Atmospheric Pollution Research xxx (xxxx) xxx–xxx

HOSTED BY Contents lists available at ScienceDirect

Atmospheric Pollution Research

journal homepage: www.elsevier.com/locate/apr

Spatial and temporal evaluation of H2S, SO2 and NH3 concentrations near
Cerro Prieto geothermal power plant in Mexico
L.C. Aguilar-Dodiera, J.E. Castilloa,∗, Penelope J.E. Quintanab, Lupita D. Montoyac,
Luisa T. Molinad, Miguel Zavalad, V. Almanza-Velozd,e, J.G. Rodríguez-Venturaa,1
a
Facultad de Ciencias Químicas e Ingeniería, Universidad Autónoma de Baja California, Calzada Universidad 14418, Parque Industrial Internacional, C.P., 22390,
Tijuana, B.C., Mexico
b
Graduate School of Public Health, San Diego State University, 5500, Campanile Dr., San Diego, CA, 92182-4162, USA
c
University of Colorado Boulder, Civil, Environmental, and Architectural Engineering Department, 1111 Engineering Drive, ECOT 441, UCB 428, Boulder, CO, 80309,
USA
d
Molina Center for Energy and the Environment, La Jolla, CA, USA
e
Massachusetts Institute of Technology, 77 Massachusetts Ave, Cambridge, MA, 02139, USA

A R T I C LE I N FO A B S T R A C T

Keywords: Power generation is associated with toxic emissions. Hydrogen sulfide (H2S) and ammonia (NH3) are toxic gases
Hydrogen sulfide emitted from geothermal power generation that have negative impacts on health and the environment.
Geothermal power plant emissions Oxidation of H2S leads to formation of SO2 that also has health and environmental effects. The Cerro Prieto
Meteorological conditions geothermal power plant (CP-GPP) is one of the largest in the world and is located 30 km south of Mexicali, a
Air pollution
border city between Mexico and the US. Power from this plant is sold to both countries. In order to assess the
Health effects
Backward trajectory model
potential effects of the emissions from the GPP on nearby low-income communities, H2S, SO2 and NH3 con-
centrations were measured at Nuevo León, El Chimi, and Estación Delta, using passive samplers from November
2013 to June 2014. The highest concentrations of H2S were recorded during periods of high atmospheric sta-
bility, and exceeded the limit recommended by the WHO (150 μg m−3, 24-h) three times downwind from the
GPP. Similarly, the mean SO2 concentrations at all three sites exceeded the limit established by the NOM-022-
SSA1-2010 (288 μg m−3, 24-h), while NH3 concentrations exceeded the limits established by the OEHHA
(100 μg m−3, 24-h) eight times at Nuevo León. A backward trajectory analysis indicated transport of air masses
from the GPP to El Chimi and Estación Delta sampling sites. The results indicate that GPP's emissions could affect
the residents in the surrounding areas, warranting public health action to investigate exposures and risks.

1. Introduction south of the city of Mexicali, on the western border between Mexico and
the United States (US) (Gallego-Ortega et al., 2000; Peralta et al.,
The generation of electricity from geothermal plants has been as- 2013). Mexicali lies in a region formed by the Imperial Valley on the US
sociated with high emissions of non-condensable gases such carbon side and the Mexicali Valley on the Mexico side of the border. With a
dioxide (CO2), hydrogen sulfide (H2S), methane (CH4), ammonia (NH3) population of about one million inhabitants (INEGI, 2010), Mexicali is a
and other greenhouse gases (Barbier, 2002; Kristmannsdóttir and dynamic city with many sources of air pollution including two ther-
Ármannsson, 2003; D'Alessandro et al., 2013; Shortall et al., 2015; moelectric power plants, industry, brick kilns, a large vehicle fleet, and
Manzella et al., 2018). These CO2 and CH4 emissions are generally intense commercial activity with the US, causing increased air pollution
lower compared to the emissions from carbon and fossil fuel-based on both sides of the border (Mendoza et al., 2010; Reyna et al., 2012).
plants, but they can severely impact the environment and contribute to In addition, a wastewater treatment plant, a common source of H2S
climate change (Peralta et al., 2013; Somma et al., 2017). (Lewkowska et al., 2016; Talaiekhozani et al., 2016; Zounemat-
The Cerro Prieto geothermal power plant (CP-GPP) is located 30 km Kermani et al., 2019), is located approximately 10 km west-southwest

Peer review under responsibility of Turkish National Committee for Air Pollution Research and Control.
∗
Corresponding author.
E-mail addresses: [email protected] (L.C. Aguilar-Dodier), [email protected] (J.E. Castillo), [email protected] (P.J.E. Quintana),
[email protected] (L.D. Montoya), [email protected] (L.T. Molina), [email protected] (M. Zavala), [email protected] (V. Almanza-Veloz).
1
Deceased.

https://doi.org/10.1016/j.apr.2019.09.019
Received 25 March 2019; Received in revised form 23 September 2019; Accepted 26 September 2019
1309-1042/ © 2019 Turkish National Committee for Air Pollution Research and Control. Production and hosting by Elsevier B.V. All rights reserved.

Please cite this article as: L.C. Aguilar-Dodier, et al., Atmospheric Pollution Research, https://doi.org/10.1016/j.apr.2019.09.019
L.C. Aguilar-Dodier, et al. Atmospheric Pollution Research xxx (xxxx) xxx–xxx

