Introduction

Some cities in Africa have been nominated as millennium cities based on the rapid urbanization and population growth (MCI 2011). Akure, Nigeria is among the Millennium Cities Initiatives (MCI) rapidly growing cities. The population of the cities doubled in the last 20 years and is currently approaching four million based on 2006 projection by the National Population census (NPC 2006). Across the world, one of the indicators of developing cities, economically and socially, is the availability of telecommunication (Sahu et al. 2015). Wireless mobile network (WMN) telecommunication was introduced into communication sector in Nigeria in 1999. Since then, number of subscribers has grown rapidly from mere 400,000 in 1999 to over 110 million in 2016 (NCC 2017). In Akure, subscribers to mobile network telecommunication have expanded from 100,000 to over 2 million in 2016. WMN in Nigeria, and indeed, in Akure are driven by non-governmental sector of the economy, which provides infrastructure and services. Some of the outdoor MWN’s facilities are the base transceiver stations (BTS) equipped with communication gadgets, which support service rendered by the providers. A major aspect of the WMN operations is generation and provision of power to outdoor BTS equipment. It is estimated that BTS equipment will require approximately 200 MW of electricity for 24 h in a year. Currently, the whole nation generates 4000 MW for consumption by approximately 170 million people. The implication of this is that the 24-h household electricity supply is currently not possible, ditto the BTS. Electricity supply to Akure metropolis stands at between 8 and 17 h in a day and may drop below this value during maintenance period or when gas supply to the generating station is interrupted. The city has expanding network of BTS powered by electricity from national grids; however, power from grids are grossly insufficient for round-the-clock operations. To make up for the short supply, BTS are fitted with diesel generators (DG). Some DGs, which may be in operation for at least 6 h in a day if are being used as back-up for grid supply or more hours if are being used as alternative source of power, emit substantial amount of noxious atmospheric pollutants that impact the local air quality, and which are potential greenhouse gases. Akure, as a designated millennium-growing city, currently has no documentation on the potential impact of emission from expanding network of BTS on air quality. This study considers the rapid expansion of BTS in Akure as case study and simulates the overall contribution emission as a result of power generation through the use of diesel generators to the air quality of the location. The effect of the emission is presented through the assessment of health risk it possesses to the local population.

