Quantifying land-based pollutant loads in coastal area with sparse data: Methodology and application in China

Ocean & Coastal Management 81 (2013) 14e28

Contents lists available at SciVerse ScienceDirect

Ocean & Coastal Management journal homepage: www.elsevier.com/locate/ocecoaman

Quantifying land-based pollutant loads in coastal area with sparse data: Methodology and application in China Jinliang Huang a, b, *, Qingsheng Li b, Zhenshun Tu c, Canmin Pan b, Luoping Zhang a, b, Pancras Ndokoye b, Jie Lin b, Huasheng Hong a, b a b c

Coastal and Ocean Management Institute, Xiamen University, Xiamen 361005, China Environmental Science Research Center, Xiamen University, Xiamen 361005, China Fujian Institute of Oceanography, Xiamen 361012, China

a r t i c l e i n f o

a b s t r a c t

Article history: Available online 22 July 2012

A systematic approach was developed to quantify land-based pollutant loads in selected bays in China where there is sparse or limited relevant data. The approach was applied to Luoyuan Bay and Xiamen Bay. Despite the data limitations in the two study areas, the approach was able to show that in Luoyuan Bay, the chemical oxygen demand (CODMn) load was mainly from soil losses, which accounted for 63 percent of the total pollutant load, whereas point sources only contributed 4 percent. Soil losses constituted the main pollution source for total nitrogen (TN) and total phosphorus (TP). Similarly, spatial variability of the source portion of land-based pollution was detected in Xiamen Bay. Non-point sources were the main source of CODMn load in all Xiamen Bay sub-sea areas, accounting for over 60 percent of the total pollutant loads. Non-point sources contributed largely to the TN and TP loads for most Xiamen Bay sub-seas. However, river discharges and point sources of pollution were also responsible for considerable TN and TP loads in some sub-seas. The application of the approach at the two bays resulted in a clear identification of the source apportionment and spatial distribution of land-based pollutant loads. Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction Over 60 percent of the world’s population lives within the coastal zone, and pollution from their activities has been causing stress to the coastal and marine environment (Bartlett and Smith, 2004; EPA, 1994). The large coastal population has exerted increasing pressure on the coastal and marine ecosystems through competition for land use and natural resources, as well as through change of consumption and use patterns as a result of rapid urbanization. This has resulted in non-point source (NPS) pollution problems, which seriously affect water quality in coastal areas. Nutrients cause eutrophication in coastal waters and estuaries, and very often stimulate the occurrence of red tides and harmful algal blooms (HABs) (Li and Dag, 2004). It is reported that enclosed and semi-closed seas in many parts of the world are suffering from habitat degradation due to increased discharges and emissions of * Corresponding author. Environmental Science Research Center, Xiamen University, Room 210, Yingxue Building, No. 422, South-siming Road, Xiamen 361005, China. Tel.: þ86 592 2182175. E-mail address: [email protected] (J. Huang). 0964-5691/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ocecoaman.2012.07.011

polluting substances, such as sewage, nutrients, organic matter, and toxic persistent substances entering the coastal waters (Melvasalo, 2000; Syvitski et al., 2005). Since 1990, land-based sources of pollution have been recognized as one of the world’s most serious marine pollution problems contributing to more than 75 percent of the pollutants entering the sea (GESAMP, 1990). Understanding and quantifying the spatial distribution of human impacts are needed in order to evaluate tradeoffs (or compatibility) between human uses of the seas and oceans and protection of coastal and marine ecosystems and the services they provide (Halpern et al., 2008). Quantification of the level of impacts is an extremely important step toward improved ocean management, since it helps transform questions of ‘yes or no’ into questions of ‘more or less’ (Peng et al., 2006). Since the late 1970s, numerous attempts have been made to investigate and quantify land-based pollution loads in coastal seas or bays around the world (Table 1). It is apparent that the major seas and bays have suffered serious ecological damage as a result of pollution principally from landbased sources (CNSPCP, 1990; SFBCDC, 2003; Ernest, 2003; Laurence, 1992; Helmer, 1977; Larsson et al., 1985; Tuncer et al., 1998; Claussen et al., 2009; Pawlak, 1980; Williams, 1996;

J. Huang et al. / Ocean & Coastal Management 81 (2013) 14e28

15

Table 1 Research efforts on land-based pollution in the major seas or bays around the world. Regions/countries

Seas/bays

Impact of land-based pollution on marine water quality

Reference/s

European

Mediterranean

A wide range of pollution sources was identified and a full account of the total pollution load was established from four sources: (a) domestic sewage; (b) industrial wastes; (c) agricultural runoff; (d) river discharges; and (e) radioactive discharges. It is estimated that millions of tons of pollutants enter the Baltic Sea from land-based every year. Over 50% of it comes from rivers. Black Sea is facing serious pollution problem and has already suffered catastrophic ecological damage primarilyfrom land-based sources.

Helmer, 1977

NPS pollution is considered to be one of the top threats to the ecological health of San Francisco Bay, which accounted for a considerable proportion of the Bay’s total pollutant load. The diffuse total phosphorus (TP) accounted for more than 60% of the total pollutant entering the Bay.

SFBCDC, 2003

In the Caribbean region, the United Nations Environment Programme (UNEP) has identified the control of domestic, industrial and agricultural land-based sources of pollution as among the most important objectives. Marine pollution from land-based activities is also threatening the coastal and marine environment in Australia which now endorses and implements UNEP Global Programme of Action for the Protection of the Marine Environment from Land-based Activities. Resulting from land-based activities, the coastal zone areas (wetlands, mangroves and lagoons) are the most affected by degradation. East Asian Seas should focus particularly on pollution arising from both land and sea-based sources. With respect to the land-based water pollutants flowing into Tokyo Bay, a large proportion comes from domestic wastewater.