of Mexicali (see Supporting Material 1, SM1) with likely, but not re- the CP-GPP; however, annual levels are made public, usually with some
ported, emissions of H2S. Similarly, the burning of tires is commonplace time delay of about one year.
in this region. The CP-GPP is situated in a region characterized by ex- Although the direct sources of SO2 include the combustion of fossil
tensive agricultural activities on communal lands along with other in- fuels by mobile sources and industrial activities in urban areas, SO2 is
dustrial services and commercial activities. Near the CP-GPP, there is an also generated from the oxidation of H2S. As the sulfur content in fuels
inactive volcano called Cerro Prieto (SEMARNAT and CFE, 2007); decreases, the SO2 concentrations measured in urban and rural areas
therefore, no contributions of H2S and SO2 are expected from this may reach similar ambient levels. In some rural areas, SO2 concentra-
source. tions exceeded those in urban areas (WHO, 2000b; Lin et al., 2012).
The CP-GPP is currently the largest geothermal field in Mexico and Airborne SO2 can be removed by dry and wet deposition or contribute
includes 13 electricity generation units grouped in four individual to the formation of strong acids when reacting with hydroxyl radicals in
power plants: Cerro Prieto 1 (CP1, 5 units), Cerro Prieto 2 (CP2, 2 the atmosphere (Newman et al., 1991; Somma et al., 2017). The
units), Cerro Prieto 3 (CP3, 2 units) and Cerro Prieto 4 (CP4, 4 units), average ambient concentration standard for SO2 in Mexico is
with electricity generation capacities of 180, 220, 220 and 100 MW, 288 μg m−3 for 24-h (NOM-022-SSA1-2010, 2010). In the US, the
respectively. Four of the five units of CP1 are officially out of service CAAQS for SO2 are 655 and 105 μg m−3 for 1-h and 24-h, respectively.
since February 2011 (Peralta et al., 2013). Power generation in the CP- Both SO2 and H2S are harmful to human health and may cause eye
GPP is based on the extraction of mixtures of steam and water, which irritation, bronchoconstriction in asthmatics, and changes in lung
are separated in a centrifuge. The steam is sent to the power plants, and function at high concentrations (U.S. Department of Health and Human
the water is carried by canals and pipelines to solar evaporation la- Services, 1998).
goons, where some of the water is evaporated, and the rest is re-injected NH3 is another important atmospheric pollutant that can be emitted
into the wells (SEMARNAT and CFE, 2007). The non-condensable gases from geothermal plants. Airborne NH3 neutralizes acids coming from
(CO2, NH3 and H2S) are emitted to the atmosphere through vents or oxides of sulfur and nitrogen to produce aerosols and smog; it acidifies
stacks installed in cooling towers, facilitating the dispersion of those the soil and water, and causes eutrophication in natural ecosystems
gases (Olafsdottir and Gardarsson, 2013). The Ministry of Environment (Roadman et al., 2003; Barrancos et al., 2013). NH3 is generally emitted
and Natural Resources (SEMARNAT) of Mexico compiles and reports by geothermal facilities at relatively low levels; however, the largest
the annual H2S emissions data from the CP-GPP. The values for 2013 sources of NH3 emissions are agricultural activities such as raising li-
through 2017 were 7.1, 8.7, 7.3, 7.1, and 7.0 Gg year−1, respectively vestock and using NH3-based fertilizer (Roadman et al., 2003; Kagel
(SEMARNAT, 2019). et al., 2007; Maasikmets et al., 2015). Atmospheric NH3 concentrations
H2S produces severe health effects including eye irritation, loss of have a marked seasonal behavior, where high concentrations are ob-
smell, pulmonary edema, and loss of respiration (WHO, 2000a). The served during the higher temperatures in spring and summer (Puchalski
odor threshold for H2S is very low and it can be detected at con- et al., 2011). The US Office of Environmental Health Hazard Assess-
centrations in the range of 1–8 μg m−3 (Powers, 2012). The World ment (OEHHA) air quality guideline for NH3 is 100 μg m−3 for 24-h
Health Organization (WHO) 24-h average air quality guideline for H2S average (OEHHA, 2003). In Mexico, there is no ambient air quality
is 150 μg m−3 (WHO, 2003). In the US, the 1-h California Ambient Air standard for NH3; the levels established by the OEHHA could be used as
Quality Standard (CAAQS) for H2S is 42 μg m−3 (CARB, 2016). Cur- a reference.
rently, there is no National Ambient Air Quality Standard (NAAQS) in The environmental impact (i.e., acidification, particulate matter and
the US or Mexico for H2S; therefore, the WHO and CAAQS guidelines terrestrial eutrophication) of H2S, SO2, and NH3 from geothermal
are used as the reference. power plants has been assessment by Parisi et al. (2019), and accord the
Atmospheric H2S can be adsorbed by soils where the elemental results obtained, the chemical interaction and the environmental im-
sulfur is retained. It can also dissolve in water, but it is preferentially pact of the pollutants released to the atmosphere should be object for
oxidized by molecular oxygen and hydroxyl radicals formed by atmo- future research. The environment impacts related to atmospheric
spheric photochemical reactions producing sulfur dioxide (SO2) and emissions are extremely related with the geographical location of the
sulfates (Newman et al., 1991; U.S. Department of Health and Human geothermal power plants and the technology used in the process of
Services, 2006). The production of oxidants is favored during periods of generation of energy.
high temperatures and intense radiation, resulting in low concentra- Acute Respiratory Infection (ARI) is a leading cause of morbidity
tions and short H2S lifetimes of about one day during the summer worldwide, especially in developing countries (Romieu et al., 2002).
(Kourtidis et al., 2008; D'Alessandro et al., 2009). During winter, higher This is also the case in the Mexicali region (Instituto de Servicios de
concentrations are observed, and its lifetime can extend up to 42 days Salud Pública de Baja California [ISESALUD], 2018) as shown in the
(Chow Pineda, 2007; Somma et al., 2017). Supporting Material 2 (SM2) Table S1. Other afflictions include Upper
Ambient concentrations of H2S emitted from geothermal plants can Respiratory Infections (URI) related to exposure to H2S like cough,
be largely influenced by meteorological conditions. Thorsteinsson et al. difficulty in breathing, runny nose, and sore throat (Heaney et al., 2011;
(2013) and Olafsdottir et al. (2014) reported high H2S concentrations Somma et al., 2017). The risk for children is higher because they have
during periods of high atmospheric stability, low wind speed and narrower airways, developing lungs and immune systems, as well as
temperature, and no precipitation. H2S concentrations have been found higher rates of breathing and longer exposures to outdoor air (Zheng
to correlate well with distance from geothermal plants and magnitude et al., 2017). The ambient air pollutants most relevant to ARI include
of emissions (D'Alessandro et al., 2009; Manzo et al., 2013); however, SO2, nitrogen oxides (NOx), primary and secondary acid aerosols and
when the sampling point is very close to the emission source and the ozone (Romieu et al., 2002). Important sources of SO2 and NOx emis-
stack is high, low concentrations of pollutants have been measured sions are small and large industries as well as power generation.
(Stathopoulos et al., 2004; Longley, 2018). In this study, H2S, SO2, and NH3 concentrations were measured
Despite its regional importance as a source of H2S emissions, few daily, for up to one week, at three low-income communities near the
studies have characterized the impacts of the CP-GPP. Peralta et al. CP-GPP from November 2013 to June 2014 (autumn-winter and spring
(2013) reported a 14-h maximum H2S concentration of 16.5 μg m−3 at seasons). NH3 concentrations were measured for the first time at these
6.7 km southeast of the CP-GPP and a 23-h maximum of 13.5 μg m−3 at communities. H2S and SO2 concentrations were measured for the first
4.3 km east of the plant. Peak night concentrations of H2S at CP1 and time at one of them. Meteorological data (temperature, wind speed, and
1.2 km east of the geothermal plant were 750 and 570 μg m−3, re- wind direction) were also collected. A backward trajectory model was
spectively, in 2010, when the boundary layer is minimum. Currently, applied to investigate the origin of the air masses at the Mexicali Valley
there is no reported information about the daily H2S levels surrounding and the sampling sites most impacted by them. The H2S, SO2, and NH3

2
L.C. Aguilar-Dodier, et al. Atmospheric Pollution Research xxx (xxxx) xxx–xxx

Fig. 1. Sampling sites and weather station around the CP-GPP, which is enclosed by the dash line. CP1-CP4 are the power plants; the communities are enclosed by
solid gray lines.

concentrations were compared with air quality guidelines. Table 2

H2S, SO2, and NH3 measurements at S1, S2, and S3.
2. Materials and method Period S1 S2 S3

2.1. Sampling sites and periods H2S SO2 NH3 H2 S SO2 NH3 H2S SO2 NH3

Nov 29 - Dec 6 ✓ ✓ ✓
Three sampling sites were selected at local rural communities lo- Dec 8 - Dec 13 ✓ ✓ ✓
cated near the CP-GPP based on the reported predominant wind di- Jan 18 - Jan 23 ✓ ✓ ✓ ✓ ✓ ✓
rection 315° (NW) and expected during the sampling periods (GobBC Jan 24 - Jan 25 ✓
and SEMARNAT, 2011). Sampling site 1 (S1) was located at Ejido Jan 26 - Jan 31 ✓ ✓ ✓
Feb 1 - Feb 6 ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Nuevo León (population 3255, 32.41° N, 115.19° W), predominantly
Feb 7 - Feb 12 ✓ ✓ ✓
upwind from the geothermal plant. Sampling site 2 (S2) was at El Chimi Feb 13 - Mar 7 ✓
(population 178, 32.37° N, 115.18° W), and sampling site 3 (S3) was at Mar 22 - Mar 25 ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Estación Delta (population 5278, 32.36° N, 115.19° W), both pre- Mar 26 - Mar 29 ✓ ✓ ✓
dominantly downwind from the geothermal plant. Fig. 1 shows the Mar 30 - Apr 2 ✓ ✓ ✓ ✓ ✓ ✓
Apr 3 - Apr 4 ✓ ✓
location of the CP-GPP, the sampling sites, and the weather station. Apr 5 - Apr 9 ✓ ✓ ✓ ✓ ✓
Samplers at S1, S2, and S3 were located outside houses at distances of Apr 10 - Apr 18 ✓ ✓ ✓
1.8, 4.0, and 5.1 km to CP4, respectively. Apr 19 - Apr 24 ✓ ✓
The stack height and emission flow characteristics of the power Apr 27 - Apr 30 ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
May 1 - May 4 ✓ ✓ ✓
plants are summarized in Table 1, adapted from Peralta et al. (2013).
May 5 - May 8 ✓ ✓ ✓ ✓ ✓ ✓
Measurements of gas concentrations were made during autumn May 10 - May 23 ✓ ✓ ✓
(November 29 to December 13, 2013), winter (January 18 to March 7, May 25 - May 26 ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
2014), and spring (March 22 to June 20, 2014). Table 2 shows details of May 27 - May 28 ✓ ✓
the H2S, SO2, and NH3 sampling campaign. A total of 104, 106, and 120 May 29 - May 30 ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Jun 7 - Jun 13 ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
days were sampled at S1, S2, and S3, respectively. Since fewer mea- Jun 14 - Jun 20 ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
surements were obtained in autumn and winter, the measurements

Table 1 were grouped as autumn-winter and spring periods for analysis.