Power needs of BTS and Carbon footprint

In February 2017, power generation hits 4043 MW in Nigeria (Thisday 2017) when gas supply to the turbine generators improved which translated to about 144.48 kWh per capita (IEA 2007). Due to increasing population, there are more demands for power; therefore, the amount of power generated is grossly inadequate for per capita consumption. Breakdown of sources of power connected to the grids in the country shows that only two sources are presently viable, namely hydropower and fossil fuel-fired turbine plants which both constitute about 98 % of available power for the population. The remaining 2% are off-grid power sources from solar panels for home and offices. Despite inadequate power supply, the total number of subscribers has increased rapidly over the past decade; at the end of 2005, there were 19,519,154 subscribers, but by the end of 2016 there were 154,529,780, which is equivalent to an increase of 12,273,693 every year (NBS 2017). In the first quarter of 2017 and relative to year 1999, the teledensity has reached capacity of 110.09% (Table 1). Teledensity is the number of subscribers with telephone connection among one hundred people in a population. It therefore varies from city to city and it is proxy measure of per capita GDP of a population. The burden of inadequate power supply in the country has impacted BTS operations, which results into network failure and poor services. The financial implication of failed service passed to the subscribers is estimated to be about one trillion Nigeria Naira annually (approximately 2.8 billion US dollars). To overcome this challenge DGs have become an alternative source of power, and so far in Nigeria more than 30, 000 DG serving at least over 36,000 BTS (Table 2) have been installed to back up power supply to the BTS (Punch News 2016). Akure alone has at least 41 DGs. The economic implication of using DG is calculated to have reached over 730 billion Nigeria Naira, approx. 2 billion US dollars (NBS 2017). In Akure, there are an estimated 30 BTS operated by different GSM network running DG for at least 6 to 12 h in a day. A typical set of BTS is shown in Fig. 1. Apart from economic implication of the running DG, harmful pollutants and greenhouse gas including particulate matters (PM), carbon monoxide (CO), black carbon (BC), sulfur dioxide (SO2), hydrocarbons (HC), carbon dioxide (CO2), and nitrogen oxides (NOx) are emitted into the atmosphere causing urban air quality problem. DG emission also produce substantial carbon footprint. A 4.8 kW DG consuming about 34 l in a day will have exhausted about 12,000 l of diesel in a year. According to Gross (2012), a liter of diesel will produce 2.68 kg of CO2 which translates to about 33.2 ton/year of CO2. GSMA (2013) estimated that, in Nigeria, an on-grid DG will consume about 1500 l of diesel in a month while an off-grid DG will use about 1700 l within the same period. With these configurations, an on-grid DG will produce about 48.2 ton/year of CO2 annual carbon footprint while an off-grid will produce 54.7 ton/year of carbon annual footprint. In Akure, the study area, due to inadequate power supply, an average of 30 DG with each having power rating of 24 kW is in operation for at least 6 h in a day, which by estimation each DG will consume 40 l of diesel. Following Gross (2012), 30 DG in use for 6 h in a day will contribute 3.37 ton/day of CO2 and 1.23 Mt of CO2 annually. This heuristic estimation underscored the contribution of DG-powered BTS to air pollution in a developing city. In addition to carbon footprint from the use of DG in powering BTS, emission of toxic gases, generated from such DG, is inimical to the health of those in the vicinity of these equipments, especially those sited in the middle of local markets and residential areas. Breathing CO can cause headache, dizziness, vomiting, and nausea. If CO levels are high enough, it can result to being unconscious or death. Exposure to moderate and high levels of CO over long periods of time has also been linked with increased risk of heart disease. Health effects of particulate matter on the other hand may include cardiovascular effects such as cardiac arrhythmias and heart attacks, and respiratory effects such as asthma attacks and bronchitis. SO2 irritates the nose, throat, and airways to cause coughing, wheezing, shortness of breath, or a tight feeling around the chest. The effects of SO2 are felt very quickly and most people would feel the worst symptoms in 10 or 15 min after breathing it in. Long-term exposure to these pollutants may result in accumulation leading to other diseases. For example, Scheers et al. (2015) found that a long-term exposure to PM10 is a risk factor for stroke event. Presently, in Nigeria, there are no regulations guiding the emission limits from use of DG in BTS operations. This study is particularly relevant as basis for setting guiding regulation for DG-powered BTS operation.

Table 1 The total number of active subscribers per one hundred people, calculated in percentage relative to 1999 (source: NCC 2017)
Full size table
Table 2 Growth of BTS in Nigeria
Full size table
Fig. 1
figure 1

A typical BTS provided with diesel generators

Full size image

Study area and methodology

Study site location and climatic description

Akure, Nigeria (latitude 7.20o–7.32o and longitude 5.13o–5.26o) has been chosen as the study location based on the status of the city based on MCI designation as a developing millennium city. The climate of the location is modulated by two contrasting wind regimes producing two seasons: dry and wet. Based on data available between 1983 and 2013, the average annual rainfall during the wet season is about 1500 mm while the dry season is characterized by hazy and cool harmattan. Harmattan is the name given to the north easterly wind blowing from Sahara desert into West Africa during the period between November and February, usually very dry and dust laden. Average temperature ranges between 21.4 and 31.3 oC per annum. Annual average humidity also ranges between 45.4 and 100%. These climatic conditions place the location as a type Aw (tropical wet-dry climate) in Köppen climatic classification. The vegetation of the study area can be described as tropical rainforest type (Balogun et al. 2014). The choice of BTS location is dictated by certain specific requirements among which are safety of equipment, high ground for better coverage, population density, and of course, the cost of acquisition of the site. Thus, a number of BTS is located very close to residential buildings allowing the minimum permissible distance of 5 m away from the building (NBS 2017) in some cases, while some are located on higher grounds for obvious reasons. In Table 3, basic measurements were taken from nine BTS located in the proximity of Oja Oba which include longitude, latitude, numbers of DGs, and their ratings (manufacturers specifications of the generators, which include fuel consumption and emission strength), number of years of operation and distance from residential buildings (when sited in residential area). The study locations and geographical positions of BTS are shown in Fig. 2 with red dots. Specific location of base stations considered in this study, based on proximity to the neighborhood of interest is given in Table 4.

