Assessing the inter-annual variability of separation distances around odour sources to protect the residents from odour annoyance

Article (PDF Available)inJournal of Environmental Sciences 79:11-24 · May 2019with 189 Reads
DOI: 10.1016/j.jes.2018.09.018
Cite this publication
Abstract
In recent years, there has been a growing concern about potential impacts on public health and wellbeing due to exposure to environmental odour. Separation distances between odour-emitting sources and residential areas can be calculated using dispersion models, as a means of protecting the neighbourhood from odour annoyance. This study investigates the suitability of using one single year of meteorological input data to calculate reliable direction-dependent separation distances. Accordingly, we assessed and quantified the inter-annual variability of separation distances at two sites with different meteorological conditions, one in Brazil and the other in Austria. A 5-year dataset of hourly meteorological observations was used for each site. Two odour impact criteria set in current regulations were selected to explore their effect on the separation distances. The coefficient of variation was used as a statistical measure to characterise the amount of annual variation. Overall, for all scenarios, the separation distances had a low degree of inter-annual variability (mean coefficient of variation values from 8% to 21%). Reasonable agreements from year to year were therefore observed at the two sites under investigation, showing that one year of meteorological data is a good compromise to achieve reliable accuracy. This finding can provide a more cost-effective solution to calculate separation distances in the vicinity of odour sources.
Assessing the inter-annual variability of separation
distances around odour sources to protect the
residents from odour annoyance
Marlon Brancher
1,2,
, Martin Piringer
3
, Davide Franco
1
, Paulo Belli Filho
1
,
Henrique De Melo Lisboa
1
, Günther Schauberger
2
1. Postgraduate Program in Environmental Engineering (PPGEA), Federal University of Santa Catarina (UFSC),88040900, Florianópolis, Brazil.
CAPES Foundation scholarship holder (main author), Brazilian Ministry of Education, Brasília, Brazil
2. WG Environmental Health, Department of Biomedical Sciences, University of Veterinary Medicine Vienna, Veterinärplatz 1, A-1210 Vienna,
Austria
3. Central Institute for Meteorology and Geodynamics, Hohe Warte 38, A-1190 Vienna, Austria
ARTICLE INFO ABSTRACT
Article history:
Received 23 November 2017
Revised 15 September 2018
Accepted 18 September 2018
Available online 28 September 2018
In recent years, there has been a growing concern about potential impacts on public health
and wellbeing due to exposure to environmental odour. Separation distances between
odour-emitting sources and residential areas can be calculated using dispersion models, as
a means of protecting the neighbourhood from odour annoyance. This study investigates
the suitability of using one single year of meteorological input data to calculate reliable
direction-dependent separation distances. Accordingly, we assessed and quantified the
inter-annual variability of separation distances at two sites with different meteorological
conditions, one in Brazil and the other in Austria. A 5-year dataset of hourly meteorological
observations was used for each site. Two odour impact criteria set in current regulations
were selected to explore their effect on the separation distances. The coefficient of variation
was used as a statistical measure to characterise the amount of annual variation. Overall,
for all scenarios, the separation distances had a low degree of inter-annual variability
(mean coefficient of variation values from 8% to 21%). Reasonable agreements from year to
year were therefore observed at the two sites under investigation, showing that one year of
meteorological data is a good compromise to achieve reliable accuracy. This finding can
provide a more cost-effective solution to calculate separation distances in the vicinity of
odour sources.
© 2018 The Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences.
Published by Elsevier B.V.
Keywords:
Environmental odour
Odour annoyance
Impact assessment
Dispersion modelling
Regulatory criteria
Separation distance
Introduction
Odour emissions have become a topic of increasing interest in
both developed and developing countries worldwide. For
many years now, environmental odour is the leading cause
of public complaints reported to authorities regarding air
quality (Hayes et al., 2014; Henshaw et al., 2006). Indeed,
environmental odour is an ambient stressor since it is
physically perceptible, negatively valued, unpredictable, un-
controllable and entails moderate adjustments (Campbell,
1983). In many jurisdictions, environmental odour is already
handled as an air pollutant subject to specific legislation
JOURNAL OF ENVIRONMENTAL SCIENCES 79 (2019) 1124
Corresponding author. E-mail: marlon.b@posgrad.ufsc.br. (Marlon Brancher).
https://doi.org/10.1016/j.jes.2018.09.018
1001-0742/© 2018 The Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences. Published by Elsevier B.V.