Stack height and emission flow characteristics.
Power Stack Stack internal Gas Gas velocity Flow rate
2.2. Measurement of H2S, SO2 and NH3
plant height (m) diameter (m) temp. (m s−1)a (m3 s−1)a
(°C)a
Sampling of H2S was performed using Radiello passive samplers
CP1 54.7 0.406 52.0 5.7 0.7 (Supelco, Bellefonte, PA, USA) with low detection limits of 45 and
CP2 40.2 0.305 112.0 24.1 1.8
1.5 μg m−3 for 1 h and 24 h exposures, respectively, and uncertainty at
CP3 40.2 0.305 115.5 33.2 2.4
CP4 54.9 0.609 56.0 8.1 2.4 2σ of 8.7% over the entire exposure range (1 h–15 days) (Radiello,
2006).
a
Average data estimated from the flow characteristics of the units. The Radiello passive sampler consists of a white diffusive body

3
L.C. Aguilar-Dodier, et al. Atmospheric Pollution Research xxx (xxxx) xxx–xxx

(RAD 120, Sigma-Aldrich, Co. LLC) with a radial design that optimizes α NH3wNH3
[NH3] =
the efficiency of the trap by allowing pollutant uptake from all direc- t (5)
tions (Venturi et al., 2016). The body contained a chemiadsorbing where [NH3] is the concentration (in ppb), wNH3 is the mass of NH3 (in
cartridge (RAD 170, Sigma-Aldrich, Co. LLC) impregnated with zinc ng), t is the exposure time (in min), and αNH3 is the conversion coef-
acetate. When in contact with H2S, zinc sulfide was formed, which was ficient (in ppb min ng−1), obtained according equation (6) (Ogawa,
then extracted with reagent water and then reacted with a ferric 2010):
chloride-amine solution acting as an oxidizing agent (Radiello, 2006).
The final solution was methylene blue, which was quantified at 665 nm 293 ⎞1.83
α NH3 = 43.8 ⎛
using a spectrophotometer (DR 5000, HACH, Loveland, CO, USA). The ⎝ 273+T ⎠ (6)
sampling rate, QK, was corrected for average temperature according to
where T is the average temperature (in °C) during the sampling period,
equation (1) (Radiello, 2006):
43.8 is the conversion coefficient (in ppb min ng−1) at 20 °C and 70%
K 3.8 relative humidity.
Q K = 0.096 ⎛ ⎞
⎝ 298 ⎠ (1) Radiello and Ogawa diffusive samplers are viable alternatives to
continuous instruments (D'Alessandro et al., 2009; Bates et al., 2013;
where K is the average temperature (in Kelvin) during the sampling Pavilonis et al., 2013; Maasikmets et al., 2015; Pope et al., 2017).
period, 0.096 is the sampling rate (in ng ppb−1 min−1) at 298 K (25 °C) Concentrations of SO2 and H2S using these samplers have been eval-
and 101.3 kPa (1 atm). Equation (1) is independent of relative humidity uated in laboratories obtaining good correlations with those obtained
in the range of 10–90% and wind speeds in the range of 0.1–10 m s−1. using reference methods (Mason et al., 2011).
All H2S concentrations were determined according to the following All samplers were placed 2 m above ground level for periods from
equation (Radiello, 2006): 24-h up to 2 weeks. Unopened Radiello cartridges and Ogawa filters
were kept at 4 °C until needed; after sampling, they were stored inside
m ⎞
C= ⎛ ⎜ × 1000
⎟
an airtight Ziploc™ bag at 4 °C until analysis. The samplers were kept on
⎝ QK t ⎠ (2)
ice during transit to sampling sites and laboratories (Salem et al., 2009;
where m is the mass of sulfide ion (in μg) obtained from the cartridge, t Adon et al., 2010; Gibson et al., 2013). Laboratory blank filters for each
is the exposure time (in min), and C is the H2S concentration (in ppb) pollutant were treated in the same way as the samples, and values were
(Radiello, 2006). used for blank corrections. No field blanks were included in this study.
Sampling of SO2 and NH3 was performed using Ogawa passive
samplers (Ogawa & Co., Pompano Beach, FL, USA), which have been 2.3. Meteorology
used successfully in several studies (Roadman et al., 2003; Salem et al.,
2009; Puchalski et al., 2011; Gibson et al., 2013). The detection limits Meteorological data were obtained from a monitoring station lo-
for SO2 are 10 and 1.4 μg m−3 for 24 h and 168 h exposures, respec- cated at the Instituto de Ciencias Agrícolas (ICA-UABC) (32.40° N,
tively. Detection limits for NH3 are 1.6 and 0.2 μg m−3 for 24 h or 168 h 115.20° W) at 8.6 m above sea level near the CP-GPP (Fig. 1). The
exposures, respectively (Ogawa, 2002). The Ogawa passive sampler monitoring station is located at 1.4 km from CP4, and it measured
consists of a Teflon cylindrical body (2 cm diameter and 3 cm length) hourly average temperature, wind speed, and wind direction during the
with two open unconnected sides. Each side contains a cellulose filter sampling period. The average wind speed, average wind direction, and
mounted between two stainless steel screens situated behind a barrier wind roses for each sampling period were processed using WRPLOT
and cap with 25 holes (Salem et al., 2009; Puchalski et al., 2011). SO2 (AERMOD View 6.2 software).
adsorbed on the cellulose filters impregnated with tri-ethanolamine Mean annual temperature in the region varies between 21 and
(PS-144, Ogawa & Co.) were extracted with reagent water. Hydrogen 23 °C; the warmest months usually occur in July and August with
peroxide was then added to the solution, and the SO2 was oxidized to temperatures above 30 °C, while the coldest months are December and
SO4−2. The final solution was sonicated for 30 min, filtered with a January, with temperatures below 13 °C (SEMARNAT and CFE, 2007).
0.45 μm filter and quantified at 450 nm by spectrophotometry ac- The predominant wind direction is northwest from December to May
cording to the SulfaVer 4 method (HACH, 2005). The SO2 concentration and southeast from June to November, with a mean wind speed of
was corrected with a concentration conversion coefficient according to about 2.2 m s−1 during both periods.
equation (3) (Ogawa, 2006):
α so2 wso2 2.4. Modeling
[SO2] =
t (3)
The Flexible Particle-Weather Research and Forecasting (FLEXP-
where [SO2] is the concentration (in ppb), wSO2 is the mass of SO2 (in ART-WRF) dispersion model (Brioude et al., 2013) was used to evaluate
ng), t is the exposure time (in min), and αSO2 is the concentration the local transport in the Mexicali Valley to identify potential air masses
coefficient conversion (in ppb min ng−1), obtained according to arriving at the measurement sites from the CP-GPP and surrounding
equation (4) (Ogawa, 2006): regions. Three-dimensional time-averaged wind data from the WRF
293 ⎞1.83 model were used to drive the Lagrangian simulations since these fields
α so2 = 3.94 ⎛ can reduce the uncertainties in the model (Brioude et al., 2012). The
⎝ 273+T ⎠ (4)
configuration of the WRF model is based on simulations performed by
where T is the average temperature (in °C) during the sampling period, Bei et al. (2013) during the Cal-Mex field campaign in 2010. A different
39.4 is the conversion coefficient (in ppb min ng−1) at 20 °C and 70% simulation domain was used for this study since a nested grid encom-
relative humidity. passing the US-Mexico border along the Mexicali Valley region was
The analysis of NH3 started with the adsorption on the cellulose added for this analysis. Given the simple topography of the region, a
filters impregnated with citric acid (PS-154, Ogawa & Co.) and ex- resolution of 3 and 1 km was used for the WRF nested domains. Back-
tracted with reagent water to yield ammonium citrate. The final solu- ward trajectories over 24-h were obtained every 15 min for the period
tion was sonicated for 30 min, filtered with a 0.45 μm filter, and from February 10 to 12, 2014. A similar approach to Bei et al. (2013)
quantified at 425 nm according to the Nessler method (HACH, 2005). was followed in this modeling exercise. Since the CP-GPP is located
The NH3 concentration was corrected with a conversion coefficient northwest of the measurement sites, the model could help identify
according to equation (5) (Ogawa, 2010): potential transport from the GPP under prevailing conditions.

4
L.C. Aguilar-Dodier, et al. Atmospheric Pollution Research xxx (xxxx) xxx–xxx

After running FLEXPART-WRF in backward mode, the potential

emission sensitivity (PES) fields were used to estimate the transport
pathways most likely to impact the measurement sites. The PES is
proportional to the residence time of a pollutant in a cell of the
Lagrangian grid, so high PES values indicate that more material can be
carried and transported downwind to the receptor site (Hüser et al.,
2017). This fact is particularly significant near the surface since most
anthropogenic emissions are released at lower altitude (Stohl et al.,
2007). Thus, a vertical shallow layer, also referred to as the footprint
layer of 100 m above ground level, was used to calculate the accumu-
lated PES for the entire simulation period for each measurement site.
The modeling results are meant to provide examples of air masses
transport patterns in the region and their link between the sampled
concentrations and the locations of the emission sources in the geo-
thermal plant, and not a quantitative association between emission
sources, concentrations and health risks; however, provides useful in-
formation on the trajectory of air masses that can impact the sampling
receptor sites.

3. Results and discussion

3.1. Spatial variability of H2S, SO2, and NH3

Fig. 2 shows the H2S, SO2, and NH3 concentrations measured during
the entire sampling period at S1, S2, and S3. H2S concentrations at S2
and S3 exceeded the 24-h limit established by the WHO (dashed lines)
once (3%) and twice (4%), respectively. Similarly, the SO2 concentra-
tions at S1, S2, and S3 exceeded the 24-h NOM-022-SSA1-2010 limit
three (30%), six (50%), and twelve times (63%), respectively. The NH3
concentrations at S1 exceeded the 24-h limit established by the
OEHHA, eight times (67%). Fig. 3. Daily average wind direction (°), wind speed (m s−1), and temperature
Fig. 3 shows the measured daily average wind direction, wind (°C) during the sampling periods.
speed, and temperature for the entire sampling period. The wind

Fig. 2. Concentrations measured at S1 (bottom), S2 (middle), and S3 (top) during the entire sampling period for: a) H2S, b) SO2, and c) NH3. Dashed lines represent
the H2S, SO2, and NH3 24-h average exposure guidelines from WHO (2003), NOM-022-SSA1-2010 (2010), and OEHHA (2003), respectively. Gray lines represent the
SO2 24-h average exposure guideline from CAAQS (2016).