Table 3 Emission characteristics of the BTS diesel powered generators used in the study
Full size table
Fig. 2
figure 2

Google map of Akure showing the locations of the BTS represented as red dots used in the study. The circled area is the approximate location of highly populated market place. Distances on both x- and y-axes are from a reference point (x,y:0,0). The shaded cross represents the approximate rooftop location of an installed AWS

Full size image
Table 4 BTS locations in the vicinity of the Major Market in Akure. The following acronyms are defined; OAD, Oba-adesida road; OJA, Oja-Oba; PAL, Palace; POS, post-office
Full size table

Field measurements and instrumentation

For the purpose of pollution modeling and on-site assessment, a two 3-month field campaign was carried out firstly, between December 2016 and February 2017 and secondly between June and August 2017 which involved measuring ambient air criteria pollutants in and around the identified BTS. The first phase was chosen to coincide with the transition period between dry and wet seasons while the second phase was during the peak of monsoon in the study location. Since distinct and different atmospheric circulation characterizes each season, it expected that pollution accumulation or dispersion during the two seasons will be significantly different. Three criteria air pollutants viz CO, NOx, and PM10 were measured using Aeroqual®Portable Particulate Monitor (PPM) which has detachable sensor heads for CO and NOx and PM10. The PM sensor head uses a laser and optical sensor to measure light scattered from particles passing through the laser beam. The optical sensor transforms scattered light into electrical signals which are processed to provide mass measurements of the pollutants. PPM is sensitive to humidity and temperature. It works best in the range of – 5 oC and 50 oC. It was used to sample pollutant concentrations at different point by moving in at least two concentric circles at a radius interval of 200 m around BTS. On each circle, at least two measurements were made during the daytime only. Measurement procedures involved activating the sensor by switching on and allowing sensors to stabilize for ten (10) minutes. According to users’ manual, the instrument should be allowed to “warm up” for three (3) minutes and allowed to stabilize for at least seven (7) minutes. After initial warm up and stabilization, measurements were taken between 9:00 am and 5:00 pm daily at interval of 30 min during the period of observation. Nighttime measurements were not considered in this study for logistics reasons. Each BTS were sampled two times during each of the campaign periods. In this way, a total of 10 measurements were obtained for each BTS during the first phase and another 10 measurements were also obtained during the second phase. All measurements were downloaded into storage for further processing. It should be noted that simultaneous measurements for all the BTS were not attainable because of limited number of monitors. In addition to the pollutants monitors, data of two automatic weather stations (AWS) obtained. One of the AWS was installed on the rooftop of the school of Earth and Mineral Science Building at Federal University of Technology campus located on latitude 7o 17’ N and longitude 5o 18’ E. The second AWS was sited approximately north of the first AWS within the city center on latitude 7o 14′ N and longitude 5o 11’ E. Wind speed/direction, temperature humidity solar radiation, rainfall measured at 1-h interval, and other meteorological parameters necessary for assessment of pollution dispersion were extracted from downloaded data on the AWS loggers. The location of AWS cited at the city center is characterized by open space without vegetation but few nearby buildings approximately between 6 m and 8 m high. The instrument supported by a 2-m stand was attached to the roof tallest building 8 m tall, making the height of the instrument approximately 10 m. Data from both PPM and AWS were quality assured by removing few bogus and unrealistic values checked against sensors performance limits. For all measuring and instrument locations, GARMIN eTrex® Vista Global Positioning System (GPS) was used to establish the longitude and latitude.