Schumacher et al., 1996

The main land-based pollutants in East China Sea are inorganic N, P, oil hydrocarbons, organic matter and heavy metals. Land-based pollution accounts for more than 60% of total pollution in the Bohai Sea. In the recent years, “Clean Bohai Sea Program” and “Sustainable Development Strategy and Implementation Plan of Bohai Sea” have been embarked so as to improve the marine water environmental of Bohai Sea

CNSPCP, 1990

Baltic Sea Black Sea North America

San Francisco Bay

Chesapeake Bay Central America

Caribbean Sea

Australia

e

West and Central Africa Asian

e

Japan

Tokyo Bay

China

East China Sea

East Asian Seas

Bohai Sea

Schumacher et al., 1996; Kouassi and Biney, 1999; CECWED, 1999; JME, 2002; Chua, 1999; SEPA, 2001; Peng et al., 2009; Zhang, 2009). However, quantifying land-based pollution loads is extremely challenging because land-based pollution sources vary and are difficult to identify. Over the last two decades, integrating Geographic Information System (GIS) with models has proven to be a viable tool to quantify NPS loads and identify critical source areas (Pullar and Spinger, 2000). Numerous studies have been conducted in watersheds and bay regions, including Chesapeake Bay (Donigian et al., 1990; Sprague et al., 2000), Santa Monica Bay (Wong et al., 1997), and the Kao-Ping River Basin in Taiwan (Ning and Chang, 2007). Human impacts on the ecosystems vary with the type and level of activities on land and at sea. Appropriate spatial mapping of pollution loads across coastal areas and bays will help improve and rationalize the management of human activities (Halpern et al., 2008). GIS has proven to be a useful tool to show the spatial variability, quantification and source apportionment of land-based pollution loads using spatial units such as sub-seas and sub-basins of marine areas and watersheds (Helmer, 1977; Chua, 1999; Yoshiaki, 2006). Models are another important tool. Taking Chesapeake Bay as an example, two types of models were used to quantify land-based pollution (Sprague et al., 2000), namely the statistical model SPAtially-Referenced Regressions On Watershed attributes (SPARROW) and the physically based model Hydrologic Simulation Program-Fortran (HSPF). Both models were developed based on the substantive monitoring data of the United States Geological Survey (USGS), National Oceanic and Atmospheric Administration (NOAA) and other stakeholders in Chesapeake Bay. To a great extent, model selection depends mainly on the goal of the simulation, the scale of the study area, the availability of data, the expected accuracy, and the temporal and financial costs (Grizzetti et al., 2005; Kovas, 2006; Kliment et al., 2008; Huang and Hong, 2010).

Pawlak, 1980 Laurence, 1992

Ernest, 2003

Williams, 1996

Kouassi and Biney, 1999 Chua, 1999 JME, 2002

SEPA, 2001; Peng et al., 2009; Zhang, 2009.

In order to quantify land-based pollution loads in the coastal areas of China, it is necessary to first identify the characteristic composition of land-based pollution before management control measures can be implemented. Unfortunately, in many cases, data covering the quality of effluent discharges and storm water runoff, streamflow and climate variables in the coastal areas are insufficient or inappropriate for the purpose of modeling. To address the above limitations, this paper focuses on the use of a GIS-based empirical model integrated with the Universal Soil Loss Equation (USLE), Sediment Delivery Ratio (SDR), and the empirical export coefficient method to quantify land-based pollutants in two selected coastal bays in China: Luoyuan Bay and Xiamen Bay. The purpose is to verify the applicability of the GIS-based model approach and to further illustrate the methods for quantifying and measuring the spatial distribution of landbased pollutant loads in coastal areas where only sparse data are available. 2. Approach A sectoral approach was used in establishing the pollution source inventory, which included the following broad categories of pollution sources: (a) point source (PS), e.g., municipal wastewater and industrial wastewater; (b) non-point source (NPS), e.g., rural domestic wastewater, soil losses, fertilizer use, and livestock and poultry breeding; and (c) river discharge. Airborne pollutants, which may reach the sea through short-distance or long-distance atmospheric transport, were not taken into consideration in this study but were the subject of a separate study (Pan, 2008). Based upon this inventory, an assessment of waste loads for each source category was made, which allowed an evaluation of the contribution by each category to the total pollution load of the coastal area

16

J. Huang et al. / Ocean & Coastal Management 81 (2013) 14e28

Fig. 1. The modeling procedure.

being studied. A similar approach for quantifying land-based pollutant loads in the coastal area, integrating GIS, remote sensing (RS) technology, empirical models and the export coefficients method was adopted. 2.1. Point source (PS) pollution loads

loads discharged into the sea were calculated using the following equation:

Lp ¼

n X

Fi  Qi  r

(1)

i

PS pollution mainly refers to domestic sewage and industrial wastewater, which are discharged into receiving water bodies after treatment in wastewater treatment plants (WTPs). In this study, the first step was to investigate the sewage treatment condition and the distribution of liquid wastes from industrial factories. In the case of municipal domestic wastewater and industrial wastewater treated in WTPs, land-based pollutants, specifically COD, N, and P pollutant

where Lp is the pollutant load from PS pollution; Fi is the flow rate for sewage outlet i (m3/s); Qi is the concentration of individual pollutants monitored at the sewage outlet i (mg/L); and r is the loss coefficient for physical, chemical and biological removal of nutrients during the transportation process, which was acquired according to related research (XUMEMCECS, 2003). It should be mentioned that untreated domestic wastewater is considered in

Table 2 Data requirements for the proposed method. Data

Data format

Application

(1) DEM

Grid

(2) Land use map (3) Soil including map and soils’ property

Raster map (Landsat TM) Vector map(polygon) Attribute data (Table, text file)

(4) Precipitation (5) Population in village (6) Livestock & poultry in village

Table or text file Table or text file Table or text file

(7) Fertilizer application (8) Industrial information (9) Sewage treatment condition (10) N and P concentration, flow at the outlet of the specific sub-watershed.

Table Table Table Table

a. To define boundary of study areas and sub-watershed b. To generate L, S factor of USLE model To generate C, P factor of USLE model a. To generate K factor of USLE model b. To calculate ER and SDR c. To calculate N, P losses due to soil losses To generate R factor of USL model To calculate domestic wastewater discharged into the sea To quantify the contribution of livestock & poultry breeding discharged into the sea To quantify the contribution of fertilizer use discharged into the sea To calculate point source pollutant loads discharged into the sea To calculate point source pollutant loads discharged into the sea To verify the calculated results

or or or or

text text text text

file file file file

Note: DEM e digital elevation model; TM e Landsat thematic mapper; USLE e Universal Soil Loss Equation; SDR e sediment delivery ratio; ER e enrichment ratio; R e rainfall erosivity factor in USLE; L e slope length factor in USLE; S e slope steepness factor in USLE; C e crop and management factor in USLE factor; P e conservation supporting practices factor in USLE; K e soil erodibility factor in USLE.

