Variations, trends and patterns of fish landings in large tropical reservoirs

Variations, trends and patterns of fish landings in large

tropical reservoirs

Tuantong Jutagate,1* Boonsong Srichareondham,2 Sovan Lek,3 Upali S. Amarasinghe4 andSena S. De Silva5

1Faculty of Agriculture, Ubon Ratchathani University, Warin Chamrab, Ubon Ratchathani, 2Department of Fisheries,

Inland Fisheries Research and Development Bureau, Chatuchak, Bangkok, Thailand, 3Laboratoire Evolution & Diversite

Biologique, Universite Toulouse, UMR 5172, CNRS – Universite Paul Sabatier, Toulouse Cedex, France, 4Department of

Zoology, University of Kelaniya, Kelaniya, Sri Lanka, and 5Network of Aquaculture Centres in Asia-Pacific, Chatuchak,

Bangkok, Thailand1

AbstractTemporal variations of fish yields in four major reservoirs in Thailand (Ubolratana; Sirindhorn; Srinakarin; Vajiralongk-

orn) were investigated with the use of long-term fish landing data (‡20 years). The long-term variations in fish yield, mea-

sured as the coefficient of variation of yearly yield, ranged mostly between 50% and 100%. For short-term variations, the

means of the relative variation (85%) were larger than the absolute variation (63%). This finding indicates that short-term

variations were inversely related to fish yield and that a higher uncertainty occurs when fish catches are low. The

stocked exotic species exhibited higher variations than the indigenous species. The trend analyses indicated some spe-

cies had sharply declined fish landings, while some species were quite stable (i.e. reservoir-adapted species). Stocked

species tended to increase in relatively shallow reservoirs, compared to the deep reservoir. Fish landing data for each

reservoir were patternized, using the self-organizing map, indicating temporal trends of chronological order. The differ-

ences among clusters in each reservoir were with respect to the weight of each species in the fish landings in each year,

and temporal changes in species composition in the reservoirs, which would primarily be attributed to the environmental

changes followed by anthropogenic pressures. The mean trophic level (s) fluctuated, resulting from changes in species

composition and weight of fish landing, as well as fish stocking programmes.

Key wordsfish landing, interannual variation, patterning, Thailand, trend analysis, trophic level.

INTRODUCTIONReservoirs are distributed widely throughout Thailand,

which contains an estimated 28 956 reservoirs ranging

from 0.01 ha to some >10 000 ha (Virapat et al. 2000).

Most large reservoirs are impounded for hydropower

generation, with fisheries being considered a secondary

benefit from the impoundments that benefits local popu-

lations (Thapanand et al. 2007). The majority of Thai res-

ervoir fishers are subsistence fishers. They use different

fishing gear, based on season, water level and fishing

ground. The fish landings are mostly indigenous species,

which form over 80% of the fish production (Jutagate

2009). The average fish yield in Thai reservoirs is esti-

mated to be 48 kg ha)1 per year, while the empirical

model between catch per unit effort and fishing effort in

Thai reservoirs indicates the estimated maximum sus-

tained yield and optimum fishing effort were 93 kg ha)1

per year and 10 fishers km)2, respectively (Moreau & De

Silva 1991). These figures are based on catch and effort

statistics, however, that rely on a constant catchability

coefficient of the nominal effort units, which rarely exists

for tropical reservoir fisheries (Amarasinghe & Pitcher

1986). Moreover, the nature of multi-species fishery and

temporal changes in fish composition would make effec-

tive implementation of management measures difficult.

In general, after a river is impounded and a reservoir

created, there are changes in fish communities resulting

from the strong alterations of physical and chemical

properties, as well as a changed ecosystem biological

*Corresponding author. Email: [email protected]

Accepted for publication 4 April 2011. 2

LR

E4

94B

Dispa

tch:

23.1.12

Jour

nal:

LRE

CE:S

ugan

ya

JournalName

ManuscriptNo.

Autho

rRec

eive

d:No.

ofpa

ges:

17PE

:Nith

ya

� 2012 The AuthorsDoi: 10.1111/j.1440-1770.2012.00494.x Journal compilation � 2012 Blackwell Publishing Asia Pty Ltd

Lakes & Reservoirs: Research and Management 2012 17: 1–17

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

productivity characteristics (Kolding & van Zwieten

2006). These eventually affect the variability of individual

species abundance and yield (Buijse et al. 1994; Ahmed

et al. 2001). Moreover, high fishing intensities will con-

tribute to a decreased biological diversity that might lead

to more unstable, and possibly lower, catches over the

long term (Kolding & van Zwieten 2006).

The differences in annual fish yield within and ⁄or

between reservoirs cannot be easily understood, likely

being due to a complex interaction of several variables

that influence biological productivity (De Silva & Amara-

singhe 2009). Several ecologists and fishery managers

have attempted to determine the yield and abundance of

fish stocks in aquatic ecosystems using physical, chemi-

cal and biological characteristics (surface area of the

river drainage basin; surface area of lakes; floodplain

areas; morphoedaphic index; depth; shoreline develop-

ment; primary production; etc.) (see Lae et al. 1999). The

utilization of one or more variables as a management

tool, however, largely depends on the nature of the fish-

eries as well as the available database (De Silva et al.

2001).

Moreover, the variability in fish yield also may be

caused by fluctuations in recruitment, and growth and

survival rates of the available target species, as well as

the fishing effort (Bayley 1988; Buijse et al. 1991). Many

studies had shown that changes in fish landings could

serve as a suitable ‘indicator’ for monitoring community

level responses to both fishing pressures and environ-

mental factors (e.g. Pauly et al. 1998; Darwall 2001; Hyun

et al. 2005; Morato et al. 2006). Variations in species com-

position of fish in reservoirs and lakes would also reflect

the variation in fish yields (i.e. individual large-sized spe-

cies contribute more weight than a small sized species),

although declines in overall fish yield may not be appar-

ent until the complete collapse of the fishery (Welcomme

2001). Thus, temporal patterns of variation in fish species

composition are one of the most important topics for fish

stock assessment of lake and reservoir fisheries

(Kubecka et al. 2009), with long-term studies of reservoir

fish communities and yields being necessary to establish

a baseline for management recommendations (Rıha et al.

2009). Similar to the reservoir fisheries elsewhere in Asia,

the variation and changes in the composition of fish

yields in Thai reservoirs are a common phenomenon.

Nevertheless, no systematic study on the trend and pat-

terns of these changes yet exists (Jutagate 2009).

The goal of this study was to examine the annual yield

variations of individual species of the selected reservoirs,

and then patternize long-term fish landing data, as well

as the changes in trophic status related to each patterned

period. The long-term fish landing data (i.e. ‡20 years) of

four major reservoirs were used in this study analyses.

On the basis of the non-selective nature of Thai inland

fisheries (Coates 2002), and the fact that fishermen catch

all species regardless of size variation (Bhukaswan &

Chookajorn 1988), the fish landings essentially represent

the actual spectrum of species composition.

MATERIALS AND METHODS

Reservoir selectionFour large reservoirs were selected for this study,

namely, Ubolratana, Sirindhorn, Srinakarin and Vajir-

alongkorn (Table 1), mainly because long-term data ser-

ies (i.e. ‡20 years) on fish landings were available for

these reservoirs. The statistical data on fish landings

were available from the Department of Fisheries (DoF),

Thailand, which have been submitted annually to the staff

of the fish conservation unit at each reservoir. Data were

collected at the fish landing sites of the reservoir by the

head of the villages in which the fish landing sites were

located. The village heads were trained on data collection

and requested to participate in weekly meetings with

DoF staff to compile the data. The catches were classi-

fied into species, recording numbers and total weight by

species.