5
L.C. Aguilar-Dodier, et al. Atmospheric Pollution Research xxx (xxxx) xxx–xxx

Fig. 4. Wind roses during: a) the entire sampling period (November 29, 2013 to June 20, 2014), b) autumn-winter season (November 29, 2013 to March 7, 2014),
and c) spring season (March 22 to June 20, 2014).

direction ranged from 337.5° to 67.5° (NNW-ENE) from November the wind direction was SE or SSE, CP-GPP emissions did not directly
2013 to January 2014, 0°–135° (N-SE) in February 2014, 315°–135° impacted these sampling sites. The CP-GPP emissions did not directly
(NW-SE) from March to April 2014, and 0°–157.5° (N-SSE) during May affect S1 due to its location upwind and the two prevailing wind di-
and June 2014. Therefore, the most frequent wind directions at Cerro rections (NW, NNW and SE, SSE). The high SO2 levels measured at S1
Prieto varied from NNW to N and from SE to SSE during the sampling were attributed to other SO2 sources from the NW, like brick kilns near
months. The wind directions obtained at Cerro Prieto during the sam- Mexicali, and local tire-burning. The high SO2 levels around the CP-GPP
pling months agreed with those reported previously from long-term were associated with the variability in the wind direction during the
measurements (SEMARNAT and CFE, 2007). A summary of meteor- sampling period and the slow process of SO2 oxidation in the atmo-
ological conditions during the sampling periods is provided in Table S2 sphere. An average oxidation rate of less than 2% h−1 in polluted air
(SM2). has been reported by others (Newman et al., 1991), and it is likely that
From November 2013 to May 2014, the wind speeds at Cerro Prieto it contributed to the SO2 background concentration.
varied from 0.3 to 2.8 m s−1, classified as calm winds and light winds, The H2S concentrations at S1 did not exceed the 24-h limit estab-
respectively. The wind speeds in June 2014 increased to 3.5 through lished by the WHO (150 μg m−3) during the entire sampling period (see
5.9 m s−1, classified as gentle and moderate winds, respectively. The Fig. 2a). Nevertheless, most of the H2S concentrations measured at S1
increases in wind speeds were associated with a linear increase in were above the WHO's recommended threshold for odor annoyance of
temperature during the sampling months. The coldest period was 7 μg m−3 (WHO, 2003). The maximum H2S concentration measured
November 2013 to February 2014, with temperatures below 18 °C; the (111 μg m−3) at S1 happened during the period Jun 8–9; this was
warmer months were from March 2014 to June 2013, with tempera- higher than those measured at S2 and S3 at the time. The prevailing
tures up to 31 °C. wind direction at this time was upwind from 157.5° (SSE); therefore, S1
The wind rose analyses in Fig. 4 indicated that the two prevailing was not impacted by emissions from the CP-GPP. This high H2S level
wind directions during the sampling period were NW and SE at calm to could be due to other unidentified emission sources. In general, most
low wind speed conditions. During the autumn-winter season (No- (60%) of the H2S samples measured at S1 were lower than those
vember 29, 2013 to March 7, 2014), the prevailing wind direction was measured at S2 and S3. These measurements confirm that S1 had up-
NW. During the spring season (March 22 to June 20, 2014), the pre- wind conditions from the power plant even though it was closest to it.
vailing wind direction was SE, which agrees with data reported by This variability of H2S concentrations at CP-GPP depends to a great
SEMARNAT and the CFE (2007). extent on meteorological conditions and not the proximity to the
The results suggest that when the wind direction was NNW, the sites emission source. Similar findings were reported by Olafsdottir et al.
most affected by the CP-GPP emissions were S2 and S3 and that when (2014), who found maximum H2S concentrations at far distances from

6
L.C. Aguilar-Dodier, et al. Atmospheric Pollution Research xxx (xxxx) xxx–xxx

the GPP at Hellisheidi and Nesjavellir, Iceland. Thorsteinsson et al. sites, suggesting that the oxidation of H2S from the GPP is an important
(2013), on the other hand, reported high H2S concentrations near the mechanism impacting SO2 background levels in the region (see Fig. 2b).
emissions sources when the wind speed was relatively high, which re- The highest mean SO2 concentrations, 421.4 and 626.5 μg m−3, were
sulted in high turbulence and low variability in H2S concentrations. measured at S2 and S3, respectively, while the mean SO2 concentration
In the period May 13–15, the maximum H2S concentration mea- at S1 was 327.4 μg m−3. The highest SO2 and H2S concentrations
sured at S2 was 192.3 μg m−3, which was higher than those measured measured at S2 and S3 coincided with a wind direction of 337.5°.
at S1 and S3. This value exceeded the WHO limit and corresponded to a Furthermore, the mean H2S concentrations showed a similar trend as
wind direction of 315° (NW). H2S concentrations at S3 exceeded the the mean SO2 concentrations, and these affected the sampling sites in
WHO limit on two occasions: Nov 30 - Dec 1 (172.6 μg m−3) and Dec the following order: S3, followed by S2, and finally by S1. Figs. S1 and
10–11 (257.2 μg m−3). These high H2S concentrations corresponded to S2 in Supporting Material (SM2) show air quality indexes at the three
wind directions of 0° (N) and 337.5° (NNW) during the sampling per- sampling sites for H2S and SO2, respectively. The air quality index was
iods. In both cases, the air masses come from the CP-GPP, carrying the calculated by dividing daily concentrations measured at each site by the
H2S emission to S2 and S3. given norm: 150 μg m−3 and 288 μg m−3 for H2S and SO2, respectively.
Even though S3 is the farthest sampling site from CP-GPP, the H2S These indexes are meant to provide a quick overview of the local air
concentrations were higher than those measured at S1 and similar to quality.
the concentrations measured at S2, with mean concentrations of 61.9 During the entire sampling period, most (66.7%) of the NH3 con-
and 64.2 μg m−3, respectively, whereas the mean concentration at S1 centrations at S1 exceeded the 24-h limit established by the OEHHA.
was only 21.9 μg m−3. Therefore, site distance from the geothermal These levels were attributed to NH3 emitted by livestock (i.e., goats)
plant alone was not sufficient to explain the high H2S concentrations of present nearby during sampling. The low NH3 concentrations measured
S2 and S3 compared to S1. Peralta et al. (2013) also reported high H2S at S1 May 25–26 and 29–30 were linked to atmospheric instability:
concentrations measured at sampling sites farthest from the CP-GPP light wind speeds (1.6 and 1.2 m s−1, respectively) and high tempera-
and low H2S concentrations at sites nearest to it. tures (27.8 and 29.2 °C, respectively). Such conditions likely caused fast
Ambient levels of SO2 can be explained by the combination of direct evaporation of NH3 near the sampling site. The low NH3 concentrations
impacts from local sources and the oxidation rate of H2S from the measured on Jun 7–13 and 14–20 were associated with high wind
geothermal plant. The oxidation of H2S is temperature dependent and speeds (3.8 and 4.9 m s−1, respectively) and high temperatures (30.0
can contribute to high regional (background) levels of SO2 under high and 28.6 °C). All NH3 concentrations measured at S1 were higher than
temperature conditions. As SO2 continues to be oxidized hetero- those measured at S2 and S3, where the concentrations did not exceed
geneously, high ambient background levels of sulfates can be formed the OEHHA limit. Fig. S3 (SM2) shows the air quality index for NH3 at
and further expand the impacts of the geothermal plant. the three sampling sites.
The SO2 concentrations at S1 during the periods Feb 1–3, Feb 4–6, Overall, the mean NH3 concentrations measured at S1, S2, and S3
and Mar 30 - Apr 2, were 3.5, 2.9, and 2.1 times higher than the 24-h were 142.9, 21.0, and 10.5 μg m−3, respectively. The higher NH3 con-
limit established by the Mexican standard NOM-022-SSA1-2010. These centrations at S1 had low correlation with the wind direction during the
SO2 concentrations also were 9.1, 7.9, and 5.7 higher than the 24-h sampling period (337.5°); therefore, it is likely that the CP-GPP does not
limit established by the CAAQS. In these sampling periods, the wind contribute significantly to levels of NH3 in the region.
directions were 337.5°, 112.5°, and 315°, respectively. In addition, the Morbidity data from the Mexicali Valley, where Cerro Prieto is lo-
temperatures were low (11.7, 13.4 and 18.2 °C, respectively), sug- cated, were available for the sampling period (2013–2014) and are
gesting that the high SO2 levels were not likely due to direct H2S oxi- provided in Table S1 (SM2). The diseases reported most frequently in
dation. These high SO2 levels are likely due to other emission sources, 2013 and 2014 that could be related to H2S and SO2 exposure include
such as brick kilns near Mexicali and a tire-burning business near the Acute Respiratory Infections (ARI), asthma, and conjunctivitis (Instituto
study zone. This result agrees with Peralta et al. (2013), who concluded de Servicios de Salud Pública de Baja California [ISESALUD], 2018).
that gases like SO2 follow a similar daily evolution, and some high ARI could be related to the SO2 concentrations found at all three sites,
concentration events were caused by local activities. which sometimes exceeded the NOM-022-SSA1-2010 and the 24-h
At S2, about half of the SO2 samples exceeded the 24-h NOM-022- CAAQS guideline. Similarly, H2S emissions from the CP-GPP may be
SSA1-2010 limit and most (83%) of these samples exceeded the CAAQS. contributing to increases in eye diseases and ARI morbidity as primary
The maximum SO2 concentration during the sampling period May emissions or as a precursor of SO2. A study by Schinasi et al. (2011)
25–26 was 3.1 times higher than the NOM-022-SSA1-2010 limit. The found that H2S concentrations were associated with eye irritation as
high SO2 concentration measured on Feb 1–3 (847.1 μg m−3) coincided well as irritation and respiratory symptoms. The morbidity associations
with the wind direction (NNW); however, in Feb 4–6 the high SO2 listed here are qualitative, but the high levels of H2S and SO2 found
concentration (551.4 μg m−3) did not coincide with the wind direction suggest that detailed exposure and health risk assessments are needed
(up wind) and low temperature, suggesting little contribution from the for this region.
CP-GPP emissions and H2S oxidation, respectively. These results sug-
gest that SO2 concentration was likely due to the accumulation of SO2 3.2. Seasonal variability of H2S and SO2
or the presence of other SO2 sources. The high SO2 concentrations
measured on May 25–26, May 29–30, and Jun 14–20 were due prob- As described above, H2S concentrations at S1 were consistently low
ably to H2S oxidation to SO2 at high temperatures (27.8, 29.2 and throughout the sampling period due to its upwind location with respect
28.6 °C, respectively). Barnes et al. (1986) reported the H2S oxidation to to the GPP and no seasonal variability was observed there. However, at
SO2 under conditions of 1 atm air and 27 °C. These values could include both S2 and S3, high H2S concentrations were found during the au-
background SO2 concentrations and other SO2 sources. At S3, most tumn-winter period coincident with an NNW wind direction. The high
(63.2%) of the SO2 samples exceeded the NOM-022-SSA1-2010 limit, concentrations measured during the spring season probably also coin-
and 89% exceeded the CAAQS. The maximum value measured, cided with the same wind direction but at lower frequency. The SO2
2251.1 μg m−3 (Jan 18–19), was related to a wind direction of 337.5°. concentrations at the three sites were likely due to high background
Another two high SO2 (1441.3 and 1010.9 μg m−3) events occurred at SO2 levels caused by SO2 accumulation (51.2% of the total samples
S3 Nov 30 - Dec 1 and Jan 22-23 with prevailing wind direction during exceed the NOM-022-SSA1-2010 limits). The highest concentrations
both periods of 0° (upwind) and low temperatures (15.2 and 15.4 °C), were measured at S2 and S3 during autumn-winter, coincident with an
suggesting impacts from local SO2 emissions sources. NNW wind direction. No seasonal variability of SO2 was observed at S1
The results show that overall, there are high SO2 levels at the three because it was not impacted by the emissions of CP-GPP.

