the stationary trawl (dai) fishery of the tonle sap-great lake

The Stationary Trawl (Dai) Fishery of the Tonle Sap-Great Lake System,

Cambodia

ISSN: 1683-1489

Mekong River Commission

MRC Technical PaperNo. 32

August 2013

C a m b o d i a . L a o P D R . T h a i l a n d . V i e t N a m

For sustainable development

The Stationary Trawl (Dai) Fishery of the Tonle Sap-Great Lake System,

Cambodia


MRC Technical PaperNo. 32

August 2013



Published in Phnom Penh, Cambodia in August 2013 by the Mekong River Commission

Cite this document as:

Halls, A.S.; Paxton, B.R.; Hall, N.; Peng Bun, N.; Lieng, S.; Pengby, N.; and So, N (2013). The Stationary Trawl (Dai) Fishery of the Tonle Sap-Great Lake, Cambodia. MRC Technical Paper No. 32, Mekong River Commission, Phnom Penh, Cambodia, 142pp. ISSN: 1683-1489.

The opinions and interpretations expressed within are those of the authors and do not necessarily reflect the views of the Mekong River Commission.

Cover Photo: J. Garrison

Editors: K.G. Hortle, T. Hacker, T.R. Meadley and P. Degen

Graphic design and layout: C. Chhut

Office of the Secretariat in Phnom Penh (OSP)576 National Road, #2, Chak Angre Krom,

P.O. Box 623, Phnom Penh, CambodiaTel. (855-23) 425 353 Fax. (855-23) 425 363

Office of the Secretariat in Vientiane (OSV) Office of the Chief Executive Officer 184 Fa Ngoum Road, P.O. Box 6101,

Vientiane, Lao PDRTel. (856-21) 263 263 Fax. (856-21) 263 264

© Mekong River CommissionE-mail: [email protected]: www.mrcmekong.org

iii

Table of contents

List of tables .... ...................................................................................................................................... vi

List of figures ......................................................................................................................................... ix

List of Plates ........................................................................................................................................ xiv

Acknowledgements .............................................................................................................................. xv

Abbreviation and acronyms ................................................................................................................. xvi

Glossary ..............................................................................................................................................xvii

Glossary of parameters (from formulas used in the paper) .................................................................. xxi

Abstract ........................... ...................................................................................................................xvii

1. Introduction ..................... ...................................................................................................................1 1.1. The Tonle Sap-Great Lake (TS-GL) System ............................................................................1 1.1.1. Location, origins and physiography .............................................................................1 1.1.2. Hydrology and hydrodynamics ....................................................................................3 1.1.3. Ecology of the Tonle Sap .............................................................................................5 Biodiversity and endemism ..........................................................................................5 The flood pulse .............................................................................................................6 1.1.4. Fisheries of the TS-GL System ....................................................................................8 Threats to fishery resources ........................................................................................12 1.2. The Tonle Sap dai fishery ......................................................................................................12 1.3. The purpose and scope of the paper .......................................................................................13 1.4. The structure of the report ......................................................................................................14

2. The Dai Fishery of Tonle Sap-Great Lake (TS-GL) System ...........................................................15 2.1. The Location of the fishery.....................................................................................................15 2.2. Target species .........................................................................................................................18 2.3. Fishing gear and operation .....................................................................................................24 2.4. Fish disposal ...........................................................................................................................30 2.4.1. Prahok ........................................................................................................................30

The Stationary Trawl (Dai) Fishery of the Tonle Sap-Great Lake System, Cambodia

iv

2.5. Economics ............................................ ...............................................................................31 2.6. Management ...........................................................................................................................35 2.6.1. Legislation ..................................................................................................................35 2.6.2. Closed season .............................................................................................................35 2.7. Effort control/Licensing .........................................................................................................36 2.8. Gear restrictions......................................................................................................................37 3. Routine monitoring activities and survey methodologies.................................................................39 3.1. Purpose of monitoring ............................................................................................................39 3.2. A history of survey methodologies and sampling regimes .....................................................39 3.3. Spatial and temporal variation in sampling effort ..................................................................43 3.4. Current survey methodology ..................................................................................................49 3.4.1. Sample stratification ...................................................................................................49 3.4.2. Data collectors and data collection process ...............................................................50 3.4.3. Variables enumerated ................................................................................................50 3.5. Catch estimation methodology ...............................................................................................52 3.5.1. Aggregated total catch................................................................................................52 3.5.2. Species-wise CPUE and catch ...................................................................................54

