Improving Non-Point Source Pollution Model Input Parameters Using Substance Flux Analysis

8/6/2019 Improving Non-Point Source Pollution Model Input Parameters Using Substance Flux Analysis

1/13

Journal of Applied Sciences 9 (14): 2519-2531,2009ISSN 1812-5654 2009 Asian Network for Scientific Information

Improving Non-Point Source Pollution Model Input ParametersUsing Substance Flux Analysis

lK. Kitbamroong, 2.3p.Sompongchaiyakul and 4.SG. Padmanabhan'National Research Center for Environmental and Hazardous Waste Management,

Chulalongkorn University, Bangkok, 10330, Thailand2Biogeochemical Research Unit, Faculty of Environmental Management,

Prince of Songkla University, Songkhla, 90112, Thailand3National Research Center for Environmental and Hazardous Waste Management, PSU Satellite Center,

"Department of Civil Engineering, North Dakota State University, Fargo, ND 58105, USA5North Dakota Water Resources Research Institute, Fargo, ND 58105, USA

Abstract: This study investigated Substance Flux Analysis (SFA) as a potential tool to obtain better estimatesof phosphorus and cadmium loadings in an on-going comprehensive research effort to model the phosphorusand cadmium transport via., surface runoff to Songkhla Lake in the southern part of Thailand. The lake is amajor producer for local fisheries. Because substantial portions of the drainage area are used for agriculture,non-point source pollution loading from the surrounding drainage area to the lake has become a concern,especially since surface runoff is the major transport mechanism for non-point source pollutants. Using SFAit is estimated that approximately 384,289 t of chemical fertilizer were applied throughout the basin in 2004. Themajor sub-watershed contributing phosphorus and cadmium was the U-Tapao and Eastern Coast Sub Basin4 Sub-watershed. Changing the fertilizers from 8-24-24, 13-13-21 and 15-15-15 (high cadmium) by 15-15-15(low cadmium) type led to a significant decrease in cadmium contribution to the lake.Key words: Songkhla Lake basin, agricultural pollution, phosphorus transport, cadmium transport, distributed

parameter models, AnnAGNPS model, TREX modelINTRODUCTION

Material reported in this study is part of acomprehensive research effort to model the non-pointsource loading of phosphorus and cadmium to SongkhlaLake, Thailand, from the surrounding drainage area. Aschematic of the overall study is shown in Fig. 1. Thestudy focuses on the land used predominantly foragriculture.

Spatially distributed parameter models are commonlyused for modeling non-point source pollution transportvia., surface runoff (Tiemeyer et al., 2007; Chen andMackay, 2004). The term lumped or distributed describesthe way in which the model handles the spatial data.Lumped models use spatially averaged parameters andperform computations over the whole catchment region.However, increases in the within variation for a catchmentmay negatively affect the accuracy of the modelpredictions (Oudin et al., 2006). Distributed models arebased on the discretization of the landscape into smallerfunctional land units. Typically, a uniform grid is used forcomputational convenience. Calculations are performed

Fig. 1: Schematic of the comprehensive studyon discrete cells and then accumulated to makepredictions over the whole catchment (Curtis et al., 2005).With advances in computer technology, distributedmodels have gained popularity. The disadvantage ofdistributed parameter models is that they require largeamount of data; the advantage is that they are able tobetter account for local variability in land and loading

Corresponding Author: Kitipan Kitbamroong, 160/19 Chaiyapruk 2, Tiwanoonwongwan 2,345 Road, Bangkoowad,Muang Pratumthani, Pratumthani, 12000, Thailand Tel: +662-9973579

2519


2/13

J.Applied Sci., 9 (14): 2519-2531,2009conditions. This ability is important for making landmanagement decisions that require a better understandingof land processes within a catchment and for applyingfarm scale management options.

The Annualized Agricultural Non-Point Source(AnnAGNPS) model is a widely used model for nutrienttransport via., runoff in agricultural watersheds. However,the Two-dimensional Runoff and Erosion and Export(TREX) model is the only model that can handle metaltransport via., runoff. Both are spatially distributedcell-based models (Yuan et aI., 2008; Polyakov et aI.,2007; Velleux et aI., 2008, 2006). Therefore, these twomodels were considered in this study: AnnAGNPS forpredicting phosphorus transport and TREX forpredicting cadmium transport. Reliable results frommodels can be expected only when good quality inputdata are used. Specifically, we need accurate estimates ofthe phosphorus and cadmium available for transport ateach cell.

