Environmental Risk Assessment of the Zhengrunzhou Water Source under the Influence of Oil Spill Accidents at the Wharf Group

Abstract

To compensate for the research defects of strong subjectivity in determining oil spill amount, insufficient consideration of wharf distribution, and incomplete indexes for reflecting the influence degree of oil spill accidents on water sources, and to enhance the supervision efficiency of the supervision department, this paper constructs a risk assessment system of water sources under the influence of the wharf group. The system includes a wharf group division method considering the wharf distribution situation; the calculation method of oil spill amount at wharves considering the oil tank capacity of main ship types and the production supervision risk at the wharves; the calculation method of the oil spill amount at the wharf group considering the wharf number, distribution density, production supervision risk and wharf oil spillage; the determination method for the influence degree of oil spill at the wharf group on the water sources and judgment method of supervision level at the wharf group, which takes the arrival time of oil slicks, the duration of over-standard petroleum concentration and the maximum over-standard multiple of petroleum concentration at the water intake as indexes; the method of determining the risk of oil spill accidents at the water source considering the cumulative effect of oil spill at the wharf group on the risk of the water sources; and the environmental risk assessment method of water sources considering oil spill accident risk and the anti-risk ability. Applying this system to the environmental risk assessment of the Zhengrunzhou water source in Zhenjiang City, we discovered that the flow field, wind field, oil spill location and oil spill amount were correlated with the influence degree of oil spill accidents on water sources, for which the flow field demonstrated the strongest correlation, while the wind field presented the weakest. The supervision level of the wharf group is mainly sub-key or non-key levels, but the level of the wharf group SD07 is approximate to the key supervision level during rising tide. Due to the strong anti-risk ability of the Zhengrunzhou water source, the environmental risks of the Zhengrunzhou water source under different working conditions are scarcely different and belong to the medium-risk level.

1. Introduction

As a water supply source for humans, the safety of water sources is crucial for regional economic development and human life [1]. With the rapid development of the social economy and the expansion of urbanization in China, the environmental risks of water sources are becoming higher and higher, triggering frequent pollution accidents at the water source [2,3,4]. Therefore, many scholars have conducted environmental risk assessments on water sources to ensure the reliability of water supply.

The environmental risk assessment of water sources at home and abroad is mainly categorized into water quality risk assessment and risk assessment, considering the influence of risk sources. For the former, the focus of water quality risk assessment has shifted from conventional pollutants to the health risk assessment of emerging pollutants. For example, with heavy metals and bacterial pathogens as the objects, J Nawab et al. [5] and Liu Songhua et al. [6] respectively implemented health risk assessments of water sources in the Malakand region of Pakistan and the Suzhou region of China. Taking drugs as the object, C J Houtman et al. [7], and Feng et al. [8], respectively conducted health risk assessments on three water sources in the Netherlands and the Chongqing water source of the Yangtze River. Moreover, Zhang Kunfeng et al. [9] even conducted risk assessment of disinfection by-products (DBPs) in the typical water sources of Wuhan under the COVID-19 pandemic situation. For the latter, the economic structure of developed countries exerts little impact on the environment, and they possess mature risk source control systems, while China is rapidly developing its industry, generating more and more risk sources threatening the safe supply of drinking water. Hence, the environmental risk assessment research of water sources considering the effects of risk sources is mainly concentrated in China.

The environmental risk assessment of water sources considering the effects of risk sources includes qualitative assessment and quantitative assessment [10]. Qualitative assessment mainly indirectly evaluates the risk level of water sources through the characteristics of risk sources such as industry type, production scale, production technology level, wastewater discharge as well as the characteristics of water sources like water quality compliance rate, flow rate and risk supervision ability. For instance, based on the risk judgment of risk sources from the indexes of industry type, the discharged amount of wastewater, the location of sewage outlets, and whether pollution accidents have happened, Yan Zhongchun [11] evaluated the water sources of Shanghai by further considering the natural conditions and management level of water sources. Focusing on the vulnerability of the water sources and the harmfulness of the toxic substances in the risk sources, Sun Hongliang et al. [12] conducted a risk assessment on the water sources of the main streams of the Yangtze River. Quantitative assessment means to establish a mathematical model of the water environment to simulate the occurrence of water pollution accidents in risk sources and quantitatively determine the degree of being affected at water sources, so as to assess the risk level of water sources. For example, Zhuang Wei et al. [13] constructed a two-dimensional unsteady model of the Yangtze River to simulate the pollution accidents caused by the secret discharge of enterprises in the Nanjing reach of the Yangtze River, so as to determine the influence degree of conventional pollutants, such as COD and ammonia nitrogen, on the water sources of the Nanjing Jiajiang water source after the accident. Employing the two-dimensional steady water-quality model to quantify the influence scope of pollution accidents in the risk sources, Ma Yue et al. [14] established a method system for identifying risk sources and determining the risk level of water sources, which was applied to the risk assessment of water sources in a chemical park of Nanjing.

According to the statistical analysis of Zhang Guanbing [2], more than 1900 water pollution accidents were reported in China from 2003 to 2019, of which 10.9% were caused by traffic accidents. Diesel oil leakage was the main source of water pollution caused by water traffic accidents. The formed oil slicks after oil spills into water seriously damaged the water ecological balance. Specifically, its toxic polycyclic aromatic hydrocarbons and toxic heavy metals can even harm human health through bio enrichment and food chain transmission [15]. However, as the centralized place that ships berth and goods are loaded and unloaded, the busy wharf operations and the entry and exit of ships, or even improper safety measures at the wharves, are prone to cause safety accidents such as ship collision and oil overturn, thus inducing oil spill accidents. Therefore, it is essential to carry out the risk assessment of water sources considering the impact of oil spill accidents for protecting the ecological environment of water sources and water quality safety. However, the current research has defects of strong subjectivity in determining oil spill amount in simulated accidents, insufficient consideration of risk increment caused by dense wharf distribution, and incomplete indexes for reflecting the influence degree of oil spill accidents on the water sources. For instance, Huang Yikang et al. [16] and Tian Wei et al. [17] simulated the oil spill amount as 10 t, which reflected the effects of oil spill accidents on water source oil by the arrival and departure time of oil slicks at the intake. However, Song Zekun et al. [18] simulated the oil spill amount as more than 300 t. Besides the arrival and departure time of oil slicks at the intake, the oil slick thickness at the intake was also considered to reflect the influence of oil spill accidents on the water sources. However, the risk increment caused by the dense distribution of wharves was not considered in the above studies.

Therefore, to compensate for the deficiencies of the above-mentioned studies, improve the accuracy of environmental risk assessment of the water source, and enhance the supervision efficiency of regulatory authorities on wharf groups surrounding the water source, this paper adopted the calculation method of the oil spill volume of wharf groups, comprehensively considering production supervision risk, tonnage of main ship type, wharf number and the distribution density, and proposed a quantitative method for the influence degree of oil spill accidents on the water source, with the arrival time of oil slicks at the intake, maximum over-standard multiple of oil at the intake, and the duration of over-standard oil as indexes. On this basis, a risk evaluation method system of the water source was established, which considered the cumulative effect of oil spill accidents at the wharf group. In view of the busy water traffic in the Jiangsu section of the Yangtze River [19] and the densely distributed wharf groups along the Yangtze River, this paper took Zhenrunzhou Water Source in the Zhenyang Section of the Yangtze River as an example to evaluate the environmental risk of a water source affected by the oil spill accidents at the wharf group.

2. Research Regions

Located in the southwest of Jiangsu Province, is the rich Delta of the Yangtze River, and the south bank in the lower reach of the Yangtze River, Zhenjiang City is at longitude 118°58′ to 119°58′ E, and latitude 31°37′ to 32°19′ N. Belonging to the southern climate zone of northern subtropics, Zhenjiang presents the climate characteristics of cold and dry winter, hot and humid summer, as well as four distinct seasons. The annual precipitation of Zhenjiang is 1427.2 mm, but the spatial and temporal distribution of precipitation is very uneven, with more than 60% concentrated from May to September. Located in Zhengrunzhou Village, Runzhou District, Zhenjiang City, the Zhengrunzhou water source lies on the upper Zhenyang reach of the Yangtze River, at the south bank of the south and north branch channel confluence of Siye Island. The intake was built in 1985, and the designed intake capacity is 600,000 m³/day, supplying raw water for Jinshan Water Plant and Jinxi Water Plant in the urban area, as well as some water plants in urban area and suburban counties. There are no sewage outlets or large-scale livestock and poultry farms in the first-grade and second-grade protection zones of water sources.