4. The Dai Fishery database: storage and processing ...........................................................................55 4.1. Database evolution .................................................................................................................55 4.2. Description of the current database ........................................................................................58 4.2.1. Data tables ..................................................................................................................58 4.2.2. Lookup Tables ............................................................................................................59 4.2.3. Length Frequency tables ............................................................................................60 4.2.4. Additional tables ........................................................................................................60 4.3. Current Database Query descriptions .....................................................................................60

5. The Ecology of the Fishery ...............................................................................................................61 5.1. Longitudinal (upstream/downstream) variation .....................................................................61 5.1.1. Aggregated catch, effort and CPUE ...........................................................................61 5.1.2. Catch diversity and composition ...............................................................................64 5.1.3. Fish size (weight) .......................................................................................................70 5.2. Variation in catch rates among individual dais .......................................................................71 5.3. Intra-annual variation .............................................................................................................77 5.3.1. Abundance (CPUE) ....................................................................................................77 5.3.2. Species diversity and similarity ................................................................................83

Page v

5.3.3. Size (weight) .............................................................................................................91 5.4. Inter-annual variation .............................................................................................................93 5.4.1. Catch, effort and CPUE ..............................................................................................93 5.4.2. Catch diversity and species composition ...................................................................93 5.4.3. Size (weight) ............................................................................................................101 5.5. Hydrological effects on fish biomass ...................................................................................103

6. Summary and Conclusions ............................................................................................................. 113 6.1. Management Implications .................................................................................................... 117 6.2. Recommendations ................................................................................................................ 119 6.2.1. Monitoring fisheries resources in the TS-GL and beyond ....................................... 119 6.2.2. Recommendations to improve the existing dai fishery monitoring programme .....120 6.2.3. Further research ........................................................................................................121

7. References .............................. ........................................................................................................123

8. Annex ................................... ..........................................................................................................131 8.1. Dai Fishery principal data tables (Cans and Ngor, 2006) ....................................................133 8.2. 2009 Dai Fishery database principal data tables ..................................................................137

Table of contents

Page vi

List of tables

Table 1 Dai locations in the Tonle Sap during the 2007–08 season .................................................. 16

Table 2 Mean (and range) dimensions, mesh sizes, and periods of operation of each dai net type .. 24

Table 3 Reasons reported by dai operators to mechanise hauling operations ................................... 29

Table 4 Dai Fishery profit estimates 1999–2000................................................................................ 32

Table 5 Dai unit licence allocation by type for the 2007–08 season . ................................................ 36

Table 6 Maximum gear dimensions and minimum mesh sizes for the three types of dai nets ......... 37

Table 7 Estimates of catches prior to and including the 1994–95 seasons ........................................ 41

Table 8 Monthly dai sampling regime planned for the 1997–98 season............................................ 42

Table 9 Monthly dai sampling regime proposed for the 2000–01 fishing season ............................. 44

Table 10 Monthly dai sampling regime planned for the 2002–03 fishing season ............................... 45

Table 11 Summary of changes made to the dai fishery survey design (1994–2008) and numbers of dais sampled each season. ........................................................................... 46

Table 12 The relative locations of High (shaded cells) and Low Yield (unshaded cells) dais . ........... 49

Table 13 Description of the main tables in the dai fishery database developed . ................................. 57

Table 14 Description of the main tables in the dai fishery database 2009 ........................................... 59

Table 15 Estimates of aggregated catch and effort by season and municipality. ................................. 61

Table 16 Estimates of fishing mortality, (F) for the dai fishery for a range of fishing efforts. ............ 64

Table 17 Species contributing up to 80% of mean cumulative percentage (Cum%) dissimilarity between Phnom Penh Municipality and Kandal Province in each season. ........................... 69