Indexing variationsCharacteristics in annual variation of fish yields were per-

formed on the basis of the method proposed by Buijse

et al. (1991). An analysis of the coefficient of variation

Table 1. Characteristic of study reservoirs

Reservoir Region

Year of

impoundment

Surface

area (ha)

Mean

depth (m)

Catchment

area (km2)

Average annual

yield ± SD (MT) since 2000

Ubolratana North-east 1965 41 000 15.8 12 160 10 007.3 ± 233.8

Sirindhorn North-east 1971 29 200 5.1 2100 657.5 ± 209.1

Srinakarin Central 1977 40 000 44.6 10 880 329.3 ± 129.1

Vajiralongkorn Central 1986 16 700 25.2 3720 378.8 ± 156.7

� 2012 The AuthorsJournal compilation � 2012 Blackwell Publishing Asia Pty Ltd

2 T. Jutagate et al.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

(CV0) of the residuals (i.e. the difference between the

true and estimated values) was carried out to express the

long-term variations in fish yields of individual species.

To investigate trends in the long-term variations, the

CV0s of the 1st (linear trend, CV1) and 2nd order polyno-

mials (CV2) were estimated by using time (i.e. years)

and fish yields (i.e. fish landings) as independent and

dependent variables, respectively.

For short-term variations (i.e. 1 year to the next or in-

terannual), the index of absolute variation (Ua) and the

relative variation (Ur) were applied, using the following

models.

Ua ¼jdyjy

¼ 100�meanjyi � yi�1jy

ð%Þ ð1Þ

where y = mean value of long-term fish landings data; yi

and yi+1 = fish landings in a given year and previous year,

respectively, and

Ur ¼ 100� 2� 1� 1

10r

� �

1þ 1

10r

� �� �

ð%Þ�

ð2Þ

where r = mean of absolute difference of log transferred

catches, as calculated by:

r ¼X

n

i¼2

log10ðyi=yi�1Þj j= n� 1ð Þ ð3Þ

where n = duration of time series data. Ua was applied to

determine the absolute differences in the annual catch

between successive years, while Ur was applied because

the fishers generally experience short-term variations in

their catch as a percentile change (Buijse et al. 1991).

Thus, if Ur > Ua, the variation is inversely related to fish

yield. If Ur < Ua, the variation is directly related to fish

yield (Buijse et al. 1991; Ahmed et al. 2001).

Trend analysesFish landing data were standardized by re-scaling the

time series data, so that each of the different time series

can be compared, thereby making the averages equal

zero and the standard deviation equals 1 (Grainger &

Garcia 1996). To visualize the trend of individual species,

the data were then fitted using the nonparametric regres-

sion, the locally weighted scatter plot smoother or Lowess

(Cleveland 1979). In this procedure, each co-ordinate is

smoothed using a defined proportion of the neighbours

nearest to the target point, over parts of their ranges

(Trexler & Travis 1993; Brosse & Lek 2002). Optimal fit-

ting is obtained by iteratively minimizing the residuals

between the observed and estimated values. Two of the

major advantages of this method are that it can accu-

rately fit both linear and nonlinear data and also that it

automatically shows the degree of dependence of the

response to the predictor. There is no equation associ-

ated with the Lowess curve, however, due to its nonpara-

metric nature, with the result being that only graphical

results are obtained (Brosse & Lek 2002).

Clustering in fish landingsTo explore the overall ‘picture’ of fish landings, temporal

patterns of fish landing were clustered, using the self-

organizing map (SOM). SOM, also called a Kohonen

map (Kohonen 2001), is an unsupervised artificial neural

network learning method for analysing, clustering and

modelling various types of large databases. Hyun et al.

(2005) demonstrated that complex datasets, such as the

long-term fisheries data, were successfully patternized

using SOM. The advantages of this method, compared to

the conventional clustering analysis (e.g. multi-dimen-

sional scaling; factorial analysis), are discussed exten-

sively elsewhere (e.g. Park et al. 2006; Kalteh et al.

2008).

The SOM methodology has also been successfully

used in many aquatic ecological applications, especially

for transforming a nonlinear relationship of multivariate

data into a lower dimension (e.g. Giraudel & Lek 2001;

Park et al. 2006). The SOM consists of input, formed by

a set of sample units (i.e. annual fish landings data), and

output layers, formed by units arranged in a two-dimen-

sional grid, connected with computational weights (i.e.

weight vector). SOM algorithm maps a set of input vec-

tors (i.e. years, onto a set of vectors of output units)

according to the characteristics of the input vector com-

ponents (i.e. taxa of fish landing in this study) (Kangur

et al. 2007). The output layer consists of two-dimensional

networks of neurons arranged on the map of a hexagonal

lattice (i.e. map unit), because it does not favour horizon-

tal or vertical directions (Park et al. 2006).

During the SOM learning process, map units topo-

graphically close in the array will activate each other,

facilitating our learning something from the same input

vector. The map units in the output layer then compete

with each other, and the winner, whose weight is the

minimum distance from the input vector (i.e. the best

matching unit, BMU), is chosen to arrange the output

layer (Kangur et al. 2007). Samples with similar species

composition and weight of individual species were classi-

fied in the same cell or in the neighbouring cells. More-

over, using weighted vectors of a trained SOM, a

clustering technique (Ward’s method) was used to divide

the main clusters and subdivide the SOM cells into sev-

eral subclusters (Giraudel & Lek 2001). The quality of

the SOM map was measured via two criteria: quantization

and topographical errors. The first is the average

� 2012 The AuthorsJournal compilation � 2012 Blackwell Publishing Asia Pty Ltd

Fish yields of large tropical reservoirs 3

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

distance between each data vector and its BMU, which

measures the map resolution, with the latter being the

proportion of all data vectors for which first and second

BMUs are not adjacent units, which measures topology

preservation (Kohonen 2001).

Each input unit for each reservoir in this study was

accounted for with the ln (fish landing + 1) of each fish

for the input layer. Log-transformation of the data was

applied to reduce the distribution skewness of the fish

landing data (Agenbag et al. 2003). The number of output

map units for the output layer (i.e. map size) was deter-

mined as 5ffiffiffi

np

, where n = number of samples. This pro-

duces the best compromise number of output map units,

as proposed by the Laboratory of Computer and Informa-

tion science, CIS, Helsinki University of Technology (Ves-

anto 2000). The SOMS software package is available at

the website (http://www.cis.hut.fi/projects/somtoolbox/3 ).

All statistical calculations and graphics were done using

the R Program (R Development Core Team 2009). Analy-

sis of similarities (ANOSIM) was used to test statistical

differences among clusters of similar SOM cells, using

the library ‘vegan’ in Program R (Oksanen et al. 2006).4

Mean trophic level of fish yieldsThe succession of fish yields also was expressed in terms

of temporal changes in the trophic level. To calculate the

mean trophic level (s), the landings (Y) for a particular

year (i) was multiplied by the trophic level of the individ-

ual species groups (j) (TLij) and then taking a weighted

mean (Pauly et al. 1998). The trophic level estimates of

the fish species were all available, being taken from Fish-

Base (http://www.fishbase.org; Froese & Pauly 2009) via

the life history tool, as follows:

s ¼P

ij TLijYijP

Yij

ð4Þ

RESULTS

Annual variation of fish landingsThere were 124 cases (i.e. species of fish landing from

four reservoirs) for which continuous time series data of

>5 years were used in this analysis. The long-term varia-

tion (i.e. CV0) normally occurred in every case, ranging

from 30 to >200, but mainly being between 50% and 100%

(Table 2). The species with the lowest variations (i.e.

most stable fish landings) and highest variations (i.e.

high fluctuations in fish landings) in each reservoir were

Barbonymus schwanenfeldii, and Channa micropeltes for

Ubolratana Reservoir, Barbonymus gonoinotus and Heni-

chorhynchus siamensis for Sirindhorn Reservoir, Hampala

sp. and Clarias batrachus for Srinakarin Reservoir, and

Hemibagrus nemurus and H. siamensis for Vajiralongkorn

Reservoir.

