Dominant sources of dietary carbon and nitrogen for shrimp reared in extensive rice-shrimp ponds

Dominant sources of dietary carbon and nitrogen for

shrimp reared in extensive rice-shrimp ponds

Michele A Burford1, Nigel P Preston1,Truong Hoang Minh2,TranThi Tuyet Hoa2, Stuart E Bunn3 &Vanessa M Fry3

1CSIROMarine Research, Cleveland, Queensland, Australia2CanTho University, CanTho,Vietnam3Centre for Riverine Landscapes, Gri⁄th University, Nathan, Queensland, Australia

Correspondence:M Burford, Centre for Riverine Landscapes, Gri⁄th University, Nathan, Queensland 4111, Australia.

E-mail: m.burford@gri⁄th.edu.au

Abstract

Stable isotope analysis was used to determine thesources of dietary nitrogen (N) and carbon (C) forshrimp during the rearing phase in extensive rice-shrimp ponds in MyXuyen and Gia Rai districts,Viet-nam. Farm-made feed was added as a food source inshrimp ponds in My Xuyen district, and based onstable isotope analyses, was generally a poor dietarysource. The commercial formulated feed used in GiaRai also appeared to contribute little directly to thenutritional needs of the shrimp. In contrast, the nat-ural biota in all ponds appeared to contribute sub-stantially. In particular, biota from beam trawls andbenthic organic matter were the most likely sourcesof nutrition in MyXuyen ponds, while benthic organ-ic matter was the main source in Gia Rai ponds. d15Nratios in the natural biota in My Xuyen farms de-creased over the growing season, suggesting in-creased N ¢xation in the case of the benthic organicmatter reaching values as lowas1m.This suggests N-limitation in the ponds and that natural biota becomeincreasingly dependent on N ¢xed by algae and/orother microorganisms. There is the potential to pro-mote the growthof the planktonand hence, the othernatural biota, by the judicious addition of fertilizer.

Keywords: shrimp, nutrition, natural biota, nat-ural feed, carbon, nitrogen

Introduction

In saline a¡ected areas of theMekong Delta,Vietnam,the traditional wet-season rice crop is often supple-mented with a dry-season crop of extensively farmed

shrimp when salt water intrudes into the waterways(Tran, Dung & Brennan 1999). The adoption ofshrimp as a second crop in the dry season has re-sulted in signi¢cant income gains for some farmers(Tran et al.1999; Brennan, Clayton & Tran 2000). In2000, 10000 t of shrimp were produced from40000ha of rice-shrimp farms in Vietnam (Bren-nan, Preston, Clayton & Tran 2002).In the rice-shrimp system, postlarval shrimp are

recruited into the ponds through water exchangefrom nearby canals. Additionally, ponds are oftenstocked with hatchery-reared Penaeus monodon (Fab-ricius). Ponds contain one or more platforms forgrowing rice, surrounded by a deeper channel forrearing shrimp. Farmers use a variety of di¡erentfeeding practices and diets for the shrimp that rangefrom relying on the natural biota as the only foodsource, to feeding with farm-made formulations, orfeeding with more expensive commercial feeds. A re-cent economic analysis of representative rice-shrimpfarms has indicated that the addition of farm-madefeeds, containing the main ingredients of rice and ri-cebran, has had little impact onproduction (Brennanet al.2000). It may be that the dietary requirements ofthe shrimp are being met by natural pond biotarather than the farm-made feed; however, the role ofnatural biota is poorly understood.Measurement of stable isotope ratios in food

sources and consumers may be used to determinethe importance of di¡erent food sources to prawn nu-trition (Gearing 1991; Shearer & Kohl 1993). d15N-ni-trogen and d13C-carbon ratios are most commonlyused. The stable isotope signature of a consumer re-£ects the isotopic signatures of material assimilated

Aquaculture Research, 2004,35,194^203

194 r 2004 Blackwell Publishing Ltd

and provides an integration of feeding over time (Pe-terson & Fry1987). Stable isotope analysis has provedparticularly e¡ective in the study of aquatic foodwebs where there are often marked di¡erences inthe isotope signatures of the major primary sources(Boon & Bunn 1994; Loneragan, Bunn & Kellaway1997).In this study, we used stable isotope analysis to de-

termine the sources of dietary N and C for shrimpgrown in rice-shrimp ponds, particularly comparingthe role of the farm-made and commercial feeds, andthe natural biota.

