Introduction

A health impact assessment (HIA) is a process to estimate the nature and probability of adverse health effects in humans who may be exposed to chemicals in contaminated environmental media, now or in the future. Protecting human health from the risk of pollutant exposure is the ultimate goal of environmental protection. Even though environmental impact assessment (EIA) has been a proven tool over the past 35 years and has a good track record in evaluating the environmental risks and opportunities of project proposals and improving the quality of outcomes, its theory and practice for HIA continues to lag behind policy expectations (Lebret 2015; Rouéle and Jabot 2018). China has experienced a worsening environmental situation in recent years due to rapid industrialization (Niu et al. 2017; Lamoreaux and Arizona 2019). While HIA is a part of the EIA system, as claimed by the China Act Environmental Impact Assessment (revised edition, 2018), a theory integrating HIA into EIA is still under development and remains inadequate in practice (Chang et al. 2017). Most local administrative requirements for HIA in project planning EIA are relatively weak and may even be thoroughly ignorant of environmental decision-making, therefore making them disconnected from HIA in practice.

There is increasing interest in merging HIA into an operational project EIA (post-EIA), which is important for decision-making when planning surrounding land-use and the regulation of protection polices for nearby populations. However, the key technology for merging HIA into EIA is still a challenge and most HIA studies have focused on the accuracy of risk assessments in toxicology and epidemiology (Ruby et al. 1999, Bari and Kindzierski 2018a, b), while there has been little focus on concepts linked to spatial factors. The area impacted by an operational project is defined by its geographic location and the emission intensity, as well as the likely level of exposure, which is associated with the spatial dispersion concentration and the population’s exposure patterns. Therefore, more scientific techniques are expected to be introduced into the theory of EIA. With insights into exposure assessments, more scientists are beginning to associate population exposure profiles on a geographic scale with the diffusion of pollutants (Wang et al. 2018a). A geographic information system (GIS) provides a perfect solution to overcome the obstacle of a scenario population’s exposure profiles based on the population’s social characteristics and the spatial distribution of the pollutants (Vu et al. 2013; Minolfi et al. 2018), which forms a linkage for an HIA scenario based on land use to a post-EIA via counterpart spatial analysis. Several similar studies on how to conduct an HIA have been outlined, some of which are inspiring (Aliyu et al. 2014; Boudet et al. 1999; Spickett et al. 2012; Lei and Hilton 2012; Lebret 2015). Of these, one study was on the health risk assessment of a modern municipal waste incinerator that proposed linking an air pollutant dispersion model to the exposure parameters using the demographic characteristics of the nearby population Boudet et al. 1999). Other studies have attempted to incorporate wind flow effects in land-use regression models to predict nitrogen dioxide concentrations for health exposure studies, with some adopting remote sensing techniques (Arain et al. 2007; Knibbs et al. 2018). This inspired us to introduce spatial factors to link a scenario HIA via land-use analyzes to a post-EIA.

Here, a case of municipal sewage treatment plant (MSTP) volatile organic compounds (VOCs) is presented. Odors released from treatment units have long been a challenge when managing MSTPs (Gruchlik et al. 2017) and are naturally linked to the adopted treatment technique. MSTP odors are regarded as typical MSTP pollutants (Wang et al. 2018a, b) and generate wide concern. Here MSTP odors are taken as an example to probe HIA theory with a post-EIA. This paper addresses a way to introduce the geographic factors of a scenario HIA via land-use to the post-EIA of an operational MSTP.

Materials and methods

Screening priority control MSTP pollutants

MSTP odors are reported widely and are classified as odors volatile pollutants, including alkanes, alkenes, aromatic hydrocarbons, halogenated hydrocarbons, sulfur-containing organic compounds, oxygenated organic compounds, hydrogen sulfide (H2S), ammonia (NH3), and carbon disulfide (Lewkowska et al. 2016; Liu et al. 2017). Additional odors pollutants have been confirmed as being released from grate, grit chamber, and sludge treatment units (Liang and Liu 2016; Meng et al. 2016; Ren et al. 2018). In this paper, two typical pollutants, H2S and NH3, were chosen as characteristic pollutants indicating odors because neither has ever been explored at high levels, that is, H2S levels of 145,700 μg/m3 and NH3 levels of 63 μg/m3 (QEP et al. 2010; Lehtinen and Veijanen 2011; Wang 2013), for an extreme situation at an operational MSTP.