Frequency distributions of CVs had modes at 40% and

60% for the 1st and 2nd order polynomials, respectively

(Table 2). Graphic plots between CV0 to CV1 and CV2

indicated all the co-ordinates were below the bisectrix

line (Fig. 1a,b), indicating significant trends on fish

landings for all four reservoirs. In the short-term (i.e.

interannual) variations, the distributions of both varia-

tions (i.e. Ua and Ur) did not conform to normality,

instead illustrating a positive skew, with mode values at

40% for Ua and 60% for Ur, respectively (Fig. 2). The

mean values of Ua and Ur were 63% and 85%. Ur was

significantly larger than Ua (t-test; P-value = 5.1 · 10)6),

indicating the short-term variation was inversely related

to yield. Comparing variations among the three groups of

fish landings (Table 2) illustrated that the stocked exotic

species had the highest variation for both the long-term

and short-term variations, with the average ± SD of CV0

and Ua being 107.3 ± 3.2 and 74.0 ± 10.1, respectively, fol-

lowed by the stocked indigenous species (95.5 ± 26.0 and

63.7 ± 18.3) and the indigenous species (94.0 ± 16.0 and

57.8 ± 15.1).

Trends in fish landings of individual speciesLanding trend profiles of 12 common species found in all

four selected reservoirs are illustrated in Fig. 3. Species

able to survive in the stagnant waterbodies, including

Channa striata (CHAS), H. nemurus (HEMN), Mast-

acembelus armatus (MASA) and Oxyeleotris marmorata

(OXYM), revealed either stable or increased trends after

a certain post-impoundment period (i.e. less fluctuation

around the average; Fig. 3a). For Vajiralongkorn Reser-

voir, however, these fish showed an increasing trend

after impoundment before peaking, and then declining,

although the fluctuations were in a narrow range. In con-

trast to the species well adapted to reservoir condition,

species such as Osteochilus hasselti (OSTH), Dangila

siamensis (DANS), Notopterus notopterus (NOTN) and

Hampala spp. (HAMS) showed continuous declines in

fish landings, being below the average fish landings for

the time series data after post-impoundment (Fig. 3b). It

was noted that, except for OSTH, the three remaining

species that inhabited Sirindhorn Reservoir apparently

illustrated upward trends after 1995.

For the stocked fish species (Fig. 3c), the indigenous

stocked H. siamensis (HENS) showed a sharp increase

after stocking commenced in 1990 for all reservoirs.

Declining fish landings of Morulius chrysophekadion

(MORC) were observed in Ubolratana and Vajiralongkorn

Reservoirs. However, a decreasing and then increasing

� 2012 The AuthorsJournal compilation � 2012 Blackwell Publishing Asia Pty Ltd

4 T. Jutagate et al.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

Table 2. Coefficient of variation (%) for zero (CV0), first (CV1)- and second (CV2)-order polynomials, and indices (%) for absolute (Ua)

and relative (Ur) short-time variations of fish landings in each reservoir 8

Scientific name Abbreviation CV0 CV1 CV2 Ua Ur

(a) Ubolratana Reservoir

Family Notopteridae

Notopterus sp. NOTN 59 26 27 26 27

Family Clupeidae

Clupeichthys aesarnensis CLUA 71 46 46 42 54

Family Cyprinidae

Barbodes gonionotus (a) BARG 81 25 32 67 59

Barbodes schwanenfeldii BARS 47 33 36 27 62

Cyprinus carpio (b) CHCS 101 77 93 131 93

Dangila siamensis DANS 134 63 77 138 194

Hampala sp. HAMS 105 67 80 47 52

Henicorhynchus siamensis (a) HENS 66 20 24 52 53

Labeo rohita (b) LABR 144 75 86 69 46

Moruluis chrysopeakadion MORC 145 103 107 64 106

Osteochilus hasselti OSTH 57 39 41 31 30

Osteochilus melanopluera OSTM 179 5 86 65 128

Puntioplites proctozysron (a) PUNP 53 15 30 31 32

Family Bagridae

Hemibragrus nemurus HEMN 73 26 38 48 80

Mystus sp. MYSS 118 85 98 40 55

Family Pangasiidae

Pangasius hypophthalamus (a) PANH 73 44 45 47 45

Family Siluridae

Kryptoperus bleekeri KRYB 91 3 9 47 55

Ompok krattensis OMPK 55 3 25 49 63

Wallago attu WALA 118 26 41 66 90

Family Clariiidae

Clarias batrachus (a) CLAB * * * * *

Family Mastacembelidae

Macrognathus siamensis MACS 111 57 66 75 89

Mastacembelus armatus MASA 95 13 35 73 61

Family Cichlidae

Oreochromis niloticus (b) OREN 93 27 66 43 64

Family Eleotridae

Oxyeleotris marmoratus OXYM 63 19 26 43 43

Family Nandidae

Pristolepis fasciatus PRIF * * * * *

Family Osphronemidae

Osphronemus gouramy (a) OSPG * * * * *

Family Channidae

Channa lucius CHAL 153 98 98 98 131

Channa micropeltes CHAM 223 33 61 33 61

Channa striata CHAS 79 52 62 33 51

Family Palaemonidae

Macrobrachium rosenbergii (b) MACR 96 42 56 88 84

� 2012 The AuthorsJournal compilation � 2012 Blackwell Publishing Asia Pty Ltd

Fish yields of large tropical reservoirs 5

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

trend for MORC was observed in Sirindhorn Reservoir,

similar to DANS, NOTN and HAMS. MORC has been

stocked in Srinakarin Reservoir since 1990, making

MORC landings for this lake quite stable. The exotic

stocked Oreochromis niloticus (OREN) and Macrobrachi-

um rosenbergii (MACR), which were originally stocked in

Table 2. (Continued)

Scientific name Abbreviation CV0 CV1 CV2 Ua Ur

(b) Sirindhorn Reservoir

Family Notopteridae

Notopterus sp. NOTN 62 26 46 36 47

Family Clupeidae

Clupeichthys aesarnensis CLUA 51 20 23 34 45

Family Cyprinidae

Barbodes gonionotus (a) BARG 50 30 30 39 35

Barbodes schwanenfeldii BARS 82 69 69 43 125

Cyclocheilichthys apogon CYCA 84 73 76 40 104

Dangila siamensis DANS 100 38 51 61 100

Hampala sp. HAMS 89 44 67 44 105

Henicorhynchus siamensis (a) HENS 136 10 100 63 103

Moruluis chrysopeakadion MORC 69 25 56 38 57

Osteochilus hasselti OSTH 95 25 88 62 152

Puntius brevis PUNB 113 33 83 58 157

Family Bagridae

Hemibragrus nemurus HEMN 51 10 31 34 60

Hemibragrus wyckoides (a) HEMW 100 89 89 39 151

Family Pangasiidae

Pangasius hypophthalamus (a) PANH 94 56 94 131 192

Family Siluridae

Kryptoperus bleekeri KRYB 71 49 55 78 139

Ompok krattensis OMPK 103 65 71 74 155

Family Clariiidae

Clarias batrachus (a) CLAB 94 32 50 80 128

Family Mastacembelidae

Mastacembelus armatus MASA 121 70 100 45 48

Family Cichlidae

Oreochromis niloticus (b) OREN 104 43 43 63 55

Family Eleotridae

Oxyeleotris marmoratus OXYM 70 51 60 47 86

Family Nandidae

Pristolepis fasciatus PRIF 66 27 54 35 52

Family Osphronemidae

Osphronemus gouramy (a) OSPG * * * * *

Family Channidae

Channa lucius CHAL 98 46 73 42 84

Channa micropeltes CHAM 130 30 96 51 93

Channa striata CHAS 67 40 59 31 41

Family Palaemonidae

Macrobrachium rosenbergii (b) MACR 103 24 86 93 156

Note: (1) letters (a) and (b) after the scientific name indicate the stocked indigenous and exotic species, respectively, and (2) * data con-

tained few years (<5 years) and were not taken into analyses of variations.

� 2012 The AuthorsJournal compilation � 2012 Blackwell Publishing Asia Pty Ltd

6 T. Jutagate et al.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

the 1980s, tended to increase since 1990. High fluctua-

tions, either above or below the averages, however, were

commonly observed.