Available online at www.sciencedirect.com
ScienceDirect
www.elsevier.com/locate/jes
(Brancher et al., 2017). In the policy context of the European
Union (EU), it is worth mention that, for the first time, odour
has been considered in the Joint Research Centre Reference
Report on monitoring of emissions to air and water from
installations covered by the Industrial Emissions Directive
2010/75/EU (Brinkmann et al., 2018).
Odour exposure has been associated with health issues,
affecting both the physiological and psychosocial status
(Sucker et al., 2009). Physiological health symptoms comprise,
for instance, headache, nausea, respiratory complications,
tiredness, eye irritation and palpitations (Schiffman and
Williams, 2005). Because dilution occurs in the atmosphere,
odours often reach the population at concentrations far below
toxicity thresholds, making direct toxicological mechanisms
unlikely to explain the association between exposures and
symptoms (Blanes-Vidal et al., 2014). Interestingly, epidemi-
ological studies have shown indirect mechanisms in which
psychosocial responses (i.e., odour annoyance) mediate phys-
ical symptom reporting (Blanes-Vidal, 2015). Hence, odour
annoyance has been identified as one of the most important
effects due to exposure to malodour (Cantuaria et al., 2017;
Shusterman, 1992).
A suitable parameter to describe the influence of an odour
source on the nearby residential area is the separation
distance intended to embrace the area within which odour
has the potential to cause annoyance (Piringer et al., 2015;
Schauberger et al., 2000). The separation distance approach is
part of an integrated multi-tool strategy recommended by
Brancher et al. (2017) to manage environmental odours. The
separation distance divides the area around odour-emitting
facilities into two zones: (i) a zone beyond the separation
distance where odour annoyance is likely to be avoided and
(ii) a zone closer than the separation distance where loss of
public amenity can be expected (Piringer et al., 2016). The
separation distance can be fixed, given by pre-established
distances; or variable, determined as direction-dependent
distances on a case-by-case basis typically using dispersion
models. In this instance, time series of ambient odour
concentrations predicted by dispersion models are evaluated
by so-called odour impact criteria OIC. As a result, separation
distances between an odour source and residential areas are
calculated in a direction-dependent manner. Thereby,
direction-dependent separation distances are the ultimate
measure accounting for the entire chain from the odour
emission rate, the dilution in the atmosphere, and the
evaluation of the time series of ambient odour concentrations
by the OIC (Sommer-Quabach et al., 2014).
Nowadays, a wide range of OIC is in force, which shows
that the assessment of odour annoyance varies greatly
(Brancher et al., 2017; Griffiths, 2014; Sommer-Quabach et al.,
2014). The OIC can be specified by three components: (i) the
odour concentration threshold C
t
(given in European odour
units per cubic metre ou
E
/m
3
or equivalent units), (ii) the
percentile rank value P(also specified as exceedance proba-
bility 100 P), and (iii) the averaging time A
t
. Typically,
dispersion models predict hourly time series of ambient
odour concentrations (De Melo Lisboa et al., 2006; Drew
et al., 2007). If the OIC are specified for an A
t
shorter than
1 hr, then a coefficient called peak-to-mean factor Fcomes
into play. The Fis used to estimate concentrations for shorter
averaging times than that equivalent to the model output, as
an attempt to mimic the odour perception of the human nose
(Schauberger et al., 2012). A pioneering concept structured in
the computation of ambient odour concentration variances to
determine this Fhas been recently presented (Ferrero et al.,
2017; Oettl and Ferrero, 2017; Oettl et al., 2018).
It is well known that, together with emissions, meteoro-
logical data play a central role in dispersion modelling (Capelli
et al., 2013). Unsurprisingly, a critical methodological step in
the calculation of separation distances using dispersion
models is the acquisition, pre-processing and validation of
meteorological data. In this regard, a key challenge is to
calculate representative distances, while the meteorological
input data is reduced. In addition, international regulatory
requirements for odour dispersion modelling differ consider-
ably in the sense that odour studies can be conducted on a
monthly, annual or multi-year basis over the meteorological
input data (Brancher et al., 2017). The year-to-year variation of
odour contour lines, in particular, has been briefly touched by
few technical reports (ERM, 2012; Featherston et al., 2014;
GHD, 2015). However, the inter-annual variability of direction-
dependent separation distances to avoid odour annoyance
has yet to be explored. This knowledge is of relevance not only
for future research but also for improving current odour
regulations.
In this work, we investigated whether one single year of
meteorological input data is enough to calculate reliable
separation distances. For this purpose, we assessed and
quantified the inter-annual variability of separation distances
at two sites with different meteorological conditions. The
calculations were undertaken for São José dos Pinhais (Brazil,
near Curitiba) and Groß-Enzersdorf (Austria, near Vienna).