Meteorological parameters

The studied MSTP is located in the Haidian District of Beijing. Meteorological data from 1992 to 2016 were surveyed and used for further analyzes. Wind directions, which play an important role in pollutant diffusion predictions, were summarized. Austal 2000 G was employed to simulate the pollutant diffusion, and meteorological data of the annual 2017 parameters were adopted.

Pollutant diffusion simulations

Environmental monitoring

The environmental monitoring included both manual and online methods. The monitoring time was set to the summer because volatile odors are known to be relatively serious during this period. Manual sampling was conducted following the requirements of China standard Air and exhaust gas monitoring and analysis methods. The H2S analysis followed the China standard GB/T11742–1989, the NH3 analysis followed the China standard HJ533–2009, and the odor analysis followed the China standard GB/T14675–93. All the procedures, including manual sampling, monitoring, and analysis, complied strictly with quality control regulations. The manual monitoring period lasted from August 28 to 30, 2018, with a frequency of two times a day. The monitoring times were 03:00–04:00 and 16:00–17:00, avoiding rainy and windy days.

The OLFOSENSE (AIRSENSE) instrument was adopted as an online monitor; this instrument is composed of a variety of hybrid sensor arrays (four metal oxide sensors, one photo-ion detector, and four types of electro-chemical sensors). The levels of the pollutants, that is, H2S, NH3, and odors, were recorded by the online instruments in real time. Instantaneous data were read five times a minute, with the average value being calculated and recorded automatically. All equipment was regularly inspected and maintained in a timely manner to ensure the quality of the readings. Abnormal MSTP operating conditions may produce extreme emissions; to obtain the frequency of such extremes, we used the maximum value monitored, divided by the average annual predicted value.

Prediction of the emission source

Two situations were taken into account to evaluate the source emissions: normal and extreme conditions that are, assuming that all deodorization measures failed. The incident probabilities of an extreme condition were weighted by the accident odds for similar industries.

Pollutant diffusion simulation

The main simulated data came from the monitor results, and the diffusion variation was simulated using the Austal 2000 G software. The odor time, that is, the duration of the odors, was chosen as the basic parameter. Odor levels higher than the value set as the olfactory threshold, cGS, were selected by the program. The hourly average values, \( \overline{c} \), of the selected odor parameters were calculated with a guarantee rate of more than 90%, c0.9. When the hourly average value of c0.9 was higher than the threshold value cGS, the odor duration time of the monitored odors in that hour was recorded. That is, when the odor value c0.9 was higher than the threshold value cGS, the time in which odors were monitored in the counterpart hour was recorded as the odor duration time. In the spatial analysis, each grid from others’ distance for prediction was 20 m, and the total number of grids was 200 × 200. The center point coordinate was near the main odor discharge cylinder in the MSTP. The research area was designated by a geographic cycle with a radius of 4 km from the center point.

Scenario exposure levels according to the land-use analysis

Exposure profile survey

The demographic characteristics and the counterpart exposure behaviors were obtained via the land-use attributes. The land-use information was obtained from the municipal land administration. According to the information offered by the local government, the land uses in the Haidian District were classified into several types, including commercial, office, residential, school, and grassland. These classes were verified by our field investigation and data analysis. The results indicate that there were no other industries near the MSTP; however, there were some residential areas and office buildings. A grid of 20 m × 20 m was set in the program. The population number in each land grid was counted, along with the social characteristics of the population. Similar land-use types were consolidated into a single type, and then the counterpart population’s exposure behavior was investigated. To obtain the average behavioral parameters, a group of users or residents on each type of land was selected randomly to answer a questionnaire; their exposure profiles were surveyed and then employed as land-use attributes.

A total of 310 residents around the MSTP were selected randomly for the questionnaire. All questionnaires were administered on the condition that informed consent forms were signed. The questionnaire content included education, occupation, lifestyle, daily life exposure patterns, and the participant’s subjective understanding of the risk of odors. The population was categorized by socioeconomic factors including stay-at-home residents, office workers, and students. The likelihood exposure times according to the demographic characteristics were summarized and assigned to counterpart land-use attributes.