As for the water source location, the wharves are mainly distributed within 20 km of the upstream water source, and the 20–40 km riverbank of the upstream water source is mainly farmland and green space. Although there are many wharves at over 40 km of the upstream water source, they are far from the water source. Once the oil spill accident occurs, oil slicks will swing back and forth along with the tides, and the north branch of Siye Island also plays the shunting effect. Therefore, oil spill accidents at over 40 km of the upstream water source have little influence on the Zhenrunzhou water source. It is stipulated in the Guidelines for Protection of Centralized Drinking Water Source Environment [10], a document from the Ministry of Ecology and Environment, that the risk-source investigation area surrounding the water source is 20 km upstream of the second-grade protection zones of the water source. Hence, considering that the Zhenyang section of the Yangtze River is the tidal section, this paper takes the range of 20 km upstream to 10 km downstream of the second-grade protection zones of the water source as the research area.

According to the data of ports and wharves in the Jiangsu reach of Yangtze River, provided by Jiangsu Maritime Affairs Department and the results of field investigation, there are 27 wharves within the range of 20 km upstream and 10 km downstream of the second-grade protection zones of the Zhengrunzhou water source. The specific locations of water sources and wharves are shown in Figure 1. Wharves are mainly distributed in the south bank of the Yangtze River, of which wharves JE01~JE16 are located in the same bank of the upstream of the water source; JE26 and JE27 are located in the same bank of the downstream of the water source; and JE17~JE25 are located in Yangzhou, the north bank opposite the water source. According to the Report of Zhenjiang Environmental Quality from 2015 to 2020, the water quality of Zhenrunzhou drinking water source satisfies the Class III water standard, with the water qualification rate of 100%. Referring to the average annual flow of the Datong Hydrological Station in the Yangtze River from 1950 to 2000, it can be found that the annual average flow of the Zhenrunzhou water source is 28,700 m³/s [20]. As the investigation results of Zhenjiang Municipal Environmental Protection Bureau in 2020 reveal, all indexes of the supervision and emergency response capacity of the Zhenrunzhou water source reach the qualification rate of 100%.

3. Methodology

According to the wharf distribution situation surrounding the Zhengrunzhou water source, the relative densely distributed wharves are divided into the wharf groups. Based on the ship berth situation at the wharf group, the risk increment of oil spill accidents caused by wharf number, wharf distribution density, and production supervision risks at various wharves are considered to determine the oil spill amount when oil spill accident occurs at the wharf group, and then the two-dimensional mathematical model of water environment in the Yangtze River is constructed, simulating the transport and weathering processes of oil slicks on the river surface after an oil spill accident occurs at the wharf group under different working conditions. Consequently, the specific values of the three influence indexes of oil spill accidents—the arrival time of oil slicks at the intake, the maximum over-standard multiple of oil at the intake, and the duration of over-standard oil—are obtained. Based on the above three indexes, the risk influence index of oil spill accidents at the wharf group is calculated by the risk index method, so as to determine the supervision level of the wharf group. Furthermore, the superimposed risk influence of different oil spill accidents at the wharf group on the water source and the anti-risk ability of water sources are comprehensively considered to determine the environmental risk level of the Zhenrunzhou water source. The specific process is shown in Figure 2.

3.1. Division of Wharf Groups

Regardless of other factors, the shorter the distance between wharves, the greater the probability of a docking collision accident between wharves, and the greater probability of oil spill accidents. Therefore, a distance threshold should be set. When the distance between two wharves is less than this threshold, the two wharves are considered to be in the same wharf group, and their probability of oil spill accident within the wharf group is greater than that outside the wharf group. As Kunz C suggests [21], the stroke distance of large ships after berthing is about 550 m, and thus this paper decides that when the distance between wharves is less than 2 km and the wharves are distributed densely, they belong to the same wharf group. The prim algorithm [22] is adopted to construct the minimum spanning tree of wharves for dividing the wharf groups. For specific methods and steps, please refer to the research results of Zhou Qi et al. [23]

3.2. Risk Assessment of Wharf Production Supervision

Considering that factors reflecting wharf production and supervision risk, such as production equipment technology level, management system level, emergency prevention system as well as environment monitoring and control system, are all qualitative indexes and it is difficult to quantitatively determine their risk levels, this paper adopts a semi-quantitative risk index method [24] to determine the production supervision risk level at wharves. The specific steps are as follows: (1) classify and score the risk levels of each assessment index of wharf production supervision risk, and divide the risk levels into four types: extremely low, low, medium and high, corresponding to the four scoring values of 1, 2, 3 and 4, respectively; (2) determine the wharf risk index that reflects the wharf production supervision risk by the weighted sum, based on the weight and score value of each index; (3) determine the risk level of wharf production supervision according to the wharf risk indexes; the risk levels are divided into extremely low, low, medium and high, corresponding to four numerical ranges of [1,1.5), [1.5,2), [2,3) and [3,4], respectively [25].

The classification of risk levels of each index is determined according to the literature [26]; the risk levels are determined according to the levels of wharf production equipment technology in the industry, including international advanced, domestic advanced, domestic average and domestic backward. The completeness and enforcement of the wharf management system is characterized by the completeness of the safety production management system, the qualification degree of the clean production audit, and the ISO14000 standard certification, so as to determine the risk level of the wharf management system. The risk level of the wharf emergency prevention system is determined by the formulation and implementation of a wharf risk accident emergency plan and the installation and operation of pollution prevention equipment. The risk level of the wharf environmental monitoring system is determined by the installation and operation of the wharf environmental risk monitoring equipment and monitoring degree of completing the pollution index. The risk level classification and scoring standard of the risk assessment index of wharf production supervision are shown in

Table 1. The risk level classification and scoring standard of the risk assessment index of the wharf.

Index	Risk Level Classification Standard and Scoring Value				Index Weight
	High Risk (Score Value: 4)	Medium Risk (Score Value: 3)	Low Risk (Score Value: 2)	Extremely Low Risk (Score Value: 1)
Production equipment and technological level	Domestic backward	Domestic ordinary	Domestic advanced	International advanced	0.287
Management system	Incomplete	Complete but unreasonable	Complete and reasonable	Complete and reasonable with good implementation effect	0.287
Emergency prevention system	Without Emergency plans or anti-pollution facilities	With either emergency plans or anti-pollution facilities	With both emergency plans and anti-pollution facilities, but without regular drill	With both emergency plans and anti-pollution facilities, and with regular drill	0.229

.

3.3. Simulation Technology of Oil Spill Accidents at Wharf Groups

As the wharves are all located along the main stream of the Yangtze River, a mathematical model of the water environment of the Yangtze River can be constructed to simulate the oil spill accidents at the wharves.

3.3.1. Simulated Working Conditions

Considering the low water level of the Yangtze River during the dry season and the relatively high probability of water traffic accidents [26], the hydrological conditions are designed to take a 90% guarantee rate during the dry season. As wharves are distributed both upstream and downstream of the Zhenrunzhou water source and the Jiangsu reach of the Yangtze River is a tidal river, the occurrence time of oil spill accidents is set as the periods of sharp ebb and rise of spring tide, considering the most unfavorable situations. Moreover, according to the meteorological and climatic data of Zhenjiang city [27], this paper selects three wind fields of adverse wind direction (west wind) with the instantaneous maximum wind speed of 12.4 m/s, dominant wind direction (northeast wind) of 6 m/s during the dry season and calm wind to design the simulated working conditions of oil spill accidents, as shown in

Table 2. Simulated working conditions of oil spill accidents.

Number	Flow Field	Wind Field
		Wind Direction	Wind Speed/(ms−1)
1	Sharp rise	W	12.4
2	Sharp ebb	W	12.4
3	Sharp rise	NE	6
4	Sharp ebb	NE	6
5	Sharp rise	calm wind
6	Sharp ebb	calm wind

.

3.3.2. Location of Accidents

The location of oil spill accidents at the wharf group is determined according to the risk index of each wharf and the distance from the wharf to the intake (or the opposite bank of the intake) along the riverbank of Yangtze River. The specific formula is as follows:

$$S_t=\frac{{\displaystyle \sum _{i=1}^m(s_{ti}·k_{ti})}}{{\displaystyle \sum _{i=1}^m(k_{ti})}}$$

(1)

In which, S_t refers to the distance (km) from the location of the oil spill accident of the tth wharf group along the Yangtze River to the intake (or the opposite bank of the intake); S_ti means the distance (km) from the ith wharf at the tth wharf group from the riverbank of the Yangtze River to the intake (or the opposite bank of the intake); k_ti denotes the risk index of the ith wharf at the tth wharf group; and m is the wharf number at the tth wharf group.