J. Huang et al. / Ocean & Coastal Management 81 (2013) 14e28

17

Fig. 2. Location of study regions.

this study as NPS pollution, and the calculation method for this will be introduced in the following section.

2.2. Non-point source (NPS) pollution loads Using integrated grid-based GIS with USLE, empirical models and export coefficient, the contribution and distribution of COD, N and P from NPS in terms of soil losses, fertilizer use, livestock and poultry breeding, and rural domestic wastewater were calculated and identified.

2.2.1. GIS application for delineating and defining boundary of study area and sub-watersheds The identification of sub-watersheds that comprise the whole watershed is commonly done in analyzing spatial variability of NPS processes and characteristics (Huang et al., 2008). In this study, the sub-watershed served as the basic unit for calculating and analyzing land-based sources of NPS. 2.2.2. Universal soil loss equation As an important form of land-based sources of NPS pollution, soil losses make an important contribution to the degradation of

Table 3 General description of Luoyuan Bay and Xiamen Bay. Study bay

Luoyuan Bay Xiamen Bay

Area (km2)

Population (cap.)

GDPa (108 Yuan)

Area of cropland

Fertilizer useb

Livestock þ poultryc

Land

Sea

Total

City

GDP1

GDP2

GDP3

(km2)

(t)

(104 cap.)

860 1696

150 145

253,183 1,672,356

61,656 1,141,606

9.6 18.5

24 737

8.8 632.4

115.6 238.26

5635 25,078

(5.8 þ 55.3)d 46.8 þ 203.1

Data sources: LSO, 2006b; XSO, 2008. a GDP means Gross Domestic Product. GDP1, GDP2, and GDP3 mean GDP from agriculture, primary industry and secondary industry, respectively. b Fertilizer use means the sum of N, P, and K chemical fertilizer. c Livestock refers to the amount of swine. d Data in Luoyuan County in 2003 (LSO, 2004). e No WTPs.

Sewage treatment ratio (%) ee 77%

18

J. Huang et al. / Ocean & Coastal Management 81 (2013) 14e28

Table 4 Input data source for proposed method. Data

Data format

Data sources

(1) DEM

Grid

(2) Land use map

Raster map (Landsat TM)

(3) Soil including map and soils’ property

(4) Precipitation

Vector map (polygon) Attribute data (Table, text file) Table or text file

DEM with 10 m resolution provided by Fuzhou EPBL; DEM with 83.3 m resolution downloaded via USGS websiteX Land use/land cover data obtained by unsupervised classification of TM in 2007 in Luoyuan Bay and 2008 in Xiamen Bay from China Remote Sensing Satellite Ground Station Soil Survey Group (SSG, 1991)

(5) Population in village

Table or text file

(6) Livestock & poultry in village

Table or text file

(7) Fertilizer application (8) Industrial information (9) Sewage treatment condition (10) N&P concentration, flow at the outlet of the specific sub-watershed

Table Table Table Table

or or or or

text text text text

file file file file

Daily precipitation in 2006, provided by Luoyuan Climate StationL; Daily precipitation in 2008, provided by Xiamen Climate StationX. Yearbook in Luoyuan (LSO, 2006b) and Lianjiang (LSO, 2005)L; Xiamen Statistics Office (XSO, 2008)X. Yearbook in Luoyuan (LSO, 2006b) and Lianjiang (LSO, 2006a)L; Xiamen Statistics Office (XSO, 2008)X. Yearbook in Luoyuan (LSO, 2006b) and Lianjiang (LSO, 2006a)L; Xiamen Statistics Office (XSO, 2008)X. Yearbook in Luoyuan (LSO, 2006b) and Lianjiang (LSO, 2006a)L; Xiamen Statistics Office (XSO, 2008)X. Yearbook in Luoyuan (LSO, 2006b) and Lianjiang (LSO, 2006a) L; Xiamen Statistics Office (XSO, 2008)X. Monitoring at Bili subwa. over the period June to Oct. 2007L; monitoring data in sewage outlet discharged into the sea in 2003, provided by FIO (FIO, 2003)X.

Note: LLuoyuan Bay; XXiamen Bay.

water bodies. In this study, USLE was applied in a GIS environment to determine the average annual soil loss for each grid of the USLE and to predict soil loss for a given site as a product of six major factors, where values at a particular location (e.g., a given grid cell) can be expressed numerically (Wischmeier and Smith, 1978). Soil losses were calculated as follows:

A ¼ RK LSCP

(2)

where A ¼ annual soil loss in t hm2 yr1; R ¼ rainfall erosivity factor (J mm m2 h1); K ¼ soil erodibility factor (t J1 mml); L ¼ slope length factor; S ¼ slope steepness factor; C ¼ crop and management factor; and P ¼ conservation supporting practices factor.

Fig. 3. Sub-watershed delineation in Luoyuan Bay.

J. Huang et al. / Ocean & Coastal Management 81 (2013) 14e28

top soil layer; Xkt in t hm2 yr1 is the average annual soil loss; Sd is sediment delivery ratio; and ER is the enrichment ratio. Sdi, the fraction of the gross soil loss from cell i that actually reaches a continuous stream system, was estimated following Ferro and Minacapilli (1995) as a function of travel time

Table 5 Land-based pollutant loads in Luoyuan Bay. Sub-seas

Land-based pollutant loads discharged into the sea (t a1) CODMn

TN

TP

Total

Sub-sea 1 Sub-sea 2 Sub-sea 3 Total

3420.58 2116.39 176.22 5713.19

975.44 763.12 47.0 1785.56

113.59 98.52 10.55 222.66

5662.3 3881.7 298.94 9842.94

Sdi ¼ expðbti Þ

Table 6 Land-based pollutants loads in Luoyuan Bay. Land-based pollutants loads

CODMn (t a1)

TN (t a1)

TP (t a1)