Materials and methods

Study sites

The study was conducted at three rice-shrimp farms,eachwith one pond, in the My Xuyen district in1998,and three rice-shrimp farms, each with one pond, inthe Gia Rai district of the Mekong Delta,Vietnam, in2000 (Fig. 1). The size of the ponds varied from 1 to1.4 ha (Table1), and all farms in theMyXuyen districthad a central platform area (80% of the total pondarea, water depth approximately 20 cm), where awet season rice crop was grown prior to a dry seasonshrimp crop, and a ditch around the perimeter of theplatform (water depth approximately1m). At Gia Rai,ponds had four platforms surrounded by deeperwater.

Prior to stocking, ponds were prepared by removalof sedimentedmaterial in the ditches, drying the plat-forms, liming the entire pond and killing predatorswith rotenone. The ponds were stocked with hatch-ery reared P. monodon postlarvae (mean total length,18mm) at stocking densities between 1 and 3 ani-mals m�2 (Table 1). Other shrimp species (Penaeusmerguiensis (de Man)) were also present in someponds in lesser numbers, having been recruited fromthe river during routine water exchanges. Theseshrimp are referred to as ‘wild shrimp’. After a periodof 9^16 weeks, all shrimp were harvested fromponds, and total weights for each pondwere recorded(Table1).There were two main types of feed added to the

ponds: commercial and farm-made. The farm-madefeed was made by combining ¢shmeal and ricebran,mixing in cooked rice to form a paste, then forcingthe feed through a mincer. The extruded paste wasair-dried. Later in the season, cornmeal was alsoadded to this feed prior to mixing with cooked riceand forming a paste. Fishmeal was only used in thefeed at the start of the season. The quality and typeof feed varied from pond to pond (Table1). There weretwo brands of commercial formulated feeds; one usedin My Xuyen farms (CF1) and another used at Gia Raiin two of the three farms (CF2). One of the ponds atGia Rai had no added feeds. In the My Xuyen farms,CF1 was used only for a short period early in theseason.

Figure 1 Map of the study sites in the Mekong Delta,Vietnam. Farms were in the My Xuyen and Gia Rai regions.

Aquaculture Research, 2004, 35, 194^203 Sources of dietary carbon and nitrogen for cultured shrimp MABurford et al.

r 2004 Blackwell Publishing Ltd, Aquaculture Research, 35, 194^203 195

Sampling and analysis

Ponds were sampled at three occasions during theshrimp growth season: one (early), two (mid) andthree (late) months after stocking in My Xuyenponds; and one (early) and two (mid) months afterstocking in Gia Rai ponds. After this time the trial atGia Rai was terminated due to an outbreak of WhiteSpot Syndrome Virus in the shrimp. Triplicate sam-ples were taken of the natural biota, feed and shrimpin each of the three ponds early, mid and late seasonsfor stable isotope analysis. The seston was sampledusing plankton net (mesh size 100 mm) towed 20min the ditch. The tow was repeated twice. The com-bined samples from the tows were washed from thenet into containers and stored on ice until returnedto the laboratory. Triplicate samples were taken.Beam trawl samples were taken by towing a trawl(mesh size 1000 mm) across both the ditch and theplatform.The tow was repeated twice. The biota fromboth tows were picked out and stored in a containeron ice. Triplicate samples were taken. Samples ofbenthic organic matter were taken with a core fromthe sediments at three sites in the ditch and com-bined, and at four sites on the platform and com-bined. Triplicate samples of sediment from both theditch and the platformwere taken. Samples were alsotaken of the commercial feeds, farm-made feeds, P.

monodon and the wild shrimp from each farm for iso-tope analysis.All feed, shrimp and biota samples were then dried

at 60 1C in the laboratory, ground with a mortar andpestle, and stored until analysed for stable nitrogenand carbon isotope ratios, and Nand C content usingan elemental analyser (Eurovector 3000) and massspectrometer (Micromass Isoprime). Three replicatesamples were analysed for shrimp and biota. Ratiosof 13C/12C and 15N/14N were expressed as the relativeper mil (m) di¡erence between the sample and con-ventional standards (PDB carbonate and N2 in air)

dX ¼ ðRsample=Rstandard � 1Þ � 1000 ðmÞ

where X513C or 15Nand R513C/12C or15N/14N (Pe-terson & Fry1987).Measurement precision was approximately 0.1m

and 0.2m for 13C/12C and 15N/14N respectively.Water quality parameters were monitored in all

ponds. Temperature and salinity were monitored atthree sites on the platform and at three sites in theditch early in the morning and mid-afternoon eachday using a datalogger (Yeokal). In the case of theditch, readings were taken at the top and bottomof the water column.Water samples were also takenfor chlorophyll a and total suspended solids (TSS)analysis at the same time as stable isotope samples.