Exposure levels based on the land-use analysis

Following the simulation of the spatial superimposed distribution of the pollutants, the grid center concentration was taken as the average inhalation level for that particular location. Exposure levels for different land uses were calculated following Eq. (1), a popular technique used to make inhalation exposure assessments.

$$ \mathrm{ADD}=\frac{\mathrm{c}\ast \mathrm{IR}\ast \mathrm{ET}\ast \mathrm{EF}\ast \mathrm{ED}}{\mathrm{BW}\ast \mathrm{AT}} $$
(1)

Here, ADD is the daily average exposure dose; C indicates the concentration of the chemical substances in the air; IR represents the amount of respiration; ET indicates the duration of daily exposure; EF designates the frequency of exposure; ED indicates the duration of the exposure; and BW represents the local population’s average weight, according to the data gathered by our investigation. The final parameter AT is the total expected exposure time.

For mixed land-use areas, the average exposure levels were estimated using Eq. (2).

$$ \mathrm{PWEL}=\frac{\sum \left(\mathrm{pi}\times \mathrm{ci}\right)}{\sum pi} $$
(2)

PWEL is the population’s average exposure level, where the population refers to people with activities in a certain geographic area, which can be divided into grid numbers i on a map; pi represents the population inside each grid and ci indicates the average concentration of air pollutants inside each grid in units of μg/m3 or mg/m3. Different types of people, categorized by their exposure patterns based on their social attributes, may experience large differences in their exposure to MSTP pollutants. To establish the PWEL values, it is important to obtain practical activity behaviors for the demographic characteristics. In our study, we attributed the social attribute types of the population on the combined gridded map because it is easy to attribute people’s exposure profiles to spatial data.

Scenario health risk analysis

After assigning the exposure amount to each geographic coordinate, a scenario health risk analysis was further conducted with respect to the traditional inhaled HIA (Zhou et al. 2011); the hazard quotient (HQ) was used to describe the risk rating:

$$ HQ=E/ RfD, $$

where E indicates the total exposure amount and RfD indicates the reference dose. If HQ ≥ 1, the HIA is unacceptable; otherwise, it is acceptable.

Some uncertainty factors that may contribute to variations in the health risk need to be considered. The collected material was assumed to have an accuracy deviation of ±5%. Different types of people, such as children, elder, and pregnant women, were assigned a health risk sensitivity scope of ±5%. In addition, the cumulative effects of social factors (e.g., economic income and professional background) contributed an uncertainty of ±10%. Considering the comprehensive uncertainty mentioned above in HIA, the final evaluated HIA should be adjusted such that the prediction value is R ± R × 20%.

Statistical analyses

Statistical analyzes were performed using the SPSS 18.0 software. All the monitoring data were submitted to a Kolmogorov–Smirnov (K–S) normal distribution test. When P > 0.05, the data are considered to be consistent with a normal distribution. ARC GIS 10.4.1 was used to map the risk contours.

Results

Meteorological analysis

The wind direction frequency (FWD) was summarized by analyzing the meteorological data (Table 1). Due to the influence of the monsoon, the FWD varies with the season. The dominant wind direction in winter is considered to be NNE, NE, and N, with appearance frequencies of 15.3%, 13.7%, and 12.3%, respectively. In spring, the FWD shifts to SSW and S, with frequencies of 12.3% and 11.0%, respectively, followed by NNE, with a reported frequency of 11.0%. The dominant wind direction in summer was confirmed to be NNE, NE, and N, with frequencies of 13.0%, 11.0%, and 10.7%, respectively, followed by S and SSW. Meanwhile, in autumn, the dominant wind direction was NNE, NE, and N, with frequencies of 14.7%, 13.7%, and 11.0%, respectively. With respect to the historic meteorological data, annual meteorological data including hourly data of the meteorological parameters in 2017 were set as the main meteorological parameters for the simulation, and a breeze rose diagram (Fig. 1) indicates that the wind frequency in 2017 was similar to that in previous years. In 2017, the dominant wind directions were NNE, NE, and N, with frequencies of 13.5%, 11.9%, and 10.8%, respectively, and rather low FWD values were seen for E, ESE, and SE, with frequencies of 1.9%, 1.2%, and 1.9%, respectively. Our findings suggest that the main FWDs affecting air pollutant diffusion were in the NNE and NE directions.

Table 1 Statistics of frequency of wind direction (%)
Full size table
Fig. 1
figure 1

Wind rose map in 2017

Full size image

Contaminant diffusion simulations

To map the spatial diffusion of the pollutants, we need to obtain the emission strength and predict the diffusion according to the geographic and meteorological parameters.