Industrial wastewater Domestic wastewater Livestock & poultry breeding Fertilizer use Soil losses Total

243.11 1258.07 623.57 e 3588.42 5713.17

e 159.26 404.82 494.92 726.56 1785.56

e 35.18 47.83 40.31 99.34 222.66

2.2.3. Empirical models and the export coefficient method 2.2.3.1. Empirical equations for calculating nutrient losses due to soil losses. Nutrient losses caused by soil losses were calculated as follows:

(3)

where LSkt in kg hm2 are the nutrient losses in particulate form; a is a constant; CSkt in kg mg1 is the N and P concentrations in the

(4)

where t is travel time (h); and b is basin-specific parameter. The time for runoff water to travel from one point to another in a watershed was determined by the flow distance and velocity along the flow path (SCS-TR-55, 1975; Bao et al., 1997). If the flow path from cell i to the nearest channel traverses Np cells, the travel time from that cell was calculated by adding the travel time for each of the Np cells located along the flow path (Jain and Kothyari, 2000):

ti ¼

Note: “e” No data.

LSkt ¼ a$CSkt $Xkt $Sd$ER

19

Np X li

v i¼1 i

(5)

where li is the length of segment i in the flow path (m) and equals the length of the side or diagonal of a cell depending on the flow direction in the cell; and vi the flow velocity for the cell (m/s). Flow velocity of overland flow and shallow channel flow can be estimated from the relationship (SCS-TR-55, 1975; Haan et al., 1994; Bao et al., 1997; Jain and Kothyari, 2000): 1=2

vi ¼ di si

(6)

where si is the slope of cell i (m/m); and di is a coefficient for cell i dependent on surface roughness characteristics (m/s).

Fig. 4. Source apportionment of non-point source total nitrogen in Luoyuan Bay.

20

J. Huang et al. / Ocean & Coastal Management 81 (2013) 14e28

Fig. 5. Source apportionment of non-point source total phosphorus in Luoyuan Bay.

The enrichment ratio was calculated as follows:

ER ¼

Cs runoff Cs soil

(7)

where Cs runoff is the percentage concentration of clay in surface soil; and Cs soil is the percentage concentration of clay in the river channel caused by soil losses. According to the ACTMO (Agricultural Chemical Transport Model), the specific surface (Ss) of the surface soil clay content (e) and clay ratio (Rc) of the soil clay content (m) caused by soil losses have a certain relationship, which can be calculated using the following formulas:

Ss; e ¼ 14:6 þ 0:84Ss; m

(8)

Rc; e ¼ 0:021 þ 1:08Rc; m

(9)

where Ss,e is specific surface for eroded material; Ss,m is specific surface for soil matrix; Rc,e is the clay ratio for eroded material; and Rc,m is the clay ratio for the soil metric. 2.2.3.2. Pollution loads from rural domestic wastewater and livestock and poultry breeding. The empirical export method for calculating pollution loads from rural domestic wastewater and livestock and poultry breeding in this study is presented as follows (Johnes, 1996):

LDþL ¼

n X

specific pollutant source i; Ni is number of livestock and poultry of type i, or of people; and Ii is the input of nutrients to source i. For animals, the export coefficient expresses the proportion of the wastes voided by the animal, which will subsequently be exported from stock houses and grazing land into the subwatersheds of the drainage network. The coefficient used in this study was derived from related literature (Zhang et al., 1997; CAEP, 2003). For human wastes, especially in the rural areas where domestic wastewater is directly discharged without treatment in WTPs, the export coefficient reflects the use of phosphate-rich detergents and dietary factors in the local population, and is adjusted to take into account any treatment of the wastes before they are discharged into the water body using the following equation:

Eh ¼ Dca  H  365  l

where Eh is annual export of N or P from the human population (kg year1); Dca is the daily output of nutrients per person (kg day1), which was determined for this study in a related research (Zhang et al., 1997; CAEP, 2003); H is the number of people in the catchment; 365 refers to number of days in a year; and l is

Table 7 Comparison of the model results and the measured values in the studied subwatershed in Luoyuan Bay. Items

Ei ½Ni ðIi Þ

(10)

i¼1

where LDþL is the loss of pollutants from rural domestic wastewater, livestock and poultry breeding; Ei is the export coefficient for

(11)

Monitored value Predicted value Deviation error (%)

Annual land-based pollutants load (t a1) TN

TP

CODMn

20.82 24.84 19.31

4.45 3.26 26.74

102.10 102.29 0.18

J. Huang et al. / Ocean & Coastal Management 81 (2013) 14e28

21

Fig. 6. Sub-watershed delineation in Xiamen Bay (Tong’an Bay and Western Sea).

a loss coefficient for physical, chemical and biological removal of nutrients during transportation process.

Therefore, fertilizer use here was viewed as an important form of NPS. N and P loads from agricultural chemical fertilizer use were calculated in this study according to Equations (12) and (13).

2.2.3.3. Pollution loads from agricultural chemical fertilizer use. N and P from excessive fertilizer use can be discharged into the receiving water via rainfall events and irrigation practices.

LfP ¼ a  ðP þ T  pÞ  R

(12)

where LfP is the loss of P discharged into receiving water; a is the

Table 8 Verifying non-point source pollutants loads using different methods and generating the transported coefficient. Study sub-watershed.

Methods

Major non-point source pollutants

Sub-watersheds # 11, 12

Proposed approach Method based on the monitored data in sewage outlet discharged into the sea Difference (%)

1944.05 1599.3 17.73

48.44

58.99

Sub-watersheds # 14, 15

Proposed approach Method based on the monitored data in sewage outlet discharged into the sea Difference (%)

5511.61 4492.3

1110.64 6,38.65

114.23 63.86

18.49

42.50

44.10

Proposed approach Method based on the monitored data in sewage outlet discharged into the sea Difference (%)

7387.74 5797.83

1109.11 636.4

73.99 29.4

21.52

42.62

60.27

19.25

44.52

54.45

CODMn

Sub-watershed # 16

Aver. deviation error (transported coefficient) (%)

TN 355.49 183.29

TP 27.67 11.35

Note: The data regarding water quantity and water quality in sewage outlet discharged into the sea is from Environmental Monitoring Station in EPB of Xiamen. Sewage outlets discharged into the sea was monitored in 2003 (FIO, 2003).