Method of analysis

Simulation of pollutant dispersion for each of the BTS was carried out using American Meteorological Society and Environmental Protection Agency Regulatory MODel (AERMOD v15181). The description and test performance of the model can be found in Cimorelli et al. (2005). The model is a Gaussian plume model refined for simple and complex domain of maximum dimension of 50 km × 50 km with users’ choice of mesh size in circular or linear arrangements. AERMOD requires surface hourly data and upper air daily data that are representative of the location’s ground surface and upper atmospheric conditions. Both the surface and upper air data must be prepared in a range of format that is acceptable to metrological processor (AERMET) component of the AERMOD system. In addition, the model also accepts on-site data prepared in users’ format. Three sets of data were explored for use in this study, namely NCDC DS3505 Integrated Surface Data download through NCDC ftp site (*.ISH format), upper air data from NOAA/ESRL Radiosonde Database (*.FSL format), and on-site data. For Nigerian stations, NCDC data have significant missing values; in addition, none of these sources has complete and coherent dataset, and alternative is sought. In this study, output from weather Research and Forecasting model (WRF) was used to generate surface and profile data file needed to run AERMOD. WRF is a mesoscale numerical weather predict model that can produce simulations based on actual atmospheric conditions; a detailed WRF description can be found in the WRF users’ guide (Skamarock et al. 2019). WRF was run over two West African domains shown in Fig. 3 with parent and daughters’ domain resolutions of 36 km, 12 km, and 4 km respectively, hereafter referred to as domain 1, domain 2, and domain 3. Some basic configuration based on best combination and performance over West Africa used in the model run is presented in Table 5. Initial boundary conditions from Global Forecast System (GSF) analyzed 3-hourly data were utilized to initialize the model. Outputs from WRF model within domain 3 were passed to the Mesoscale Model Interface Program (MMIF), which converts prognostic meteorological model output fields to the parameters and formats required for direct input into dispersion models. As one of USEPA’s pollution models input data processing software, MMIF is specifically designed to process geophysical and meteorological output field from Fifth Generation Mesoscale Model (MM5, version 3) and the WRF model (advanced research WRF (ARW]) core (Brashers and Emery 2012). The final output from MMIF consists of year, month, day, yearday (JD), hour, heat flux H (W m2), friction velocity u* (m s−1), deardorff velocity w* (m s−1), vertical potential temperature gradient above zi (K km−1), (convective) mixed-layer depth zi(m), (mechanical) boundary layer depth zm(m), Obukhov length L (m), roughness length zo(m), Bowen ratio, Albedo, hourly mean wind speed U (m s−1), wind direction (°), reference height for wind speed and direction (m), temperature (K), reference height for temperature (m) were fed directly into AERMOD bypassing AERMET. Finally, pollutant concentrations simulations were carried out for each BTS using the combined data covering the two campaign periods (January 2–March 31, 2018 and June 1–August 31, 2018) following the flow chart in Fig. 4. The treatment of NOx/NO2 conversion in AERMOD is quite different from other pollutants because of rapid conversion of NO to NO2 in the atmosphere. In this study, tier II mechanism algorithm in AERMOD was adopted. The method uses an ambient ratio technique that computes the NO2/NOX ratios based on a sixth-order regression of ambient NO2 and NOX data (Podrez 2015). Utilizing this method, NO2/NOX ratios have a maximum of 0.9 at lower NOX levels and a minimum of 0.2 at higher NOX levels. In the AERMOD run, wet and dry deposition loss parameterizations were included. Both WRF and MMIF handle orography internally and the results are transferred to AERMOD to utilize. Building dimensions were not included in this simulation. Furthermore, health risk assessments (HRA) was determined according to various available methods in the literature (Barnes and Dourson 1988; Gaylor and Kodell 2002; Foronda et al. 2007; Mutahharah et al. 2014). A more convenient index for assessing health risks pose by a toxicant is the so-called hazard quotient (HQ), which is defined as follows:

$$ \mathrm{HQ}=\mathrm{Dose}/\mathrm{RfD} $$
(1)
Fig. 3
figure 3

WRF simulation domains

Full size image
Table 5 WRF model configuration used in simulating atmospheric data
Full size table
Fig. 4
figure 4