Five years of hourly meteorological observations were used
for each site. Modelling scenarios consider a point source with
constant odour emission rate (annual mean value). Two
national OIC were selected as references to calculate the
separation distances.
1. Material and methods
1.1. Description of sites
The investigation was carried out at two sites, one in Brazil
and the other in Austria, where yearly datasets of meteoro-
logical observations are available. Furthermore, we chose
these sites because they meet the terrain requirements for
performing modelling studies using a Gaussian plume model,
and are representative of the odour sources found in the
surrounding areas. São José dos Pinhais (25.555° S, 49.132° W,
906 m ASL; close to Curitiba, the capital of the state of Paraná)
is the location of the odour source in Brazil. This site is within
flat and elevated terrain. Land uses such as farmland,
remaining forest, woody wetlands, low residential areas, and
a few industries can be found scattered around the emission
source in nearly all directions. The Austrian site is located in
Groß-Enzersdorf (48.203° N, 16.564° E, 151 m ASL), district of
Gänserndorf in Lower Austria, and east of Vienna. It is within
mainly flat terrain, typically farmland. However, surrounding
residential dwellings and a few industries (mainly in the
12 JOURNAL OF ENVIRONMENTAL SCIENCES 79 (2019) 1124
southwesterly and southeasterly directions) are present about
350500 m from the source.
1.2. Characterisation of the odour source
Among the sources of uncertainty in dispersion modelling,
algorithms that deal with the source typology are prominent.
According to Pullen and Vawda (2007), predicted concentra-
tions are fundamentally more accurate for single stacks, first
hand. So, we chose a single point source for the investigation.
The odour emission rate is constant, continuous, and station-
ary in time, with an annual mean value of 17,500 ou
E
/sec. A
variety of emission factors can be found in the German
guideline VDI 3894 Part 1 (2011) to translate this odour
emission rate into a typical livestock building. The geometry
of the source is presumed circular, with a height of 6 m from
the ground, inner diameter of 1.2 m, and vertical release. The
exit velocity is 3.0 m/s, and the gas temperature is 35 °C. This
source configuration attempts to replicate the emission from
a typical mechanically ventilated livestock building. Table 1
summarises the odour source parameters assumed for the
dispersion calculations.
1.3. Atmospheric dispersion modelling
The U.S. Environmental Protection Agency (U.S. EPA) regula-
tory air quality model, AERMOD Modelling System, was used.
The model has also been adopted worldwide and accepted for
regulatory demonstrations by several other environmental
agencies. Essentially, the modelling system consists of three
modules: the AERMOD dispersion model itself, the AERMET
meteorological processor, and the AERMAP terrain processor.
AERMOD is fundamentally a steady-state Gaussian plume
model with algorithms based on planetary boundary layer
turbulence structure and scaling concepts. AERMOD is the U.S.
EPA preferred/recommended software for demonstrating
regulatory compliance for short-range transport of air pollut-
ants (<50 km), including treatment of surface and elevated
sources for simple and complex terrain. The steady-state
concept assumes that over the model time step, the emis-
sions, meteorology, and other model inputs, are constant all
over the model domain. This assumption results in a resolved
plume with the emissions distributed throughout the plume
according to a Gaussian distribution (U.S. EPA, 2017). Com-
prehensive model principles and formulation can be found
elsewhere (Cimorelli et al., 2005; Perry et al., 2005; U.S. EPA,
2016a). The model is used with the graphic user interface
AERMOD View 9.4.0, version 16216r (Lakes Environmental
Software, Ontario, Canada). The suitability of AERMOD for the
scenarios of this work is justified based on: (i) the topographic
features and meteorological conditions of the areas being
modelled, (ii) the detail and accuracy of the primary inputs
(meteorology and emission) required for a refined model, (iii)
the way complexities of atmospheric processes are handled
by the model, (iv) the need to apply a recognised model
typically used in the permitting process, (v) the efficiency
relationship between computational time and reasonable
accuracy; and finally, (vi) the resources available to apply
such desktop software. The modelling protocol follows
current default regulatory options consistent with the Guide-
line on Air Quality Models (U.S. EPA, 2017), except where
stated and justified otherwise.