7
L.C. Aguilar-Dodier, et al. Atmospheric Pollution Research xxx (xxxx) xxx–xxx

Table 3
Spearman correlation coefficients between pollutants concentrations and meteorology during the sampling period: Temperature (T), Wind Speed (WS) and Wind
Direction (WD).
Pollutant S1 S2 S3

T WS WD T WS WD T WS WD

(°C) (m s−1) (°) (°C) (m s−1) (°) (°C) (m s−1) (°)

H2S 0.132 −0.042 −0.078 −0.079 0.124 0.147 −0.418 −0.157 −0.011
SO2 −0.782 −0.661 0.130 −0.126 0.063 −0.047 −0.479 −0.667 −0.198
NH3 −0.860 −0.280 0.061 −0.588 −0.273 0.196 −0.005 −0.043 0.448

Significant results are shown in bold with p < 0.05.

At S1, the NH3 concentrations depended on the season, with the 3.4. Backward trajectory analysis
highest and lowest values measured during autumn-winter and spring,
respectively, as shown in Fig. 2. This seasonal variability was attributed Fig. 6 shows the accumulated air mass residence time in the foot-
to differences in the mean temperature and wind speeds, which caused print layer in the period Feb 10–12, 2014. The retro plume remained
atmospheric instability, leading to high dispersion of NH3 emissions. narrow near the measurement sites, and it spanned a wider area en-
The measurements of NH3 at S2 and S3 were similar during the sam- compassing the US-Mexico border as it dispersed backward in time.
pling period; therefore, there was no seasonal variability at these 2 This indicates a relatively high spatial variability during the simulation
sites. The low NH3 concentrations measured at S2 and S3 indicated low period. Nevertheless, the residence times of northwest areas in either
NH3 contributions from the CP-GPP. measurement site were higher during the entire simulation period. This
Table 3 presents Spearman correlations between H2S, SO2, and NH3 indicated that the WRF model reproduced the observed northwesterly
concentrations and local meteorological conditions (i.e., temperature, winds, and these winds were the main geographical pathways that
wind speed and wind direction) during the sampling period. H2S con- transported any emission that could have been released within these
centrations correlated well (p < 0.05) with temperature at S3 regions.
(r = −0.418). SO2 concentrations correlated well with temperature In the footprint PES maps, high residence times indicate more
and wind speed at S1 (r = −0.782 and r = −0.661) and S3 emission sensitivity, and hence, a higher potential contribution of a
(r = −0.479 and r = −0.667). NH3 concentrations correlated well source area. Fig. 6 indicates that the local transport within the region
with temperature at S1 (r = −0.860) and S2 (r = −0.588). greatly determines if air masses carry material originating in CP-GPP to
the measurement sites. For S1, the model results suggest that there is no
contribution from CP-GPP emissions, instead there is from the sur-
3.3. Ratio of SO2 to H2S rounding areas. In contrast, S3 was the location most likely impacted
from the power plant units’ emissions (CP1-CP4). S2 is a unique case
In order to assess the impact of the emissions from CP-GPP, it is because the results suggest only a partial contribution to this site due to
important to understand the chemical transformations that sulfur un- the retro plume. Therefore, the corresponding residence time partially
dergoes in this region. H2S is only emitted directly from the plant. SO2 covers the emission sources from the CP-GPP, indicating that not all the
is then produced from chemical reactions of H2S in the atmosphere but emissions from the power plants impact S2.
it is also produced by local and urban sources in Mexicali that are then Fig. 6 shows that CP1 and CP2 emissions did not impact S2. This
transported by air masses to the study region. may be due to 1) four of the five units of CP1 being out of service since
Airborne H2S, with an atmospheric lifetime of 17–70 h, is oxidized 2011, and 2) if CP2 was operating, the change of wind direction during
to SO2 via reactions with the hydroxyl (OH) radical (Thiemens, 1977; the modeling period would prevent S2 from being impacted by the
Seinfeld and Pandis, 2006). Once formed, SO2 further reacts with OH in emissions from CP2. However, this is speculative since the location of
a series of steps involving intermediate sulfur compounds that even- S2 and S3 within the output grid from FLEXPART-WRF are very close,
tually form H2SO4, which rapidly condenses on aerosol particles due to making it difficult to attribute this behavior, because they are within
its very low vapor pressure. The typical atmospheric lifetime of SO2 the model's spatial resolution limits. Therefore, even though the con-
based on reactions with the OH radical is about 13 days (Finlayson-Pitts tribution from Cerro Prieto is likely to impact S3, it is not clear to what
and Pitts, 2000) although dry and wet depositions are also efficient SO2 extent it might be at S2 for this specific simulation period.
removal pathways. The results also suggest that the measurement sites could be im-
As a first approximation, it was considered that the SO2 con- pacted not only by the emissions from the CP-GPP, but also by emis-
centrations measured in this study mostly derived from the oxidation of sions from other sources since the air parcels can originate south of the
H2S caused by the increase of the temperature (although wind direction US-Mexico border. Nevertheless, the regions up to west Mexicali (dia-
was the most prevalent parameter). Thus, the potential impacts of H2S mond) and the corresponding anthropogenic sources along that
emissions from the geothermal plant could be assessed by comparing pathway, like the Mexicali wastewater treatment plant, are likely to be
the sulfur-content molar ratio of SO2/H2S, as shown in Fig. 5. transported first and then mixed with Cerro Prieto's emissions.
Results show that in the region surrounding the geothermal plant, Consequently, emissions from the Mexican side of the border had a
sulfur was substantially more abundant as SO2, particularly at S1. This higher probability of being transported towards the measurement sites,
behavior was probably due to the shorter lifetime of H2S due to oxi- particularly S1, during this period. In addition, the model suggests that
dation via OH in comparison to SO2. These results suggest that S1 was emissions from the wastewater treatment plant farther to the west, near
the site most impacted by the SO2 emissions derived from the waste- Santa Isabel (square), could be transported as well, mostly to S2 and S3.
water treatment plant or other local SO2 sources. The low ratios at S2
and S3 compared to S1, suggest that these sites were first impacted by
the H2S derived from CP-GPP emissions. In general, the ratios greater 4. Conclusions
than 1 at all sites, indicate that sulfur is primarily in the form of SO2.
Furthermore, it is likely that H2S is continually oxidized to SO2, which For the first time, ambient levels of H2S, SO2, and NH3 were con-
increases the SO2 concentrations in the Cerro Prieto area. currently measured at three sites (S1, S2, and S3) in the region near CP-