Flow in AERMOD modeling system for the study. SFC and PFL are surface and upper air data respectively. RCM is Regional Climate Model

Full size image

Equation (1) is used to characterize non-carcinogenic health risk as a result of inhalation of toxicant (US EPA 2009). HQ is also a dimensionless quantity with critical value that of unity depending on the RFD values at the denominator. When HQ < 1, this implies no-adverse health effect, but HQ > 1 this implies adverse health effect (US EPA 2009).

Dose intake was calculated using Eq. 2; it is measured in μg of the contaminant per kg of body weight per day (Nazaroff et al. 2001) as follows:

$$ \mathrm{Dose}=c\times \frac{\mathrm{CR}\times \mathrm{EF}\times \mathrm{ED}}{\mathrm{BW}\times \mathrm{AT}} $$
(2)

where C is the concentration in μg/m3 of simulated or observed pollutant concentrations in air; CR is the contact rate (30m3/day air inhalation); EF is the exposure frequency (90 day/year); ED is the exposure duration (30 years); AT is the average time. For non-carcinogens, AT = ED = 30 years × 365 days × 24 h and carcinogens, AT = 70 years × 365 days × 24 h; BW is the body weight (70 kg for adults)

RFD, which is the reference dose factor, is the ratio of the no adverse effect level over the uncertainty factor, which is given as in Eq. 3. In the determination of RFD for respiratory diseases considered in this study, we scaled RFD by the morbidity ratio (MR) for humans, to help characterize the risk of different respiratory diseases associated with inhaling dose of each toxicant.

$$ \mathrm{RFD}=\frac{\mathrm{NOAEL}\times \mathrm{MR}}{\mathrm{UF}} $$
(3)

Results and discussion

Wind pattern in Akure during the dry and wet period is show as wind rose in Fig. 5. As described in the previous session, two seasonal wind regimes dominate over the location: southwesterly wind during the wet period and north easterly wind during the dry. In Fig. 5a, the prevailing wind over Akure was southwesterly emanating from the Gulf of Guinea as part of the monsoon systems, and therefore is laden with moisture. The maximum daily winds at 10 m height during this season ranges between 5.6 and 5.2 ms−1 while the minimum range is between 1.2 and 1.0 ms−1. The moist characteristics of the wind modify atmospheric conditions and thus pollutants dispersion during the period. One of such characteristics is the frequent strong wind embedded in thunderstorms associated with the season (Oluleye and Okogbue 2013). Similarly in Fig. 5b, northeasterly prevailed. Well-established in the easterly winds are a deep layer of dry and dusty harmattan winds and the presence of these winds over the location implies complete modification of atmospheric variables over the location. Maximum winds at 10 m height during this period range between 4.5 and 3.2 ms−2 while the minimum wind range is between 1.2 and 1.0 ms−2. Therefore, the dissimilar atmospheric characteristics play a major role in air quality of the location. Prevailing average meteorological variables are compared with climatological values during the study period as presented in Table 6. There are slight variations in both measured and simulated values of the meteorological variables compared to climatology (long-term mean) of the study period. Measured rainfall shows a drier than “normal” situation during the dry period. Simulated rainfall was also lower than “normal.”

Fig. 5
figure 5

Wind rose during the a wet b dry periods chosen for this study

Full size image
Table 6 Mean value of meteorological variables during the study period compared with long-term climatology
Full size table