The model domain, at both sites, consists of a circular area
of 750 m radius centred on the source. The domain is
discretised using a polar grid network. Receptors are distrib-
uted along 72 radial directions, with the initial direction at 0°
and with moves of 5° clockwise, over 20 concentric rings. The
nearest and the last ring are placed 50 m and 750 m from the
source, respectively. The distance of the nearest ring from the
source allows for the satisfactorily accurate calculation of
odour concentrations because to date Gaussian plume models
are inherently more uncertain for receptors very close to the
source. 1440 receptor points are placed for the calculation of
odour concentrations for each site. The design of the receptor
network is supported by the receptor density and location and
not because of the total number of receptors. The receptor grid
is progressively more resolved near the source, which proves
to be the hotspot of maximum impact for our scenarios
(highest predicted concentrations). Both the receptor grid and
the size of the domain influence the computational model
time. Consequently, their assumptions reflect the level of
detail needed for the output. In other words, the choices
capture the extent of the odour impact adequately. Receptors
are positioned 1.5 m above the ground at the average height of
the human nose. No background concentrations are assumed.
The influence of a possible building downwash effect is not
considered. Both sites are classified as rural, so the rural
dispersion option was selected.
Terrain elevation data are obtained from the Shuttle Radar
Topography Mission (SRTM) conducted by the National
Aeronautics and Space Administration (NASA). The data for
the model domains are in SRTM1 (spacing for individual data
points is 1 arc-second), which corresponds to about 30 m
resolution. Accordingly, the digital elevation models are built
using the AERMAP terrain processor, version 11103 (U.S. EPA,
2016b). The domain at the Brazilian site has elevations from
near 883 m to 911 m ASL. For the Austrian site, elevations
from near 147 to 153 m ASL can be found (Fig. 1). The model
can account for elevated orographic effects. This is performed
by inputting elevated receptor heights to model the effects of
terrain above or below stack base. Because of the orography of
the sites, the elevated option is used to characterise the
terrain effects. Elevations/hill heights are assigned to recep-
tors and the odour source by AERMAP.
Land surface characteristics (i.e., albedo, Bowen ratio, and
surface roughness length) around the meteorological towers
Table 1 Characteristics of odour source assumed for
dispersion calculations.
Parameter Description Unit
Source type Point .
Release type Vertical .
Geometry Circular .
Release height 6 m
Inner diameter 1.2 m
Gas temperature 35 °C
Exit velocity 3.0 m/sec
Volume flow rate 3.39 m
3
/sec
Odour emission rate 17,500 ou
E
/sec
13JOURNAL OF ENVIRONMENTAL SCIENCES 79 (2019) 1124
were determined by the procedures of AERSURFACE (U.S. EPA,
2008) and AERMET User's Guides (U.S. EPA, 2016c) using the
AERSURFACE utility (version 13016). For the Brazilian site,
surface characteristics were extracted from the Global Land
Cover Characterisation GLCC for South America, with a
resolution of 1 km. For the Austrian site, the CORINE
CLC2006 database with 0.1 km resolution was used. In order
to improve the resolution and to homogenise the resolution of
the surface characteristics data, a refinement was performed
using the tool called Land Use Creator available for AERMOD-
View. The surface roughness was determined by sectoring (12
angular sectors of 30°) with a default upwind distance of 1 km
radius relative to the meteorological tower location. Albedo
and the Bowen ratio values were determined based on a
default area of 10 × 10 km also centred on the meteorological
tower. Monthly values were assigned to account for a
temporal change of surface characteristics.
The adjusted surface friction velocity technique (ADJ_U*) is
currently considered a default regulatory option in the
AERMOD (U.S. EPA, 2016c). As a previous diagnostic evalua-
tion, we ran the model for both sites to verify the sensitivity
that the ADJ_U* option could exert on the predicted concen-
trations. The model results at both sites showed high linear
correlation (R
2
0.99) for ADJ_U* turned on against ADJ_U*
turned off, for the two OIC tested. Residual plots further
exhibited the goodness of the linear fit. Moreover, by visual
scrutiny on the shape of the separation distances (contour
plots), no changes in the envelope were found. Once model
outputs at both sites were well-correlated for ADJ_U* turned
on against ADJ_U* turned off for the selected OIC, the usage of
the ADJ_U* option becomes non-compulsory for our modelling
scenarios. In response, we ran the model with ADJ_U* turned
off to demonstrate the full range of atmospheric stability
estimated by AERMET.