8
L.C. Aguilar-Dodier, et al. Atmospheric Pollution Research xxx (xxxx) xxx–xxx

Fig. 5. Boxplots of [SO2]/[H2S] ratios at S1, S2, and S3. The blue dotes represent the mean.

Fig. 6. Time integrated footprint PES for a) S1, b) S2, and c) S3 (denoted as the tip of each air mass) from Feb 10–12, 2014. Cerro Prieto power plants are the symbols
clustered toward the bottom: CP1 (open left-pointing triangle), CP2 (open circle), CP3 (open downward-pointing triangle), and CP4 (closed circle). The closed
diamond, right-pointing triangle, and square toward the top represent the urban region of Mexicali, the Imperial Valley, and Santa Isabel, respectively. The solid
black line indicates the US-Mexico border.

9
L.C. Aguilar-Dodier, et al. Atmospheric Pollution Research xxx (xxxx) xxx–xxx

GPP in Mexico. These measurements were then used as a basis for di- Atmos. Chem. Phys. 10, 7467–7487. https://doi.org/10.5194/acp-10-7467-2010.
agnosing the potential impacts of emissions from the geothermal plant. Barbier, E., 2002. Geothermal energy technology and current status: an overview. Renew.
Sustain. Energy Rev. 6, 3–65. https://doi.org/10.1016/S1364-0321(02)00002-3.
During the entire sampling period, 51% of the samples exceeded the Barnes, I., Bastian, V., Becker, K.H., Fink, E.H., Nelsen, W., 1986. Oxidation of sulfur
SO2 concentrations limit established by the NOM-022-SSA1-2010 while compounds in the atmosphere: I. Rate constants of OH radicals reactions with sulfur
20% of the samples exceeded the NH3 concentrations established by dioxide, hydrogen sulphide, aliphatic thiols and thiophenol. J. Atmos. Chem. 4,
445–466. https://doi.org/10.1007/BF00053845.
OEHHA. Only 3% of the samples exceeded the H2S concentration limit Barrancos, J., Briz, S., Nolasco, D., Melián, G., Padilla, G., Padrón, E., Fernández, I., Pérez,
established by the WHO. Overall, H2S and SO2 concentrations were N., Hernández, P.A., 2013. A new method for estimating greenhouse gases and am-
higher during autumn-winter than during spring coinciding with lower monia emissions from livestock buildings. Atmos. Environ. 74, 10–17. https://doi.
org/10.1016/j.atmosenv.2013.03.021.
mean temperatures and wind speeds as well as higher atmospheric Bates, M.N., Garrett, N., Crane, J., Balmes, J.R., 2013. Associations of ambient hydrogen
stabilities. Measured SO2 concentrations were higher than the H2S sulfide exposure with self-reported asthma and asthma symptoms. Environ. Res. 122,
concentrations during the rest of the year. The results suggest that 81–87. https://doi.org/10.1016/j.envres.2013.02.002.
Bei, N.F., Li, G.H., Zavala, M., Barrera, H., Torres, R., Grutter, M., Gutierrez, W., Garcia,
background SO2 levels in the region are largely impacted by the
M., Ruiz-Suarez, L.G., Ortinez, A., Guitierrez, Y., Alvarado, C., Flores, I., Molina, L.T.,
emissions from the geothermal plant via H2S oxidation to SO2 and by 2013. Meteorological overview and plume transport patterns during Cal-Mex 2010.
direct SO2 emissions from other unreported sources around the geo- Atm. Environ. Times 70, 477–489. https://doi.org/10.1016/j.atmosenv.2012.01.
thermal field, like brick kilns and a wastewater treatment plant near 065.
Brioude, J., Angevine, W.M., McKeen, S.A., Hsie, E.-Y., 2012. Numerical uncertainty at
Mexicali. Consequently, the contribution from other sources of H2S and mesoscale in a Lagrangian model in complex terrain. Geosci. Model Dev. (GMD) 5,
SO2 need to be further investigated by doing concurrent measurement 1127–1136. https://doi.org/10.5194/gmd-5-1127-2012.
campaigns at different sites around the CP-GPP, and during all seasons. Brioude, J., Arnold, D., Stohl, A., Cassiani, M., Morton, D., Seibert, P., Angevine, W.,
Evan, S., Dingwell, A., Fast, J.D., Easter, R.C., Pisso, I., Burkhart, J., Wotawa, G.,
Similarly, more detailed modeling can be done using more specific in- 2013. The Lagrangian particle dispersion model FLEXPART-WRF version 3.1. Geosci.
formation on the emission sources, as well as more detailed meteor- Model Dev. 6, 1889–1904. https://doi.org/10.5194/gmd-6-1889-2013.
ology information, topography features, land use, and chemical com- CARB, 2016. Ambient air quality standards. Calif. Air Resources board. https://ww3.arb.
ca.gov/research/aaqs/aaqs2.pdf?_ga=2.253901836.587531124.1565663701-
position of emissions. 1417475160.1564692278.
The NH3 concentrations measured at S1 were higher than those Chow Pineda, I.G., 2007. Gaussian modelling of the dispersion of hydrogen sulphide from
measured at S2 and S3 likely due to the presence of livestock grazing Hellisheidi power plant. Iceland. Geotherm. Train. Program. 5, 55–78. http://www.
os.is/gogn/unu-gtp-report/UNU-GTP-2007-05.pdf.
near that location. The results suggest that NH3 levels in the region are D'Alessandro, W., Brusca, L., Kyriakopoulos, K., Michas, G., Papadakis, G., 2009.
not substantially impacted by emissions from the CP-GPP. Hydrogen sulphide as a natural air contaminant in volcanic/geothermal areas: the
The FLEXPART-WRF model was used to investigate pollution case of Sousaki, Corinthia (Greece). Environ. Geol. 57, 1723–1728. https://doi.org/
10.1007/s00254-008-1453-3.
transport from the CP-GPP. It was found that during February 10–12,
D'Alessandro, W., Aiuppa, A., Bellomo, S., Brusca, L., Calabrese, S., Kyriakopoulos, K.,
2014, the industrial complex could have contributed only to the pol- Liotta, M., Longo, M., 2013. Sulphur-gas concentrations in volcanic and geothermal
lution levels at S2 and S3, with no apparent impact to S1. S2 and S3 are areas in Italy and Greece: characteristing potential human exposures and risks. J.
populated areas with vulnerable inhabitants such as young children. In Geochem. Explor. 131, 1–13. https://doi.org/10.1016/j.gexplo.2012.08.015.
Finlayson-Pitts, B.J., Pitts Jr., J.N., 2000. Chemistry of the Upper and Lower Atmosphere.
this regard, emissions from surrounding neighborhoods on the Mexican United State of America, Academic Press.
side of the border were most likely to impact S1 during this period, Gallego-Ortega, R., Quintero-Nuñez, M., García-Cueto, O.R., 2000. H2S Dispersion model
contributing to the total emissions transported to the receptor sites. at Cerro Prieto geothermoelectric power plant. In: Proceeding World Goethermal
Congress, pp. 579–584. May 28 - June 10, 2000, Kyushu-Tohoku, Japan. https://
Even though S1, S2, and S3 are relatively close to each other, the local www.geothermal-energy.org/pdf/IGAstandard/WGC/2000/R0898.PDF.
transport dynamics greatly determine the impact of the geothermal Gibson, M.D., Heal, M.R., Li, Z., Kuchta, J., King, G.H., Hayes, A., Lambert, S., 2013. The
plant and anthropogenic emissions in the region. spatial and seasonal variation of nitrogen dioxide and sulfur dioxide in Cape Breton
Highlands National Park, Canada, and the association with lichen abundance. Atmos.
Simultaneous sampling of H2S and SO2 should be conducted to Environ. 64, 303–311. https://doi.org/10.1016/j.atmosenv.2012.09.068.
verify if the highest concentrations of SO2 are due to the oxidation of GobBC, ISESALUD, 2018. Morbilidad en el Valle de Mexicali, años 2013-2014. Secretaría
H2S. Longer sampling periods of H2S, SO2, and NH3 encompassing all de Salud.
GobBC, SEMARNAT, 2011. Programa para mejorar la calidad del aire en Mexicali 2011-
seasons need to be performed. It is also recommended that sampling at
2020. Gob. del Estado Baja Calif. 32–208. http://www.semarnat.gob.mx/
different sites and distances from CP-GPP be performed, including a archivosanteriores/temas/gestionambiental/calidaddelaire/Documents/Calidad
‘background’ location as reference. Finally, a comprehensive air quality %20del%20aire/Proaires/ProAires_Vigentes/ProAire%20Mexicali%202011-2020.
pdf.
modeling study on the evaluation of potential health risks in the region
HACH, 2005. DR5000 spectrophotometer. Procedures manual. file:///C:/Users/carol/
will require longer sampling periods near the exposed communities Downloads/DR5000%20Spectrophotometer%20Procedures%20Manual%20(1).pdf.
downwind, as well as detailed information on the emission character- Heaney, C.D., Wing, S., Campbell, R.L., Caldwell, D., Hopkins, B., Richardson, D., Yeatts,
istics of the geothermal plant. K., 2011. Relation between malodor, ambient hydrogen sulfide, and health in a
community bordering a landfill. Environ. Res. 111, 847–852. https://doi.org/10.
1016/j.envres.2011.05.021.
Acknowledgments Hüser, I., Harder, H., Heil, A., Kaiser, J.W., 2017. Assumptions about footprint layer
heights influence the quantification of emission sources: a case study for Cyprus.
Atmos. Chem. Phys. 17, 10955–10967. https://doi.org/10.5194/acp-17-10955-2017.
The authors acknowledge the support of residents from Ejido Nuevo INEGI, 2010. Instituto Nacional de Estadística y Geografía.Censo de población y vivienda.
León, El Chimi, and Estación Delta and their valuable help during the Mexicali, Baja California. https://www.inegi.org.mx/app/areasgeograficas/?ag=02.
sampling campaign. This research was partially supported by a scho- Kagel, A., Bates, D., Gawell, K., 2007. A guide to geothermal energy and the environment.
Geothermal energy association. pp. 1–75. http://www.charleswmoore.org/pdf/
larship from Consejo Nacional de Ciencia y Tecnología (CONACYT). Environmental%20Guide.pdf.
Kourtidis, K., Kelesis, A., Petrakakis, M., 2008. Hydrogen sulfide (H2S) in urban ambient
Appendix A. Supplementary data air. Atmos. Environ. 42, 7476–7482. https://doi.org/10.1016/j.atmosenv.2008.05.
066.
Kristmannsdóttir, H., Ármannsson, H., 2003. Environmental aspects of geothermal energy
Supplementary data associated with this article can be found, in the utilization. Geothermics 32, 451–461. https://doi.org/10.1016/S0375-6505(03)
online version, at https://doi.org/10.1016/j.apr.2019.09.019. 00052-X.
Lewkowska, P., Cieślik, B., Dymerski, T., Konieczka, P., 2016. Characteristics of odors
emitted from municipal wastewater treatment plant and methods for their identifi-
References cation and deodorization techniques. Environ. Res. 151, 573–586. https://doi.org/
10.1016/j.envres.2016.08.030.
Adon, M., Galy-Lacaux, C., Yoboué, V., Delon, C., Lacaux, J.P., Castera, P., Gardrat, E., Lin, W., Xu, X., Ma, Z., Zhao, H., Liu, X., Wang, Y., 2012. Characteristics and recent trends
Pienaar, J., Al Ourabi, H., Laouali, D., Diop, B., Sigha-Nkamdjou, L., Akpo, A., Tathy, of sulfur dioxide at urban, rural, and background sites in North China: effectiveness of
J.P., Lavenu, F., Mougin, E., 2010. Long term measurements of sulfur dioxide, ni- control measures. J. Environ. Sci. 24, 34–49. https://doi.org/10.1016/S1001-
trogen dioxide, ammonia, nitric acid and ozone in Africa using passive samplers. 0742(11)60727-4.
Longley, I., 2018. Technical paper: TP05: road tunnel stack emissions. https://