Substance Flux Analysis (SFA) is a technique usedfor tracking and assessing inputs, stocks and outputs ofa particular substance for a defined system boundary. Themethod establishes a mass balance of goods and selectedsubstances (Kleijn et al., 2008, 2000). SFA is a widelyestablished method that has been used to supportdecision making in various fields such as wastemanagement, nutrient management and urban metabolismanalysis (Doberlet al., 2002; Hug andBaccini, 2002;Lofts,2007). Compared with conventional mass balanceapproaches, SFA is more comprehensive since it relatessubstance flows to processes (Lofts, 2007). By providingearly recognition of environmentallyrelevant material fluxchanges, it enables the identification of potentialenvironmental problem, helps tracing the origins ofpollution problems and, therefore, can be used to supportenvironmental policy developments (Lofts, 2007). SFAcan be applied at different geographical scales and atdifferent locations. Chen et al. (2008) carried out a SFAstudy in China and found that the technique can still beused effectively even in situations where poor dataavailability and quality are encountered. In data richcountries, the application of SFA for various decisionsupport purposes is becoming increasingly common(Tangsubkul et aI., 2005; Kwonpongsagoon et aI.,2007a, b).

Songkhla Lake is located in southern Thailand.Intense agriculture-related activities in the lakecatchments have raised concerns about the potential forsurface runoff to transport significant amounts ofphosphorus and cadmium into the lake. In recent years,macrophyte blooms lasting for several months have beenobserved in the middle of the lake. Phosphorus causesconcern because it is the prime cause of eutrophication in

tropical lakes (Drolc and Koncan, 2002). Cadmium wasalso found in the lake sediments (Sirinawin andSompongchaiyakul, 2005). Cadmium is a highly toxic,bio-accumulative heavy metal that can cause kidneydisease and prostate cancer if ingested. Cadmium iscommonly associated with rock phosphates used as rawmaterial for producing fertilizers. Prolonged use ofcadmium-contaminated products can lead to unacceptablyhigh concentrations of cadmium in agricultural soil, asituation inwhich cadmium can accumulate to high levelsin food crops such as rice and root vegetables. Cadmiumis subject to health standards in most importing nations.High concentration of cadmium, if found in crops, couldlead to the rejection of those crops for humanconsumption or even for animal feed. Contaminated soilsover large areas are difficult or impossible to remedy andeventually result in severe public health and ecologicalproblems. Increasing agricultural activities in thecatchments have the potential to increase cadmiumaccumulation in the lake (Sae-Eong et aI., 2002). Though,cadmium is not of great concern under the presentconditions, it could pose problems under potential futurescenarios of agricultural activities in the basin. Publishedliterature on the pollutant transport to Songkhla Lake isscarce. Recent studies are mostly published in Thailanguage journals. Sereewatthanachai et al. (2004)performed a preliminary study of phosphorus andcadmium in the agricultural soil of the catchment andstated that estimates carmot be precisely made on thearmual rate of accumulation. Liu et al. (2004) created astatistical model by balancing the physical quantities ofphosphorus flows using SFA. Chen et al. (2008)investigated the nutrient flows in agricultural systems inChina using a partial substance flow analysis (SFA)method and an Agricultural Phosphorus Flow Analysis(AgiPhosFA) model. However, they make no attempt tocompare SFA estimateswith other measurements direct orindirect. None of the authors intended to use their resultsto improve input parameters of non-point source pollutanttransport models.

The objective of this study is to investigate thepotential of SFA to improve estimates of phosphorus andcadmium available on the land for transport by surfacerunoff into Songkhla Lake, Thailand. Data for the year2004 were used. The use of SFA to improve estimates ofinput loading of phosphorus and cadmium to non-pointsource pollution models is the focus of this study.

MATERIALS AND METHODSThe study was conducted in the Songkhla Lake basin

(Fig. 2) located in the Southern part of Thailand duringJanuary-December, 2004.

2520


3/13

J . A, pp li ed SC i. , 9 (14): 2519-2531, 2009

Fig. 2: Sub-watersheds of the Songkhla Lake Basin

Models usedAnnu al iz ed a gr i; u.l tl lr .l l n on -p o in t s o uree (A nnA G NP S)model: Transport of phosphorus via runoff is one of theconcerns in this study. The predominant land use in thestudy area is ae.;iculture. The Annf..GNPS model is widelyused for modeling nutrient transport via., runoff inagri cul.tural water she ds, The basi c modeling componentsin the model are hydrology, sediment, nutrient andpesticide transport. The Jl.nnP....GNPS model requires alarge amount of input data: soil, land use, managementand weather. The m anagem ent data include s crop,non- crop, fertiliz er, fi e1d m anagem ent, sche dulem anagem ent and op e ration m anagem ent data. The modeltracks nutrients such as nitrogen and phosphorustransport, but does not track transport of metals such ascadmium (Yuanetai., 2008; PolyakovetaI., 2007). One ofthe obj ective s of this study is to estim ate inputpho sphorus am ounts as ac cur ate 1y as p ossi bl e f or theAnnAGNPS mode1.T wo-d im ensionalrw wffand en siln and e :\ }l '011 (TRE X)model: Another concern in this study is the transportof cadmi urn via., runoff. TREX is the on! Y sp ati allydistributed cell-based model which can handle metaltransport via., runoff (IIell eux eid,2008, 2006) . TREXhas four major submode1s: (1) hydrology, (2 ) sedimenttransport, (3) chemic el transp ort and (4 ) metal tax icity. Theam aunt of cadmiurn avail able for transp ort is an im p ortantinput to the model. Another objective of this study is toestim ate input cadmi urn am ounts as accur ate1y as possiblefor the TREX model.Study area: Songl:dl1a Lake and its basin are spread over3 provinces of Thailand: Phettalung, Songkhla andN akhon Si Thamm arat (Fig 2). The 1ek e coversappr oxim ate1y an area of 1,042 km ~with a drainage are a of