1.4. Meteorology
1.4.1. Brazilian site
Surface meteorological observations of 1 hr time-steps for wind
direction (W
d
), wind speed (W
s
), air temperature (T), atmospheric
pressure (P
atm
), and cloud cover (CC) were obtained from the
National Oceanic and Atmospheric Administration (NOAA)
database for Afonso Pena International Airport SBCT (25.531° S,
49.167° W). At airports, W
d
is typically recorded to the nearest 10°
for that hour. The weather years selected to perform the
modelling were 2004, 2008, 2013, 2014, and 2015. Due to not
meeting the U.S. EPA minimum requirements of completeness,
the additional annual datasets from 2005 to 2016 were
disregarded. Preprocessing was conducted to fix unordered
times, non-uniformly spaced times, missing data, and duplicate
records. Gaps were filled using established data substitution
protocols (U.S. EPA, 2000, 2016c). Representativeness regarding
both spatial and temporal resolution of the meteorological data is
mandatory. The SBCT station dataset is representative of the
meteorological conditions at the source location and in adjacent
areas because of the (i) proximity of the meteorological tower to
the area being modelled: (~4.5 km), (ii) complexity of the terrain:
topography between the surface station and the source location
is not complex, (iii) surface characteristics: comparable land use
characteristics around the meteorological tower tothe area being
modelled, and (iv) period of the data collection and completeness:
recent and valid years of weather data are used. Upper air data for
the weather years previously selected were obtained for SBCT
from the NOAA/ESRL Radiosonde Database. Both surface and
Fig. 1 Elevations at (left) São José dos Pinhais/Brazil and (right) Groß-Enzersdorf/Austria: Model domains are within the
circular areas of 750 m radius centred on the source (red star marker). Legends denote elevations in metres and their
associated colours; scale bars indicate metres.
14 JOURNAL OF ENVIRONMENTAL SCIENCES 79 (2019) 1124
upper air data were inspected using quality assurance procedures
and validation and were processed using AERMET (version 16216).
Atmospheric pressure is used within the model basically to
calculate dry air density, and cloud cover is a necessary input to
AERMET to derive the micrometeorological parameters. The
model uses the Monin-Obukhov similarity theory to estimate
the stability of the planetary boundary layer. This theory is
grounded on the Obukhov stability length, which is an estimation
of the height where the shear production of turbulent kinetic
energy is comparable with the buoyancy production of turbulence
kinetic energy (Temel and van Beeck, 2017).
1.4.2. Austrian site
Primary surface meteorological data in 1 h time-steps for W
d
,
W
s
, T, and P
atm
were made available by the Central Institute
for Meteorology and Geodynamics (ZAMG, Vienna, Austria)
for Groß-Enzersdorf GE (48.199° N, 16.559° E). Wind direction
was recorded to the nearest 1°. However, GE station does
not have CC observations; then, this variable was also
provided by ZAMG for Schwechat Vienna International
Airport LOWW (48.110° N, 16.569° E) which is situated
~10 km from the source. Minor missing data were filled
using recommended procedures. The period of data collec-
tion is associated with the Brazilian site to harmonise the
meteorological years used for model calculations. For that
reason, we selected five years of weather data for each site.
The GE station dataset is representative of the spatial and
temporal conditions at the odour source location and in
adjacent areas for the same conditions previously described
(e.g., the meteorological tower is located ~ 0.6 km from the
source). Upper air data for the corresponding surface
weather years were obtained from the NOAA/ESRL Radio-
sonde Database for Wien-Hohe Warte WHW (located
~16 km from the source, 48.248° N, 16.356° E).
Table 2 summarises the information on the surface and upper
air meteorological stations at the Brazilian and Austrian sites.
1.5. Selection of odour impact criteria
The calculation of the direction-dependent separation dis-
tances was performed for two national OIC, as follows:
OIC
1
:C
t
= 0.25 ou
E
/m
3
,P= 90th, A
t
= 1 hr;
OIC
2
:C
t
=1ou
E
/m
3
,P= 98th, A
t
= 1 hr.
The OIC
1
is presently used in Germany (GOAA, 2008;
TA-Luft, 2002). The OIC
2
is used, for example, in Flanders
(Belgium) for new geographically isolated livestock farms
(LNE, 2008; Willems et al., 2015). Criteria of this type are
often used for odour impact assessment purposes. A detailed
description of OIC in several jurisdictions throughout
the world can be found in recently published papers
(Brancher et al., 2017; Brancher et al., 2016; Sommer-
Quabach et al., 2014).
For all simulations, the same source data and modelling
assumptions were considered. This arrangement enables the
calculated separation distances to deviate mainly because of
the length of the meteorological input data and the selected
OIC. Meteorological data were combined into a single model
run encompassing the whole period of meteorology (i.e.,
multiple-year modelling configured by concatenating the
five individual years of meteorological data for each site).
The results of these multiple-year model runs were named
5 years. Furthermore, individual model runs were con-
ducted for each meteorological year separately. In this case,
each output was named according to the year in which the
meteorological data collection occurred.