10
L.C. Aguilar-Dodier, et al. Atmospheric Pollution Research xxx (xxxx) xxx–xxx

chiefscientist.nsw.gov.au/__data/assets/pdf_file/0004/196645/TP05_Road-Tunnel- Stathopoulos, T., Lazure, L., Saathoff, P., Gupta, A., 2004. The effect of stack height, stack
Stack-Emissions.pdf. location and rooftop structures on air intake contamination. A laboratory and full-
Maasikmets, M., Teinemaa, E., Kaasik, A., Kimmel, V., 2015. Measurement and analysis of scale study. Studies and research projects. http://www.irsst.qc.ca/media/
ammonia, hydrogen sulphide and odour emissions from the cattle farming in Estonia. documents/PubIRSST/R-392.pdf.
Biosyst. Eng. 139, 48–59. https://doi.org/10.1016/j.biosystemseng.2015.08.002. Salem, A.A., Soliman, A.A., El-Haty, I.A., 2009. Determination of nitrogen dioxide, sulfur
Manzella, A., Bonciani, R., Allansdottir, A., Botteghi, S., Donato, A., Giamberini, S., Lenzi, dioxide, ozone, and ammonia in ambient air using the passive sampling method as-
A., Paci, M., Pellizzone, A., Scrocca, D., 2018. Environmental and social aspects of sociated with ion chromatographic and potentiometric analyses. Air Qual. Atmos.
geothermal energy in Italy. Geothermics 72, 232–248. https://doi.org/10.1016/j. Heal. 2, 133–145. https://doi.org/10.1007/s11869-009-0040-4.
geothermics.2017.11.015. SEGOB, 2010. Diario Oficial de la Federación. Norma Oficial Mexicana NOM-022-SSA1-
Manzo, C., Salvini, R., Guastaldi, E., Nicolardi, V., Protano, G., 2013. Reflectance spectral 2010. Salud ambiental. Criterio para evaluar la calidad de aire ambiente, con re-
analyses for the assessment of environmental pollution in the geothermal site of Mt. specto al dióxido de azufre (SO2). Valor normado para la concentración del dióxido
Amiata (Italy). Atmos. Environ. 79, 650–665. https://doi.org/10.1016/j.atmosenv. de azufre (SO2) en el ambiente, como medida de protección a la salud de la población.
2013.06.038. Primera Sección. pp. 47–54. http://dof.gob.mx/nota_detalle.php?codigo=5158348&
Mason, J.B., Fujita, E.M., Campbell, D.E., Zielinska, B., 2011. Evaluation of passive fecha=08/09/2010.
samplers for assessment of community exposure to toxic air contaminants and related Seinfeld, J.H., Pandis, S.N., 2006. Atmospheric, Chemistry and Physics. United State of
pollutants. Environ. Sci. Technol. 45, 2243–2249. https://doi.org/10.1021/ America. John Wiley and Sons.
es102500v. SEMARNAT, CFE, 2007. Proyecto geotermoeléctrico Cerro Prieto V. Manifestación de
Mendoza, A., Pardo, E.I., Gutierrez, A.A., 2010. Chemical characterization and pre- Impacto ambiental. Modalidad Particular. I-1-VIII-10. http://sinat.semarnat.gob.
liminary source contribution of fine particle matter in the Mexicali/Imperial Valley mx/dgiraDocs/documentos/bc/estudios/2007/02BC2007E0001.pdf.
border area. Air Waste Manage. Assoc. 60, 258–270. https://doi.org/10.3155/1047- SEMARNAT, 2019. Registro de Emisiones y transferencias de Contaminantes (RETC), Baja
3289.60.3.258. California. Mexico. https://apps1.semarnat.gob.mx:445/retc/retc/index.php.
Newman, L., Krouse, H.R., Grinenko, V.A., 1991. Sulphur isotope variations in the at- Schinasi, L., Horton, R.A., Guidry, V.T., Wing, S., Marshall, S.W., Morland, K.B., 2011. Air
mosphere. Stable isotopes in the assessment on natural and anthropogenic sulphur in pollution, lung function, and physical symptoms in communities near concentrated
the environment (Chapter 5). 92, 133–176. https://www-legacy.dge.carnegiescience. swine feeding operations. Epidemiology 22 (2), 208–215. https://doi.org/10.1097/
edu/SCOPE/SCOPE_43/SCOPE_43_5_Chp5.pdf. EDE.0b013e3182093c8b.
OEHHA, 2003. Ammonia. Technical Support Document: Toxicology Clandestine Drug Shortall, R., Davidsdottir, B., Axelsson, G., 2015. Geothermal energy for sustainable de-
Labs: Methamphetamine. vol. 1. pp. 1–11. http://citeseerx.ist.psu.edu/viewdoc/ velopment: a review of sustainability impacts and assessment frameworks. Renew.
download?doi=10.1.1.615.1012&rep=rep1&type=pdf. Sustain. Energy Rev. 44, 391–406. https://doi.org/10.1016/j.rser.2014.12.020.
Ogawa, 2002. Product Specification Sheet. Model 3300. OgawaTM Passive Sampler. . Somma, R., Granieri, D., Troise, C., Terranova, C., De Natale, G., Pedone, M., 2017.
Ogawa, 2006. NO, NO2, NOx and SO2 Sampling Protocol Using the Ogawa Sampler*. Modelling of hydrogen sulfide dispersion from the geothermal power plants of
Yokohama City Research Institute for Environment Science. pp. 1–28. http:// Tuscany (Italy). Sci. Total Environ. 583, 408–420. https://doi.org/10.1016/j.
ogawausa.com/wp-content/uploads/2017/11/prono-noxno2so206_206_1117.pdf. scitotenv.2017.01.084.
Ogawa, 2010. NH3 Sampling Protocol Using the Ogawa Sampler*. Yokohama City Stohl, A., Burkhart, J.F., Eckhardt, S., Hirdman, D., Sodemann, H., 2007. An Integrated
Research Institute for Environment Science. pp. 1–9. http://ogawausa.com/wp- Internet-Based System for Analyzing the Influence of Emission Sources and
content/uploads/2017/11/pronh3_1117.pdf. Atmospheric Transport on Measured Concentrations of Trace Gases and Aerosols.
Olafsdottir, S., Gardarsson, S.M., 2013. Impacts of meteorological factors on hydrogen Norwegian Institute for Air Research, Kjeller, Norway. https://folk.nilu.no/
sulfide concentration downwind of geothermal power plants. Atmos. Environ. 77, ~andreas/publications/web_based_tool.pdf.
185–192. https://doi.org/10.1016/j.atmosenv.2013.04.077. Talaiekhozani, A., Bagheri, M., Goli, A., Talaei Khoozani, M.R., 2016. An overview of