22

J. Huang et al. / Ocean & Coastal Management 81 (2013) 14e28

2.3. River discharges

Table 9 Land-based source pollutants loads in XMB (unit: t a1). Sub-Bays

Sub-seas

CODMn

TP

Total

Tong’an Bay

Sub-sea TD Sub-sea TX Sub-sea TI Total

8020.1 1120.8 2955.7 12,096.6

5220.5 358.5 903.3 6482.3

487.6 25.2 63.5 576.3

13,728.2 1504.5 3922.5 19,155.2

Western Sea

Sub-sea WN Sub-sea WS Total

4784.9 4420.4 9205.3

1559.6 2038.2 3597.8

111.6 151.8 263.4

6442.1 6610.4 13,052.5

21,301.9

10,080.1

839.7

32,207.7

Total

TN

coefficient of transferring P2O5 to P, namely 0.44; P represents the amount of P fertilizer applied for crops; T represents the amount of compound fertilizer applied into soil; p is the proportion of P2O5 in the compound fertilizer; R is the rate of P loss from source to the receiving water.

LfN ¼ ðN þ T  nÞ  R

(13)

where LfN is the loss of N discharged into the receiving water; N represents the amount of N fertilizer applied to the soil; T represents the amount of compound fertilizer applied to the soil; n is the proportion N fertilizer in the compound fertilizer; and R is the rate of N loss from the source to the receiving water.

River discharges are an important part of land-based pollutant sources. In this study, the assessment of the quantity of land-based pollutant load of the rivers was realized by multiplying water quantity by water quality based on the data monitored at the outlet of the river. 2.4. The modeling procedure An approach for quantifying land-based pollutant loads in coastal areas where there are insufficient field data was adopted by integrating GIS, RS technology, empirical models (pollutant loss equations) and export coefficients. The modeling procedure is shown in Fig. 1. 2.5. Data requirements The data requirements for this proposed approach are listed in Table 2. As shown in Table 2, some data, such as population, numbers of livestock and poultry, fertilizer use, sewage treatment conditions, can be obtained through literature survey (including local yearbooks); while other data, such as N and P concentrations, and flow at the outlet of the specific sub-watersheds, can be obtained via in situ

Fig. 7. Source apportionment of land-based pollutant (COD) in Tong’an Bay and Western Sea.

J. Huang et al. / Ocean & Coastal Management 81 (2013) 14e28

surveys and sampling. Additional data, such as a digital elevation model, RS imagery [e.g., Landsat Thematic Mapper (TM)], soil and precipitation are available from relevant government departments. 3. Case studies In this study, Luoyuan Bay and Xiamen Bay were chosen to illustrate and verify the methodology proposed. 3.1. General description of Luoyuan Bay and Xiamen Bay and data sources Luoyuan Bay and Xiamen Bay are both located in the coastal area of southeast China (Fig. 2). As the sixth largest bay in Fujian Province, Luoyuan Bay is located northeast of Fuzhou, the capital of Fujian Province, and is surrounded by mountains, with a mean elevation of 215 m (Huang et al., 2010). On the other hand, Xiamen Bay, which is located in the southern part of Fujian Province and on the west coast of the Taiwan Strait, supports a traditional trading port and serves as a tourist coastal city in southeast China. Table 3 is a general description of the two bays. As illustrated in Table 3, Xiamen Bay suffers from the impacts of intensive

23

urbanization compared to Luoyuan Bay. The Xiamen Bay region is characterized by a higher GDP and urban population growth and the rate of sewage treatment is comparatively higher. Table 4 lists the input data for the models and equations in the proposed method. 3.2. Case study 1: Luoyuan Bay Twenty-five sub-watersheds and three sub-sea areas were delineated in Luoyuan Bay (Fig. 3). The annual land-based pollutant loads in Luoyuan Bay are summarized in terms of sub-seas. As shown in Table 5, CODMn is the largest contributor, followed by total nitrogen (TN), while total phosphorus (TP) is the least. Additionally, sub-sea 1 receives the greatest total loading of CODMn, TN and TP. Land-based sources of major pollutants were further identified in terms of industrial wastewater, livestock and poultry breeding, fertilizer use and soil losses. As shown in Table 6, and Figs. 4 and 5, soil losses largely contributed to the CODMn, TN and TP loads. The major pollution sources of CODMn are soil losses (62.81 percent), followed by domestic wastewater (22.02 percent), and industrial wastewater with the smallest proportion (4 percent). Livestock and poultry breeding are the second largest sources of TP, while the second largest source of TN is fertilizer use. It should be noted that

Fig. 8. Source apportionment of land-based pollutant (total nitrogen) in Tong’an Bay and Western Sea.

24

J. Huang et al. / Ocean & Coastal Management 81 (2013) 14e28

the main sources of NPS pollutants TN and TP vary in different subseas. For TP, the contribution of soil losses in sub-sea 3 accounts for more than 50 percent, whereas, for TN, the contribution of soil losses in all the sea areas except for sub-sea 2, accounts for more than 40 percent. In order to verify the results derived from the above model, the data from the Environmental Monitoring Station in Luoyuan County, which was generated through in situ sampling of water quality and water quantity over the period JuneeOctober 2007, were analyzed. Based on the water quantity and water quality data at the outlet of the study sub-watershed, the annual pollutant loads were calculated and compared with the results derived from the model. As shown in Table 7, the deviation errors of TN, TP and CODMn pollutant loads are 19.31 percent, 26.74 percent and 0.18 percent, respectively. As such, the model calculation deviation errors are within an acceptable error range. Thus, the modeling result can be used as scientific support for controlling and managing land-based pollution in Luoyuan Bay. 3.3. Case study 2: Xiamen Bay Twenty-seven sub-watersheds and five sub-seas (namely, subsea TD, TI, TX, WN and WS) were delineated in Xiamen using GIS

(Fig. 6). Given the geographic distribution of sewage outlets, subwatersheds 11 and 12; 14 and 15; and 16 were selected to verify the calculated results. Based on the monitoring data of water quantity and water quality in the sewage outlets (except for the PS at the outlet of the WTP) discharging into Xiamen Bay, pollutant loads discharged into the sub-watersheds (Fig. 6) were calculated. The results were compared with the results obtained using the proposed methods in this study, integrating GIS with the empirical export coefficient, so as to verify NPS pollution load calculations and further generate the transported coefficient. Intuitive knowledge of flow regime and nutrient export indicates that the results obtained via the proposed approach will exceed the result obtained from multiplying the data for water quantity by that of water quality in the sewage outlet discharged into the sea. This implies that the NPS pollution loads obtained using GIS and the empirical export coefficient method take into account the potential amount of NPS pollution emission from all sources, but ignore pollutant losses during the transport and migration processes. To calculate the actual amount of pollutants discharged into the sea, the reduction coefficient must also be taken into consideration. In this study, the actual amount of NPS pollutants discharged into the sea is the difference between the

Fig. 9. Source apportionment of land-based pollutant (total phosphorus) in Tong’an Bay and Western Sea.