Table 1 Shrimpand feed data for farms inMyXuyenandGia Rai districts. Shrimpharvest includes Penaeusmonodonandwildshrimp

My Xuyen Gia Rai

Farm1 Farm 2 Farm 3 Farm 4 Farm 5 Farm 6

Pond size (ha) 1 1.3 1 1 1.2 1.4

P. monodon stocking 1.6 1.6 1.6 3 3 1

density (m� 2)

Grow-out period (weeks) 16 16 16 9 9 9

Shrimp harvest (kg ha� 1) 340 400 440 19 160 130

P. monodon survival (%) 83 87 94 10 23 57

FCR 2.1 1.7 2.2 5.4 1.3 �Feed

Commercial feed (kg)

CF1 67 67 75 � � �CF2 � � � 102 206 �

Total added 67 67 75 102 206 �Ingredients in farm-made feed (kg)

Rice 400 400 300 � � �Ricebran 125 125 150 � � �Cornmeal 50 50 27 � � �Fishmeal 60 60 400 � � �

Total added 635 635 877

FCR5 food conversion ratio.

Sources of dietary carbon and nitrogen for cultured shrimp MABurford et al. Aquaculture Research, 2004, 35, 194^203

196 r 2004 Blackwell Publishing Ltd, Aquaculture Research, 35, 194^203

Water for chlorophyll a analysis was ¢ltered throughglass ¢bre ¢lters (GF/F), extracted in acetone andmeasured spectrophotometrically (Je¡rey & Welsh-meyer1997). For TSS analysis, water was ¢ltered ontopre-weighed, pre-combusted glass ¢bre ¢lters (GF/F),dried at 60 1C for 24 h, then reweighed at room tem-perature.

Results

The shrimp survival and harvest values were similarin all three My Xuyen ponds with between 340 and440 kg (mean individual weight 25.6 g) being har-vested from each farm (Table 1). All three farms usedfarm-made feeds with rice, ricebran, cornmeal and¢shmeal but the ratio of ingredients and totalamount added varied from farm to farm. Fishmealwas only used in the ¢rst month, cornmeal only to-wards the end of the season. Commercial feed (CF1)was only added early in the season. Food conversionratios (FCRs), calculated from the addition of farm-made and commercial feeds, ranged from1.7 to 2.2.In contrast, shrimp survival and harvest values for

Gia Rai farms were lower (19^160 kg, mean indivi-dual weight 6.0, 23.0 and 22.6 g for farms 4, 5 and 6respectively) despite the higher stocking densities intwo of the three ponds. The low values were due toan outbreak of white spot syndrome virus in theshrimp resulting in ponds being harvested midwaythrough the grow-out season (N. Preston, pers.comm.). Prior to the viral outbreak, shrimp growthrates for the My Xuyen and Gia Rai farms were simi-lar, that is, 1.6 and 2.0 gweek�1 respectively. Nofarm-made feed was used in the Gia Rai farms but acommercial feed (CF2) was added to two of the threefarmswith FCRs ranging from1.3 to 5.4. Farm 6, withthe highest shrimp survival, had no feed additionandshrimp were reliant on the natural biota.

Temperature and salinity in My Xuyen ponds ran-ged from 24.9 to 40.0 1C, and 4.8 to17.5 respectively(Table 2). In Gia Rai ponds, temperatures were lowerand the salinity was higher (23.3^33.0 1C, and9.1^29.9 respectively). Chlorophyll a concentrationswere generally higher in Gia Rai ponds than MyXuyen ponds (6^46 and 1^29 mg L�1 respectively)(Table 2). TSS concentrations were also higher in GiaRai ponds than My Xuyen ponds (78.6^380.2 and42.6^135.6mg L�1respectively).Carbon:nitrogen (C:N) ratios in the seston and bio-