1.6. Statistical analysis
Separation distances are typically drawn to the opposite side
of the W
d
because this is the direction to which emissions
spread. This direction is called transport direction T
d
(given by
T
d
=W
d
+ 180°). For instance, when the wind blows from the
South (180°) the corresponding separation distance is located
to the North (T
d
= 360°) (Schauberger et al., 2006; VDI 3894 Part
2, 2012). The calculated direction-dependent separation dis-
tances are given in increments of 10° using the stack position
as the reference point for the distance determination. The
separation distances are given in full metres. The contour
method used to draw the separation distances is B-spline
smoothing. Statistical analyses were performed using the
Guide to the expression of Uncertainty in Measurement
(GUM), which recommends a standardised approach for
expressing the uncertainty of measurements (BIPM et al.,
2008). We determined the mean separation distance values
(sample size = 5), the standard deviations (degrees of free-
dom = 4), in addition to the upper and lower confidence
interval boundaries (k= 2, level of confidence = 95%) over the
single meteorological years. The coefficient of variation CV,
Table 2 Meteorological stations selected for the modelling applications. The meteorological years used are 2004, 2008,
2013, 2014, and 2015.
Site Station code Type Coordinates Elevation
ASL (m)
Distance from
the source (km)
Hourly meteorological parameters
Brazil SBCT Surface 25.531° S, 49.167° W 908 4.5 W
d
,W
s
,T,P
atm
,CC
SBCT Upper air 25.531° S, 49.167° W 908 4.5 W
d
,W
s
,T,P
atm
Austria GE Surface 48.199° N, 16.559° E 154 0.6 W
d
,W
s
,T,P
atm
LOWW Surface 48.110° N, 16.569° E 183 10 CC
WHW Upper air 48.248° N, 16.356° E 198 16 W
d
,W
s
,T,P
atm
SBCT: Afonso Pena International Airport; GE: Groß-Enzersdorf; LOWW: Schwechat Vienna International Airport; WHW: Wien-Hohe Warte; W
d
:
wind direction; W
s
: wind speed; T: air temperature; P
atm
: atmospheric pressure; CC: cloud cover; ASL: above sea level.
15JOURNAL OF ENVIRONMENTAL SCIENCES 79 (2019) 1124
  • Article
    Full-text available
    Governments are increasingly introducing odour impact criteria (OIC) to determine separation distances between odour sources and residential areas. Previous studies have shown the wide range of OIC available for this purpose, depending on the desired level of protection against odour annoyance. However, it is unclear whether OIC with similar levels of protection can ensure analogous separation distances, which would reasonably be expected. This study presents a comparative analysis of separation distances calculated at two sites for different OIC, but all related to an equivalent level of protection. Here, the equivalent level of protection was defined for urban residential areas (land use), swine odour (hedonic tone) and new facilities (facility type). In this manner, the regulatory criteria currently enforced in Germany, Ireland, and Queensland (Australia) were selected as references for the investigation. The results clearly show that, even for an equivalent level of protection, disparate separation distances can be obtained. Differences in separation distances were found to be greater in prevailing wind directions compared to distances in additional wind directions. Overall, the results demonstrate a risk of poor conclusions in odour assessments. This means that care must be taken when adopting OIC for decision making, principally in those countries that have not yet established specific regulations to manage environmental odours. Concomitantly, the results stress the need for better harmonisation of the concept of the odour impact criterion and components thereof. By using perturbation analysis, it has also been found that the stack exit temperature influences the separation distances in a distinct way, reliant on the criteria used to determine the distances. This finding is of significance for input data collection in future odour modelling studies. Furthermore, approaches used to derive OIC, equivalence between dispersion modelling and field inspections (European standard EN 16841-1), as well as implications of the findings for regulatory practice are summarised and discussed.
  • Article
    Full-text available
    Concentrated animal feeding operations (CAFOs) such as dairy farms are a source of odorous compound emissions. In this study, by identifying relevant odour sources within a 300-head dairy farm and quantifying their emissions, we determined the separation distances to avoid odour annoyance around the dairy farm with two empirical models (Austrian and German Verein Deutscher Ingenieure (VDI) model) and a dispersion model (AERMOD). Besides, this study ponders on the selection of an appropriate meteorological station that best represents the area surrounding the farm. Results show that the maximum separation distances of an exceedance probability of P = 15% determined by the two empirical and the dispersion models are 524 m, 440 m and 655 m, while the minimum values are 202 m, 135 m, and 149 m, respectively. The NE–SW stretching separation distances match well with the wind rose. The mean ratios of separation distances determined by the two empirical models to that of the dispersion model are 1.23 and 0.95. Moreover, statistics of the separation distances indicate good accordance between the empirical models and the dispersion model.