J. Huang et al. / Ocean & Coastal Management 81 (2013) 14e28

results from the GIS-based empirical method, and the results obtained by multiplying the data on water quantity by that of the water quality in the sewage outlets. This difference is called the transported coefficient and is shown in Table 8. As Table 8 shows, the average difference of CODMn is the smallest, i.e., only 19.25 percent, followed by TN, with 44.52 percent; while TP is the largest, with 54.45 percent. Considering the pollutant transport processes of CODMn, TN and TP and the results of research on Xiangshan Bay (XUMEMCECS, 2003), the authors believe that the results of NPS pollution load based on the proposed method are basically reasonable, credible and can be further used to estimate land-based sources of NPS pollution loads of Xiamen Western Sea and Tong’an Bay. Consequently, the difference between the two methods for land-based pollutants CODMn, TN, and TP was used as the transported coefficient. As shown in Table 8, the reduction coefficients for CODMn, TN, and TP in Tong’an Bay and the Western Sea are 19.25 percent, 44.52 percent, and 54.45 percent, respectively, which form the basis for estimating the landbased pollutant loads discharged into Xiamen Bay. The results of the annual pollutant loads from land-based pollution in Xiamen Bay are summarized in terms of sub-seas in

25

Table 9. From the table, it can be seen that the order of land-based pollutant loads entering Tong’an Bay and the Western Sea is CODMn, TN and followed by TP. In addition, CODMn, TN and TP loads in Tong’an Bay are much more than those of the Western Sea. This phenomenon is in part due to the fact that Tong’an Bay has a larger land area than that of the Western Sea. Comprehensive analysis of the contribution from PS, NPS and river discharges, CODMn, TN, and TP pollutant loads entering Tong’an Bay and the Western Sea is presented in Figs. 7e9. Source apportionment of the land-based pollutants (CODMn, TN and TP) shows a large difference among the five sub-seas in Tong’an Bay and the Western Sea (Figs. 7e9). NPS is the main source for CODMn in all sub-seas, accounting for over 60 percent of the total loads. NPS contributed largely to TN and TP for most sub-seas. Interestingly, river discharge and PS were responsible for considerable TN and TP loads in the sub-sea TD and sub-sea WS, respectively. Further analysis of source apportionment of diffuse N and P loads entering Tong’an Bay and the Western Sea was conducted. The results are illustrated in Figs. 10 and 11. Although some spatial variability of source apportionment of NPS pollution of TN and TP exhibited among the five sub-seas

Fig. 10. Source apportionment of non-point source total nitrogen in Tong’an Bay and Western Sea.

26

J. Huang et al. / Ocean & Coastal Management 81 (2013) 14e28

Fig. 11. Source apportionment of non-point source total phosphorus in Tong’an Bay and Western Sea.

Table 10 Comparison of source apportionment for major land-based pollutants among other coastal areas. Study area

Pollutants Pollution load from different sources (103 ton/year) Percentage of pollution load from different sources (%)

Mediterranean (Helmer, 1977)

BOD 500 COD 1100 Phosphorus 22 Nitrogen 110

Domestic

Black Sea Coast of Turkey (Tuncer et al., 1998) Phosphorus Nitrogen TSS

Industrial

Agricultural

River

Domestic

Industrial

Agricultural

River

900 2400 5 25

100 1600 30 65

1000 2700 260 600

20.00 14.10 6.94 13.75

36.00 30.77 1.58 3.13

4.00 20.51 9.46 8.13

40.00 34.62 82.02 75.00

e e e

3.53 36.30 2151.58

3.99 8.56 1.53

e e e

96.01 91.44 98.47

0.15 3.40 33.40

Baltic Sea (Larsson et al., 1985)

BOD5/7 TP TN

228.76 18.20 87.83

367.48 3.58 13.93

e e e

1094.11 50.17 640.50

13.53 25.29 11.83

21.74 4.98 1.88

e e e

64.73 69.73 86.29

Baltic Sea (Pawlak, 1980)

BOD7 TP TN

314.10 2.77 24.14

268.47 0.89 19.40

e e e

506.96 18.41 258.23

28.83 12.57 8.00

24.64 4.03 6.43

e e e

46.53 83.41 85.57

16.54 3.42 0.66

12.37 0.48 0.84

e e e

e e e

57.22 87.62 44.01

42.78 12.38 55.99

e e e

e e e

Niger Delta, Nigeria (Ajao and Anurigwo, 2002) BOD5 Phosphorus Nitrogen Note: “e” no data.