ta from beam trawls were above Red¢eld (1958) ratiosand higher than the shrimp in the My Xuyen pondsearly in the growth season (Fig. 2). By mid-season,the C:N ratios of beam trawl and seston samples haddecreased, becoming closer to the C:N ratios ofshrimp. The C:N ratios of biota from beam trawls re-mained low late in the growth season. In the Gia Raiponds, the C:N ratios of the natural biota were lowand similar to the shrimp throughout the growthseason (Fig. 2).There was little di¡erence in the d13C and d15N ra-

tios of the potential food sources and shrimpbetweenthe three ponds at My Xuyen. Therefore, the datafrom all ponds were combined for each growth stagefor ease of comparison between growth stages andfarming districts. Early in the growth season thed13C and d15N ratios of the beam trawl biota and thefarm-made feed were similar to those of the shrimpwhile the isotope signatures of seston and the sedi-ment were higher in the case of d15N and lower inthe case of d13C (Fig. 3). Di¡erences in the formula-tions of the farm-made feed between farms are re-£ected in the d13C and d15N ratios. By mid-season,the farm-made feed and seston had lower d13C ratioscomparedwith the shrimpwhile the beam trawl, ¢la-mentous algae and benthic organic matter had simi-lar ratios. d15N ratios for these food sources were alllower by 3^7m than for the shrimp. The commercial

Table 2 Ranges in temperature (1C), salinity, chlorophyll a (mg L�1) and total suspended solids (TSS) (mg L�1) concentrationsat three farms in the My Xuyen district (1997/1998) and three farms in the Gia Rai district (2000)

Temperature (1C) Salinity Chlorophyll a (lg L� 1) TSS (lg L� 1)

My Xuyen

Farm 1 24.9–38.6 5.9–17.4 3–14 42.6–104.7

Farm 2 25.0–40.0 6.0–17.5 1–29 51.7–91.2

Farm 3 26.6–39.9 4.8–16.2 3–19 44.8–135.6

Gia Rai

Farm 1 23.6–33.0 9.2–26.9 20–46 131.1–380.2

Farm 2 23.3–32.2 14.1–29.9 7–29 81.6–192.4

Farm 3 23.8–32.4 9.1–28.7 6–41 78.6–209.3

Aquaculture Research, 2004, 35, 194^203 Sources of dietary carbon and nitrogen for cultured shrimp MABurford et al.

r 2004 Blackwell Publishing Ltd, Aquaculture Research, 35, 194^203 197

feed, CF1 had isotopic signatures similar to theshrimp, although it was only added early in thegrowth season.The d13C and d15N ratios of the poten-tial food sources and shrimp remained much thesame late in the growth season. The %nitrogen inthe farm-made feeds changed over the season from6.2% early in the season, whichwas comparablewiththe commercial feed CF1 (6.3%), decreasing to 3.2%by mid-season and 2.5% by the end of the season.Early in the growth season in Gia Rai ponds, the

benthic organic matter and seston had d13C ratios si-milar to the shrimp while the commercial feed, CF2and biota from beam trawls had higher d13C ratios(Fig. 3). By mid-season, there was little change withthe exception of an increase in the d13C ratio ofseston. d15N ratios of benthic organic matter were2^3m lower than for shrimp. The %nitrogen in CF2feed was 6.3%.If a primary source of carbon is important to the

aquatic food web, then the spatial and temporalvariability in d13C values of consumers should track

the observed variability of the source. The d13C ratiosof P. monodon and wild shrimp were compared withthe d13C ratios of potential food sources for all farms(Fig. 4). There was a signi¢cant correlation betweenthe d13C ratios of the shrimp and the biota frombeamtrawls (Fig.4a, R250.58, Po0.001). There was also asigni¢cant correlation between benthic organic mat-ter and shrimp (Fig. 4c, R250.86, Po0.001). In con-trast, there was no correlation between the addedfeed sources or seston and the shrimp (Fig.4b and d).There was a trend of decreasing d15N ratios in the

pond biota over the grow-out season in both MyXuyen and Gia Rai ponds (Fig. 5a and b). This wasmost pronounced in the benthic matter where valuesdecreased to between1m and 2m.

Discussion

Farm-made feed used in My Xuyen ponds was highlyvariable in quality and composition and, in general,

Figure 2 Percent carbon and nitrogen in shrimp and natural biota in (a) early, (b) middle and (c) late growth seasonponds in My Xuyen district, and (d) combined early and middle growth season ponds in Gia Rai district inVietnam. Solidline is the Red¢eld (1958) ratio. BOM5 benthic organic matter.