Environmental monitoring

Air samples were monitored in August, when extremely hot weather conditions are ideal for VOC emissions. The hourly average levels of the pollutants were monitored (Table 2), as well as the values lower than the standard (0.01 mg/m3). The duration of the online monitoring was from 0:00 on August 24 to 12:00 on September 2 (Fig. 2, including Figs. 2a–d). The maximum pollutant period was observed to be between 14:00 and 18:00 (Table 3). The maximum value during this period was taken assuming that the extremely risk from source emission may took place. Here, the odor concentration was estimated to be 1318 (nondimensionalize value), the H2S emission rate was approximately 6.94 × 10−5 kg/h, and the NH3 emission rate was 6.75 × 10−2 kg/h. The extreme pollution levels were set assuming that all deodorization measures had failed; these values were assumed to be 4169 for odors, 1.96 × 10−4 kg/h for H2S, and 1.96 × 10−2 kg/h for NH3.

Table 2 Hourly average levels of pollutants
Full size table
Fig. 2
figure 2

Levels of H2S and NH3. Note: Air samples were monitored in the August, the days an extremely hot weather condition for the VOCs emission. Duration monitor online had been performing from 0:00 August 24–12:00 September 2 (a–d). The emission levels fell down to zero during the night-time because of all running chambers closed. The maximum value were chosen for the extremely scenario risk calculation

Full size image
Table 3 Prediction of the emission sources
Full size table

Prediction of the emission source and spatial simulation

Diffusion levels of the average annual H2S emissions were plotted spatially (Fig. 3). The maximum was 2.8 × 10−3 μg/m3, appearing 100 m southwest of the emission source, with a 0.14% ratio to the standard value (0.01 mg/m3). The annual average H2S levels around the spatial grids for the sensitivity subjects whereas the population was concentrated changed from 1.39 × 10−4 to 6.10 × 10−4 μg/m3 (not shown). The levels of the H2S prediction were far lower than the standard, and the maximum H2S value appeared southwest of the emission source. The annual average levels of NH3 (Fig. 4) showed a diffusion trend similar to H2S with respect to the geography (Figs. 3 and 4). The average odor levels exhibited the frequency of annual odor occurrences in each grid (Fig. 5). More than 4% of the high-frequency odors occurred at a distance of 1000 m from the emission source. The extreme condition revealed that odors occurred frequently around the sensitivity points, varying from levels of 3.2% to 10.8% (not shown) and at levels of more than 6% in some locations. It is suggested that the MSTP should take measures to mitigate odors and avoid occurrences of extreme conditions. The current findings, the prediction of the emission source, and the simulation on the spatial grids can be used to associate the pollutants with potentially sensitive segments of the population that may be exposed to these pollutants (Fig. 6).

Fig. 3
figure 3

Average annual H2S concentration distribution diagram (Unit: μg/m3)

Full size image
Fig. 4
figure 4

Average annual NH3 concentration distribution diagram (μg/m3)

Full size image
Fig. 5
figure 5

Average annual odor hours frequency (Total releases 4169 OU/s, Unit: %)

Full size image
Fig. 6
figure 6

Geographical distribution of the sensitivity people

Full size image

Exposure parameters of sensitive populations according to the land-use analysis

Exposure parameters

Based on the land-use analysis, the main geographic grids affected by pollutant diffusion were residential estates, a kindergarten, and parts of office buildings. Other areas were grassland. Similar land-use function grids (20 m × 20 m) were combined for the spatial analysis (Fig. 6), and the counterpart population were classified into several types according to their social attributes. According to our investigation, younger people paid more attention to MSTP risks and hazardous materials because they were better educated with respect to the potential risks of odors and were more exposed to social media than the older population. Office workers preferred to spend less time outdoors than children. Older residents mostly preferred to stay home with their windows closed. The socio-classification types of the population were sorted, and their exposure profiles are summarized in Table 4 (including 4a–4f). The behaviors of the exposure types were queried and investigated statistically; however, there was no difference on prevention odor (p > 0.05).

Table 4 The exposure profiles were summerized
Full size table

Estimation of the exposure amount

To protect the population as much as possible, the scenario emission amounts were assumed to be at the extreme condition for the following exposure evaluation. The annual exposure levels of sensitive populations (Table 5) were estimated taking into account parameters reported previously (Zhou et al. 2011; Du et al. 2014).