J. Huang et al. / Ocean & Coastal Management 81 (2013) 14e28

located in Tong’an Bay and the Western Sea, domestic wastewater and fertilizer use are the two main forms of NPS pollution (Figs. 10 and 11). The contribution of soil losses, livestock and poultry breeding is small. This phenomenon shows a large difference with that in Luoyuan Bay. Consequently, more attention should be given to domestic wastewater pollution and chemical fertilizer use in the management of Tong’an Bay and the Western Sea. 4. Discussion and conclusion In this study, a systematic approach for quantifying and measuring the spatial distribution of land-based pollutant loads in Louyuan Bay and Xiamen Bay was demonstrated through the integration of Raster GIS, USLE, SDR, and the empirical export coefficient method. The methodology and data now available, despite some limitations, enabled the authors to quantify land-based pollution loads. The findings can be compared with those of other bays so as to secure some management pointers for further reference in controlling land-based pollution. Table 10 shows the range of source apportionment for land-based pollutants in different coastal areas of the world. Rivers play a dominant role in land-based pollutant discharge into the Mediterranean Sea, Black Sea and Baltic Sea (Table 10). This is understandable since rivers receive domestic, industrial and agricultural pollutant loads from the entire drainage basins and act as vehicles for major pollutants transport into the sea (Helmer, 1977). In our study, the river contribution is similarly high in N and P loads for sub-sea TD in Xiamen Bay, as shown in Figs. 7 and 8. However, for Luoyuan Bay, the contribution of the river was not significant partly due to the fact that the river in this region is relatively small and partly due to lack of monitoring data. The current study showed that CODMn in Luoyuan Bay was derived mainly from soil losses and rural domestic wastewaters, amounting to 63 percent and 22 percent of the total load respectively. On the other hand, soil losses are the main source for TN and TP, accounting for over 40 percent while point source pollution contributed only 4 percent. In the Xiamen Bay, non-point source was the main source for CODMn in all sub-seas, accounting for over 60 percent of the total load, as well as contributing to a large proportion of TN and TP loads for most sub-seas. These pollutants arise mainly from rural domestic wastewater and chemical fertilizers from agriculture practices. However, river discharge and point source pollution were responsible for a considerable TN and TP load in some sub-seas such as sub-sea TD and sub-sea WS. Application of the proposed approach at the two bays resulted in clear identification of the source apportionment and spatial distribution of land-based pollutant loads despite the sparse data available. The approach quantifies the source apportionment of land-based pollution in terms of river discharge, point and nonpoint sources (including soil losses, fertilizer use, livestock/ poultry breeding, and rural domestic wastewater discharge). The approach enables classification and identification of the critical areas for land-based pollution in coastal areas, thereby providing scientific support for coastal management measures. Acknowledgments This research was supported by both the State Oceanic Administration “908” project (Grant No. 908-02-02-03) of China and the Fuzhou Environmental Protection Bureau. Gratitude is extended to Professor John Hodgkiss for his assistance with the English language.

27

References Ajao, E.A., Anurigwo, S., 2002. Land-based sources of pollution in the Niger Delta, Nigeria. Ambio 31 (5), 442e445. Bao, J., Maidment, D., Olivera, F., 1997. Using GIS for Hydrologic Data-processing and Modeling in Texas. CRWR Online Report No. 97-4. Center for Research in Water Resources, University of Texas at Austin, TX. Bartlett, D., Smith, J., 2004. GIS for Coastal Zone Management. CRC Press, London. CAEP (Chinese Academy on Environmental Planning), 2003. Guideline for Surface Water Environmental Capacity Checking Technology in China. CECWED (Central Environmental Council Water Environmental Division, Japan), 1999. Basic Policy for the Fifth Areawide Total Pollutant Load Control. Interim report of areawide pollutant load control ad hoc committee. Chua, T.E., 1999. Marine pollution prevention and management in the East Asian Seas: a paradigm shift in concept, approach and methodology. Mar. Pollut. Bull. 39, 80e88. Claussen, U., Zevenboom, W., Brockmann, U., Topcu, D., Bot, P., 2009. Assessment of the eutrophication status of transitional, coastal and marine waters within OSPAR. Hydrobiologia 629, 49e58. CNSPCP (Coastal Nonpoint Source Pollution Control Program), 1990. Section 6217 of the Coastal Zone Act Reauthorization Amendments of 1990. www.epa.gov/nps/ coastnps.html (accessed 29.12.00). Donigian, A.S., Bicknell, B.R., Linker, L.C., Hannawald, J.E., Chang, C., Reynolds, R., 1990. Chesapeake Bay Program e Watershed Model Application to Calculate Bay Nutrient Loading. Environmental Research Laboratory, Office of Research and Development, USEPA, Athens, GA. EPA, 1994. EPA’s Polluted Brochure. EPA-841-F-94-005. www.epa.gov/owow/nps/ qa.html (accessed 07.03.08). Ernest, H.R., 2003. Chesapeake Bay Blues, Science, Politics, and Struggle to Save the Bay. New York, Oxford. Ferro, V., Minacapilli, M., 1995. Sediment delivery processes at basin scale. Hydrolog. Sci. J. 40, 703e717. FIO (Fujian Institute of Oceanography), 2003. Technical report on surveying and monitoring land-based pollution sources in Xiamen municipals. GESAMP, 1990. The State of the Marine Environment. IMO/FAO/UNESCO/WMO/ IAEA/UN/UNEP Joint Group of Experts on the Scientific Aspects of Marine Pollution. UNEP Regional Seas Reports and Studies No. 115. UNEP, Nairobi. Grizzetti, B., Bouraoui, F., Marsily, G.D., 2005. Modelling nitrogen pressure in river basins: a comparison between a statistical approach and the physically-based SWAT model. Phys. Chem. Earth 30, 508e517. Haan, C.T., Barfield, B.J., Hayes, J.C., 1994. Design Hydrology and Sedimentology for Small Catchments. Academic Press, San Diego. Halpern, B.S., Walbridge, S., Selkoe, K.A., Kappe, C.V., Micheli, F., D’Agrosa, C., Bruno, J.F., Casey, K.S., Ebert, C., Fox, H.E., Fujita, R., Heinemann, D., Lenihan, H.S., Madin, E.M.P., Perry, M.T., Selig, E.R., Spalding, M., Steneck, R., Watson, R., 2008. A global map of human impact on marine ecosystems. Science 319, 948e952. Helmer, R., 1977. Pollutants from land-based sources in the Mediterranean. Ambio 6 (6), 312e316. Huang, J., Hong, H., 2010. Comparative study of two models to simulate diffuse nitrogen and phosphorus pollution in a medium-sized watershed, Southeast of China. Estuar. Coast. Shelf Sci. 86 (3), 387e394. Huang, J., Hong, H., Zhang, L., 2008. Research on control division of agricultural nonpoint source pollution at a medium-sized watershed scale in Southeast China. Front. Environ. Sci. Eng. China 2 (3), 333e339. Huang, J., Lin, J., Tu, Z., 2010. Detecting spatiotemporal change of land use and landscape pattern in a coastal gulf region, southeast of China. Environ. Dev. Sustain. 12 (1), 35e48. Jain, M.K., Kothyari, U.C., 2000. Estimation of soil losses and sediment yield using GIS. Hydrolog. Sci. J. 45, 771e786. JME (Japanese Ministry of the Environment), 2002. Environment White Paper of Heisei, 13 (2001). Johnes, P., 1996. Evaluation and management of the impact of land use change on the nitrogen and phosphorus load delivered to surface waters: the export coefficient modelling approach. J. Hydrol. 183, 323e349. Kliment, Z., Kadlec, J., Langhammer, J., 2008. Evaluation of suspended load changes using AnnAGNPS and SWAT semi-empirical erosion models. Catena 73, 286e299. Kouassi, A.M., Biney, C., 1999. Overview of the marine environmental problems of the West and Central African region. Ocean Coast. Manage. 42, 71e76. Kovas, A., 2006. Comparative study of two watershed scales to calculate diffuse phosphorus pollution. Water Sci. Technol. 53 (2), 281e288. Larsson, U., Elmgren, R., Wulff, F., 1985. Eutrophication and the Baltic Sea: causes and consequences. Ambio 14 (1), 9e14. Laurence, D.M., 1992. The Black Sea in crisis: a need for concerted international action. Ambio 21 (4), 278e286. Li, D., Dag, D., 2004. Ocean pollution from land-based sources: East China Sea, China. Ambio 33 (1e2), 107e113. LSO (Luoyuan Statistics Office), 2004. Luoyuan Statistics. Luoyuan Statistics Office. LSO (Lianjiang Statistics Office), 2005. Lianjiang Statistics. Lianjiang Statistics Office. LSO (Lianjiang Statistics Office), 2006a. Lianjiang Statistics. Lianjiang Statistics Office. LSO (Luoyuan Statistics Office), 2006b. Luoyuan Statistics. Luoyuan Statistics Office.