Sources of dietary carbon and nitrogen for cultured shrimp MABurford et al. Aquaculture Research, 2004, 35, 194^203

198 r 2004 Blackwell Publishing Ltd, Aquaculture Research, 35, 194^203

had a su⁄ciently di¡erent d13C ratio from the shrimpto suggest that it was not an important source of car-bon for the shrimp. Additionally, the d13C and d15Nratios of the shrimp did not change markedly duringthe season, despite the large £uctuations in ratios inthe farm-made feed, and the reduction in the %nitro-gen in the feed below levels generally considered ne-cessary to meet the protein requirements of theshrimp (Guillaume 1997). This reduction is likely tobe due to the removal of ¢shmeal, which has a highprotein level, from the ingredients early in the sea-son. Similarly, the commercial feed used in Gia Raiponds did not appear to be an important nutritionalsource since the shrimp were too d13C and d15N de-pleted compared with the feed. At the farm with nofeed addition, there was no obvious di¡erence ind13C and d15N ratios. Given this ¢nding, it is not sur-prising that an earlier economic analysis of rice-shrimp ponds found that at this low level of stockingdensity, farm-made feeds had little impact on shrimpproduction (Brennan et al.2000).Much of the organic matter added as farm-made or

commercial feed is likely to have entered a microbialdead-end (Lewis, Hamilton, Rodr|¤ guez, Saunders III& Lasi 2001) or have been discharged into adjacent

waterways during water exchange episodes. Promo-tion of high bacterial respiration may lead to low dis-solved oxygen in ponds with little direct transfer ofbiomass to higher order consumers. A previous studyin extensive shrimp ponds showed that a high pro-portion of the N and C inputs were ultimately dis-charged from ponds (Alongi, Johnston & Xuan2000). Eutrophication of coastal waters caused byshrimp farming is of increasing concern globallyand has the capacity to promote algal blooms, causehypoxia, reduce biodiversity and negatively impacton ¢sh nursery sites (Dierberg & Kiattisimkul 1996;Pa¤ ez-Osuna 2001). The use of farm-made or poor-quality commercial feeds is, therefore, an ine⁄cientfarming practice and a potential contributor to coast-al eutrophication.The similarity of the d13C isotopic ratios of the

shrimp, benthic organic matter and biota from beamtrawls, suggests that the natural biotawere an impor-tant carbon source for the shrimp in all My Xuyenponds. Assuminga 3m increase in the d15N ratiowithtrophic level (Shearer & Kohl 1993), benthic matterand biota from beam trawls appear to be the main ni-trogen sources for shrimp nutrition. Filamentous al-gae may also have contributed to shrimp nutrition,

My Xuyen

Middle

P.monodonfarm-madefeed

beam trawl

wild shrimp

CF1

algae

Late

P.monodonfarm-madefeed beam trawlseston

wild shrimp

Early

δ13C (0/00)

δ15C

(0/ 00

)

P.monodon

beam trawlwild shrimp

CF2

Gia Rai

Middle

P.monodonbeam trawl

wild shrimpCF2

10

12 Early

2

0

4

6

8

10

12

2

0

4

6

8

10

12

2

0

4

6

8

10

12

2

0

4

6

8

10

12

2

0

4

6

8

−30 −28 −26 −24 −22 −20 −18

−30 −28 −26 −24 −22 −20 −18 −30 −28 −26 −24 −22 −20 −18

−30 −28 −26 −24 −22 −20 −18 −30 −28 −26 −24 −22 −20 −18

P.monodon

farm-madefeed2

farm-madefeed1

beam trawl

seston BOM

BOM

seston

BOM

BOM

seston

BOM

seston

(a)

(a) (b)

(b) (c)

Figure 3 d13C and d15N values (m) of shrimp, feed and natural biota in (a) early, (b) middle and (c) late growth seasonponds in My Xuyen district, and (a) early and (b) middle growth season ponds in Gia Rai district in Vietnam. BOM5

benthic organic matter.