Predicted concentration of pollutants

The first highest predicted concentration of CO for average period of 1 h, 3 h, and 8 h during the wet season is shown in Fig. 6. Predicted concentration for ranges from maximum of 1013.4 μg/m3 to a minimum of 1.5 μg/m3, and the maximum concentration lie at the center of the market very close to the sources of the emission at about 300 m away. As expected, a 1-h average is more polluted than the 3 h and 8 h because during the first few hours of release, pollutants dispersion is usually slow. However, longer residence time allows mixing and transformation of pollutants in the atmosphere. Longer residency also enhances deposition and scavenging. Hence, range of predicted concentration in the entire domain of the study area for 3 h is between the maximum of 480.5 μg/m3and minimum of 0.9 μg/m3 while the range for an 8-h period is between 390.9 and 1.0 μg/m3. Of significant import is the pattern of pollutant dispersion during the dry period in the study location. It is noted that during this period spreading is restricted to a few meters away from the points of release due to characteristics of the atmospheric factor controlling the spreading. Wet period in the study location is characterized by intermittent rain and strong winds resulting from convective activities with high dynamic instability (Oluleye and Okogbue 2013). Strong wind, Brownian diffusion, and high dynamic instability could lead to strong turbulence and vigorous mixing which account for vigorous spreading and quick dilution (Foret et al. 2006). The very high moisture content of the atmosphere during this period facilitates wet removal; deposition of pollutants and, in case of PM10, aerosol absorption due to solubility also takes place. Apparently, the situation during the dry season is in sharp contrast to that of wet season presented in Fig. 7. During the dry season, atmospheric is characterized by high pollution content in the study area due to slow removal from the atmosphere. For example, the range of a 1-h first highest CO concentration during the period is between maximum of 669.3 μg/m3 and minimum of 1.2 μg/m3 for a 1-h averaging period. This condition changes slightly for a 3-h averaging period with a range between 425.7 and 0.3 μg/m3. Similarly, the range for an 8-h averaging period is between 90.0 and 0.2 μg/m3. Unlike the wet season when the dominating wind is south westerly, the prevailing wind during the dry season is the north easterly wind which characteristically is laden with aerosols in form of suspended particulate matters mainly dust transported from far location of sahelian origin (Omotosho 1985). AERMOD performance during both seasons shows that the model was able to capture the distinctive characteristic dispersion of pollutants peculiar to each period. Comparing dispersion during the wet and dry periods, it is obvious that removal rate is higher during the wet season than the dry season. This result has a far-reaching health implication on the occupants of the study area. Ishak and Everard (2017), Gunawan et al. (2018) reported that air pollution related diseases such as acute bronchitis and coughing are prevalent during the dry season in West Africa. The report of Ishak and Everard (2017) was largely based on long-range air pollution transport connected to spreading of mineral dust from dust emitting areas and gaseous pollutants from biomass burning across the region. Air pollution-related diseases is therefore expected to become more aggravated in a highly polluted location where effluent from generators are rampant in addition to already high “background” values. In Table 7 we compare the maximum predicted concentration of CO, NOx, and PM10 with observation during the dry period. It is generally noted that the measured concentrations are higher than the model predicted values. Although, model biases are an issue in modeling exercise, the net difference in the measured and model results could be recognized as the background values. AERMOD modeling bias has been discussed by various authors (e.g., Cimorelli et al. 2005; Rood 2014), they compared AERMOD with observations and other models and found that the AERMOD showed very strong positive bias of between 0.3 and 1.46 μg/m3 at shorter distances from a reference point and at longer distances the positive bias is between 0.54 and 0.96 μg/m3. Thus, the 1-hourly background concentration of NOx, PM10, and CO during the dry period is respectively 11 μg/m3, 32 μg/m3, and 35 μg/m3, which translate to 14.3%, 19.3%, and 5.2% relative to predicted values. The “background concentrations” is assumed here to be the combination of model bias concentrations and actual background values. The latter is yet to be documented for the study area. Table 8 also shows similar results but for wet period. Measured concentrations of pollutants are still higher than the predicted values except for PM10, which exhibited reduced atmospheric load during the wet period. The reduction PM10 is attributable to constant moisture scavenging. The dust particles dissolve easily in wet environment therefore restricting their dispersion and accumulation. The model bias (reduction due to atmospheric removal) between predicted and measured PM10in an average of 1 h, 3 h, and 8 h respectively are10.9, 4.1, and 1.2 μg/m3.