Temporal patterns of fish landingsFor each SOM map, neurons located physically close to

each other indicated similar input patterns. Each SOM

map exhibited low final quantization and topographical

errors, which made them authentic (Table 3). Through

the SOM learning process, the annual fish landing for

each reservoir was patterned according to the similarity

of catch compositions and the weight of individual

catches during the period considered, when the SOM

maps also revealed trends among yearly catches. Accord-

ing to the U-matrix distances retrieved from the trained

SOM and incorporated with a hierarchical cluster analy-

sis, four clusters of yearly fish landings were observed in

all selected reservoirs, being significantly different among

clusters in all maps (ANOSIM test, P-value <0.001, based on

1000 permutations) (Table 3). Except for the SOM map

of Vajiralongkorn Reservoir (Fig. 4d), where the 1993

and 1994 fish landings were included in the same cluster

to those in 2003 and afterwards, SOM maps of the

remaining reservoirs exhibited explicit temporal trends of

chronological order with the annual fish landings

(Fig. 4a–c).

Twenty-nine fish species were used for the Ubolratana

Reservoir analysis (Fig. 5a), using time series data avail-

able from 4 years after impoundment, the first phase

being from 1969 to 1978 (i.e. Cluster Ia). In this cluster,

the majority of catches were the riverine species, domi-

nated by HENS, B. schwanenfeldii (BARS), OSTH, MORC

(a) (b)

Fig. 1. 9Plots between the CV0 and higher order CVs of fish landings in study reservoirs (a: CV0 vs. CV1; b: CV0 vs. CV2).

Fig. 2. 10Distributions of absolute (Ua)

and relative (Ur) short-time variations

of fish landings in study reservoirs.

POOR

QUALIT

YFIG

POOR

QUALIT

YFIG

� 2012 The AuthorsJournal compilation � 2012 Blackwell Publishing Asia Pty Ltd

Fish yields of large tropical reservoirs 7

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

and Puntioplites proctozysron (PUNP). Among these fish,

the BARS and MORC contributions to fish landings were

then regressed, being minimal in the last phase (i.e.

2001–2008; cluster IIb). The species exhibiting substantial

catches in the early phases (i.e. cluster Ia and Ib) and

then tended to decrease were Channa micropeltes

(CHAM), Pristolepis fasciatus (PRIF) Osteochilus melano-

pluera (OSTM), Channa lucius (CHAL) and DANS.

Meanwhile, it was observed that the limnophilic species

viz., HEMN, MASA and Macrognathus siamensis (MACS)

exhibited an increasing trend after impoundment. The

stocked species illustrated a significant contribution to

A

B

Fig. 3. 11Average standardized landings

of common species found in study

reservoirs (Lowess curves were used to

fit the data; for the trend lines:

blue, Ubolratana Reservoir; orange,

Sirindhorn Reservoir; brown, Srinakarin

Reservoir; black, Vajiralongkorn

Reservoir; for the X-axes: blue, Ubolra-

tana Reservoir; orange, Sirindhorn

Reservoir; black, Srinakarin and

Vajiralongkorn reservoirs).

Colouronline,B&W

inprint

POOR

QUALITY

FIG

� 2012 The AuthorsJournal compilation � 2012 Blackwell Publishing Asia Pty Ltd

8 T. Jutagate et al.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

fish landings between 1992 to the present time (i.e. clus-

ter IIa and IIb). Among them, Oreochromis niloticus

(OREN) and Pangasius hypopthalmus (PANH) were domi-

nant.

On the basis of 26 landing species in Sirindhorn Res-

ervoir (Fig. 5b), it was observed that the contributions of

the riverine species after 14 years of impoundment, such

as Clupeichthys aesarnensis (CLUA), HAMS, P. fasciatus

(PRIF), NOTN, OXYM, HEMN, B. gonionotus (BARG)

and MORC, were substantially high in cluster Ia (1985–

1992). They were relatively constant in others, implying

they can adapt behaviourally to the reservoir conditions.

Cluster Ib was characterized by a decline of OSTH,

Cyclochelichthys apogon (CYCA), Puntius brevis (PUNB)

and CHAM, C. batrachus (CLAB). Two distinct pelagic

species (CLUA and DANS) contributed to fish landings

in cluster IIa (1999–2000). Cluster IIb (2001–2007) was

obviously seen in the contribution to fish landings of

the stocked species, such as Hemibragrus wyckoides

(HEMW), OREN, MACR, Kryptopterus bleekeri (KRYB)

and PANH.

A total of 48 species were analysed for Srinakarin Res-

ervoir, with and the available data being from 1987 (i.e.

10 years after impoundment; Fig. 5c). Contributions of

riverine species, such as HAMS, HEMN, PRIF and BARG

in every cluster, were similar to the results observed for

Sirindhorn Reservoir. CYCA, OSTH and Mystus singarin-

gan (MYSS) were species exhibiting substantial landings

in cluster Ia (1987–1994), but declined thereafter. Similar

results in the decline of some species in other reservoirs,

such as CYCA and OSTH, indicated they were obligatory

riverine fish. Landings in cluster Ib (1995–2000) were

dominated by Chitala ornata (CHIO) and OREN. Channa

striata (CHAS), HEMW, CLUA and Kryptoperus bleekeri

C

Fig. 3. 11(Continued).

Table 3. Inputs and results of SOM analysis in study reservoirs

Reservoir Time series Fish species Map site

Quantization

error

Topographical

error

No. of

clusters ANOSIM test

Ubolratana 1969–2008 29 6 · 5 0.709 0.000 4 R = 0.839; P < 0.001

Sirindhorn 1985–2007 26 5 · 4 0.920 0.000 4 R = 0.834; P < 0.001

Srinakarin 1987–2007 49 5 · 4 1.426 0.000 4 R = 0.861; P < 0.001

Vajiralongkorn 1987–2007 33 5 · 4 0.799 0.000 3 R = 0.783; P < 0.001

SOM, self-organizing map.

Colouronline,B&W

inprint

POOR

QUALITY

FIG

� 2012 The AuthorsJournal compilation � 2012 Blackwell Publishing Asia Pty Ltd

Fish yields of large tropical reservoirs 9

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

(KRYB) started to increase in cluster Ia and subsequently

continuing. Clusters IIa (2001–2003) and IIb (2003–2007)

involved the landings of stocked species, with the most

obvious species being PANH. It is noteworthy that, for

Srinakarin Reservoir, except for OREN, contributions of

stocked species, such as PANH, Chinese carps (CHCS)

Cirrhinus mrigala (CIRM) and MACR, were substantial

since 2000.

Data collection for 33 species in Vajiralongkorn Reser-

voir commenced immediately after 1 year of impound-

ment. The temporal fluctuations of key species generally

were relatively similar to the results obtained for the

above three reservoir (Fig. 5d), although there were two

noteworthy observations. First, in regard to four snake-

head species (Channidae) found in this lake, contribu-

tions of CHAM and CHAS in landings were quite

constant, while those for Channa aurolineatus (CHAA)

and Channa lucius (CHAL) declined after impoundment.

Second, cluster analysis indicated that the composition of

fish landings in 1993 and 1994 were involved with those

in 2003–2007 (cluster III). The latter result was likely to

be due to the balance in the proportion of fish landings

in cluster II (1995–2002), and the decrease in some

stocked fish landings (e.g. HENS, MACR, OREN, KRYB

and Puntius orphoides (PUNO)) found in clusters III.

Trends in mean trophic levelThere was a concordant trend between s and fish land-

ings. The reservoir fisheries targeted high trophic level

species, with s values ranging between 2.5 and 3.4. Vajir-

alongorn Reservoir was an exceptional case, with the s

value being around 3 and above. Figures 6a–d illustrate

high fluctuations in the mean trophic levels of fish land-

ings for all reservoirs and a gradual transition from land-

ings of large carnivorous fishes (s � 3) to herbivorous

fishes (s � 2).

A sharp decrease was observed for Ubolratana Reser-

voir in ‘cluster Ia’ period (Fig. 6a), and the s value nar-

rowly varying between 2.5 and 2.8 from the late 1970s to

2000. It exhibited a recently declining s trend between

2002 and 2008, however, coincided with changes in land-

ings composition (Fig. 6a). The latter highlighted the

occurrence of a declining catch of many species

occurred, although not for herbivorous species such as

HENS and PUNP. A peak of s was observed for Sirind-

horn Reservoir during the first year of data collection

(a) (b)

(c) (d)

Fig. 4. 12Patterning of fish landing in each reservoir by year using SOM network (a, Ubolratana Reservoir; b, Sirindhorn Reservoir; c,

Srinakarin Reservoir; d, Vajiralongkorn Reservoir; the similarity among SOM cells of each model was studied using hierarchical clustering

agglomerate by Ward method to identify the cluster number; bold and dashed lines indicate main and subclusters, respectively). SOM, self-

organizing map.