3.3.3. Oil Spill Amount

Oil spill accidents are mainly manifested in two forms: oil spill from oil tanks of general cargo ships and oil spill from cargo tanks of tanker ships. The type and tonnage of main ships berthed at the wharves are determined by field research on the type, quantity and tonnage of ships berthed. The capacity of oil tanks or cargo oil tanks of main ships is determined by the interpolation method as well as the reference to the corresponding statistical data of the ship type, tonnage, and tank capacity [28]. The oil spill amount is estimated at 4%/h of tank capacity [14]; the oil spill time is assumed to be 0.5 h. On the other hand, there is still an occurrence probability of wharf oil spill accidents. For wharves with low production supervision risk, the occurrence probability of oil spill accidents is low, and even if they happen, they can be detected and blocked in a timely manner, with low oil spill amount. On the contrary, for wharves with high production supervision risk, they suffer a relatively high oil spill amount. Therefore, the wharf production supervision risk should also be taken into account for calculating the wharf oil spill amount, and the specific calculation formula is as follows:

$$PY=4\%·y·\frac{k-1}{4-1}$$

(2)

where, PY refers to the amount of wharf oil spill (t); y denotes the capacity of oil tank of the main ship berthed at the wharf (t); k means wharf risk index; 4 and 1 are the maximum and minimum values of wharf risk index.

However, the probability of ship collision is higher for the wharf groups with large quantities of wharves and centralized distribution than those of independent wharves. In addition, if the production supervision risk of two wharves in the wharf group is high, the probability of ship collision between the two wharves will also be increased. Hence, the distribution density, distribution quantity and production supervision risk of wharves in the wharf group should be considered comprehensively based on the oil spill amount at the wharves, to determine the oil spill amount of oil spill accidents at the wharf group. The specific calculation formula is as follows:

$$DY=P{Y}_{av}·(1+N·C)$$

(3)

$$P{Y}_{av}=\frac{{\displaystyle \sum _{i=1}^{n}P{Y}_i·k_i}}{{\displaystyle \sum _{i=1}^n{k}_i}}$$

(4)

In which, DY refers to the oil spill amount of the wharf group; PY_av denotes the weighted average value of the oil spill amount of each wharf at the wharf group. n means the number of wharves; PY_i and k_i are the oil spill amount and risk index of the ith wharf respectively. C indicates the risk distribution density index of the wharf group (WRDI for short), which comprehensively reflects the production supervision risk and distribution density of wharves within the wharf group, with the maximum value of 1. N symbolizes the wharf quantity index (WQI for short), reflecting the increment of oil spill risk caused by the increase of wharf quantity, with the maximum value of 1.

Regardless of other factors, the greater the quantity of wharves at the wharf group, the greater the probability of oil spill accidents. When the wharf quantity is small, the probability of oil spill caused by the increase of wharf quantity is greater, but with the increase of wharf quantity, the probability of increment should be smaller and smaller. As the investigation reveals, the average quantity of wharves surrounding the 31 water sources at or above the county level along the Yangtze River in Jiangsu Province is about 12. According to the above description, the WQI calculation principle is determined as: when the surrounding risk sources are fewer than 12, WQI increases linearly with the quantity of wharves, but the increment becomes smaller when the number exceeds 12. The specific calculation formula is as follows:

$$N=\{\begin{array}{l}(n-1)/12,\hspace{1em}\mathrm{if}\hspace{1em}n<12;\\ 1-1/n,\hspace{1em}\mathrm{if}\hspace{1em}n\ge 12\end{array}$$

(5)

where n and N share the same meanings as above.

WRDI can be manifested by the respective risk index of each of the two wharves and the distance between them. The larger the risk index and the smaller the distance, the greater the probability of ship collision between the two wharves, resulting in greater probability of oil spill accidents. As shown above, the maximum value of the risk index at the wharf is 4, and thus the ratio of the wharf risk index to 4 can be utilized to indirectly represent the relative probability of command operation errors at a single wharf. As Kunz C suggests [21], the stroke distance of large ships after berthing is about 550 m, and hence this paper sets: the distance between wharves is less than 2 km, and the wharves are distributed densely. If a command operation error occurs at one of the wharves, collisions may occur between ships berthed at the wharves and wharves adjacent to them, triggering the accident chain between wharves. When the wharf accident chain is within 0.5 km, and the wharves are distributed densely, if a command operation error occurs at one of the wharves, the probability of ship collision accident between wharves is 1. When the distance exceeds 0.5 km, the probability decreases with the increase of the distance. When the distance is more than 2 km, and the wharves are not distributed densely, the probability of ship collision accidents between wharves is 0, without the accident chain between wharves. Based on the above analysis, the calculation formula of WRDI is as follows:

$$C=\frac{{\displaystyle \sum _{i=1}^m{w}_i·\sqrt{(k_{iu}-1)·(k_{id}-1)}}}{(4-1)m}$$

(6)

$$w_i=\{\begin{array}{l}1,\hspace{1em}\mathrm{if}\hspace{1em}{d}_i\le 0.5;\\ (2-d_i)/(2-0.5),\hspace{1em}\mathrm{if}\hspace{1em}0.5<d_i\le 2\end{array}$$

(7)

In which, m denotes the number of existing accident chains; w_i signifies the distribution density of wharves at both ends of the ith accident chain, and the maximum value is 1. d_i means the length of the ith accident chain (km); K_iu and K_id refer to risk indexes of wharves at both ends of the ith accident chain. 4 and 1 refer to the maximum and minimum values of wharf risk index; C shares the same meanings as above.

3.4. Supervision Level Determination of Wharf Groups

Although the three indexes of the arrival time of oil slicks at the intake, the duration of over-standard petroleum concentration, and the maximum over-standard multiple of petroleum concentration at the intake can all reflect the influence degree of oil spill accidents at the wharf group on the water sources, the influence degree manifested by each index is different. Therefore, to comprehensively reflect the influence degree of oil spill accidents on water sources, this paper first determines the degree of being affected at water sources corresponding to each influence index, and subsequently employs the risk index method to comprehensively determine the influence degree of oil spill accidents on water sources at wharf groups, so as to further determine the supervision level of wharf groups.

As the investigation of the water treatment capacity of Zhenjiang Jingxi Water Plant reveals, the removal rate of pollutants is mainly distributed between 60% and 90%, which indicates that when the water quality exceeds the standard about two to five times, the water plant can still treat it to reach the standard. Hence, when the over-standard multiple of petroleum concentration is less than three times, it is considered as low-risk influence. According to the standard for design of outdoor water supply engineering (GB50013-2018) [29], the designed regulation and storage capacity of the water plant is 10–20% of the maximum daily water supply. After conversion, the emergency water supply capacity of the water plant is between 2.4 to 4.8 h, and thus the duration of over-standard petroleum concentration below 2.4 h is regarded as low-risk influence. The emergency capacity of the water source determines the threshold of the arrival time of the oil slicks. Through the field investigation of the above-mentioned water plant, it can be observed that the water plant can feedback the pollution information and take corresponding measures within 1 h after the oil spill accident occurs. Therefore, the arrival time of oil slicks above 1 h is defined as low-risk influence. Relying on the risk influence level classification standard of the risk influence index of oil spill accidents, based on the above low-risk influence threshold, this paper adopts the analytic hierarchy process combined with the expert evaluation method to determine the weight of each index, as shown in

Table 3. Risk influence level classification standard for the influence index of oil spill accidents.

Influence Index of Oil Spill Accidents	Weight	Risk Influence Level Classification Standard
		High	Medium	Low	Extremely Low
Arrival time of oil slicks at the intake	0.325	≤0.5	(0.5,1]	(1,3]	>3
Duration of over-standard petroleum concentration at the intake	0.247	>6	(2.4,6]	(1.2,2.4]	≤1.2
Maximum over-standard multiple of petroleum concentration at the intake	0.428	>6	(3,6]	(1,3]	≤1

.

Through simulating oil spill accidents at the wharf group, the specific values of the above three indexes can be obtained. The risk influence level of each index is determined by referring to Table 3, and the risk influence is scored 1, 2, 3 and 4, respectively, according to the four risk influence levels of extremely low, low, medium, and high. According to the weight and the risk influence score value of each index, the oil spill risk influence index (OSRI, which represents the influence degree of oil spill accidents at the wharf group on the water source) of the wharf group is determined by weighted sum, and the supervision level of the wharf group is further determined. As for the grade classification, some scholars determined the classification standard according to the results, such as the classification of energy systems by Zhu et al. [30], and Kabir et al. [31]. Considering that OSRI ∈ [1,4], which is the same as the wharf risk index, the classification standard of risk supervision level is determined by referring to the classification standard of the wharf risk level, as shown in

Table 4. Classification standard of risk supervision level at the wharf group.

Risk Supervision Level	Key Supervision	Sub-Key Supervision	Non-Key Supervision
OSRI	≥3	[1.5,3)	<1.5

.

3.5. Environmental Risk Assessment of Water Sources

On the one hand, the environmental risk of water sources is influenced by the oil spill accidents surrounding the wharf group. On the other hand, the anti-risk ability of the water sources will also affect the environmental risk level of water sources. Therefore, the environmental risk assessment of water sources should also consider the two factors of environmental risk caused by oil spill accidents at the wharf group and the anti-risk ability of water sources.