Fig. 6
figure 6

AERMOD Predicted concentrations of pollutant in the study area during the wet period. a Predicted 1-h average CO concentration. b Predicted 8-h average CO concentration. c Predicted 24-h average CO concentration

Full size image
Fig. 7
figure 7

Same as in Fig. 6 but for dry period. a Predicted 1-h average CO concentration. b Predicted 8-h average CO concentration. c Predicted 24-h average CO concentration

Full size image
Table 7 Measured and first highest time average-predicted pollutants concentrations (PC) during the dry period. Percentage difference (% diff) was calculated relative to predicted values
Full size table
Table 8 Same as in Table 7 but for wet period
Full size table

Health risk assessment

Comparison with standards

We compared predicted and measured pollutants concentrations with World Health Organization (WHO), European Union (EU) and local authority, and the National Environmental Standard and Regulation Enforcing Agency (NESREA) standards, and the result is presented in Table 9. The purpose of comparing the “in situ” pollutants concentrations, in our case both the model predicted and measured, with some regulated set standards is to evaluate air quality of the study area based on permissible atmospheric loads of these pollutants in line with the existing regulations to protect the health of the people. Air quality standard limits are quite variable from one country to another, and some countries place priority on some pollutants than others depending on objectives and current loads of the pollutants in the atmosphere; therefore, no complete standards are available. However, WHO standards are stricter than country’s limits. The predicted and measured CO, NOx, and PM10concentrations are precariously below limits in all cases of time average.PM2.5 are complex mixture of carbonaceous particles that have diameter that is less than 2.5 μm (Debrock et al. 2009). PM2.5 result in respiratory tract irritation, cardiovascular, and lung diseases and can also increase the frequency and intensity of asthmatic attacks (Mbiaké et al. 2017). The mean concentration threshold value for PM2.5 set by WHO is 25 μg/m3 for 24 h. NOx compounds generally cause ozone depletion and higher concentration of NOx leads to the production of ozone in the troposphere (He et al. 2014a, b). The concentrations of NOx in the ambient air are usually considered non-toxic, which is the reason while it is not subjected to any confining standard (Fourmeaux et al. 2000). The mean standard threshold value for NO is 200 μg/m3 for an hour. Similarly, the accepted mean threshold value for CO by WHO is 30 mg/m3(approximately 30,000 μg/m3) for an hour and 10 mg/m3 for 8 h. The simulated and measured values of CO are lower that the WHO limits for both 1 h and 8 h.

Table 9 comparison of highest predicted and measured pollutants concentrations with local and international standards
Full size table

Health hazard quotients

In this section, we investigate the effects of CO, NOx, and PM10 concentrations on human health exposed to daily dose from the generator source. Extreme cases were chosen to illustrate the severity of human exposure to these pollutants. Thus 1st, 2nd, and 3rd highest values of the model results were chosen in addition to highest measured concentrations and the potential effects were determined against air pollution-related diseases such as asthma, aggravated asthma, eye irritation, and reproductive and development toxicity. The effects of human exposure to air pollution begin to manifest when a critical acceptable pollutant dose is exceeded (Deshpande et al. 2017). This threshold varies from person to person; however, when no-observed-adverse-effect-level (NOAEL) or low-observed-adverse-effect-level (LOAEL) threshold level is exceeded, certain discomfort will become noticeable resulting into human morbidity. In Table 10, the ratio of NOAEL/LOAEL referred to as MR is shown for the aforementioned diseases. These are the standard value of MR for the different non-carcinogenic diseases used for the HRA based on different literatures.

Table 10 No/low observed adverse effect level (NOAEL/LOAEL) reference value and the uncertainty factor (UF) used for the calculation of health risk in human for 1-h exposure. These values are based on the references cited
Full size table

These values are based on experimental results of the authors cited. Using these MRs and the uncertainty factor the health HQs were obtained as presented in Tables 11, 12, and 13. HQ for CO in Table 11 shows a high level greater than unity in both dry and wet seasons indicating high-risk level asthma, aggravated asthma, eye irritation, and reproductive and developmental toxicity of a 1-h exposure for persons working or residing where these accumulated pollutant concentrations are located. In Table 12, HQ for NOx are consistently lesser than unity for dry and wet seasons, and thus poses not apparent health issues related to diseases considered in this study. Similarly, QH for PM10 shows risk value lesser than unity during the wet season and greater during the dry seasons implying that health risks related to diseases considered here are higher during the dry season.