Colouronline,B&W

inprint

POOR

QUALITY

FIG

� 2012 The AuthorsJournal compilation � 2012 Blackwell Publishing Asia Pty Ltd

10 T. Jutagate et al.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

(i.e. 14 years after impoundment), which then continued

to exhibit a downward trend. In recent years, however,

the trend appeared to be slightly inverted, attaining a

value of 2.88 in 2007. This was likely due to the high pro-

duction of small carnivorous fish (i.e. CLUA).

High fluctuations in s values were observed for Srinak-

arin and Vajiralongkorn reservoirs (Fig. 6c,d). Sharp

increases ofs, after an initial decline, were observed in

1994–1996 and in 1991–1994 in Srinakarin and Vajir-

alongkorn reservoirs, respectively. Although carnivorous

species landings were relatively constant in both reser-

voirs, it can be explained by the reduced landing of

herbivorous fish (e.g. CYCA; MORC) in Srinakarin Reser-

voir, and OSTH and DANS in Vajiralongkorn Reservoir

during these periods.

DISCUSSION

Variations of Thai reservoir fish landingsAlthough obtaining an accurate picture of the fish stocks

in reservoirs is a difficult challenge, it also is an appropri-

ate goal for scientific development (Kubecka et al. 2009).

(a) (b)

(c) (d)

Fig. 5. 13Contribution of yield (i.e. ln (fish landing + 1)) of each fish species in each cluster (i.e. period) of study reservoirs (a, Ubolratana

Reservoir; b, Sirindhorn Reservoir; c, Srinakarin Reservoir; d, Vajiralongkorn Reservoir).

POOR

QUALIT

YFIG

� 2012 The AuthorsJournal compilation � 2012 Blackwell Publishing Asia Pty Ltd

Fish yields of large tropical reservoirs 11

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

The design of the present study provided a ‘picture’ of

yield variations in Thai reservoirs. Moreover, because of

the non-selective nature of Thai inland fisheries (Coates

2002), variations in fish landings of individual species

were relatively directed to abundance of the species in

communities. Higher relative annual variation of (Ur),

than index of absolute variation (Ua), indicated the

annual variations in Thai reservoir fish landings was

inversely related to yield, resulting in a higher uncer-

tainty when catches are low (Buijse et al. 1991). The CV0

values obtained indicate that few species could exhibit a

stable yield after impoundment (i.e. low CV0). These spe-

cies included Channa striata (CHAS), Hemibragrus nemu-

rus (HEMN) and Oxyeletris marmoratus (OXYM). Short-

lived species, such as the small clupeid Corica soborna,

also exhibited high production, low CV0 and small inter-

annual differences in their catches (Ahmed et al. 2001).

The same results were obtained for the small clupeid

Clupeichthys aesarnensis (CLUA), which is among the

core fishery target in Thai reservoirs (Jutagate et al.

2003). High CV0 of CLUAin Srinakarin Reservoir is likely

attributable to the ban of the luring lift net from 1989 to

1999, mainly because a large number of fish larvae were

also caught (Amornchairojkul & Sricharoendham 1997),

followed by the re-approval in 2000 of restrictions in fish-

ing grounds in the open pelagic zone (Jutagate & Matt-

son 2003). Scales of CV0 and Ua could also have

indicated the variability of individual species in communi-

ties (Blanchard & Boucher 2001). The species exhibiting

small short- and long-term variations indicate a stable bio-

mass and wide fluctuation for the species with high short-

and long-term variations (Buijse et al. 1991).

The results of the trend analysis also indicated that

species with high CV0 or non-consistent CV0 exhibited

either declining trends (Fig. 3b) or wide fluctuations in

fish landings (i.e. downward trends in the early phase of

post-impoundment, followed by an apparently upward

trend), as observed for Sirindhorn Reservoir. These spe-

cies could be declared riverine specialists, which adapted

themselves poorly to reservoir conditions. Some of these

specialists could proliferate in a reservoir, however,

where it could at least spawn in the connected river or

upstream area during the early part of the rainy season

(De Silva 1983). Examples from Sirindhorn Reservoir,

where fish landings of many species trended upward,

were like attributable to the declaration of closed fishing

areas in the upstream end of the reservoir since 1995

(Jutagate & Mattson 2003). Thus, limited spawning areas

and ⁄ or high fishing pressures during the spawning per-

iod could impact the annual recruitment of these species

which, in turn, might result in large variation in the

yields of these specialists (Ahmed et al. 2001).

(a) (b)

(c) (d)

Fig. 6. 14Temporal changes in mean trophic level (,s) of fish landings in study reservoirs (a, Ubolratana Reservoir; b, Sirindhorn Reservoir; c,

Srinakarin Reservoir; d, Vajiralongkorn Reservoir; horizontal dashed line indicates average s in each cluster).

POOR

QUALIT

YFIG

� 2012 The AuthorsJournal compilation � 2012 Blackwell Publishing Asia Pty Ltd

12 T. Jutagate et al.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

High fluctuations in fish landings of reservoir-adapted

species in Vajiralongkorn Reservoir (Fig. 3a) could be

related to fishing pressures and the reservoir characteris-

tics per se. In the early post-impoundment phase, the fish-

ers mostly targeted large carnivorous species that could

survive in reservoir conditions (Petr 1985), examples

being CHAS, HEMN, MASA and OXYM, which prefer lit-

toral habitats. The characteristics of a deep reservoir with

very limited littoral zone, however, such as Vajiralongk-

orn Reservoir (Table 1), could be the main reason for the

declining production of these fishes and, subsequently, in

their landings, which is similar to other deep reservoirs

such as the Rımov Reservoir in Czech Republic (Prcha-

lova et al. 2009).

Kolding and van Zwieten (2006) stated that, as a reser-

voir ages, a shift in species composition apparently takes

place. An obvious shift would be pristine riverine species

forced to move upstream, and perhaps eventually disap-

pear, while the open water environment provides a

favourable habitat for lacustrine species, thereby facilitat-

ing their increase. By using SOM, the long-term complex

data of fish landings could be efficiently utilized to inves-

tigate temporal patterns of change (Hyun et al. 2005).

The present analysis indicated temporal changes in spe-

cies composition from riverine to lacustrine species,

which could primarily be due to environmental changes

(i.e. from a running or flowing water environment, to a

stagnant or pooled water environment), followed by

human activities (e.g. fishing and fish stocking pro-

grammes). The riverine species that commonly occupy

the middle to lower sectors of the river course can gener-

ally constitute basic colonizers when rivers are

impounded and converted to reservoirs (Welcomme et al.

2006). Nevertheless, most riverine species are adapted to

changing environmental fluctuations from hydrological

oscillations, with their breeding typically being seasonally

defined in coincidence with the floods in tropical regions

(Junk & Wantzen 2004), in which there is decreased

flood pulse variation in the reservoir (Wantzen et al.

2008). The obvious example in Thai reservoirs is the

absence of Pangasiid and some Silurid fishes, except

when they are stocked (Jutagate et al. 2005). In contrast,

the species that can adapt behaviourally to lacustrine con-

ditions (such as many cyprinids) can thrive in high abun-

dances (Welcomme et al. 2006).