3.5.1. Oil Spill Accident Risks of Water Sources

If there is only one wharf group around the water source, the effects of oil spill accidents at the wharf group on the water source will be equivalent to the oil spill accident risks suffered by the water source, and the oil spill accident risk index of the water source (WORI for short, which represents the influence degree of oil spill accidents on the water source) means the OSRI of the wharf group. However, in most cases, there are multiple wharf groups around the water source, and thus the cumulative effects of oil spill accidents of multiple wharf groups on the risk of the water source should be considered while determining the risks of oil spill accidents suffered by the water source. Due to a certain distance between the wharf groups, water traffic accidents will not affect each other. Therefore, it is not appropriate to simulate the risk effects of the extreme case when oil spill accidents occur simultaneously at all wharf groups on the water source.

Referring to the significance of OSRI mentioned above, this paper sums up the OSRI of each wharf group around the water source. Based on the maximum value, it then considers the proportion of OSRI of other wharf groups in the sum to determine the cumulative effects of oil spill accident risks, and finally the WORI is obtained. The specific formula is as follows:

$$B=\underset{1\le t\le q}{\max }(B_t)+[4-\underset{1\le t\le q}{\max }(B_t)]·\gamma$$

(8)

$$\gamma ={\displaystyle \sum _{t=1}^q\left[\frac{B_t}{4}·B_t/{\displaystyle \sum _{t=1}^q{B}_t}\right]}-\frac{\underset{1\le t\le q}{\max }(B_t)}{4}·\underset{1\le t\le q}{\max }(B_t)/{\displaystyle \sum _{t=1}^q{B}_t}$$

(9)

In which, B denotes WORI, where 4 is the maximum value that B can reach; q and B_t refer to the number of wharf groups around the water source and the OSRI of the tth wharf group; and γ means the cumulative coefficient of risk effect.

The risk level of oil spill accidents in the water source can be determined by referring to the risk level standard of wharf production supervision. For details, please refer to Section 3.2.

3.5.2. Anti-Risk Ability of Water Sources

The anti-risk ability of water sources is generally manifested in the natural anti-risk ability of water quality and hydrology, and the risk control abilities, including pre-monitoring ability and post-emergency ability.

Based on the research results of Zhou Qi et al. [32], this paper uses the water qualification rate to characterize the water quality of the water source. The higher the qualification rate, the better the environmental quality of the water source and the stronger the anti-risk ability. The annual average discharge is applied to characterize the hydrological status of the water source. The larger the annual average discharge, the stronger the water diffusion ability and self-purification ability, and the stronger the anti-risk ability of the water source. Moreover, the scale qualification rate of command and monitoring equipment, scale qualification rate of personnel, training rate of personnel, coverage rate of automatic monitoring ability, completion rate of monitoring indexes, as well as daily supervision and case execution rate are used to characterize the supervision ability of the water source. The completion rate of isolation and protection projects at the protection zone, water supply guarantee rate of standby water sources, emergency equipment, material support rate and emergency plan improvement rate are adopted to characterize the emergency response capacity of water sources.

Each index is scored as 1, 2, 3 and 4, respectively, according to the four risk levels of extremely low, low, medium, and high, and weighted summation is adopted to determine the anti-risk ability index (which represents the anti-risk ability level) of the water source. The weight of each index and the classification standard of risk level are based on the relevant research results [25], as shown in

Table 5. Risk level classification standards for the evaluation index of the anti-risk ability of water sources.

Evaluation Index	Index Weight	Risk Level Classification Standards
		Extremely Low	Low	Medium	High
Water qualification rate (%)	0.188	≥95%	[80%,95%)	[60%,80%)	<60%
Annual average discharge (m3/s)	0.188	≥1500	[150,1500)	[15,150)	<15
Scale qualification rate of command and monitoring equipment (%)	0.063	≥90%	[80%,90%)	[60%,80%)	<60%
Scale qualification rate of personnel (%)	0.023	≥90%	[80%,90%)	[60%,80%)	<60%
Training rate of personnel (%)	0.017	≥90%	[80%,90%)	[60%,80%)	<60%
Monitoring section coverage (%)	0.082	≥90%	[80%,90%)	[60%,80%)	<60%
Completion rate of monitoring indexes (%)	0.114	≥90%	[80%,90%)	[60%,80%)	<60%
Daily supervision and case execution rate (%)	0.013	≥95%	[85%,95%)	[60%,85%)	<60%
Completion rate of isolation and protection projects (%)	0.082	≥90%	[80%,90%)	[60%,80%)	<60%
Water supply guarantee rate of standby water sources (%)	0.126	≥80%	[50%,80%)	[20%,50%)	<20%
Emergency equipment, material support rate (%)	0.065	≥90%	[80%,90%)	[60%,80%)	<60%
Emergency plan improvement rate (%)	0.036	≥90%	[80%,90%)	[60%,80%)	<60%

.

Based on the level standard of production supervision risk at wharves, the anti-risk ability standard of water sources can be determined, as shown in

Table 6. Anti-risk ability standard of water sources.

Level	Weak	Medium	Strong	Extremely Strong
Index	≥3	[2,3)	[1.5,2)	[1,1.5)

. According to the anti-risk ability index of water sources, the anti-risk level of water sources is determined as in Table 6.

3.5.3. Assessment of Environmental Risk Level of Water Source

According to the risk index of oil spill accidents at water sources and the anti-risk index of water sources, the weight of 0.65:0.35 is taken to sum up the environmental risk index of water sources (which represents the degree of environmental risk of water sources) to determine the environmental risk level of water sources, as shown in

Table 7. Environmental risk level standard of water sources.

Level	High	Medium	Low	Extremely Low
Index	≥3	[2,3)	[1.5,2)	[1,1.5)

.

4. Construction of Mathematical Models

Based on the hydrodynamic model, this paper adopts the “oil particle” model to simulate oil spill accidents under various working conditions, scattering the spill oil as a large number of oil particles, which represent a certain amount of oil. Firstly, the position of each “oil particle” is calculated. Subsequently, the weathering model is employed to calculate oil slick composition, moisture content, density, viscosity, and other changes. Finally, the spatial and temporal distribution of oil slick concentration can be simulated by counting the quantity and mass of oil particles in a hydrodynamic discrete grid.

4.1. Hydrodynamic Model

According to the characteristics of a tidal channel in the Jiangsu section of the Yangtze River, a two-dimensional tidal current hydrodynamic model is constructed in the Nanjing-Jiangyin section of the Yangtze River on the premise of verifying the hydrodynamic data such as water level and discharge, to simulate the changes of flow field and water level in the study area, which is taken as the driving conditions of the hydrodynamic field in the oil particle model.

The underwater elevation of the model selects the measured underwater topography data of the Yangtze River in 2010 and applies the unstructured grid, with 18,991 grid numbers. The average side length of the grids is 300 m, and the grids are encrypted in local areas. The Yangtze River discharge is large and there is no discharge station between the Nanjing and Jiangyin sections, while the discharge of the Datong station accounts for about 97% of the Yangtze River discharge into the sea, which can represent the flow of this section of the Yangtze River. Therefore, the daily average discharge of Datong hydrological station in December 2013 is taken as the upper boundary of the model, and the water level of Jiangyin hydrological station in December 2013 is taken as the lower boundary of the model. The initial water level of the model is set as 2.5 m (taken from the average water level of the water level yearbook data), the temperature as 20 °C, and the initial flow rate as 0. The measured water levels of Nanjing and Zhenjiang from 21 to 31 December 2013 are selected to calibrate the model, determining the value range of roughness of the main channel of the Yangtze River between Nanjing and Jiangyin as 0.01 to 0.02. The comparison between the calculated results of the model and the measured values of hydrodynamic data of each hydrological station is shown in Figure 3. The maximum absolute error of water level calculated by the model is 19 cm, and the maximum relative error of flow is 4.51%. The Nash efficiency coefficient (NSE) of the measured value and the calculated value of the three groups is 0.922, 0.940 and 0.837, respectively, and as all are close to 1, the model is appropriate to the hydrodynamic simulation of the calculation area.

4.2. Oil Particle Model

After entering the water body, the spilled oil will start the transport process of spreading, drifting and turbulent diffusion under the action of gravity, viscous force, surface tension, hydrodynamic force, and wind stress, accompanied by weathering processes of evaporation, dissolution, and emulsification, which change the oil body components.