LegendDL~eDub-W>J1mh!dbomdzry1 Kl.ongP~~'%L a end ~2 N 1 J! :h om .3T~4 ~ BO l 'J .5 PhIIl Poh mdRlJ. t t l Jpl rum.6 U-T~~ mdE:l s t .em CON Sub Basin41 ElstmtC~~Sub Basin 2 mi38 ElstmtC oNSub Basin l

7,687 km". The basin spans approximately 150 km fromnorth to south and65 km from east to west. Songkhlaisthe only lake in Thailand that is a large lagoon system:fre sh water fr om pr eci pi teti on drain into the 1eke throughstreams and overland flow and mixes with saline waterfrom the sea. The basin is bounded by two mountainranges. The higher gr ounds of both the m ountain range sare covered with rainfor e sis. Undul ating p 1 ains alternatingwith low hills run parallel to the north-south mountainrange in the basin. The area towards the east approachingthe 1ek e is a large flat plain containing m 0 st1Yr i ce farm s.North of Songkhla Lake is a large wetland that coversappr oxim atel y 137 km". A 1arge fl a t plain lie s east of theS ongkhla basin between the lake and the sea The basinis divide d into 12 sub- basins. The ma jori ty of S ongkhlaLake Basin (SLB) 1and, 5,660 km ~is use d for growing ri c eand rubb er tr e es (60 and 30%, r e sp e c ti ve10. Fore st landoccupies 1,164 km", most of which is rainforest and theremaining is mangrove forest and swamp. Other land usecategori es include natural water bo dy (12.5%), re si denti alarea (2.6%), industrial, man-made water body, roads andundeveloped land. The average rainfall in the basin is2,043 mm.

Current status 0f the lake water q uality: S ongkhla Lakec onsi sts of 4 parts: Thai e N oi, upp er S ongkhl a Lake,middle Songl:dl1a Lake and lower S ongkhl a Lake. Thelatter is connected to the Gulf of Thailand at MuangDistrict, S ongkhl a Pr ovince. 0ne of 2 criteria isnec e ssary for eutr ophi c ati o n of a 1ake: pho sphate must bemore thanO.05 mg L-I or Chlorophyll amust be more than10 ug L-I (Drolc and Koncan, 2002; Smith et al., 1999).Nutrient concentrations in the upper, middle and lowerSongkhl a Lake indic ate a hi gher l eve1 of pho sphate at thelower Song}:hla Lake, whereas Chlorophyll a is higher in

2521


4/13

J.Applied Sci., 9 (14): 2519-2531,2009the upper andmiddle Songkhla Lakes. The U-Tapao canallocated in the lower Songkhla Lake contributes to theheavy loading of nutrients and wastewater to the lake dueto the industries and dense population along the canal.Although, the general water quality in the lake is withinacceptable ranges, eutrophication and nutrient levels inthe lake are suspected to be higher than stated. The areaof most critical concern is the upper lake system,consisting of Thale Noi, Thale Luang and Thale Sap.Cadmium was also found in the lake sediments(Sirinawin and Sompongchaiyakul, 2005). Althoughcadmium is not a concern under the present conditions,future agricultural activity could pose problems; therefore,methods are needed to accurately estimate the nutrientsand metals transported by runoff into the lake.Methods: This study investigates the potential of the SFAmethod to estimate the phosphorus and cadmium inputloads into Songkhla Lake; data will be gathered throughin-depth interviews with fertilizer importers, wholesalers,retailers and farmers and through laboratory testing of soilsamples from all the three provinces in the basin. As afirst step, the SFA was applied to the whole basin as alumped system for evaluating the applicability of themethod. For the purposes of this study, amounts ofphosphorus and cadmium were estimated with andwithout applying SFA and compared. The improvedestimates were subsequently used to model one selectedsub-watershed.