Although Kolding and van Zwieten (2006) stated that

as the ‘fishing-down-the-food-web’ (Pauly et al. 1998) is

not likely to occur in a reservoir system because of the

high resilience of freshwater ecosystems and because of

this property, the s of the community could gradually

return towards its original state (Darwall 2001). Our

results suggest that ecosystem structure changes were

influenced by fisheries in Thai reservoirs, resulting in a

declining s regarding fish landings. After impoundment

occurs, fishing efforts usually target larger fish, which

often are predatory fishes. In addition to a low intrinsic

growth rate, and longer period of resilience of the preda-

tory species, this would eventually result in a decreased

abundance of high trophic level species, relative to low

trophic level ones in the ecosystem (Baeta et al. 2009).

s could be a useful indicator to describe the state of fish-

eries because size-selective mortality causes decreases in

the relative abundance of larger species (Welcomme

2001; Baeta et al. 2009). Except for fishing intensity, and

natural oscillations in species abundance, s may be

affected by changes in fishing technology and economic

factors, which are always taken into account in regard to

declining high trophic level species (i.e. top predators)

and subsequent decreases in s(Caddy et al. 1998). This is

not likely the case in the present study, however,

because inland fishing gears are mostly traditional, the

most common gear being the monofilament gillnet,

which was introduced in the 1960s (Jutagate & Mattson

2003). Furthermore, almost all, if not all, fishes were uti-

lized for home consumption, regardless of their economic

value. Fluctuation ins for a complex ecosystem, such as a

large lake, would result from the complex interactions of

the communities, as well as changing food resources

(Njiru et al. 2005). As this study indicated, however, in

the less complex situation, such as a reservoir, fluctua-

tion of s in fish landings in Thai reservoirs was also

caused by the fish stocking programme.

Role of the stocking species in Thaireservoir fish landings

Contributions of the stocked species, either exotic or

indigenous, have caused significant changes in the pat-

terns of fish landings in Thai reservoirs. Fish stocking is

regularly practiced in Thai reservoirs, with the clear

understanding that it is for the general benefit of the

open-access fishers that continue to rely on these

resources (De Silva & Funge-Smith 2005). In the 2009 fis-

cal year, the total number of stocked fish and giant fresh-

water prawn in the inland waterbodies countrywide was

estimated to be �2500 million individuals, being com-

prised of exotic and indigenous species ranging from the

carnivores Clarias spp. to the herbivores Barbonymus

spp. (DoF 2008). The most common stocked exotic spe-

cies are the Chinese and Indian major carps. Although

these exotic species can grow in size, and are consis-

tently found in the fish landings, they contributed less

significantly to the landings, compared to the indigenous

� 2012 The AuthorsJournal compilation � 2012 Blackwell Publishing Asia Pty Ltd

Fish yields of large tropical reservoirs 13

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

species. De Silva and Funge-Smith (2005) mentioned that

the primary reason most exotic stocked species do not

tend to significant influence the yields in large lacustrine

waterbodies, although they can grow to a large size, is

because they are generally unable to reproduce in these

types of waterbodies, therefore being unable to form

large populations that would compete for common

resources. Nevertheless, among the exotic stocked spe-

cies, only Nile tilapia exhibited the potential to self-recruit

(Virapat 1993), forming a sufficient stock to sustain the

reservoir fisheries.

However, the indigenous species have proven to be

successful, as evidenced from the high yields, as exhib-

ited in Ubolratana Reservoir. Puntioplites proctozystron

(PUNP) and Pangasius hypopthalmus (PANH) are the

dominant indigenous species in this reservoir since they

have been stocked in 1969 and 1977 (Petr 1989), and

since the 1990s, respectively. Furthermore, Henicorhyn-

chus siamensis (HENS), PANH and Osphronemus gourami

(OSPG) are among the successfully colonized species in

Sirindhorn, Srinakarin and Vajiralongkorn reservoirs, hav-

ing been harvested regularly since the continuous stock-

ing programmes began in 1980 (Department of Fisheries

2007, unpubl. data). Moreover, Barbonymus gonionotus

(BARG) also has been recognized as a successfully

stocked species in many Thai reservoirs (Pawaputanon

1992). There is also an exceptional case in Thai reservoir

fisheries wherein the giant freshwater prawn Macrobrach-

ium rosenbergii (MACR) occurs in fish landings. MACR

has been regularly stocked since 1990, and their varia-

tions on stocking rates and yields have made temporal

differences in fish landings from Thai reservoirs. The

stocking of this prawn is relatively uncommon and must

be performed regularly as it requires brackish water in

the initial stages (De Silva & Funge-Smith 2005). The

popularity of this prawn is because of its high price.

Renunual and Silapachai (2005) provided an example

illustrating that a recapture rate of 1.8% of stocked

M. rosenbergii in Bangpra Reservoir resulted in an eco-

nomic profit of 721.6%. Meanwhile, 97% of the economic

value of the fish landings in Pak-Mun Reservoir was also

attributed to this prawn (Sripatrprasite & Lin 2003).

Implications for fisheries managementAlthough the optimization of fishing efforts to maximize

fish yields (e.g. Moreau & De Silva 1991; Jutagate et al.

2003) or economic and social values (e.g. Thapanand

et al. 2007) in Thai reservoirs has been studied, its suc-

cess in implementation is still unclear, mainly because of

the nature of multi-gears and multi-species fisheries. The

fisheries are mostly developed in the littoral zone, being

based on the lacustrine adapted species (Jutagate 2009),

as can be seen from the composition of the landings. The

sustainability of the fishery implies that protecting the lit-

toral zones is particularly important because of their key

role as a main habitat for several species, either as feed-

ing grounds or reproductive areas (Thapanand et al.

2007), as well as the upstream riverine areas of the reser-

voir, which are important for most cyprinids during the

spawning season (De Silva 1983).

Fish stocking programmes in reservoirs are consid-

ered good options for enhancing fish production. Stock-

ing should favour indigenous species, however, because

of the consistently high returns in fish landings, as illus-

trated in this study. Stocking of exotic species should be

accordingly halted. Moreover, exotic species have small

niche breadth and have illustrated they cannot compete

for food sources with indigenous fishes (Villanueva et al.

2008), further reducing their value for stocking purposes.

The low value (<0.0015) of the gross efficiency transfer

of primary production through the fish catches in Thai

reservoirs (Villanueva et al. 2008; Thapanand et al. 2009) 5,

compared to the other tropical inland waterbodies

(� 0.005; Christensen & Pauly 1993), suggests a large

excess production of phytoplankton and plants. This

implies that large proportions of zooplankton and herbiv-

orous fish can be added to these water systems, which

can support higher trophic level species. At the same

time, the impacts of stocking fish from hatchery popula-

tions should be of major concern, because they can lead

to decreased genetic variation and genetic identity of wild

populations (Kamonrat 2008).

ACKNOWLEDGEMENTSThis research article was made possible as part of the

NACA-ICEIDA Project ‘Strategies for Development of

Asian Reservoir and Lake Fisheries Management &

Development’ (ICE ⁄SL ⁄FIS ⁄ 2007 ⁄ 02 – NACA). Statistical

analyses were conducted during T. Jutagate time as a vis-

iting researcher at the Laboratoire Evolution and Diver-

site Biologique, Universite Toulouse, France in 2009. We

are grateful to Dr. Gael Grenouillet, Universite Toulouse,

for advising in statistics. We also thank Pisit Phomikong,

Department of Fisheries, for details on the stocked spe-

cies in the representative reservoirs. We also thank the

anonymous reviewers of the manuscript for their invalu-

able comments and suggestions.

REFERENCESAgenbag J. J., Richardson A. J., Demarcq H., Freon P.,

Weeks S. J. & Shillington F. A. (2003) Estimating envi-

ronmental preferences of South African pelagic fish

� 2012 The AuthorsJournal compilation � 2012 Blackwell Publishing Asia Pty Ltd

14 T. Jutagate et al.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

species using catch size- and remote sensing data.

Prog. Oceanogr. 58, 275–300.

Ahmed K. K. U., Hambrey J. B. & Rahman S. (2001)

Trend in inter-annual yield variation of reservoir fisher-

ies in Bangladesh, with special reference to Indian

major carps. Lakes Reserv.: Res. Manage. 6, 85–94.

Amarasinghe U. S. & Pitcher T. J. (1986) Assessment of

fishing effort in Parakrama Samudra, an ancient man-

made lake in Sri Lanka. Fish. Res. 4, 271–82.

Amornchairojkul S. & Sricharoendham B. (1997) Fisher-

ies Activities and Fish Catch in Srinakarin Reservoir,

Kanchanaburi Province. National Inland Fisheries Insti-

tute, Bangkok.

Baeta F., Costa M. J. & Cabral H. (2009) Changes in the

trophic level of Portuguese landings and fish market

price variation in the last decades. Fish. Res. 97, 216–

22.