Aquaculture Research, 2004, 35, 194^203 Sources of dietary carbon and nitrogen for cultured shrimp MABurford et al.

r 2004 Blackwell Publishing Ltd, Aquaculture Research, 35, 194^203 199

based on the d13C and d15N ratios in mid-growth sea-son, while seston did not appear to contribute sub-stantially to the Nand C requirements of the shrimp.In contrast, shrimp in Gia Rai ponds weremore13C

depleted than those in My Xuyen ponds and shrimpappeared to be deriving carbon from the benthicmat-ter but not signi¢cantly from the beam trawl biota orseston. Assuming a 3m increase in the d15N ratio

with trophic level, the benthic organic matter mayhave provided at least some of the nitrogen require-ments of the shrimp. Other studies have shown thatpostlarval and juvenile shrimp derive much of theirnutrition from benthic sources (Newell, Marshall, Sa-sekumar & Chong 1995; Dittel, Epifanio, Cifuentes &Kirchman 1997). Studies in extensive shrimp pondshave shown that the zooplankton and macrobenthos

benthic organic matter

−30

−28

−26

−24

−22

−20

−30

−28

−26

−24

−22

−20

−30

−28

−26

−24

−22

−20

−30

−28

−26

−24

−22

−20

−30 −28 −26 −24 −22 −20 −18 −30 −28 −26 −24 −22 −20 −18

−30 −28 −26 −24 −22 −20 −18−30 −28 −26 −24 −22 −20 −18

beam trawl feed

seston

δ13C of source (0/00)

δ13 C

of s

hrim

p (0

/ 00)

My Xuyen – Wild shrimp

My Xuyen-P.monodon O

R2= 0.58

R2= 0.15

R2= 0.09

P < 0.001

R2= 0.86

P < 0.001

ns

ns

Gia Rai – Wild shrimp

Gia Rai-P.monodon

(a) (b)

(c) (d)

Figure 4 A comparison of d13C values (m) of shrimp and food sources: (a) beam trawl; (b) farm-made and commercialfeeds; (c) benthic organic matter; and (d) seston in all ponds in My Xuyen and Gia Rai districts.

Vietnam time scale data

δ15 N

(0/

00)

My Xuyen

0

2

4

6

8

10

12

Early Middle Late0

2

4

6

8

10

Early Middle Late

Gai Rai

P.monodonWild shrimpBeam trawlSestonBOM

(a) (b)

Figure 5 d15N values (m) of the natural biota and shrimp in ponds: early; mid; and late growth seasons in (a) My Xuyenand (b) Gia Rai districts. BOM5 benthic organic matter.

Sources of dietary carbon and nitrogen for cultured shrimp MABurford et al. Aquaculture Research, 2004, 35, 194^203

200 r 2004 Blackwell Publishing Ltd, Aquaculture Research, 35, 194^203

densities can be su⁄cient to support low stockingdensities of shrimp (Johnston, Lourey, Tien, Luu &Xuan 2002).The highly turbid waters, as indicated by the high

TSS concentrations, are likely to have limited thegrowth of primary producers in the rice-shrimpponds, particularly in the ditches. This is re£ected inthe lowchlorophyll a concentrations in the pond. Stu-dies in extensive shrimp ponds have found low pri-mary productivity, as a result of the high turbidityreducing light availability (Alongi, Dixon, Johnston,Tien & Xuan1999; Johnston et al.2002).Turbid waterenters the pond system during water exchanges andthe resulting build-up of sediments is environmen-tally and ecologically unsustainable (Brennan et al.2002). However, reducing the turbidity of the watermay result in other water quality problems, for exam-ple, excessive growth of benthic ¢lamentous algaecan cause hypoxia and fouling the gills of shrimpthereby negatively a¡ecting shrimp growth andsurvival.The decrease in d15N ratios in the natural biota

throughout the season in My Xuyen ponds suggeststhat N ¢xation is an important process in providing Nto the natural biota. In the case of the shrimp N iso-tope signatures, the observed decline is counter to an

expected increase in d15N ratios as shrimp grow andfeed at a higher trophic position (Peterson & Fry1987). d15N ratios in the benthic organic matter atthe end of the season were close to 1m, consistentwith the presence of N ¢xing algae (Yoshioka,Wada& Hayashi 1994). This contrasts with extensiveshrimp ponds where N ¢xation rates were low (Alon-gi et al.2000). However, these ponds lacked a centralplatform with high light availability and rice plantdetritus creating a high surface area for epiphyticgrowth. The addition of N fertilizer early in thegrowth seasonmay therefore be bene¢cial in promot-ing the growth of the natural biota. However, caremust be taken to ensure that over-fertilizing doesnot occur as this will increase coastal eutrophicationand may promote blooms of nuisance algae, for ex-ample, ¢lamentous algae, in the ponds.The ¢ndings from this studycan be summarized in

a conceptual model of the major pathways of carbonand nitrogen £ow that highlights the likely di¡er-ences between shrimp ponds in the two districts(Fig.6). In MyXuyen, biota from beam trawls contrib-uted directly to shrimp nutrition, while benthic or-ganic matter and ¢lamentous algae contributedboth directly and indirectly. Farm-made feeds did notcontribute signi¢cantly to shrimp nutrition and the