Several methods can be used for improving thenutrient and metal input estimates, including SubstanceFlux Analysis (SFA) and Material Flow Analysis (MFA)developed by Baccini and Brunner (1991), Life CycleAnalysis (LCA) (Tukker, 2002) and Partial Economicequilibrium Analysis (PEA); however, method usedshould be able to determine the problem-causingmechanism in an operational fashion in the sense ofTable 1: Material flow related analysis and associated issues of concern

having a low data demand and being easy to constructand run in practice. SFA was selected in this researchbecause it uses an input-output analysis of material orsubstance compared to LCA, which focuses on thefunction of a product, not on the amount of the product.In addition, SFA tracks the processes within the system,whereas MFA treats the system as a black-box. Drolc andKoncan (2002) developed a phosphorus balance usingMFA in a river basin and evaluated different scenarios forpollution reduction. Montangero and Belevi (2008) alsoused MFA to assess the impact of environmentalsanitation systems on the phosphorus load dischargedinto surfacewater inHanoi, Vietnam. However, the intentof both the studies was not to derive improved estimatesof input to non-point source models. The PEA methodalso tracks the process within the system, but it was notchosen because the method also performs an economicanalysis, which makes the process more complex.Therefore, in this study, SFA is used to refine thepreviously available estimates of loading,which ultimatelyimproves the quality of the major input to non-pointsource pollution transport models. Bouman et al. (2000)evaluated the differences and similarities of the methodsand results of the model in a practical way. Kleijn et al.(2008, 2000) categorized the usage ofmeasurement tools(Table 1)based on the objectives and primary interests oftheir studies. Table 1 clearly shows that SFA is thepreferable tool to estimate the amounts of nutrients andmetals available for transport via runoff. SFA is preferredfor its simplicity and for not treating the system as a blackbox.Substance flux analysis: Substance Flux Analysis (SFA)uses an input-output analysis of substances. The firststep for applying SFA is to develop a conceptual modelof the system under study. The conceptual modeldeveloped for SongkhlaLake Basin ispresented in Fig. 3,

I ssues o fconcern

Speci fi c concerns rela ted to environmental impacts ,supply security, technology developmentWith cer ta in business, economics activiti es , count ri es, regionsAssociated with

Objects of Substances Materialprimary interest

Chemical elements Raw material andor compounds e.g., semi-finished goodsCd, Cl, Pb, Zn, e.g., energy carriers,Hg,N,P, C, metal (ferrous,CO" CFC non-ferrous), sand

and gravel, t imber,plastics

Type of Substance flow Material systemanalysis analysis analysisType of Substance Individualmeasurement flux materialtool accounts flow accountsKleijn et ai. (2002, 2008)

Manufacturedgoodse.g., bat teries , cars,computers,

Life cycleanalysisLifecycleinventories

General environment and economic conc ems rel at edto the throughputOf substances , materi al, manufactur ing goodAt the level ofBusiness Economic

activitiese.g ., product ion sectors ,chemical industry, i ronand steel industry,constructi on

Countries,regionse.g., aggregatemass of materialsand relatedmaterials mix,groups ofmaterial, selectedmaterialsEconomy-wide!\!IF analysisEconomy-widematerial flowaccounts

e.g., firms, comp anies ,pl ant s, med ium sizedand big enterprises,MNEs

Business level!\!IF analysisBusiness materialflowaccounts

Input-outputanalysisPhysicalinput-outputtable

2522


5/13

J. Applied Sci., 9 (14): 2519-2531,2009

Cereals:lI_IIIII~."fruits.vegetables

Fig. 3: Conceptual model of Songkhla Lake Basin system for SFAwhich shows the subsystems, processes and flows ofsubstances. Goods containing phosphorus and cadmiumthat had passed through relevant processes in the SLBover the one-year period of 2004 were examined.Accumulations of phosphorus and cadmium onagricultural soil via., phosphate fertilizer, feed for swine,manure from swine farming and precipitation wasconsidered potentially significant. Phosphorus andcadmium stored in soil can be taken upby plants, releasedinto the air, leached into groundwater and/or washedaway by surface nmoff. Even in predominantlyagricultural areas, only trace amounts of phosphorusbelow detectable limit leach into groundwater. Similarlythe levels of cadmium in lower soil layers are also belowdetection limit. However, nmoff due to rain events can beamain carrier of phosphorus and cadmium transport fromthe system. The only significant pathways leading to thelake are nmoff and direct precipitation on the surface ofthe lake.

Data from both primary and secondary sources needto be compiled to establish a database of sources andamounts of phosphorus and cadmium in the systemproducts and processes (Table 2-4). The non-availabilityof different categories of data for the same periods hasbeen a problem. For example, the concentration ofcadmium inphosphate fertilizer was investigated in 2004,but the corresponding figures for the concentration inanimal feed were only available for the year 2001.Moreover, neither governmental bodies nor researchcenters possess adequate records of the rate of increaseof these substances each year. That such factors areoften obstacles to the application of the SFA approachin most developing nations, compared to developedones, is not unexpected (Baccini and Brunner, 1991;Tangsubkul et al., 2005). The paucity of such data is a

