ. Other Respir. Viruses 2013, 7, 6–10. [CrossRef]
55. Gonzales, J.L.; Hennen, W.H.G.J.; Petie, R.; de Freitas Costa, E.; Beerens, N.; Slaterus, R.; Kuiken, T.; Stahl, J.; Elbers, A.R.W.
Risicofactoren Voor Introductie van HPAI-virus op Nederlandse Commerciële Pluimveebedrijven, 2014–2022. Wageningen
Bioveterinary Research Report 2211632. 2022. Available online: [Link] (accessed on 9 February 2024).
(In Dutch)
56. Brown, J.D.; Stallknecht, D.E.; Valeika, S.; Swayne, D.E. Susceptibility of wood ducks to H5N1 highly pathogenic avian influenza
virus. J. Wildl. Dis. 2007, 43, 660–667. [CrossRef]
57. Pantin-Jackwood, M.J.; Spackman, E.; Leyson, C.; Youk, S.; Lee, S.A.; Moon, L.M.; Torchetti, M.K.; Killian, M.L.; Lenoch, J.B.;
Kapczynski, D.R.; et al. Pathogenicity in Chickens and Turkeys of a 2021 United States H5N1 Highly Pathogenic Avian Influenza
Clade 2.3.4.4b Wild Bird Virus Compared to Two Previous H5N8 Clade [Link] Viruses. Viruses 2023, 15, 2273. [CrossRef]
58. Spackman, E.; Pantin-Jackwood, M.J.; Lee, S.A.; Prosser, D. The pathogenesis of a 2022 North American highly pathogenic clade
2.3.4.4b H5N1 avian influenza virus in mallards (Anas platyrhynchos). Avian Path 2023, 52, 219–228. [CrossRef]
59. Swayne, D.E.; Slemons, R.D. Using mean infectious dose of high- and low-pathogenicity avian influenza viruses originating from
wild duck and poultry as one measure of infectivity and adaptation to poultry. Avian Dis. 2008, 52, 455–460. [CrossRef]
60. Leyson, C.; Youk, S.S.; Smith, D.; Dimitrov, K.; Lee, D.H.; Larsen, L.E.; Swayne, D.E.; Pantin-Jackwood, M.J. Pathogenicity and
genomic changes of a 2016 European H5N8 highly pathogenic avian influenza virus (clade [Link]) in experimentally infected
mallards and chickens. Virology 2019, 537, 172–185. [CrossRef]
Pathogens 2024, 13, 571 18 of 18

61. Seo, I.-H.; Lee, I.-B.; Moon, O.-K.; Jung, N.-S.; Lee, H.-J.; Hong, S.-W.; Kwon, K.-S.; Bitog, J.P. Prediction of the spread of highly
pathogenic avian influenza using a multifactor network: Part 1—Development and application of computational fluid dynamics
simulations of airborne dispersion. Biosys Eng. 2014, 121, 160–176. [CrossRef]
62. Zhao, Y.; Aarnink, A.J.A.; De Jong, M.C.M.; Groot Koerkamp, P.W.G. Airborne Microorganisms From Livestock Production
Systems and Their Relation to Dust. Crit. Rev. Environ. Sci. Technol. 2014, 44, 1071–1128. [CrossRef]
63. Van Leuken, J.P.G.; Swart, A.N.; Havelaar, A.H.; Van Pul, A.; Van der Hoek, W.; Heederik, D. Atmospheric dispersion modelling
of bioaerosols that are pathogenic to humans and livestock—A review to inform risk assessment studies. Microb. Risk Anal. 2016,
1, 19–39. [CrossRef]
64. Hagenaars, T.; Gonzales, J.; de Vos, C.; Aarnink, A. Modelling Emission of Bio-Aerosols Carrying Zoonotic Microorganisms from
Livestock Houses: Quantification Data and Knowledge Gaps; Report WBVR-1712754; Wageningen University & Research: Lelystad, The
Netherlands, 2017; 34p. Available online: [Link] (accessed
on 9 February 2024).
65. Belser, J.A.; Pulit-Penaloza, J.A.; Brock, N.; Creager, H.M.; Gustin, K.M.; Tumpey, T.M.; Maines, T.R. Inherent Heterogeneity of
Influenza A Virus Stability following Aerosolization. Appl. Environ. Microbiol. 2022, 88, e0227121. [CrossRef]
66. Keawcharoen, J.; van den Broek, J.; Bouma, A.; Tiensin, T.; Osterhaus, A.D.; Heesterbeek, H. Wild birds and increased transmission
of highly pathogenic avian influenza (H5N1) among poultry, Thailand. Emerg. Infect. Dis. 2011, 17, 1016–1022. [CrossRef]
[PubMed]
67. Soliman, A.; Saad, M.; Elassal, E.; Amir, E.; Plathonoff, C.; Bahgat, V.; El-Badry, M.; Ahmed, L.S.; Fouda, M.; Gamaleldin, M.; et al.
Surveillance of avian influenza viruses in migratory birds in Egypt, 2003–2009. J Wildl Dis 2012, 48, 669–675. [CrossRef] [PubMed]
68. Ip, H.S.; Dusek, R.J.; Bodenstein, B.; Torchetti, M.K.; DeBruyn, P.; Mansfield, K.G.; DeLiberto, T.; Sleeman, J.M. High Rates of
Detection of Clade [Link] Highly Pathogenic Avian Influenza H5 Viruses in Wild Birds in the Pacific Northwest During the
Winter of 2014-15. Avian Dis. 2016, 60 (Suppl. 1), 354–358. [CrossRef] [PubMed]
69. Lee, E.K.; Kang, H.M.; Song, B.M.; Lee, Y.N.; Heo, G.B.; Lee, H.S.; Lee, Y.J.; Kim, J.H. Surveillance of avian influenza viruses in
South Korea between 2012 and 2014. Virol. J. 2017, 14, 54. [CrossRef] [PubMed]
70. Caliendo, V.; Kleyheeg, E.; Beerens, N.; Camphuysen, K.C.J.; Cazemier, R.; Elbers, A.R.W.; Fouchier, R.A.M.; Kelder, L.; Kuiken,
T.; Leopold, M.; et al. Effect of 2020–21 and 2021–22 Highly Pathogenic Avian Influenza H5 Epidemics on Wild Birds, the
Netherlands. Emerg. Infect. Dis. 2024, 30, 50–57. [CrossRef] [PubMed]
71. EFSA (European Food Safety Authority); ECDC (European Centre for Disease Prevention and Control); EURL (European
Reference Laboratory on Avian Influenza); Brown, I.; Kuiken, T.; Mulatti, P.; Smietanka, K.; Staubach, C.; Stroud, D.; Therkildsen,
O.R.; et al. Scientific report: Avian influenza overview September–November 2017. EFSA J. 2017, 15, 5141. [CrossRef]
72. Racicot, M.; Venne, D.; Durivage, A.; Vaillantcourt, J.-P. Description of 44 biosecurity errors while entering and exiting poultry
barns based on video surveillance in Quebec, Canada. Prev. Vet. Med. 2011, 100, 193–199. [CrossRef] [PubMed]
73. Velkers, F.C.; Blokhuis, S.J.; Veldhuis Kroeze, E.J.B.; Burt, S.A. The role of rodents in avian influenza outbreaks in poultry farms:
A review. Vet. Quart. 2017, 37, 182–194. [CrossRef] [PubMed]
74. Racicot, M.; Cardinal, A.-M.; Tremblay, D.; Vaillancourt, J.-P. Technologies monitoring and improving biosecurity compliance in
barn anterooms. Front. Vet. Sci. 2022, 9, 1005144. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.