Table 5 Estimation of the sensitivity people’s average exposure amount
Full size table

HIA scenarios

Next, the non-carcinogenic risk was assessed. The RfD value for the risk assessment was derived from data on the Internet (https://rais.ornl.gov/cgi-bin/tools/TOX_seaScenariorch?select=chem). The sub-chronic oral reference dose for H2S was 3.00 × 10−2 mg/kg-day and the chronic oral reference dose for NH3 was 3.40 × 10 mg/kg-day. The HIA contour map is displayed in Fig. 7. This map demonstrates that both the H2S and NH3 health risks were acceptable for all grids, even for the extreme condition, and that both odds were less than one in a hundred. Therefore, the post-EIA of the MSTP verified the security of the surrounding population. Given that uncertainty factors were taken into account, the HIA is within acceptable levels.

Fig. 7
figure 7

Contour map of HRA

Full size image

Discussion

EIA has been successfully promoted to integrate environmental considerations into development projects. The practical applications of EIA have led to the emergence of two spin-off assessment approaches: social impact and cumulative impact assessments. Such assessments not only identify the strategic choices that might lead to more sustainable health outcomes but also reduce risks to an already operational project.

Many researchers have attempted to integrate HIA into EIA (Fehr 1999; Demidova and Cherp 2005; Bhatia 2007; Corburn and Bhatia 2007; Rajiv and Aaron 2008; Valeberg et al. 2009; Harris et al. 2015; Chanchitpricha and Bond 2018), and some have tried to employ geographic parameters to identify the scope of the health effects and the emission source sites (Johnston et al. 2019); however, a standard guideline for how to integrate HIA into EIA has not yet come into being due to academic complications with respect to this integration. The lack of HIA in EIA practices is a difficult obstacle to overcome when human health protection is concerned; however, this is less of a problem in post-EIA. The usual HIA procedure contains four steps: (1) hazard identification, (2) dose response, (3) exposure assessment, and (4) risk characterization. Exposure assessment, a crucial step in HIA, is usually performed using monitoring data with weak spatial identification (Bari and Kindzierski 2018a, b; Huang et al. 2019). The reason for the HIA deficiency in EIA in practice is the lack of a spatial factor in its overall theoretical system compared to EIA. To integrate HIA into EIA, we connected the simulation of the pollutant diffusion with the scenario population’s exposure via spatial land-use parameters. A case examining odors from an MSTP was addressed here.

Three aspects to this problem need to be considered. Traditionally, the first step involves identifying the types of adverse health effects that can be caused by exposure to the pollutants in question and characterizing the quality and weight of evidence supporting this identification. Complying with this principle, the priority control pollutants of the MSTP were weighed by their emission amounts, which are linked closely with the treatment units and techniques (Jianting et al. 2015; Lewkowska et al. 2016). H2S and NH3 were both chosen as characteristic pollutants indicating odors because they have both been released in large amounts, causing wide concern (Bandosz et al. 2000). Both are often produced from the microbial breakdown of organic matter by sulfate-reducing microorganisms in the absence of oxygen gas, in a process commonly known as anaerobic digestion. These gases can cause an unpleasant smell and produce long-term adverse effects.

The second problem relates to defining the dose–response parameters. Usually, for available toxic dose–response parameters that have been approved and deduced by previous studies, the dose–response parameters can be cited directly. In this study, a sub-chronic oral reference dose of H2S of 3.00 × 10−2 mg/kg-day and a chronic oral reference dose of NH3 of 3.40 × 10 mg/kg-day were used.

The third aspect deals with how to conduct exposure assessments by incorporating spatial factors into the HIA theory, the most important procedure to incorporate HIA into post-EIA. Two aspects need to be coordinated using the same geographic factors. First, the diffused contaminants need to be obtained based on the geography, and then the exposure amount, which is linked closely with the population’s exposure profiles on the counterpart lands, needs to be estimated. People’s activities, which are determined by their social characteristics, are correlated closely with the land-use areas in which they are located.

Meteorological parameters are critical factors in the contaminant diffusion simulation. In our study, historic FWD statistics were analyzed (Table 1) and used to evaluate the main wind directions in the prediction. The dominant historic wind directions were NNE, NE, and N, with frequencies of 13.5%, 11.9%, and 10.8%, respectively, in keeping with the annual meteorological parameters for 2017, the trends of which are shown on a wind rose map (Fig. 1). The results indicate that the main wind directions affecting the air pollutant diffusion were NNE and NE.