  • Technical Report
    Full-text available
    The monitoring of emissions to air and water represents an important element in preventing and reducing pollution from industrial installations and in ensuring a high level of protection of the environment taken as a whole. Therefore, the Industrial Emissions Directive 2010/75/EU (IED) addresses the monitoring of emissions in a number of instances. This JRC Reference Report on Monitoring (ROM) summarises information on the monitoring of emissions to air and water from IED installations, thereby providing practical guidance for the application of the Best Available Techniques (BAT) conclusions on monitoring in order to help competent authorities to define monitoring requirements in the permits of IED installations. Moreover, the information and recommendations provided by this document may help the Technical Working Groups (TWGs) to derive BAT conclusions during the drawing up and review of BAT Reference documents (BREFs).
  • Article
    The assessment of odour annoyance varies vastly among countries even within the European Union. Using so-called odour-hour frequencies offers the distinct possibility for either applying dispersion models or field inspections , both generally assumed to be equivalent. In this study, odour-hours based on field inspections according to the European standard EN 16841-1 (2017) in the vicinity of a pig-fattening farm have been compared with modelled ones using the Lagrangian particle model GRAL, which uses odour-concentration variances for computing odour hours as recently proposed by Oettl and Ferrero (2017). Using a threshold of 1 ou m −3 (ou = odour units) for triggering odour hours in the model, as prescribed by the German guideline for odour assessment, led to reasonable agreements between the two different methodologies. It is pointed out that the individual odour sensitivity of qualified panel members, who carry out field inspections, is of crucial importance for selecting a proper odour-hour model. Statistical analysis of a large number of data stemming from dynamic olfactometry (EN 13725, 2003), that cover a wide range of odorants, suggests that the prescribed method in Germany for modelling odour hours may likely result in an overestimation, and hence, equivalence with field inspections is not given. The dataset is freely available on request.
  • Article
    Most environmental epidemiology studies have examined pollutants individually. Multi-pollutant approaches have been recognized recently, but to the extent of our knowledge, no study to date has specifically investigated exposures to multiple air pollutants in rural environments. In this paper we characterized and quantified residential exposures to air pollutant mixtures in rural populations, provided a better understanding of the relationships between air pollutant mixtures and annoyance responses to environmental stressors, particularly odor, and quantified their predictive abilities. We used validated and highly spatially resolved atmospheric modeling of 14 air pollutants for four rural areas of Denmark, and the annoyance responses considered were annoyance due to odor, noise, dust, smoke and vibrations. We found significant associations between odor annoyance and principal components predominantly described by nitrate (NO3−), ammonium (NH4+), particulate matter (PM10 and PM2.5) and NH3, which are usually related to agricultural emission sources. Among these components, NH3 showed the lowest error when comparing observed population data and predicted probabilities. The combination of these compounds in a predictive model resulted in the most accurate model, being able to correctly predict 66% of odor annoyance responses. Furthermore, noise annoyance was found to be significantly associated with traffic-related air pollutants. In general terms, our results suggest that emissions from the agricultural and livestock production sectors are the main contributors to environmental annoyance, but also identify traffic and biomass burning as potential sources of annoyance.
  • Article
    A novel methodology for estimating odour hours in the frame of licencing procedures is presented. In contrast to the widely used constant-factor-4 model, which is the prescribed method in Germany, a model based on computing concentration variances is proposed. It is derived upon the advection-diffusion equation for the concentration variance, but is strongly simplified by neglecting the transport and diffusion terms. In this way, the method becomes extremely efficient with regard to computation times. Furthermore, the model is independent on the type of dispersion model used to calculate average concentrations, which are necessary for subsequently computing concentration variances. In a second step, simulated concentration variances are used in combination with a slightly modified two-parameter Weibull probability density function to get the 90th percentile of the cumulative frequency distribution of odour-concentration fluctuations, which is required for computing a so-called odour hour. The model is operated in post-processing mode and can, thus, easily be implemented in existing dispersion models. It's validity has been tested against two tracer tests carried out in Germany and the U.S.