4.2.1. Transport Model

(1)

Spreading process

After entering the water body, the spilled oil starts the spreading process under the action of gravity, adhesion, and surface tension, forming a large area of oil slicks. Fay [33] proposed the theory of self-expansion of oil slicks in a still water environment, but the impact of wind stress and other conditions were not considered. Hence, the calculated spreading range of oil slicks was usually small. This paper adopts the modified Fay’s formula by Lehr et al. [34] after considering wind stress to calculate the area of oil slicks:

$$A_o=2270{(\frac{\rho -{\rho }_y}{{\rho }_y})}^{2/3}{V}_0^{2/3}{t}^{1/2}+40{(\frac{\rho -{\rho }_y}{{\rho }_y})}^{1/3}{V}_0^{1/3}{U}_w^{4/3}t$$

(10)

where, A_o means the oil slick area (m²); ρ_y denotes oil density; ρ refers to the water density; t signifies time (s); V_o is oil spill volume (m³); U_w is the wind speed at 10 m above the water surface (m/s).

(2)

Drifting process

The acting force of oil particle drift mainly originates from water surface velocity and wind stress. According to Lagrange, the calculation formula for the total velocity of oil particle drift is as follows [35]:

$$U_t=C_w{U}_w+U_s$$

(11)

In which, U_t is the total drift velocity of oil particles (m/s); U_w denotes the wind speed at 10 m above the water surface (m/s); C_w means wind conductivity coefficient, setting as 0.03–0.04; U_s refers to the surface velocity of water (m/s), which is calculated according to the two-dimensional hydrodynamic model.

(3)

Turbulent diffusion process

The turbulent diffusion process of oil slicks is a random diffusion process. Assuming the horizontal oil diffusion is isotropic, the random walking method is adopted to simulate the random diffusion distance of oil particles in the unit time step [36], and the specific formula is as follows:

$$S_{\alpha }={\left[R\right]}_{-1}^1\sqrt{6D_{\alpha }·\Delta t_p}$$

(12)

where, S_α refers to the random diffusion distance (m); ${\left[R\right]}_{-1}^1$ denotes a random number between 1 and −1; D_α means the diffusion coefficient in α direction (m²/s); and Δt_p signifies the time step (s).

Based on the two-dimensional hydrodynamic model, the horizontal diffusion coefficient is calibrated by utilizing the investigation data of the oil spill accident on the east side of the Yangtze River at the law enforcement base of Taizhou Marine Department from 26 to 27 November 2009. The specific data are as follows: the oil spill volume is 1 ton; the duration is 30 min, the wind field condition is the west wind of 2.92 m/s, and the initial oil concentration is 0 mg/L. The value range of the horizontal diffusion coefficient of the oil spill model is 0.15~0.2 m²/s. The Nash efficiency coefficient (NSE) of the measured value and the model calculated values of oil concentration is 0.921, which is close to 1, and thus the values of the horizontal diffusion coefficient of oil particles can better simulate the transport process of oil slicks after oil spill.

4.2.2. Oil Spill Weathering Model

The weathering processes of oil particles include evaporation, dissolution, and emulsification. The composition of oil particles will change in the weathering process, but the horizontal position will not change.

(1)

Evaporation

As the main component of the oil spill weathering process, evaporation is the most important factor affecting the residual oil spill on the water surface. The evaporation of most crude oil can reach more than 40%, and the diesel oil can reach even more than 80%. Currently, the calculation methods for oil slick evaporation include the analytical method proposed by Stiver and Mackay [37] and the multi-component calculation method proposed by Audunson [38]. The former relies on empirical parameters, while the latter requires detailed oil spill data. Based on the actual situation, this paper selects the analytical method proposed by Stiver and Mackay [37], and the specific formula is as follows:

$$F_v=\ln \left[1+B(\frac{T_G}{T})\theta e^{(A-B\frac{T_0}{T})}\right]\frac{T}{B{T}_G}$$

(13)

$$\theta =K_e{A}_{o}t/V_0$$

(14)

$$K_e=2.5\times {10}^{-3}·U_w^{0.78}$$

(15)

where, F_v denotes the evaporation rate of oil slicks (%); T₀ is the initial boiling temperature of oil (K); T_G refers to the gradient of distillation curve (K), which is the data obtained in the process of crude oil distillation. As empirical parameters, the empirical values of A and B are 6.3 and 10.3 respectively. T means temperature (K); K_e is the material transfer coefficient; V₀ signifies the initial volume of oil spill (m³); A_o denotes oil slick area (m²); t is time (s); and U_w shares the same meanings as above.

T₀ and T_G values are calculated by considering the viscosity, wax content, asphaltene content and other properties of the oil [39].

$$T_0=944.51-445.82/{\rho }_{20}-170/\nu +37/(0.37w+2a)$$

(16)

$$T_G=2103.31-1481.23/{\rho }_{20}+1.34w\nu$$

(17)

In which, a and w are the mass fraction (%) of asphaltene and wax in the oil respectively; ρ₂₀ refers to the density of oil at 20 °C; and ν denotes the dynamic viscosity of oil at 15 °C.

(2)

Emulsification

With the spreading of the oil slicks, oil slicks are torn into oil droplets by the turbulent energy of the water flow, forming oil-in-water emulsification. These emulsions can be stabilized by surfactants to prevent oil droplets from returning to the oil slicks. The leading diffusion force is wave-breaking in bad weather, while the major diffusion stress is oil slick stretching and compression in calm weather. The emulsification degree of the oil spill is manifested by water content. This paper employs the formula proposed by Mackay et al. [40] to reflect emulsification degree with water content:

$$Y_w=\frac{1}{K_B}(1-e^{-K_A{K}_B{(1+U_w)}^2}t)$$

(18)

where, Y_w is water content (%); K_A = 4.5 × 10⁻⁶; K_B = 1.25; t denotes time (s); and U_w shares the same meanings as above.

(3)

Density and viscosity changes

According to the model of Mackay et al., oil density can be determined by assuming a linear relationship between the density change, evaporation rate and temperature change, and the specific formula is as follows [41]:

$$\rho =(1-Y_w)[(0.6{\rho }_0-0.34)F_v+{\rho }_0]+Y_w·{\rho }_w$$

(19)

In which, ρ is the oil density; ρ₀ means the initial oil density; ρ_w denotes the density of water; F_v and Y_w share the same meanings as above.

Due to the small change of environmental temperature in the process of oil spill accident, the temperature factor on viscosity is neglected, but the effects of evaporation and emulsification on oil are considered. The specific formula is as follows [38]:

$$\nu ={\nu }_0·{10}^{4F_v}·\exp [2.5Y_w/(1-0.654Y_w)]$$

(20)

where, ν refers to the kinematic viscosity of oil; ν₀ means the initial kinematic viscosity of oil; F_v and Y_w share the same meanings as above.

4.2.3. Vertical Distribution of Oil Concentration in Water under Oil Slicks

The formula of Mackay is adopted to estimate the vertical change of oil concentration in water under oil slicks, and the specific formula is as follows [40]:

$$c=(V_0\rho \sqrt{\frac{\pi }{4D_Z·t}}/A_0)\exp (-y\sqrt{\frac{\pi }{4D_Z·t}})$$

(21)

In which, c is the oil concentration (mg/L) at the water depth y under the oil slicks; V₀ denotes the volume of oil spill (m³); ρ means oil film density (kg/m³); D_Z refers to the vertical diffusion coefficient of oil particles, with the empirical value as 0.01 m²/s. A_o signifies oil slick area (m²); and t is time (s).

5. Results and Discussion

5.1. Division Results of Wharf Groups

Taking the wharves within 20 km upstream and 20 km downstream of the second-grade protection area of Zhenrunzhou water source as nodes, the Prim algorithm [22] is adopted to construct the minimum spanning tree of wharves for wharf division, as shown in Figure 4. Specifically, the 27 wharves in the surrounding investigation area of the water source can be divided into seven wharf groups from SD01 to SD07, of which wharf group SD03 contains a maximum of nine wharves. The corresponding area of SD03 is the Dantu Economic Development Zone, about 11 km upstream of the water source. With numerous enterprises in the area, their attached wharves are densely distributed along the river.

5.2. Location and Amount of Oil Spill at Wharf Groups

The risk index method [24] is employed to evaluate the production supervision risk of each wharf, and the results are shown in

Table 8. Risk index of wharves and calculation results of oil spill amount.