28

J. Huang et al. / Ocean & Coastal Management 81 (2013) 14e28

Melvasalo, T., 2000. Regional marine environmental management and the GPA-LBA: perspectives and the need for scientific support. Ocean Coast. Manage. 43, 713e724. Ning, S.K., Chang, N.B., 2007. Screening the relocation strategies of water quality monitoring stations by compromise programming. J. Am. Water Resour. Assoc. 41 (5), 1039e1052. Pan, C., 2008. Estimation of marine pollution loads in Xiamen Bay. Master thesis in Xiamen University. Pawlak, J., 1980. Land-based inputs of some major pollutants to the Baltic Sea. Ambio 9 (3e4), 163e167. Peng, B., Hong, H., Xue, X., Jin, D., 2006. On the measurement of socioeconomic benefits of integrated coastal management (ICM): application to Xiamen, China. Ocean Coast. Manage. 49, 93e109. Peng, B., Jin, D., Burroughs, R., 2009. Regional ocean governance in China: an appraisal of the clean Bohai Sea. Coast. Manage. 37 (1), 70e93. Pullar, D., Spinger, D., 2000. Toward integrating GIS and catchments models. Environ. Modell. Softw. 15, 451e459. Schumacher, M., Hoagland, P., Gaines, A., 1996. Land-based marine pollution in the Caribbean: incentives and prospects for an effective regional protocol. Mar. Policy 20 (2), 99e121. SCS-TR-55, 1975. Urban Hydrology for Small Watersheds. Technical Release No. 55. USDA Soil Conservation Service, Washington, DC. SEPA (State Environmental Protection Administration), 2001. Blue Bohai Sea Action Plan. SEPA, Beijing. SFBCDC (San Francisco Bay Conservation and Development Commission), 2003. Final Staff Report: Water Quality Protection and Non-point Source Pollution Control in San Francisco Bay. Sprague, L.A., Langland, M.J., Yochum, S.E., Edwards, R.E., Blomquist, J.D., Phillips, S.W., Shenk, G.W., Preston, S.D., 2000. Factors Affecting Nutrient Trends

in Major Rivers of the Chesapeake Bay Watershed. United States Geological Survey. SSG (Soil Survey Group), 1991. Soils in Fujian. Fujian Science and Technology Press, Fuzhou. Syvitski, J.P.M., Vorosmarty, C.J., Kettner, A.J., Green, P., 2005. Impact of humans on the flux of terrestrial sediment to the global coastal ocean. Science 308 (5720), 376e380. Tuncer, G., Karakas, T., Balkas, T.I., Gokcay, C.F., Aygnn, S., Yurteri, C., Tuncel, G., 1998. Land-based sources of pollution along the Black Sea coast of Turkey: concentrations and annual loads to the Black Sea. Mar. Pollut. Bull. 36 (6), 409e423. Williams, C., 1996. Combating marine pollution from land-based activities: Australian initiatives. Ocean Coast. Manage. 33 (1e3), 87e112. Wischmeier, W.H., Smith, D.D., 1978. Predicting Rainfall Erosion Losses. A Guide to Conservation Planning. The USDA Agricultural Handbook No. 537. Wong, K.M., Strecker, E.W., Stenstrom, M.K., 1997. GIS to estimate storm-water pollution mass loading. J. Environ. Eng. 123 (8), 737e745. XSO (Xiamen Statistics Office), 2008. Xiamen Statistics. Chinese Statistics Press, Beijing. XUMEMCECS (Xiamen University and Marine Environmental Monitoring Center of East China Sea) State Oceanography Administration, 2003. Report on Marine Environmental Capacity and Total Amount Control in Xangshan Bay of Ningbo. Yoshiaki, T., 2006. An index directly indicates land-based pollutant load contributions of domestic wastewater to the water pollution and its application. Sci. Total Environ. 370, 425e440. Zhang, D., Zhang, X., Zhang, J., Shen, G., 1997. Integrated research and evaluation on non-point source pollutions in Shanghai suburbs. Acta Agr. Shanghai 13 (1), 31e36. Zhang, Z., 2009. Sustainable Development Strategy and Implementation Plan of Bohai Sea, Status and Progress. National Marine Environmental Monitoring Center, State Oceanic Administration, China, Manila, Philippines.