  • Article
    Full-text available
    Exposure to environmental odour can result in annoyance, health effects and depreciation of property values. Therefore, many jurisdictions classify odour as an atmospheric pollutant and regulate emissions and/or impacts from odour generating activities at a national, state or municipal level. In this work, a critical review of odour regulations in selected jurisdictions of 28 countries is presented. Individual approaches were identified as: comparing ambient air odour concentration and individual chemicals statistics against impact criteria (maximum impact standard); using fixed and variable separation distances (separation distance standard); maximum emission rate for mixtures of odorants and individual chemical species (maximum emission standard); number of complaints received or annoyance level determined via community surveys (maximum annoyance standard); and requiring use of best available technologies (BAT) to minimize odour emissions (technology standard). The comparison of model-predicted odour concentration statistics against odour impact criteria (OIC) is identified as one of the most common tools used by regulators to evaluate the risk of odour impacts in planning stage assessments and is also used to inform assessment of odour impacts of existing facilities. Special emphasis is given to summarizing OIC (concentration percentile and threshold) and the manner in which they are applied. The way short term odour peak to model time-step mean (peak-to-mean) effects is also captured. Furthermore, the fundamentals of odorant properties, dimensions of nuisance odour, odour sampling and analysis methods and dispersion modelling guidance are provided. Common elements of mature and effective odour regulation frameworks are identified and an integrated multi-tool strategy is recommended.
  • Article
    A model for the evaluation of the concentration fluctuation variance is coupled with a one-particle Lagrangian stochastic model and results compared to a wind-tunnel simulation experiment. In this model the concentration variance evolves along the particle trajectories according to the same Langevin equation used for the simulation of the velocity field, and its dissipation is taken into account through a decay term with a finite time scale. Indeed, while the mean concentration is conserved, the concentration variance is not and our model takes into account its dissipation. A simple parametrization for the dissipation time scale is proposed and it is found that it depends linearly on time and on the ratio between the size and the height of the source through a scaling factor of 1 / 3.
  • Article
    Full-text available
    Environmental odour exposure that results from anthropic activities may cause an adverse effect on human health. The exposure is frequently classified as airborne pollutant and therefore mainly regulated by federal or state jurisdictions. The goal of the regulations is the limitation of odour exposure to avoid nuisance. For this purpose, odour impact criteria (OIC) are formed by an odour concentration threshold and an accepted probability of exceedance of this concentration (i.e., percentile) to define compliance. Dispersion models generally calculate time series of hourly-mean concentration values at receptors surrounding odour sources. To mimic odour perception of the human nose, short-time peak concentrations derived from one-hour mean values may also be considered in OIC by using the peak-to-mean concept. In this paper, a review of OIC applied in jurisdictions throughout South America is provided according to the desired level of protection. The focus is on OIC which are used predominantly for the evaluation of the time series of ambient odour concentration, calculated by dispersion models. Particularly, the odour framework in countries such as Brazil, Chile and Colombia is presented. Sampling and analysis methods of odours and technical guidance associated with odour modelling will also be presented when mentioned in the regulations or technical guides. Furthermore, the importance of establishing exposure limits is discussed, to provide regulatory agencies and industry an effective and objective tool for the realization of odour impact assessment studies. Undeniably, olfactometric methodologies combined with criteria for impact assessment are essential instruments for managing conflicts arising from odour episodes and to achieve compliance for new, existing or expansion of facilities.
  • Article
    The present study is devoted for the development of two equation Reynolds-Averaged Navier–Stokes (RANS) closures for computational fluid dynamics (CFD) modelling of the atmospheric boundary layer. By using the inflow conditions based on the Monin–Obukhov similarity theory, the closure coefficients of the proposed models are derived from the analytical solutions of simplified turbulent transport equations. Modifications are conducted for three different turbulence models, which are standard and Re-Normalisation Group (RNG) . Numerical experiments are performed for the homogeneous atmospheric boundary layer and the results are compared with the theoretical values in comparison to the standard versions of modified turbulence models. Developed models are implemented to open source CFD toolbox OpenFOAM.
  • Article
    Full-text available
    Direction-dependent separation distances to avoid odour annoyance, calculated with the Gaussian Austrian Odour Dispersion Model AODM and the Lagrangian particle diffusion model LASAT at two sites, are analysed and compared. The relevant short-term peak odour concentrations are calculated with a stability-dependent peak-to-mean algorithm. The same emission and meteorological data, but model-specific atmospheric stability classes are used. The estimate of atmospheric stability is obtained from three-axis ultrasonic anemometers using the standard deviations of the three wind components and the Obukhov stability parameter. The results are demonstrated for the Austrian villages Reidling and Weissbach with very different topographical surroundings and meteorological conditions. Both the differences in the wind and stability regimes as well as the decrease of the peak-to-mean factors with distance lead to deviations in the separation distances between the two sites. The Lagrangian model, due to its model physics, generally calculates larger separation distances. For worst-case calculations necessary with environmental impact assessment studies, the use of a Lagrangian model is therefore to be preferred over that of a Gaussian model. The study and findings relate to the Austrian odour impact criteria.