Olafsdottir, S., Gardarsson, S.M., Andradottir, H.O., 2014. Spatial distribution of hy- principles of odor production, emission, and control methods in wastewater collec-
drogen sulfide from two geothermal power plants in complex terrain. Atmos. tion and treatment systems. J. Environ. Manag. 170, 186–206. https://doi.org/10.
Environ. 82, 60–70. https://doi.org/10.1016/j.atmosenv.2013.10.013. 1016/j.jenvman.2016.01.021.
Parisi, M.L., Ferrara, N., Torsello, L., Basosi, R., 2019. Life cycle assessment of atmo- Thiemens, M.H., 1977. The Atmospheric Oxidation of H2S and the Oxygen Isotopic Ratios
spheric emission profiles of the Italian geothermal power plants. J. Clean. Prod. 234, of Tropospheric Sulfates (thesis Ph. D). Florida State University, Tallahassee. file:///
881–894. https://doi.org/10.1016/j.jclepro.2019.06.222. C:/Users/carol/Desktop/Mendeley/The%20atmospheric%20oxidation%20of
Pavilonis, B.T., O’Shaughnessy, P.T., Altmaier, R., Metwali, N., Thorne, P.S., 2013. %20H2S%20and%20the%20oxygen%20isotopic%20ratios%20of%20tropospheric.
Passive monitors to measure hydrogen sulfide near concentrated animal feeding html.
operations. Environ. Sci Process Impact. 15, 1271–1278. https://doi.org/10.1039/ Thorsteinsson, T., Hackenbruch, J., Sveinbjörnsson, E., Jóhannsson, T., 2013. Statistical
c3em00082f. assessment and modeling of the effects of weather conditions on H2S plume dispersal
Peralta, O., Castro, T., Durón, M., Salcido, A., Celada-Murillo, A.T., Navarro-González, R., from Icelandic geothermal power plants. Geothermics 45, 31–40. https://doi.org/10.
Márquez, C., García, J., de la Rosa, J., Torres, R., Villegas-Martínez, R., Carreón- 1016/j.geothermics.2012.10.003.
Sierra, S., Imaz, M., Martínez-Arroyo, A., Saavedra, I., Espinosa, M.D.L.L., Torres- U.S. Department of Health and human Services, 1998. Toxicological Profile for Sulfur
Jaramillo, A., 2013. H2S emissions from Cerro Prieto geothermal power plant, Dioxide. Agency for Toxic Substances and Disease Registry (ATSDR). pp. 1–118.
Mexico, and air pollutants measurements in the area. Geothermics 46, 55–65. http://www.atsdr.cdc.gov/toxprofiles/tp116.pdf.
https://doi.org/10.1016/j.geothermics.2012.12.001. U.S. Department of Health and Human Services, 2006. Toxicological Profile for Hydrogen
Pope, K., So, Y.T., Crane, J., Bates, M.N., 2017. Ambient geothermal hydrogen sulfide Sulfide. Agency for Toxic Substances and Disease Registry (ATSDR). pp. 1–208.
exposure and peripheral neuropathy. Neurotoxicology (Little Rock) 60, 10–15. https://babel.hathitrust.org/cgi/pt?id=mdp.39015084063778;view=1up;seq=3.
https://doi.org/10.1016/j.neuro.2017.02.006. Venturi, S., Cabassi, J., Tassi, F., Capecchiacci, F., Vaselli, O., Bellomo, S., Calabrese, S.,
Powers, W., 2012. Perception and physiological responses to odors. Factsheet, pork in- D'Alessandro, W., 2016. Hydrogen sulfide measurements in air by passive/diffusive
formation gateway. http://porkgateway.org/resource/science-of-smell-part-1- samplers and high-frequency analyzer: a critical comparison. Appl. Geochem. 72,
perception-and-physiological-responses-to-odors/. 51–58. https://doi.org/10.1016/j.apgeochem.2016.07.001.
Puchalski, M.A., Sather, M.E., Walker, J.T., Lehmann, C.M.B., Gay, D.A., Mathew, J., WHO, 2000a. Air quality guidelines-second edition. Chapeter 6.6. Hydrogen sulfide. J.
Robarge, W.P., 2011. Passive ammonia monitoring in the United States: comparing Appl. Physiol. 3, 1–7. http://www.euro.who.int/__data/assets/pdf_file/0019/
three different sampling devices. J. Environ. Monit. 13, 3156–3167. https://doi.org/ 123076/AQG2ndEd_6_6Hydrogensulfide.PDF.
10.1039/c1em10553a. WHO, 2000b. Air quality guidelines-second edition. Chapeter 7.4. Sulfur dioxide. J. Appl.
Radiello, 2006. Radiello® centro di Ricerche ambientali-padova. Fondazione salvatore Physiol. 1–12. http://www.euro.who.int/__data/assets/pdf_file/0020/123086/
maugeri. www.radiello.com. AQG2ndEd_7_4Sulfurdioxide.pdf.
Reyna, M.A., Bravo, M.E., López, R., Nieblas, E.C., Nava, L.M., 2012. Relative risk of WHO, 2003. Concise International Chemical Assessment Document No.53. Hydrogen
death from exposure to air pollutants: a short-term (2003-2007) study in Mexicali, Sulfide: Human Health Aspects. Geneva, Switzerland. pp. 1–35. http://www.who.
Baja California, Mexico. Int. J. Environ. Health Res. 22, 370–386. https://doi.org/10. int/ipcs/publications/cicad/en/cicad53.pdf.
1080/09603123.2011.65015. Zheng, P., Wang, J., Zhang, Z., Shen, P., Chai, P., Li, D., Jin, M., Tang, M., Lu, H., Lin, H.,
Roadman, M.J., Scudlark, J.R., Meisinger, J.J., Ullman, W.J., 2003. Validation of Ogawa Chen, K., 2017. Air pollution and hospital visits for acute upper and lower respiratory
passive samplers for the determination of gaseous ammonia concentrations in agri- infections among children in Ningbo , China : a time-series analysis. Environ. Sci.
cultural settings. Atmos. Environ. 37, 2317–2325. https://doi.org/10.1016/S1352- Pollut. Res. 24, 18860–18869. https://doi.org/10.1007/s11356-017-9279-8.
2310(03)00163-8. Zounemat-Kermani, M., Stephan, D., Hinkelmann, R., 2019. Multivariate NARX neural
Romieu, I., Samet, J.M., Smith, K.R., Bruce, N., 2002. Outdoor air pollution and acute network in prediction gaseous emissions within the influent chamber of wastewater
respiratory infections among children in developing countries. Outdoor Air Pollut. treatment plants. Atmos. Pollut. Res., Journal Pre-proof. https://doi.org/10.1016/j.
Respir. Infect. 44, 640–649. https://doi.org/10.1097/01.jom.0000023383.41727.32. apr.2019.07.013.

11