Product/processesTable 2: System parameters and characteris tics

Consumption/quantityChemical fertilizer- 8-24-24,13-13-21, 15-15-15and 16-16-16 formulaNumber of swineFeed for swinePrecipitationAgricul tural soi l areaRunoff

4,965226,3901.51,8805,6914,896xlO '

Unit

Headglheadldaymm year :'km-m3year-1

Sereewatthanachai et al. (2004)Table 3: Phosphorus in p roduct systems and proces sesProduct/processes Content/concentration UnitChemical fertilizerSLC: 8-24-24, 13-13-21,15-15-15 and 16-16-16 formula

Feeds for swineGeneralManure f rom swine farming

PrecipitationAgr icul tural a reas ofAmphoe Hat-Yai , Songkhla

Agri cul tural soil Agr icultural soilMajor soi l in ThailandPeninsular ThailandSLC

Plant cultivationGeneral

RunoffKlong U-TapaoRajjaprabha Dam reservoir0-25m>25 mTapi-Pumduang basin

9,8, 6.4, 6,8 and 7,3,respectively750-800,03

1-76 (ava ilable P)38-1,137 ( tot al P)2-3,5 (avail ab le P)39-238 (tota l P)24-288 (tota l P)5-100,002-0AOND-0,360,14-0,68


6/13

Table 4: Cadmium in product systems and processes

J.Applied Sci., 9 (14): 2519-2531,2009

Product/processes Contents/concentration UnitChemical fertilizerSLC : ( 82424, 1313-21, 1.4,1.4,30.1 and 1.4,15-15-15 and 16-16-16 formula) respectively

Feeds for swine 0.18-0.32Manure from swine fanning 0.32PrecipitationSLC 1.63xl 0-4-2.23 xl 0-3Mauritius ND


7/13

J.Applied Sci., 9 (14): 2519-2531,2009This study attempted to estimate the input amounts

for transport models using SFA. A total of 86 interviewscovering all of the 3provinces including farmers, fertilizerresellers, cooperatives and government officials werecarried out. Thirty-five fertilizer samples were alsocollected from wholesalers, resellers and farmers forlaboratory testing of phosphorus and cadmium amounts.The weak acid (2M HJ"-.JO])eaching method was used toextract metals from the fertilizer samples. Subsequently, aninductively coupled plasma-optical emissionspectroscopy on aPerkin-Elmer ICP-OESOptima 2000DVinstrument was used to determine cadmium levels.Phosphorus is extracted from the soil using Bray 2solution, ascorbic acid (CORS0), as extractant. Table 5indicates fertilizer utilization for each province from eachsource. Table 6 indicates planting area of various crops.Table 7 contains the fertilizer application rates for eachcrop. The resulting data on plantation area and fertilizerutilization are presented in Fig. 4 and 5, respectively.Figure 4 shows planting areas where fertilizer was applied(affected and non-affected areas). Amounts offertilizerTable 5: Fertilizer utilization

Phatthalung (t year:")Fertilizer -------------------------------------------------formula Survey Resellers Co-ops16-8-4 4,402 4,519 5,00015-7-8 4,402 4,912 5,70820-8-20 52,829 53,761 53,94516-20-0 50,623 51,069 52,20925-7-7 92 150 15015-15-15 2,996 3,766 5,0968-24-24 313 465 1,91513-13-21 313 757 1,360Total 115 970 120231 126987

utilization (Kg/rai/year) in the planting area are shown inFig. 5. The entry of phosphorus and cadmium intoagricultural soil was analyzed via primary and secondarydata. The results found that the entry of phosphorus andcadmium into the agricultural soil was potentiallysignificant via., three products and one process line: (1)phosphate fertilizer, (2) feed for swine, (3) manure fromswine farming and (4) precipitation. The entry points andpathways of phosphorus and cadmium enteringagricultural soil are illustrated in the SFA schematic inFig. 6. These entry points are discussed below.Phosphate fertilizer: This study found that approximately384,289 t of chemical fertilizer was applied in the SLBagricultural soil in the year 2004, whereasSereewatthanachai et aZ . (2004) reported 55,307 t in2002. Interview data from SLBwholesalers revealed thatthe fertilizer formulae(N-P-K)used in the areawere 16-8-4,15-7-8,20-8-20,16-20-0,25-7-7, 15-15-15, 18-24-24andl3-l3-21. The main formulae used in paddy fieldsand old rubber farming were 20-8-20 and 16-20-0.