Bayley P. B. (1988) Accounting for effort when compar-

ing tropical fisheries in lakes, river flood-plains and

lagoons. Limnol. Oceangr. 33, 963–77.

Bhukaswan T. & Chookajorn T. (1988) Present status of

the reservoir fisheries in Thailand. In: Reservoir Fish-

eries Management and Development in Asia (ed. S. S.

De Silva), pp. 154–7. Development Research Center,

Ottawa, ON.

Blanchard F. & Boucher J. (2001) Temporal variability of

total biomass in harvested communities of demersal

fishes. Fish. Res. 49, 283–93.

Brosse S. & Lek S. (2002) Relationships between environ-

mental characteristics and the density of age-0 Eur-

asian Perch Perca fluviatilis in the littoral zone of a

lake: a nonlinear approach. Trans. Am. Fish. Soc. 131,

1033–43.

Buijse A. D., van Densen W. L. T. & Machiels M. A. M.

(1991) Characteristics in the annual variation of yield

from professional fisheries in freshwater bodies of the

temperate and the tropical zones. In: Catch and Effort

Sampling Strategies: Their Applications in Freshwater

Fisheries Management (ed. I. G. Cowx), pp. 305–22.

Fishing News Book, London.

Buijse A. D., van Densen W. L. T. & van Breukelen S. W.

F. (1994) The scope of stabilizing yield for various

types of fish stocks and fisheries. In: Rehabilitation of

Freshwater Fisheries (ed. I. G. Cowx), pp. 57–68. Fish-

ing News Book, London.

Caddy J. F., Csirke J., Garcia S. M. & Grainger R. J. R.

(1998) How pervasive is ‘‘fishing down marine food

webs?’’ Science 282, 1383–5.

Christensen V. & Pauly D. (1993) Trophic models of

aquatic ecosystems. ICLARM Conference Proceedings

No. 26, ICLARM, Manila.

Cleveland W. S. (1979) Robust locally weighted regres-

sion and smoothing scatterplots. J. Am. Statist. Assoc.

74, 829–36. 6

Coates D. (2002) Inland Capture Fishery Statistics of

Southeast Asia: Current Status and Information Needs.

Asia-Pacific Fishery Commission, FAO, Bangkok.

Darwall W. (2001) Preliminary investigations of commu-

nity level responses to benthic trawling in the demersal

fish fauna of Lake Malawi ⁄Niassa, Africa. In: Proceed-

ings of the Lake Malawi Fisheries Management Sym-

posium (eds O. L. F. Weyl & M. V. Weyl), pp. 189–94.

Department of Fisheries, Malawi.

De Silva S. S. (1983) Reproductive strategies of some

major fish species in Parakrama Samudra reservoir

and their possible impact on the ecosystem: a theoreti-

cal consideration. In: Limnology of Parakrama Samu-

dra, Sri Lanka: A Case Study of An Ancient Man-made

Lake in the Tropics (ed. F. Schiemer), pp. 186–91. Junk

Publisher, The Hague.

De Silva S. S. & Amarasinghe U. S. (eds) (2009) Reser-

voir Fisheries in Asia and Pacific. Network of Aquacul-

ture Centres in Asia and Pacific, Bangkok.

De Silva S. S. & Funge-Smith S. J. (2005) A Review of

Stock Enhancement Practices in the Inland Water Fish-

eries of Asia. RAP Publication No. 2005 ⁄ 12. Asia-Pacific

Fishery Commission, Bangkok.

De Silva S. S., Amarasinghe U. S., Nissanka C., Wi-

jesooriya W. A. D. D. & Fernando M. J. J. (2001) Use

of geographical information systems as a tool for pre-

dicting fish yield in tropical reservoirs: case study on

Sri Lankan reservoirs. Fish. Manag. Ecol. 8, 47–60.

Department of Fisheries (DoF) (2008) Action Plans of

Department of Fisheries for the Fiscal Year 2009.

Department of Fisheries, Bangkok. (In Thai)

Froese R. & Pauly D. (2009) ‘‘FishBase’’. Available from

URL: http://www.fishbase.org. Accessed July to

December 2009.

Giraudel J. L. & Lek S. (2001) A comparison of self-orga-

nizing map algorithm and some conventional statistical

methods for ecological community ordination. Ecol.

Model. 146, 329–39.

Grainger R. J. R. & Garcia S. M. (1996) Chronicles of

marine fishery landings (1950–1994): trend analysis

and fisheries potential. FAO Fisheries Technical Paper.

No. 359, FAO, Rome.

Hyun K., Song M. Y., Kim S. & Chon T. S. (2005) Using

an artificial neural network to patternize long-term fish-

eries data from South Korea. Aquat. Sci. 67, 382–9.

Junk W. J. & Wantzen K. M. (2004) The flood pulse con-

cept: new aspects, approaches and applications – an

update. In: Proceedings of the Second International

� 2012 The AuthorsJournal compilation � 2012 Blackwell Publishing Asia Pty Ltd

Fish yields of large tropical reservoirs 15

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

Symposium on the Management of Large Rivers for

Fisheries: Volume 2. pp. 335–57. FAO Regional Office

for Asia and the Pacific, Bangkok.

Jutagate T. (2009) Reservoir fisheries in Thailand. In:

Reservoir Fisheries in Asia and Pacific (eds S. S. De

Silva & U. S. Amarasinghe), pp. 96–113. Network of

Aquaculture Centres in Asia and Pacific, Bangkok.

Jutagate T. & Mattson N. S. (2003) Optimizing fishing

gear operation in Sirinthorn Reservoir, Thailand. Nat.

Hist. Bull. Siam Soc. 51, 105–22.

Jutagate T., De Silva S. S. & Mattson N. S. (2003) Yield,

growth and mortality rate of the Thai river sprat (Clup-

eichthys aesarnensis Wongratana, 1983) in Sirinthorn

Reservoir, Thailand. Fish. Manage. Ecol. 10, 221–31.

Jutagate T., Krudpan C., Ngamsnae P., Lamkom T. &

Payooha K. (2005) Changes in the fish catches during

the trial of opening the sluice gates of a run-of-the river

reservoir in Thailand. Fish. Manage. Ecol. 12, 57–62.

Kalteh A. M., Hjorth P. & Berndtsson R. (2008) Review

of the self-organizing map (SOM) approach in water

resources: analysis, modelling and application. Environ.

Model. Softw. 23, 835–45.7

Kamonrat W. (2008) Genetic diversity of Barbonymus gon-

ionotus and Henicorhynchus siamensis populations in

the lower Mun River after restocking programme: a

case study in Pak-Mun Dam. Thai Fish. Gazet. 62,

493–7. (In Thai).

Kangur K., Park Y. S., Kangur A., Kangur P. & Lek S.

(2007) Patterning long-term changes of fish community

in large shallow Lake Peipsi. Ecol. Model. 203, 34–44.

Kohonen T. (2001) Self-Organizing Maps. Springer-Ver-

lag, Berlin.

Kolding J. & van Zwieten P. A. M. (2006) Improving Pro-

ductivity in Tropical Lakes and Reservoirs. WorldFish

Center, Cairo.

Kubecka J., Hohausova E., Matena J. et al. (2009) The

true picture of a lake or reservoir fish stock: a review

of needs and progress. Fish. Res. 96, 1–5.

Lae R., Lek S. & Moreau J. (1999) Predicting fish yield of

African lakes using neural networks. Ecol. Model. 120,

325–35.

Morato T., Watson R., Pitcher T. J. & Pauly D. (2006)

Fishing down the deep. Fish Fish. 7, 24–34.

Moreau J. & De Silva S. S. (1991) Predictive fish yield

models for lakes and reservoirs of the Philippines, Sri

Lanka and Thailand. FAO Fishery Technical Paper No.

319, FAO, Rome.

Njiru M., Waithaka E., Muchiri M., van Knaap M. &

Cowx I. G. (2005) Exotic introductions to the fishery of

Lake Victoria: what are the management options? Lakes

Reserv.: Res. Manage. 10, 147–55.