Wharf Group	Wharf	Location	Risk Index	Risk Level	Oil Spill Amount (t/h)
SD01	JE01	119.378° E, 32.213° N	2.949	medium	10.778
	JE02	119.374° E, 32.210° N	2.509	medium	3.706
	JE03	119.375° E, 32.212° N	2.509	medium	2.489
SD02	JE04	119.350° E, 32.194° N	2.831	medium	25.604
	JE05	119.334° E, 32.189° N	2.745	medium	2.591
SD03	JE06	119.311° E, 32.187° N	2.541	medium	2.074
	JE07	119.308° E, 32.187° N	2.831	medium	34.374
	JE08	119.282° E, 32.188° N	2.949	medium	16.592
	JE09	119.274° E, 32.191° N	2.831	medium	24.663
	JE10	119.260° E, 32.197° N	2.831	medium	24.757
	JE11	119.257° E, 32.199° N	2.702	medium	1.244
	JE12	119.255° E, 32.199° N	2.702	medium	1.659
	JE13	119.291° E, 32.187° N	3.057	high	16.592
	JE14	119.288° E, 32.187° N	3.057	high	10.778
SD04	JE15	119.236° E, 32.211° N	2.702	medium	1.244
	JE16	119.228° E, 32.211° N	2.702	medium	1.244
SD05	JE17	119.325° E, 32.240° N	2.59	medium	1.244
	JE18	119.328° E, 32.239° N	2.751	medium	0.207
	JE19	119.291° E, 32.241° N	2.509	medium	6.188
	JE20	119.296° E, 32.241° N	2.595	medium	12.965
	JE21	119.306° E, 32.241° N	2.595	medium	27.954
SD06	JE22	119.238° E, 32.235° N	2.595	medium	18.451
	JE23	119.242° E, 32.235° N	2.429	medium	2.074
	JE24	119.242° E, 32.236° N	2.702	medium	1.244
	JE25	119.230° E, 32.234° N	2.595	medium	93.636
SD07	JE26	119.409° E, 32.244° N	2.863	medium	0.415
	JE27	119.406° E, 32.235° N	2.541	medium	2.074

. Among the 27 wharves, 25 belong to the high-risk level, while JE13 and JE14 belong to the high-risk level, with the risk index of 3.057, distributed in the wharf group of SD03. The two high-risk wharves are the respective logistics wharves of two petrochemical enterprises, respectively, with the berthing capacity of above 40,000 t. Their loading and unloading of goods are mainly dangerous goods like benzene, xylene, phenol, cyclohexanone, and oil. However, the anti-pollution facilities of the two wharves are relatively simple, with only concise oil absorption felt, and the two wharves have not installed automatic water quality detectors, and thus their sensitivity to sudden environmental accidents is poor, resulting in the highest risk index. Other wharves are mostly bulk cargo, general cargo and outfitting wharves, whose loading and unloading goods do little harm to the water environment. Simple anti-pollution facilities and automatic water quality detection equipment make their risks relatively low.

The type and tonnage of main ships berthed at each wharf are obtained through field research. Referring to the statistical data of tonnage and oil tank capacity of different ship types, the oil tank capacity is obtained by the interpolation method, and the oil spill amount of oil spill accidents at wharves is determined by 4%/h of the oil tank capacity, as shown in Table 8. To be specific, there are huge differences in the oil spill amount of 27 wharves. Wharf JE25 is the outfitting wharf of Jinling Shipyard, with the 100,000-t tonnage of its main ship type and its maximum oil spill amount of 93.636 t/h, which is nearly three times that of wharf JE07 (the secondary large one). As the main ship type of wharf JE18, a bulk cargo wharf, has only 400 t tonnage, its possible oil spill amount is the least, at only 0.207 t/h.

WQI, WRDI and oil spill position of the wharf group are calculated according to the number, location, and risk index of the wharves in the wharf group, and subsequently the oil spill amount of each wharf is considered, and the oil spill amount of each wharf group has been calculated, as shown in

Table 9. Calculation results of oil spill amount at wharf group.

Wharf Group	WRDI	WQI	Weighted Average Value of Wharf Oil Spill Amount (t/h)	Oil Spill Amount of Wharf Groups (t/h)	Distance to the Intake (m)
SD01	0.549	0.167	5.94	6.484	2449
SD02	0.422	0.083	14.275	14.777	6382
SD03	0.375	0.667	15.016	18.773	12,272
SD04	0.498	0.083	1.244	1.296	16,925
SD05	0.347	0.333	9.625	10.738	8278
SD06	0.443	0.25	28.996	32.207	14,010
SD07	0.393	0.083	1.195	1.234	1948

. It can be observed from Table 9 that WRDI varies widely among various wharf groups, ranging from 0.3 to 0.6. The main reason is that the length of wharf accident chains within each wharf group varies greatly. For instance, Wharf SD01 is densely distributed, and the length of three accident chains is less than 500 m, with an average length of only 307 m. Consequently, its WRDI, 0.549, is the largest. Relatively speaking, although the average risk index of SD03 is the largest (2.833), due to its relatively sparse distribution, its 1041 m average length of the accident chain, and less than 500 m length for only four of the 15 accident chains, and thus its WRDI is relatively small at 0.375. Owing to the largest number of wharves at SD03, its WQI is two to eight times that of the other wharf group. Therefore, although its WRDI is relatively small, its risk increment is the largest, which is reflected in the oil spill amount, increasing from 15.016 t/h of the weighted average value of wharf oil spill amount to 18.773 t/h of the oil spill amount at the wharf group. However, for SD06, with a small number of wharves, although the oil spill amount of JE23 and JE24 is very small, the oil spill amount of JE25 reaches 93.636 t/h, resulting in the largest oil spill amount of SD06 wharf group, reaching 32.207 t/h. Comparing the oil spill volume of each wharf group, it is observed that compared with the wharf number and distribution location, the wharf oil spill volume has the greatest influence on the oil spill volume of the wharf group, and the determination of the oil spill volume of the wharf group is related to the main ship tonnage and wharf production supervision risk. In the long run, the main ship tonnage seldom changes, but production regulatory risk will fluctuate. Consequently, the supervision risk of wharf production is the most important in determining the oil spill volume of the wharf. To reduce oil spill accidents and the oil spill volume, the risk of wharf production supervision should be kept at a low risk level in the long term.

5.3. Simulation Results of Oil Spill Accidents at Wharf Groups

Based on the two-dimensional hydrodynamic model, the oil particle model is adopted to simulate the drift and diffusion process of oil slicks after oil spill accidents occur in different working conditions of each wharf group (Table 2). After the oil spill accident, the oil slick drifts and spreads along the coastline of the Yangtze River. Affected by the ebb and flow tide, the oil slick oscillates back and forth and drifts downstream. The drift trend of the oil slick is roughly the same under different flow and wind fields. The simulated changes of petroleum concentration at the intake of the water source are shown in Figure 5, where the time refers to the experience time after the oil spill accident, FRT represents sharp rise, and FET denotes sharp ebb.

As Figure 5 reveals, the water quality standard (Class II) of water source intakes stipulated in the Environmental quality standards for surface water (GB3838-2002) [42] is used as the threshold value to judge the calculation results, and the five wharf groups SD01~SD04 and SD07, which are on the same bank as the water source, present influences on the water source intakes after oil spill accidents, respectively. However, for SD05 and SD06, located in the north bank of the north branch of the Yangtze River, where the south branch flow is approximately twice the north branch flow, although the width of the river surface at the water intake is relatively small, the huge flow of the south branch makes it hard for oil slicks formed after oil spill accidents to affect intake. Moreover, after the oil spill accident occurs at SD06, due to its large oil spill amount, the petroleum concentration at the intake fluctuates slightly, but not exceeding the water quality standard of Class II. Located downstream of the water source, SD07 only affects the water source during rising tide.

According to the simulation results of the model, the specific values of the influence index on the water source after oil spill accidents in various wharf groups under different working conditions can be obtained, as shown in

Table 10. Simulation calculation results of influence index of oil spill accidents.

Influence Index of Oil Spill Accidents	Wharf Group	Working Conditions
		Sharp Rise			Sharp Ebb
		W	NE	Calm Wind	W	NE	Calm Wind
maximum over-standard multiple of petroleum concentration at the intake	SD01	16.61	17.07	16.98	58.03	58.30	58.31
	SD02	38.63	46.94	46.81	77.97	78.17	75.66
	SD03	42.81	52.69	50.51	65.50	75.86	66.94
	SD04	2.07	2.28	2.20	3.00	3.39	3.15
	SD05	0	0	0	0	0	0
	SD06	0	0.14	0.14	0	0.14	0.29
	SD07	37.64	36.84	36.95	0	0	0
arrival time of oil slicks at the intake	SD01	7.16	7.22	7.22	1.44	1.44	1.44
	SD02	7.72	7.83	7.83	2.72	2.83	2.83
	SD03	8.89	9	9	4.44	4.56	4.56
	SD04	11.17	11.39	11.39	6.94	7.06	7.06
	SD05	-	-	-	-	-	-
	SD06	-	15.34	15.28	-	15.5	10.72
	SD07	1.67	1.67	1.67	-	-	-
duration of over-standard petroleum concentration at the intake	SD01	0.67	0.67	0.67	0.78	0.67	0.78
	SD02	0.89	0.67	0.67	1	0.78	0.78
	SD03	1	0.89	1	0.78	2.67	1.33
	SD04	1.11	0.89	0.89	2.67	2.33	2.11
	SD05	0	0	0	0	0	0
	SD06	0	0	0	0	0	0
	SD07	1.33	1.21	1.22	0	0	0

.