My Xuyen ponds

Gia Raiponds

Beam trawl

Shrimp

AlgaeBOM

*13C-depleted component

*?Beam trawlHomemadefeed

Shrimp

AlgaeBOM

N2 N2

CO2/CH4 CO2/CH4

Microbialdead end

Microbialdead end

Commercialfeed

Commercialfeed

N N

Figure 6 Conceptual diagram of the £ow of nitrogen and carbon through ponds in My Xuyen and Gia Rai districts.BOM5 benthic organic matter.

Aquaculture Research, 2004, 35, 194^203 Sources of dietary carbon and nitrogen for cultured shrimp MABurford et al.

r 2004 Blackwell Publishing Ltd, Aquaculture Research, 35, 194^203 201

main source of nitrogen was from ¢xation of atmo-spheric nitrogen. In Gia Rai ponds, benthic matterand ¢lamentous algae contributed more directly toshrimp nutrition than in My Xuyen ponds with lesscontribution from beam trawl biota. Commercialfeeds did not contribute directly to shrimp nutrition;however, they may have provided a nutritionalsource for beam trawl biota, based on the similarityof the 13C ratios for the commercial feed and biota.As in My Xuyen ponds, much of the nitrogen wassupplied by nitrogen ¢xers. It is not clear to what de-gree di¡erences inwater quality parameters betweenthe two districts, namely temperature, salinity andTSS, a¡ected the £ows of nitrogen and carbonwithinthe pond systems, or shrimp production.In conclusion, our study has shown that there is

scope to improve dietary nitrogen and carbonsources for shrimp in rice-shrimp ponds. The domi-nant role of the natural biota in these low intensitysystems may be further enhanced with the judiciousaddition of fertilizers to stimulate productivity. Fu-ture work is warranted to ascertain the trophic path-ways of Nand C to determinewhere the natural biotaare gaining their nutrition, and ultimately the sourceof nutrients for the shrimp. Additionally, this studydid not quantify the d15N and d13C ratios in the phy-toplankton or algae on the rice stalks or ditch walls.They may also be signi¢cant sites of algal productionthat ultimately underpin the food web supportingshrimp. The bene¢ts of the added feed were highlyvariable and seem to be dependent on the feed formu-lation. As farmers move towards intensi¢cation, theuse of higher-quality feeds may improve yields aswell as minimize nutrient wastes that are ultimatelydischarged into coastal waterways. It is likely that thewater stabilityof the air-dried feeds was a key issue inthe poor utilization of feed. Additionally, the substan-tial reduction in %nitrogen in the feed over the sea-son suggests that it would not supply the proteinrequirements of the shrimp. These results coupledwith those of Brennan et al. (2000) suggest that thereis little economic value in the production of low-qual-ity farm-made feeds for extensive rice-shrimp farm-ing inVietnam.

Acknowledgments

We wish to thank the farmers in My Xuyen and GiaRai districts,Vietnam, who took part in the study. Da-nielle Johnston and Frank Coman provided construc-tive criticism of this paper. This work was funded by

the Australian Centre for International AgriculturalResearch (ACIAR) and CSIRO Marine Research.

References

Alongi D.M., Dixon P., Johnston D.J., Tuan D.V. & Xuan T.T.(1999) Pelagic processes in extensive shrimp ponds of theMekong delta,Vietnam. Aquaculture175,121^141.

Alongi D.M., Johnston D.J. & Xuan T.T. (2000) C and N bud-gets in shrimp ponds of extensive mixed shrimp-man-grove forestry farms in the Mekong delta, Vietnam.Aquaculture Research 31,387^399.

Boon P.I. & Bunn S.E. (1994) Variations in the stable-isotopecomposition of aquatic plants and their implication forfood-web analysis. Aquatic Botany 48,99^108.

Brennan D.C., Clayton H. & TranT.B. (2000) Economic char-acteristics of rice shrimp farms in the Mekong Delta,Viet-nam. Aquaculture Economics andManagement 4,1^13.