Oksanen J., Kindt R., Legendre P. & O’Hara R. B. (2006)

vegan: Community Ecology Package. R package ver-

sion 1.8–3.

Park Y. S., Tison J., Lek S., Giraudel J. L., Coste M. &

Delmas F. (2006) Application of a self-organizing map

to select representative species in multivariate analysis:

a case study determining diatom distribution pattern

across France. Ecol. Inform. 1, 247–57.

Pauly D., Christensen V., Dalsgaard J., Froese R. & Tor-

res F. J. Jr. (1998) Fishing down marine food webs.

Science 279, 860–3.

Pawaputanon O. (1992) Inland Capture Fisheries in Thai-

land. FAO Fisheries Report No. 458, FAO, Rome.

Petr T. (1985) Inland fisheries in multi-purpose river

basin planning and development in tropical Asian coun-

tries: three case studies. FAO Fishery Technical Paper

No. 265, FAO, Rome.

Petr T. (1989) The use of cyprinids in fisheries manage-

ment of large inland water bodies of the Indo-Pacific.

In: Papers Contributed to the Workshop on the Use of

Cyprinids in Fisheries Management of Large Inland

Water Bodies of the Indo-Pacific (ed. T. Petr), pp. 1–5.

FAO, Rome.

Prchalova M., Kubecka J., Cech M. et al. (2009) The

effect of depth, distance from dam and habitat on spa-

tial distribution of fish in an artificial reservoir. Ecol.

Freshw. Fish. 18, 247–60.

R Development Core Team (2009) The R Manuals. Avail-

able from URL: http://cran.r-project.org/manuals.html.

Accessed 14 May 2009.

Renunual A. & Silapachai W. (2005) Evaluation on the

Stocking Program of the Giant Freshwater Prawn

(Macrobrachium rosenbergii de Man) in Bang Phra Res-

ervoir, Chonburi Province. Chonburi Inland Fisheries

Research and Development Centre, Chonburi (In Thai).

Rıha M., Kubecka J., Vasek M. et al. (2009) Long term

development of fish populations in the Rımov Reseer-

voir. Fish. Manag. Ecol. 16, 121–9.

Sripatrprasite P. & Lin C. K. (2003) Stocking and recap-

ture of freshwater prawn (Macrobrachium rosenbergii

De Man) in a run of- river type dam (Pak Mun Dam)

in Thailand. Asian Fish. Sci. 16, 167–74.

Thapanand T., Moreau J., Jutagate T. et al. (2007)

Towards possible fishery management strategies in a

newly impounded man-made lake in Thailand. Ecol.

Model. 204, 143–55.

Trexler J. C. & Travis J. (1993) Non-traditional regression

analyses. Ecology 74, 1629–37.

Vesanto J. (2000) Neural network tool for data mining:

SOM Toolbox. Proceedings of Symposium on Tool

Environments and Development Methods for Intelligent

� 2012 The AuthorsJournal compilation � 2012 Blackwell Publishing Asia Pty Ltd

16 T. Jutagate et al.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

Systems (TOOLMET2000), pp. 184–96. Oulun yliopist-

opaino, Oulu.

Villanueva M. C., Moreau J., Amarasinghe U. S. & Schi-

emer F. (2008) Foodweb and the trophic structure of

two Asian reservoirs using Ecopath with Ecosim and

Ecospace: a comparative study. In: Aquatic Ecosystems

and Development: Comparative Asian Perspectives

(eds F. Schiemer, D. Simon, U. S. Amarasinghe & J.

Moreau), pp. 413–34. Backhuys Publishers, Leiden.

Virapat C. (1993) Bionomics of Fish Stocking in Two

Thai Reservoirs; Biological, Management and Develop-

ment Perspectives. Ph.D. Thesis, Dalhousie University,

Nova Scotia.

Virapat C., Phimonbutra U. & Chantarawaratid C. (2000)

Fishery and Fisheries Management in Thai Reservoirs:

Review and Assessment. Mekong River Commission,

Vientiane.

Wantzen K. M., Junk W. J. & Rothhaupt K. O. (2008) An

extension of the flood-pulse concept (FPC) for lakes.

Hydrobiologia 613, 151–70.

Welcomme R. L. (2001) Inland Fisheries: Ecology and

Management. Fishing News Books, Blackwell, Oxford.

Welcomme R., Winemiller K. O. & Cowx I. G. (2006)

Fish environmental guilds as a tool for assessment of

ecological condition of rivers. River Res. Appl. 22, 377–

96.

� 2012 The AuthorsJournal compilation � 2012 Blackwell Publishing Asia Pty Ltd

Fish yields of large tropical reservoirs 17

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

Author Query Form

Journal: LRE

Article: 494

Dear Author,

During the copy-editing of your paper, the following queries arose. Please respond to these by marking up your

proofs with the necessary changes/additions. Please write your answers on the query sheet if there is insufficient

space on the page proofs. Please write clearly and follow the conventions shown on the attached corrections sheet.

If returning the proof by fax do not write too close to the paper’s edge. Please remember that illegible mark-ups

may delay publication.

Many thanks for your assistance.

Query reference Query Remarks

Q1 AUTHOR: Please check that authors and their affiliations are cor-

rect.

Q2 WILEY-BLACKWELL: Please check date of acceptance.

Q3 AUTHOR: Please check this website address and confirm that it is

correct. (Please note that it is the responsibility of the author(s) to

ensure that all URLs given in this article are correct and useable.)

Q4 AUTHOR: Oksanen 2006 has been changed to Oksanen et al. 2006

so that this citation matches the Reference List. Please confirm that

this is correct.

Q5 AUTHOR: Thapanand et al. 2009 has not been included in the Ref-

erence List, please supply full publication details.

Q6 AUTHOR: Please check the abbreviated journal title.

Q7 AUTHOR: Please check the abbreviated journal title.

Q8 AUTHOR: Please check the content of Table 2.

Q9 AUTHOR: Figure 1 is of poor quality. Please check required art-

work specifications at http://authorservices.wiley.com/submit_illus-

t.asp?site=1

Q10 AUTHOR: Figure 2 is of poor quality. Please check required art-

work specifications at http://authorservices.wiley.com/submit_illus-

t.asp?site=1

Q11 AUTHOR: Figure 3 is of poor quality. Please check required art-

work specifications at http://authorservices.wiley.com/submit_illus-

t.asp?site=1

Q12 AUTHOR: Figure 4 is of poor quality. Please check required art-

work specifications at http://authorservices.wiley.com/submit_illus-

t.asp?site=1

Q13 AUTHOR: Figure 5 is of poor quality. Please check required art-

work specifications at http://authorservices.wiley.com/submit_illus-

t.asp?site=1

Q14 AUTHOR: Figure 6 is of poor quality. Please check required art-

work specifications at http://authorservices.wiley.com/submit_illus-

t.asp?site=1

MARKED PROOF

Please correct and return this set

Instruction to printer

Leave unchanged under matter to remain

through single character, rule or underline

New matter followed by

or

or

or

or

or

or

or

or

or

and/or

and/or

e.g.

e.g.

under character

over character

new character

new characters

through all characters to be deleted

through letter or

through characters

under matter to be changed

under matter to be changed

under matter to be changed

under matter to be changed

under matter to be changed

Encircle matter to be changed

(As above)

(As above)

(As above)

(As above)

(As above)

(As above)

(As above)

(As above)

linking characters

through character or

where required

between characters or

words affected

through character or

where required

or

indicated in the margin

Delete

Substitute character or

substitute part of one or

more word(s)Change to italics

Change to capitals

Change to small capitals

Change to bold type

Change to bold italic

Change to lower case

Change italic to upright type

Change bold to non-bold type

Insert ‘superior’ character

Insert ‘inferior’ character

Insert full stop

Insert comma

Insert single quotation marks

Insert double quotation marks

Insert hyphen

Start new paragraph

No new paragraph

Transpose

Close up

Insert or substitute space

between characters or words

Reduce space betweencharacters or words

Insert in text the matter

Textual mark Marginal mark

Please use the proof correction marks shown below for all alterations and corrections. If you

in dark ink and are made well within the page margins.

wish to return your proof by fax you should ensure that all amendments are written clearly