Page vii

Table 18 ANOVA test results for the null hypothesis that the average weight of fish (irrespective of species) landed by the fishery does not change with dai row number after accounting for the effects of month, season and mesh size ......................................... 71

Table 19 Estimates of water velocity and water depth below each dai recorded in February 2008 ... 72

Table 20 The seasonal frequency of upward (↑) and downward (↓) changes in the abundance of common species between consecutive months, 1997–98 to 2008–09. ............................. 88

Table 21 Estimated catches of the most abundant species reported for the years 1938–39 ................. 94

Table 22 Principal families and species making up 99% of the dai catches listed in descending order of their total percentage contribution ........................................................................... 98

Table 23 Summary of the indices used to describe the various attributes of the hydrograph hypothesised to affect fish biomass ..................................................................................... 106

Table 24 ANOVA table for the GLM model ...................................................................................... 107

Table 25 Summary from the existing dai database of the number of distinct dais sampled in each of the different strata between 1994 and 2008 ....................................................... 131

Table 26 tbl_MainWi&Dos ................................................................................................................ 133

Table 27 tbl_SpeciesWin&Dos .......................................................................................................... 133

Table 28 tbl_Effort ............................................................................................................................. 134

Table 29 tlkp_SeasonYear .................................................................................................................. 134

Table 30 tlkp_GearCode ..................................................................................................................... 134

Table 31 tlkp_Species ......................................................................................................................... 134

Table 32 tlkp_SpeciesStandard .......................................................................................................... 135

Table 33 tbl _Phnom Penh Port Water level ....................................................................................... 136

Table 34 tbl _MoonFace ..................................................................................................................... 136

List of tables


Page viii

Table 35 Alphanumeric gear codes in the Access database incorporating to lunar phase and dai yield (B0001–B004) ............................................................................................... 136

Table 36 tbl_MainWi&Dos ................................................................................................................ 137

Table 37 tbl_SpeciesWin&Dos .......................................................................................................... 138

Table 38 tbl_Effort ............................................................................................................................. 138

Table 39 tlkp_GearCode ..................................................................................................................... 138

Table 40 tlkp_Species ......................................................................................................................... 139

Table 41 tlkp_SpeciesStandard .......................................................................................................... 139

Table 42 Leng_tbllengthFreq (Asterisks indicate name changes or new fields) ................................ 140

Table 43 LengtblSpecies (Asterisks indicate name changes or new fields) ....................................... 141

Table 44 Leng_tblLocation (Asterisks indicate name changes or new fields) ................................... 141

Table 45 tbl_LunarAge&Phase* (Asterisks indicate name changes or new fields) ........................... 141

Table 46 tblOtherInfo* (new table) (Asterisks indicate name changes or new fields) ....................... 141

Table 47 Annual Hydrological Indices* (Asterisks indicate name changes or new fields) ............... 142

Table 48 tbl_LunarAge&Phase* (Asterisks indicate name changes or new fields) ........................... 142

Page ix

List of figures

Figure 1 The Tonle Sap–Great Lake System ........................................................................................2

Figure 2 Mean, maximum and minimum inflow (negative values) and outflow (positive values) discharges from the Tonle Sap measured in m3 s-1 at Prek Kdam Between 2004 and 2008. ........................................................................................................3

Figure 3 Monthly mean water level (amsl) and inundated lake area ...................................................4

Figure 4 Mean monthly TSS flux, concentration and water levels in the Lake for the seasons 1997–2003 ..........................................................................................................7

Figure 5 Fishing Lots in Cambodia ...................................................................................................10

Figure 6 The location of dai fishery in the Tonle Sap .........................................................................17

Figure 7 The species composition of the dai catch in 2009-10 ..........................................................18

Figure 8 Main structures of the dai fishery in the Tonle Sap .............................................................26

Figure 9 Reported mean minimum and maximum mesh sizes of each net type by dai row. ............ 27

Figure 10 Cumulative number of dai operators using diesel engines to close and haul their nets (top) and (bottom) cumulative total engine horsepower (HP) employed in the dai fishery

since 1999. ............................................................................................................................29