SurveyNakhon Si Thammarat (tyear:")

Co-ops7,7787,77893,33068,2311,71619,5022,9712,971204 277

Songkhla (t year:")Resellers Co-ops Survey8,224 9,718 11,2708,005 9,320 11,27093,622 94,386 135,23668,977 69,909 41,4982,608 3,868 76619,958 21,186 15,6053,392 4,125 13,0593,173 3,505 13,059207959 216017 241763

Resellers11,47411,979135,39141,8531,33615,70413,87413,228244 839

12,94512,206135,86142,9261,55816,73515,24814,600252079

Table 6:Planting area of crops treated with fertilizer application in 3 provincesSub-province Phatthalung (acre) Nakhon Si Thammarat (acre) Songkhla (acre)Rubber 281,753Rice 202,493Mangosteen 6,704Palm oil 732RambutanDurianLonganPamelo

3,6722,8287,520880

497,762272,92530,26413,72825,88817,86011,3797247

721,260165,9941,6056,1301,3525,5805,2742453

Table 7: Fertilizer application rate for each cropApplication rate Spatial distribution Spatial distribution Spatial distribution

Crop Fertilizer formula' (kgIacrelyear) in Phatthalung in Nakhon Si Thammarat in SongkhlaYoung rubber 16-8-4 125 Pabon Tungsong Sadaow6 years) 15-7-8Old rubber 20-8-20 250 Pabon Chaaud SadaowRice 16-20-0 250 Muang, Kwankanoon Haisai Ranote, SatingpraMangosteen 15-15-15 250 TamotePalm oil 25-7-7 125 Bangkaew Tungsong KlonghoikongRambutan 15-15-15 250 Piboon TasalaDurian 15-15-15 250 Piboon TasalaLongan 15-15-15 125 Piboon Tasala8-24-24 125 Piboon Tasala

13-13-21 125 Piboon TasalaPamelo 15-15-15 250 Kauankanoon Hwaisai SinghanakomInterviewing of farmers and fertilizer dealers. 'The formula represents the value ofN-P-K concentration

2525


8/13

J . A, pp li ed SC i. , 9 (14): 2519-2531, 2009

Fig. 4: Planting area

Fig. 5: F ertiliz er uWizati on

Legmd=::JS u b - W ? J l . e I ' 5 h ! dbOU'l .dJJry

R D : r e rF ertilizer e f f e c t ; : ,r e~_ N rn.e ffect~ e~_ E Ifect;:,re~

LegmdoS u b - W ? J l . e I '5 h ! dbOU'l .dJJry

R D : r e rEIfect;:,re~(}(~~)_ N rn. e ffect_ Verybw: 1-1200

Low: 1201-2400_ M edium .: 2401-30)0_ H:$: 30) HmO- Very~: 4801-6000

~ - - - - - - ~ - - - - - - ~ - - - - ~ ~ ~ ~ilt,IiIP

~~/ttf'~

~

Fig. 6: SF A of pho sphorus (diam ond) and cadmi urn (rounded re ctengl e) in SLB (t ye ar-)

2 5 2 6


9/13

J.Applied Sci., 9 (14): 2519-2531,2009Sereewatthanachai et al. (2004) reported 4 specificformulas (8-24-24,13-13-21,15-15-15 and 16-16-16) andnoted that other grades of fertilizers had been used butnot analyzed. This study found phosphorus contents inthe investigated grades were 8.3, 6.5, 7A, 16.8, 6A, 13.2,20.2, 10.1%, respectively; cadmium amounts in phosphatefertilizers (PPs) were 1.5, 1.6,


10/13

J . A, pp li ed SC i. , 9 (14): 2519-2531, 2009

P


11/13

J.Applied Sci., 9 (14): 2519-2531,2009Table 11: Scenario results ofphosphorus in U-Tapao and Eastern Coast Sub Basin 4 Sub-watershedCoordinate Current phosphorus Scenarios_.____________________ loading observed _.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._.._X Y (gg kg-1 dry soil) 1.1 1.2 1.3

124.92195.9994.9872.21528.26

1.4 2.1 3.115.65 15.03 14.4336.96 31.69 27.1765.97 58.96 52.6942.82 42.30 41.7922.52 19.64 17.12140.46 116.51 96.64

652479 781907 24.43 28.52 37.61 21.26649857641771651714654446662859

779751769116760350737358740376

76.05122.0262.8144.35316.22

102.10156.2670.3958.44439.52

60.31100.0955.9635.65244.15

X and Yare coordinates inWorld Geodetic System of1984 (WGS84 system). SBL is inZone 47 in this system

Coordinate Current cadmium-------------------------------- loading observedX Y (mg kg-' dry soil)652479 781907 17.15649857 779751 49.49641771 769116 6.30651714 760350 7.26654446 737358 1.72662859 740376 10.25

ScenariosTable 12: Scenario results of cadmium in U-Tapao and Eastern Coast Sub Basin 4 Sub-watershed

1.1 1.2 1.3 1.4 2.1 3.120.60 22.96 14.29 9.05 7.05 7.1758.62 65.77 41.61 27.34 20.83 20.588.31 8.87 4.92 2.23 2.15 2.728.24 9.44 6.26 4.54 3.27 2.982.35 2.47 1.31 0.49 0.54 0.7512.50 13.84 8.46 5.13 4.10 4.30