In general, pollutants released from the MSTP during normal conditions need to be known, and while assessing HIA, the abnormal condition in which all deodorization measures failed should be taken into account.

The monitoring of the hourly average pollutant levels via manual sampling was performed (Table 2) and indicated that all results were lower than the standard (0.01 mg/m3). The maximum values were defined as the top values from the online duration monitor (Fig. 2), which had a fluctuating scope. That is, the extreme risk was assumed at the maximum value. The predictions of the emission source contained both conditions (Table 3). The extreme odor level was 4169, and the emission rates for H2S and NH3 were 1.96 × 10−4 and 1.96 × 10−2 kg/h, respectively. The extreme HIA health risk scenario was based on the conditions necessary to eliminate most of the health risk.

More challenging techniques were employed to develop the spatial HIA. The spatial variations in the pollutant diffusion were simulated using Austal 2000 G (Figs. 3, 4, and 5). All pollutant distributions had trends similar to the FWD (Fig. 1). The maximum annual average of H2S concentration was found to be 2.8 × 10−3 μg/m3, appearing southwest of the emission source with a 0.14% ratio to the standard (0.01 mg/m3). This finding implies that this area is of primary concern. A further analysis revealed that the annual average H2S level around the sensitive points where the population was concentrated changed from 1.39 × 10−4 to 6.10 × 10−4 μg/m3. Our results indicate that the predicted H2S level was far lower than the standard and the maximum levels southwest of the emission source. However, both H2S and NH3 had the same trends as the odors, and the meteorological data played an important role in their diffusion. The dependency of the diffusion conditions of the air pollutants on the meteorological parameters is supported by previous studies (Ilten and Selici 2008; Sabetghadam and Ahmadigivi 2014).

Based on the diffusion map (Figs. 3, 4, and 5), the land uses in each grid were investigated. Crowded areas were defined as sensitive points, and the geographic distribution of their coordinates was surveyed (Fig. 6); most residents are located north of the emission source in the upwind direction. Subsequently, similar land-use grids were merged. Most areas were classified as living areas, residential houses, elderly care centers, offices, and a kindergarten. Other areas were designated as grassland. Population classes and exposure profiles were summarized based on the land-use (Table 4). Mixed buildings indicated the presence of different social types of people in the same building, including office workers, retired individuals, and students. Multi-function buildings were composed of accommodations and business offices. The behaviors of the exposure types were further queried, and the results indicate that there is no difference on prevention odor (p > 0.05). The younger generation paid more attention to the risks of the MSTP. It is important to define a timeframe for HIA, and all lifetimes were set to 70 years. Combined with the exposure parameters reported previously (Bai et al. 2007; Huang et al. 2017), the annual exposure amounts of the sensitive populations were calculated (Table 5) assuming the extreme condition.

In the final step, the spatial risk scenario characterizations provided a risk appraisal calculated on the basis of the exposure and hazard. An HIA contour map was created accordingly, showing the spatial factors linking HIA to the pollutant diffusion (Fig. 7). This map summarizes the information integrated from the proceeding steps of the risk assessment, making it easy to assess the HIA spatially. In this case, both the H2S and NH3 health risk levels were acceptable on all grids, even for the extreme condition. Including the uncertainty factors, the HIA is still at an acceptable level. Therefore, the HIA of H2S and NH3 for the MSTP post-EIA indicated that the surrounding population is not in danger.

In this paper, we proposed a systematic and thorough approach to connect HIA with post-EIA and to motivate the integration of HIA into post-EIA using the case of an MSTP. This approach will provide environmental evidence to support decision-making related to the protection of human health.

Conclusions

Human health protection is considered to be a major target of EIA, by both decision-makers and the public. However, in current operational projects, human health protection is constrained by the lack of techniques for post-EIAs. This paper presents an approach to develop an HIA scenario for post-EIA systems by introducing a spatial factor including the spatial diffusion of pollutants and the population’s activity characteristics based on land-use at the geographic scale. A specific case, the linkage of HIA to the EIA of an MSTP, was discussed here, addressing an approach to resolve the lack of HIA in the current EIA. It is important to complete EIA theory and to promote the most conducive strategy to enhance human health protection and reduce negative impacts. Further theoretical and practical studies still need to be conducted.