By comparing the calculation results of the three wind field conditions in Table 10, it can be observed that either during sharp rise or sharp ebb, the maximum over-standard multiple of petroleum concentration at the intake after an oil spill accident is the largest in the northeast wind condition, while it is the smallest in the west wind condition, which is mainly attributed to the influence of wind field on flow field and oil slick drift and diffusion velocity. Compared with the other two wind fields, the west wind direction is the same as ebb tide direction and opposite to rising tide direction, and thus the velocity is relatively large during sharp ebb and relatively small during sharp rise in the west wind condition. Simultaneously, wind force also accelerates the drift and diffusion of oil slicks. Relevant studies prove that the motion speed of oil slicks is about 3% of the wind speed above 10 m on the water surface [43]. Hence, compared with the northeast wind and calm wind, the oil slick drifts and spreads faster in the west wind. With a larger area and relatively sparse distribution of oil particles, it has a smaller maximum petroleum concentration at the intake, which also leads to an earlier arrival time of oil slicks at the intake and a longer duration of over-standard petroleum concentration at the intake. On the contrary, the direction of the northeast wind is opposite to the direction of the ebb tide but the same as the rising tide. Therefore, after the oil spill accident occurs in the northeast wind, the maximum petroleum concentration at the intake is larger, the arrival time of the oil slicks at the intake is later, and the duration of over-standard petroleum concentration at the intake is shorter.

However, there are exceptions for the above situations. For example, SD03 is far away from the intake, but when the oil spill accident occurs during sharp ebb, the duration of over-standard petroleum concentration at the intake under the northeast wind condition is apparently longer than the west wind condition, and under the calm wind the condition is longer than the west wind condition. The main reason is that when the oil slicks drift to the intake, the intake and the downstream flow field are in rising tide. Consequently, under the calm wind condition, the oil slick oscillates back and forth here. Under the northeast wind condition, the oil slick oscillates back and forth here for a longer time. However, under the west wind condition, due to the faster drift speed of oil slick downstream, the oil slick has drifted downstream of the intake. Although affected by the rising tide, some oil slicks drift upstream, which does not affect the intake. SD04 is farther away from the intake. When the oil spill accident occurs during sharp ebb fall, the oil slicks drift to the intake. The intake and the downstream flow field are in rising tide. Therefore, no matter the kind of wind field condition, the oil slick will oscillate here. In the west wind condition, the oil slick has a larger area and longer length, thus longer influence on the water source. By contrast, SD01 and SD02 are close to the water source, and when oil spill accidents occur during sharp ebb, the oil slick drifts away a long distance from the intake with ebb tide and will not reach the intake when it drifts upstream with rising tide. Hence, the location of oil spill and the flow field are also important factors influencing the differences of oil spill accidents on the water source, of which SD01~SD04 upstream of the same bank of the water source are typical.

According to Figure 4 and Table 10, compared with the sharp rise, when SD01~SD04 upstream of the water source suffer oil spill accidents during sharp ebb, the petroleum concentration at the intake further exceeds the standard, the location of oil spill is closer to the intake, and their difference is more obvious. For instance, under the west wind condition, SD01~SD04 suffer oil spill accidents during sharp ebb and rise, respectively. For their maximum over-standard multiple of petroleum concentration at the water intake, the former is 3.49, 2.02, 1.53 and 1.45 times the latter. The farther away from the water source, the smaller the multiple will be. The main reason for the trend is that the closer the location of oil spill to the water source, the shorter the time of oil slick diffusion and weathering. When reaching the intake, the oil slick area is relatively small, the oil particles are relatively concentrated, and thus the oil slick concentration is relatively high. However, when the oil spill accident occurs during rising tide, the oil slick drifts upstream and then downstream, with enough diffusion and weathering time. When the oil slick reaches the intake, it has spread to a large area, with small thickness and low concentration of oil slick, leading to the shorter duration of over-standard petroleum concentration at the intake for oil spill accidents occurring during the sharp rise compared with that during the sharp ebb.

As shown in Figure 6, the closer the oil spill location to the intake, the shorter the drift distance of the oil slick, and thus the arrival time of oil slicks at the intake will be less. In addition, when the oil spill accident occurs during sharp rise, the oil slick drifts upstream first, making the arrival time of oil slicks at the intake longer than that during sharp ebb. By comparing the time difference between the two, it can be concluded that the closer the oil spill location to the water intake, the greater the time difference will be. This is because the upstream Yangtze River suffers fewer jacking effects of sea tide, which is characterized by the shorter time of rise in a tidal cycle [44]. Therefore, when SD01, closest to the intake, suffers an oil spill accident during sharp rise, its oil slick drifts upstream for the longest time, resulting in the largest time difference in the arrival time of oil slicks at the intake than that during the sharp ebb.

The oil spill amount also affects the maximum over-standard multiple of petroleum concentration at the intake. As Figure 7 reveals, for the oil spill accidents occurring during the sharp rise under the three wind field conditions, the maximum over-standard multiple of petroleum concentration at the intake is basically in a linear relationship with the oil spill amount, but it is not the case for the oil spill accidents occurring during the sharp ebb. The main reason is that SD01 and SD02 are close to the intake, and the oil slick diffusion and weathering time is relatively short.

Located at the downstream of the water source, SD07 only affects the intake when its oil spill accidents occur during sharp rise. Since the rising tide direction is in the same direction as the northeast wind, but opposite to the west wind, under the northeast wind condition, the oil slick area is larger, the petroleum concentration in the middle of the oil slick is lower, and the duration of over-standard petroleum concentration is longer after the oil spill accidents occur during sharp rise, compared with the west wind and calm wind conditions. In the west wind condition, it is opposite. Close to the intake, the arrival time of oil slicks at the water intake has little difference in the three wind field conditions for SD07. In addition, similar to the oil spill accidents occurring in SD03 and SD04, when oil spill accidents occur at SD07 during the sharp rise, the oil slick drifts upstream to the intake with the tide, and then passes through the intake with the ebb tide. Therefore, the duration of over-standard petroleum concentration at the intake is long.

5.4. Judgment Results of Supervision Level at Wharf Groups

According to the simulation results of the influence indexes of oil spill accidents shown in Table 10, the risk influence level of each influence index and the risk influence score are determined based on Table 3. According to the weight of the indexes shown in Table 3, the OSRI of each wharf group under different working conditions is obtained by weighted sum. Subsequently, the supervision level of the wharf group is determined by referring to Table 4. Please refer to Figure 8 for detailed results.

As Figure 8 reveals, the OSRI of each wharf group under different working conditions are all less than 3, some of which are even less than 1.5, indicating that the supervision level is sub-key or non-key. The OSRI of SD07 during sharp rise is the largest, at 2.856, approaching the key supervision standard level. Therefore, the management department of the water source should appropriately strengthen the corresponding regulation. However, the OSRI of SD07 during sharp ebb is the smallest, at 1. The main reason is that located downstream of the water source, SD07 is close to the intake. When oil spill accidents occur during sharp rise, petroleum concentration of oil slick is high, which can reach the intake quickly, and the duration of over-standard petroleum concentration at the intake is long. The score values of the three influence indexes of oil spill accidents are 4, 2 and 2, respectively. However, when the oil spill accidents occur during sharp ebb, the oil slick will not reach the water source, and the score values of the three influence indexes of oil spill accidents are all 1. Located on the opposite side of the water source, SD05 and SD06 basically exert no influence on the intake. Therefore, the risk index of oil spill accidents in six working conditions is 1.

The influence index of oil spill accidents at SD01~SD03 of the same upstream bank of the water source among the four wharf groups, which is the maximum over-standard multiple of petroleum concentration at the intake, has a high-risk influence with a score of 4 under all working conditions, either due to close location to the intake or the large oil spill amount. However, SD04 is farthest away from the water source with the smallest oil spill amount, and thus its influence index of oil spill accidents—the risk influence level of the maximum over-standard multiple of petroleum concentration at the intake is high-risk influence during sharp ebb working conditions and low risk influence during sharp rise working condition, with the score values of 3 and 2, respectively. The influence index of oil spill accidents from SD01 to SD04—the risk influence level for the arrival time of oil slicks at the intake varies little, which is low or very low risk influence, with the risk value of 2 or 1. As SD01 and SD02 are close to the intake, the risk influence level for the arrival time of oil slicks at the intake under sharp ebb working conditions is the low-risk influence. Being far away from the intake, when SD03 and SD04 suffer oil spill accidents during sharp ebb, the oil slick will oscillate back and forth at the intake. Consequently, the risk influence level for the duration of over-standard petroleum concentration at the intake under sharp ebb working condition is bigger, belonging to low-risk influence or high-risk influence, with the risk value of 2 or 3. However, when SD01 and SD02 suffer oil spill accidents during sharp ebb, the oil slick will quickly pass through the intake without oscillating back and forth. Therefore, the risk influence level for the duration of over-standard petroleum concentration at the intake under sharp ebb working condition is extremely low risk influence. When the oil spill accidents occur at SD01~SD04 during sharp rise, the duration of over-standard petroleum concentration at the intake is also very short, and the risk influence level is also extremely low, with the score value of 1. Comprehensively considering the score values of the three indexes of oil spill accidents, we can determine that the OSRI of SD01~SD04 under each working condition is 2~3. Only the OSRI of SD04 during the sharp rise working condition is 1.428, and the supervision level is non-key supervision, because the arrival time of oil slicks at the water intake and the duration of over-standard petroleum concentration at the intake are extremely low risk influences.

