Page 1
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
Page 2
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
Page 3
(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
Page 4
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
Page 5
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
Page 6
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
Page 7
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
Page 8
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
Page 9
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
Page 10
(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
Page 11
(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
Page 12
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
Page 13
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
Page 14
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
Page 15
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
Page 16
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
Page 17
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
Page 18
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
Page 19
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
Page 20
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