Table 11 Health-related hazards quotient (HQ) of the predicted and measured pollutants concentrations of carbon monoxide (CO) in the study area. R&D is the reproductive and developmental toxicity, asthma and INC asthma is the tendency to cause asthma and increased asthma respectively as a result of exposure, EI&BE is the quotient for eye irritation and burning eye
Full size table
Table 12 Same as in Table 11 but for NOx
Full size table
Table 13 Same as in Table 11 but for PM10
Full size table

The result obtained here could be discussed based on the type of fuel being used in the generator and the effects of weather variables on the dispersion of the pollutants. Health risks due to CO are all-time consistently higher because of the generators are run on diesel fuel. Engine running on diesel fuel are known to generate high level of CO due to incomplete combustion (Lippmann and Morton 2009). This explains the higher concentrations of CO emanating from the generators. Furthermore, health risks due to CO are particularly higher during the wet season when quick dispersion of pollutant is hindered by weak winds and static stability of the tropical atmosphere (Oluleye 2017). However, NOx human health risks are minimal because fuel in use contains very insignificant quality of nitrogen compounds. PM10 higher HQ during the dry season shows that particulate matters pose greater health risk to human during this period than wet season. In a tropical location such as Akure, the study area, dust is a prominent feature of the atmosphere; however, the lower risks during the wet season is imbedded in strong wet scavenging of PM10 during the very frequent rainy phases of the season. Thus, it can be surmised that weather elements have significant role in appropriate dispersion or otherwise of pollutant depending on seasons. In general, pollutants pose higher health risks to human during the dry season than the wet season in a tropical location.

Conclusions

In this study, we have simulated the dispersion of pollutants mainly CO, NOx, and PM10 emitted from telecommunication mast powdered by generator set in tropical location of Ojo Oba, Akure, using the AERMOD model. We have used both the surface and upper air data generated from a low-resolution configuration of WRF model to drive the pollution dispersion model. This offered a significant advantage in overcoming missing and sparse upper air data problems. Accordingly, the air quality assessment of the study area was carried out using “on-site” data of the location.

Then, results show that carbon monoxide CO which is a product of incomplete hydrocarbon combustion and one of the important elements in the atmospheric chemical process and which has an adverse effect on both the central nervous system and other sensory organs is the highest emitted pollutants of all the three pollutant simulated. The 1-h highest CO concentration simulated was 1013.4 μg/m3 during the wet period of the year, while measured value of CO for the same period was 900.8 μg/m3. Similarly, 1-h highest NOx and PM10 of 78.77 μg/m3 and 58.69 μg/m3 respectively were obtained from simulated results during the dry season.

Some air pollution related diseases such as asthma, aggravated asthma, eye irritation, and reproductive and developmental toxicity were evaluated based on measured and simulated concentrations of the pollutants. Human health risks due the simulated pollutants were assessed using the hazard quotient, the results of which indicated that CO poses greater health risks during both the dry and wet seasons, followed by PM10 during the dry season only. However, NOx was observed to have hazard quotient lesser than unity during both the wet and dry season, thereby posing relatively low health risks to human population in the study area. This assessment is particularly important because of the increasing numbers of telecommunication mast base station which support ever expanding network of service provider in the West African subregion. These base stations are powered using a generator running on diesel fuel, which produces obnoxious gases due to incomplete combustion. The impact of these gases on air quality and by extension, on human health is very severe. It is therefore recommended that renewable sources of power to support the electrical needs of the base stations be sought in solar and wind energy which emit zero pollutants to safeguard the health of population around the vicinity of these base stations.