Brennan D., Preston N., Clayton H. & Tran T.B. (2002) Anevaluation of rice-shrimp farming systems in the mekongdelta. Report prepared under the World Bank, NACA,WWFand FAO Consortium Program on Shrimp Farmingand the Environment.Work in Progress for Public Discus-sion. Published by the Consortium. http://192.150.250.134/ShrimpWebsite/Case/Vietnam/FinalVietnam.pdf.

Dierberg F.E. & Kiattisimkul W. (1996) Issues, impacts, andimplications of shrimp aquaculture in Thailand. Environ-mental Management 20,649^666.

Dittel A.I., Epifanio C.E., Cifuentes L.A. & Kirchman D.L.(1997) C and N sources for shrimp postlarvae fed naturaldiets from a tropical mangrove system. Estuarine, Coastaland Shelf Science 45,629^637.

Gearing J.N. (1991) The study of diet and trophic relation-ships through natural abundance 13C. In: Carbon IsotopeTechniques (ed. by D.C. Coleman & B. Fry), pp. 201^218.Academic Press, San Diego, USA.

Guillaume J. (1997) Protein and amino acids. In: CrustaceanNutrition. Advances inWorld Aquaculture,Vol. 6 (ed. by L.R.D|¤ Abramo, D.E. Conklin & D.M. Akiyama), pp. 26^50.World Aquaculture Society, Baton Rouge, LA, USA.

Je¡rey S.W. & Welshmeyer N.A. (1997) Spectrophotometricand £uorometric equations in commonuse in oceanogra-phy. In: Phytoplankton Pigments in Oceanography, Mono-graphs on Oceanographic Methodology No. 10 (ed. by S.W.Je¡rey, R.F.C. Mantoura & S.W.Wright), pp.597^615. UNES-CO Publishing, Paris, France.

Johnston D., Lourey M.,Tien D.V., LuuT.T. & XuanT.T. (2002)Water quality and plankton densities in mixed shrimp-mangrove forestry farming systems inVietnam. Aquacul-ture Research 33,1^14.

LewisW.M., Hamilton S.K., Rodr|¤ guez M.A., Saunders J.F. III& Lasi M.A. (2001) Foodweb analysis of the Orinoco £ood-plain based on production estimates and stable isotopedata. Journal of the North American Benthological Society20, 241^254.

Sources of dietary carbon and nitrogen for cultured shrimp MABurford et al. Aquaculture Research, 2004, 35, 194^203

202 r 2004 Blackwell Publishing Ltd, Aquaculture Research, 35, 194^203

Loneragan N.R., Bunn S.E. & Kellaway D.M. (1997) Are man-grove and seagrasses sources of organic carbon for pe-naeid prawns in a tropical Australian estuary? Amultiple stable isotope study. Marine Biology 130, 289^300.

Newell R.I.E., Marshall N., SasekumarA. & ChongV.C. (1995)Relative importance of benthic microalgae, phytoplank-ton, and mangroves as sources of nutrition for penaeidprawns and other coastal invertebrates from Malaysia.Marine Biology123,595^606.

Pa¤ ez-Osuna F. (2001) The environmental impact of shrimpaquaculture: causes, e¡ects and mitigating alternatives.Environmental Management 28,131^140.

Peterson B.J. & Fry B. (1987) Stable isotopes in ecosystem stu-dies.Annual Reviews of Ecological Systematics18,293^320.

Red¢eld A.C. (1958) The biological control of chemicalfactors in the environment. American Science 46,205^222.

Shearer G.S. & Kohl D.H. (1993) Natural abundance of 15N:fractional contribution of two sources to a common sinkand use of isotope discrimination. In: Nitrogen IsotopeTechniques (ed. by R. Knowles & T.H. Blackburn), pp. 89^125. Academic Press, San Diego, USA.

Tran T.B., Dung L.C. & Brennan D.C. (1999) Environmentalcosts of shrimp culture in the rice growing regions of theMekong Delta. Aquaculture Economics and Management 3,31^43.

Yoshioka T.,Wada E. & Hayashi H. (1994) A stable isotopestudy on seasonal food web dynamics in a eutrophic lake.Ecology 75,835^846.

Aquaculture Research, 2004, 35, 194^203 Sources of dietary carbon and nitrogen for cultured shrimp MABurford et al.

r 2004 Blackwell Publishing Ltd, Aquaculture Research, 35, 194^203 203