X and Y are coordinates inWorld Geodetic System of 1984 (W GS84 system). SBL is in Zone 47 inthis system

organic matter, high erosion, acidity, high clay percentageand high total metal) could promote the transport ofcadmium through the watershed. The scenario resultsdemonstrate that changing the fertilizer type from high-cadmium (3mg Cd kg-I) to low-cadmium (1mg Cd kg")leads to 50-60% reduction in the cadmium load to thelake. However, with other scenarios, changing the typesof crops grown in the area or lowering the fertilizer ratecould further reduce cadmium contribution to the SLB asa whole. The results of this study could be used todevelop an effective approach to minimize cadmiumloadings by changing the phosphate fertilizer type. Thescenario results for phosphorus from this study alsoindicates a potentially favorable scenario of adoptingrubber cultivation instead of horticultural crops. Decisionmakers and/or planners can use this information for cropplanning and/or agricultural extension.

CONCLUSIONDistributed-parameter cell-based models are widely

used in studies related to non-point source pollutiontransport problems. Though these are sophisticatedmodels, results will be reliable only if the data input to thespatially distributed cells are of good quality and faithfulto reality. The lack of a well-established database ofspatially distributed data is a severe problem in manysituations similar to the Songkhla Lake Basin. Methodsused and approximations made in earlier studies onestimating phosphorus and cadmium loads to the Lakewere often crude and could be sources of inaccuracy.Data from early studies should, therefore, be consideredonly as a starting point. The wide range of phosphorus

and cadmium content recorded in the literature suggeststhe need for additional measurements specific to the SLB.The study reported in this study is an attempt to improvethe input data for modeling the phosphorus and cadmiumtransport to the lake via., surface runoff using substanceflux analysis. The parameter values are then used as inputfor the spatially distributed cell-based models,AnnAGNPS and TREX The major phosphorus andcadmium contributing sub-watershed was then testedwith different scenarios to identify the impact of changesin agricultural practices to phosphorus and cadmiumloading. Though, SFA is used in this study to obtainestimates of the input parameters as lumped quantities,the idea is to use this approach on a cell-by-cell basis fordetailed modeling. The study confirms that SFA is aviable method for improving estimates of pollutantamounts available for transport via., runoff, estimates thatcan be used as input to spatially distributed pollutanttransport models.

ACKNOWLEDGMENTSThe authors acknowledge the National Research

Center for Environmental and Hazardous WasteManagement at Chulalongkom University in Bangkok,Thailand, for sponsoring this research and theDepartment of Civil Engineering at North Dakota StateUniversity, USA, for hosting the first author as avisiting scholar during the preparation of this study. Theauthors also acknowledge the help from Mary Pull,Director, Center for Writers, North Dakota StateUniversity, USA, to improve the clarity oflanguage of themanuscript.

2529


12/13

J.Applied Sci., 9 (14): 2519-2531,2009REFERENCES

Baccini, P. and P. Brunner, 1991. Metabolism ofthe Anthrosposphere. Springer, New York,ISBN-lO: 0387537783.Bouman, M., R Heijungs, E. van der Voet, I.C.I.M. Vanden Bergh and G. Huppes, 2000. Material flow andeconomic models: An analytical comparison of SFA,LCA and partial equilibrium models. Ecol. Econ.,32: 195-216.

Chen, E. and D.S. Mackay, 2004. Effects of distribution-based parameter aggregation on a spatiallydistributed agricultural nonpoint source pollutionmodel. I. Hydrol., 295: 211-224.

Chen, M., I. Chen and F. Sun, 2008. Agriculturalphosphorus flow and its environmental impacts inChina. Sci. Total Environ., 405: 140-152.

Curtis, JA, L.E. Flint, e.N. Alpers and S.M.Yarnell, 2005.Conceptual model of sediment processes in theupper Yuba River watershed, Sierra Nevada, CAGeomorphology, 68: 149-166.

Doberl, G., R Huber, P.H. Brunner, M. Eder, R Peirrard,W. Schonback, W. Fruhwirth and H. Hutterer, 2002.Long term assessment of waste management options:A new integrated and goal-oriented approach. WasteManage. Res., 20: 311-327.

Drolc, A and I.Z. Koncan, 2002. Estimation of sources oftotal phosphorus in a river basin and assessment ofalternatives for river pollution reduction. Environ.Int., 28: 393-400.

Greenwood, D.I., D.A Stone and T.V. Karpinets, 2001.Dynamic model for the effects of soil P and fertiliserP on crop growth, P uptake and soil P on arablecropping: Experimental test of the model for fieldvegetables. Ann. Bot. Lond., 88: 293-306.

Hug, F. and P. Baccini, 2002. Physiological interactionsbetween highland and lowland regions in the contextof long-term resource management. Mountain Res.Dev.,22: 168-176.