5.5. Results of Environmental Risk Assessment of Water Sources

According to OSRI calculation results of each wharf group under different working conditions, this paper employs Formulas (9) and (10) to calculate WORI and determine the risk level, as shown in

Table 11. Environmental risk assessment results of the water source.

Working Conditions	Oil Spill Accident Risk		Anti-Risk Ability		Environmental Risk of the Water Source
	Index	Level	Index	Level	Index	Level
FRT, W	3.285	high	1	Extremely strong	2.485	medium
FRT, NE	3.285	high	1	Extremely strong	2.485	medium
FRT, Calm	3.285	high	1	Extremely strong	2.485	medium
FET, W	3.165	high	1	Extremely strong	2.407	medium
FET, NE	3.269	high	1	Extremely strong	2.475	medium
FET, Calm	3.157	high	1	Extremely strong	2.402	medium

.

As the 2010~2020 Environmental Bulletin of Zhenjiang City reveals, the water quality of the Zhenrunzhou water source of the Yangtze River satisfies the water standard of class III, with 100% of the water qualification rate. Given that there are no large tributaries in the Jiangsu section of the Yangtze River, the annual average flow of the water source can be taken as 28,600 m³/s by referring to the annual average flow of Datong Hydrological Station in the upstream of the Jiangsu section of the Yangtze River [20]. According to the field investigation results, the daily supervision capacity and emergency response capacity of Zhenrunzhou water source have reached 100%. Referring to the risk level classification standard and index weight (Table 5) of the anti-risk ability assessment index of the water source, the value of each index can be obtained, and the anti-risk ability index of the water source can be obtained by weighted sum, as shown in Table 11.

Based on the risk index of oil spill accidents and the anti-risk ability index of the water source, the environmental risk index of the water source is determined by the weighted sum of 0.65:0.35, and the environmental risk level of the water source is determined, as shown in Table 11.

As Table 11 reveals, owing to the excellent water quality and hydrology conditions of the water source, its supervision and emergency indexes reach 100%, and the anti-risk index of the water source under various working conditions is 1, with extremely strong anti-risk ability. Under each working condition, the risk index of oil spill accidents at the water source ranges from 3.1 to 3.3, which are all high-risk levels. However, due to the strong anti-risk ability of the water source, the environmental risk index of the water source under each working condition is less than 3, within 2.4~2.5, and the risk level is medium. Since the influence index of oil spill accidents at SD07 is the largest, which is 2.856, the risk index of oil spill accidents during the sharp rise and the environmental risk index of the water source at SD07 are slightly larger than those during the sharp ebb.

6. Conclusions and Recommendations

6.1. Conclusions

Aiming at the existing research defects of strong subjectivity in determining oil spill amount in simulated oil spill accidents, insufficient consideration of risk increment caused by dense wharf distribution, and incomplete indexes for reflecting the influence degree of oil spill accidents on the water sources, this paper constructs a risk assessment system of the water sources under the influence of the wharf group and applies it to the environmental risk assessment of the Zhenrunzhou water source in Zhenjiang.

The method system includes: (1) wharf group division method determined by existing research results after considering the distribution of wharves surrounding the water source; (2) the calculation method of oil spill amount at wharves, determined by considering the oil tank capacity of main ship types at the wharves and wharf risk index, reflecting production supervision risk at wharves; (3) the calculation method of the oil spill amount at the wharf group considering the wharf number, distribution density, risk index and oil spill amount of each wharf; (4) the calculation method of the influence degree of oil spill accidents at the wharf group on the water sources, which are reflected by the maximum over-standard multiple of petroleum concentration, the arrival time of oil slicks at the intake, and the duration of over-standard petroleum concentration at the intake, further determining the judgment method of the supervision level at the wharf group; (5) the calculation method of the risk index of oil spill accidents at water sources, considering the cumulative effect of oil spill accidents at different wharf groups on the risk of the water sources, based on the risk influence index of oil spill accidents at the wharf group; (6) the environmental risk assessment method of water sources, comprehensively considering oil spill accident risk and anti-risk ability of water sources.

Applying this method system to the environmental risk assessment of the Zhengrunzhou water source in Zhenjiang City, we discovered that: (1) the flow field, wind field, oil spill location, and oil spill amount affect the influence degree of oil spill accidents on water sources, and the flow field has obvious influence on the arrival time of oil slicks at the intake, the maximum over-standard multiple of petroleum concentration at the intake and the duration of over-standard petroleum concentration at the intake. The oil spill location also exerts obvious influence on the arrival time of oil slicks at the intake, and the oil spill amount shows obvious influence on the maximum over-standard multiple of petroleum concentration at the intake. Relatively speaking, although the wind field affects the above three indexes, its influence degree is small. Only when in a specific oil spill location (SD03, SD04), and a specific flow field (sharp ebb), different wind fields will have a significant influence on the duration of over-standard petroleum concentration at the intake. (2) When oil spill accidents occur at the same bank upstream of the water source during sharp ebb, the wharf groups SD01~SD04 present similar influences on the water source, and their supervision level mostly belongs to the sub-key supervision level. Only SD04 during the sharp rise is of a non-key supervision level. SD07 at the same bank downstream of the water source is close to the key supervision level during the sharp rise. However, SD05 and SD06 at the opposite bank of the water source have almost no influences on the water source, belonging to the non-key supervision level. (3) There is little difference in the environmental risks of the water source under different flow fields and wind fields, all of which belong to the high-risk level. Relatively speaking, the environmental risk index of the water source is slightly higher during the sharp rise, mainly because SD07 at the same bank downstream of the water source is very close to the water source, which exerts a great influence on the water source after an oil spill accident occurs during the sharp rise.

Compared with the research results of Huang Yikang et al. [16], Tian Wei et al. [17], and Song Zekun et al. [18], the oil spill volume of this paper, determined according to real situations such as the production supervision risk at the wharves, the situation of ships berthing at the wharves, the wharf number and the distribution of the wharves, is more objective. Moreover, the indexes to reflect the influence degree of oil spill accidents on the water source take the changes of oil concentration at water intake into account. Therefore, compared with the previous risk evaluation method system, the one proposed in this paper is adopted to evaluate the risk of the water source, which is more accurate.

6.2. Recommendations

(1) The supervision department of the water source should focally strengthen the risk supervision over the high-risk wharves listed in Table 8 and urge them to equip themselves with pollution-prevention devices such as oil booms and oil absorption machines, and also install automatic water control monitors to enhance the sensitivity of high-risk wharves to the oil spill accidents. In the meantime, considering the difficulty of subsequent processing for oil spill accidents and the higher water-pollution degree than other types of pollution accidents, the supervision department of the water source should also improve the inspection frequency of high-risk wharves, urge the responsible units for the wharves to promote the production level of wharves and strengthen the daily safety management of wharves, so as to eliminate potential safety accidents such as oil overturn and prevent oil spill accidents caused by safety accidents.

(2) The supervision department of the water source should reinforce the daily inspection of the wharf group at the same bank of the water source, namely the sub-key supervision wharf groups shown in Figure 8, especially for SD07 downstream of the water source during the rising tide. Although it belongs to the sub-key supervision level, the actual supervision intensity is close to the standards of the key supervision level.

(3) Considering that the wharves are concentrated within the wharf group, and the ships are densely distributed, water traffic accidents such as ship collisions are prone to occur, increasing the occurrence risk of oil spill accidents. Therefore, the supervision department of the water source should cooperate with the maritime, transportation and other departments to strengthen the inspection of water traffic command operation situations such as shipping encounters, arrivals and berthing in the same wharf group, at the same bank as the water source, so as to eliminate potential ship collision accidents and prevent the occurrence of oil spill accidents caused by water traffic accidents.

(4) The supervision department of the water source should cooperate with the environment, traffic, customs, maritime and other departments to construct an information-sharing mechanism among multiple departments, so as to share real-time information such as ship tonnage, transported product variety, transportation volume and destination wharves within the Zhenyang section and even Jiangsu section of the Yangtze River, as well as conduct real-time key monitoring of the whole process for ships with potentially strong harmfulness and wide influence scope, and eliminate the hidden danger of oil spill accidents in advance.