Kleijn, R, R Huele and E. van der Voet, 2000. Dynamicsubstance flow analysis: The delaying mechanism ofstocks, with the case of PVC in Sweden. Ecol. Econ.,32: 241-254.

Kleijn, R., E. van derVoet andHA U. deHaes, 2008. Theneed for combining IEA and IE tools: The potentialeffects of a global ban on PVC on climate change.Ecol. Econ., 65: 266-281.

Kwonpongsagoon, S., D.T. Waite, S.J. Moore andP.H.Brunner, 2007a. A substance flow analysis in thesouthern hemisphere: Cadmium in the Australianeconomy. Clean Technol.Environ. Policy, 9: 175-187.

Kwonpongsagoon, S., H.P. Bader and R Scheidegger,2007b.Modelling cadmium flows inAustralia on thebasis of a substance flow analysis. Clean Technol.Environ. Policy, 9: 313-323.

Liu, Y., J. Chen and AP.J. Mol, 2004. Evaluation ofphosphorus flows in the dianchi watershed,Southwest of China. PopuloEnviron., 25: 637-656.

Lofts, S., 2007. Critical loads of metals and other traceelements to terrestrial environments. Environ. Sci.Technol., 41: 6326-6331.

Montangero, A and H. Belevi, 2008. An approach tooptimise nutrient management in environmentalsanitation systems despite limited data. I. Environ.Manage., 88: 1538-1551.

Oudin, L., e. Perrin, T. Mathevet, V. Andreassian ande. Michel, 2006. Impact of biased and randomlycorrupted inputs on the efficiency and the parametersofwatershed models. I. Hydrol., 320: 62-83.

Polyakov, V., A Fares, D. Kubo, I. Jacobi and C. Smith,2007. Evaluation of a non-point source pollutionmodel, AnnAGNPS, in a tropical watershed. Environ.Modell. Softw., 22: 1617-1627.

Radcliffe, I.S., RS. Pleasant and E.T. Kornegay, 2006.Estimating equivalency values of microbial phytasefor amino acids in growing and finishing pigs fittedwith steered ileo-cecal valve cannulas. I. Anim. Sci.,84: 1119-1129.

Sae-Eong, T., C. Tjell, H. Mosbaek andP. Sompongchaiyakul, 2002. Metals concentration inSongkhla lake basin's soils. Proceedings of the 28thCongress on Science and Technology of Thailand,Oct. 24-26, IEEE, pp: 1-1.

Sereewatthanachai, W., C. Ratanachai,P. Sompongchaiyakul and P. Sangganjanavanich,2004. Preliminary study of substance flows ofphosphorus and cadmium in the agricultural soilof songkhla lake catchment. Thai Environ. Eng.I., 18: 9-19.

Sirinawin,W. andP. Sompongchaiyakul, 2005.Nondetritaland total metal distribution in core sediments fromtheU-Tapao Canal, Songkhla, Thailand. Mar. Chem.,94: 5-16.

Smith, v.n., G.D. Tilman and r.c Nekola, 1999.Eutrophication: Impacts of excess nutrient inputs onfreshwater, marine and terrestrial ecosystems.Environ. Pollut., 100: 179-196.

Tangsubkul, N., S. Moore and T.D. Waite, 2005.Incorporating phosphorus managementconsiderations into wastewater management practice.Environ. Sci. Policy, 8: 1-5.

Tiemeyer,B.,R. Moussa, B.Lennartz andM. Voltz, 2007.MHYDAS-DRAIN: A spatially distributed model forsmall, artificially drained lowland catchments. Ecol.Model., 209: 2-20.

2530


13/13

J.Applied Sci., 9 (14): 2519-2531,2009Tukker, A, 2002. Life-cycle assessment and the

precautionary principle. Environ. Sci. Teclmol.,36: 70A-75A

Velleux, M.L., P.Y Julien, R. Rojas-Sanchez,W.H. Clements and IF. England, 2006. Simulation ofmetals transport and toxicity at a mine-impactedwatershed: California Gulch, Colorado. Environ. Sci.Technol., 40: 6996-7004.

Velleux,M.L., IF. England and P.Y Julien, 2008. TREX:Spatially distributed model to assess watershedcontaminant transport and fate. Sci. Total Environ.,404: 113-128.

Yuan, Y, MA. Locke andR.L. Bingner, 2008. Armualizedagricultural non-point Source model application forMississippi Delta Beasley Lake watershedconservation practices assessment. I Soil WaterConservat., 63: 542-551.Montangero, A and H. Belevi, 2008. An approach tooptimise nutrient management in environmentalsanitation systems despite limited data. I Environ.Manage., 88: 1538-1551.

2531

Date post:	08-Apr-2018
Category:	Documents
Upload:	kitipan-kitbamroong
View:	219 times
Download:	0 times

Improving Non-Point Source Pollution Model Input Parameters Using Substance Flux Analysis

Documents