Figure 11 Fish disposal pathways for low value fish (LVF) or small size fish species ........................30

Figure 12 Actual (upper line) and inflation adjusted (lower line) unit price of trey riel .................... 33

Figure 13 Mean values of key economic variables for the Cambodian dai fishery plotted as a function of dai row number . .........................................................................................34

Figure 14 License costs for 25 dais plotted as a function of their annual profit ..................................35

Figure 15 Number of licensed dais ......................................................................................................37


Page x

Figure 16 Changes in sampling effort between the 1994–95 and 2008–09 expressed as the total number of hauls sampled in each season ..............................................................................46

Figure 17 Total number of hauls sampled each (a) month and (b) lunar phase (Quarters 1–4) each season .................................................................................................. 47

Figure 18 (a) Total number and (b) mean number of hauls per dai (±SD) sampled by row each season ................................................................................................48

Figure 19 Outline of the stratification of the dai fishery sampling regime ..........................................53

Figure 20 Structure of the database files held in the DOS version ARTFISH showing folders tiered by Administrative Province, season and month ............................................55

Figure 21 Structure of the database files held in the Windows version ARTFISH showing folders tiered by Season, Administrative province, month and file type . ............................56

Figure 22 Entity-relationship diagram for the dai fishery database developed ...................................57

Figure 23 Entity relationship diagram for the Dai fishery database in use from 2009. ........................60

Figure 24 Mean sampled catch rates (1997–2009) plotted by row . .....................................................62

Figure 25 The Delury depletion model fitted to mean sampled catch rates (1997–2009) expressed as the number of fish caught per dai unit per day, and cumulative fishing effort measured in dai units from the most upstream row 15. ..............................................63

Figure 26 Estimates of (a) species richness (S) (Mean±SE) and (b) Shannon Diversity Index (H) (Mean±SE) for each row from Row 1 (downstream) to 15 (upstream)

for the seasons 1997–98 to 2008–09 averaged across months and lunar phases. ................66

Figure 27 MDS ordinations based on Bray-Curtis indices of similarity derived from fourth-root transformed mean CPUE for individual dais in January during the peak lunar phase (second quarter), 1997–98 to 2002–03 seasons ....................................................................67

Figure 28 MDS ordinations based on Bray-Curtis indices of similarity derived from fourth-root transformed mean CPUE for individual dais in January during the peak lunar phase

(second quarter), 2003–04 to 2008–09 .................................................................................68

Figure 29 Estimates of mean weight of the fish assemblage sampled from the dai fishery by row and season ............................................................................................................... 70

Page xi

List of figures

Figure 30 Estimates of mean depth below each dai row ......................................................................73

Figure 31 Estimates of mean water velocity at each dai row ...............................................................74

Figure 32 The relationship between water velocity and depth ............................................................74

Figure 33 Mean loge-transformed catch rates during each lunar quarter (1–4) of the survey month (February 2008) plotted as a function of the estimated water depth below the dai unit. .....75

Figure 34 Mean loge-transformed catch rates during each lunar quarter (1–4) of the survey month (February 2008) plotted as a function of the estimated water velocity at each dai unit.......76

Figure 35 Mean dai CPUE (kg) calculated for each day of the fishing season during (a) 1997–98, (b) 1998–99 and (c) 1999–2000 .....................................................................78

Figure 36 Mean dai CPUE (kg) calculated for each day of the fishing season during (a) 2000–01, (b) 2001–02 and (c) 2002–03. ..............................................................................................79

Figure 37 Mean dai CPUE (kg) calculated for each day of the fishing season during (a) 2003–04, (b) 2004–05 and (c) 2005–06 .........................................................................80

Figure 38 Mean dai CPUE (kg) calculated for each day of the fishing season during (a) 2006-07, (b) 2007–08 and (c) 2008–09 .........................................................................81

Figure 39 Loge-transformed average daily catch rates (kg/dai/day) by (a) month and (b) lunar phase for the seasons 1997–98 to 2008–09 ..........................................................82

Figure 40 Estimates of mean (±95% CI) (a) species richness (S) and (b) Shannon Diversity Index (H) (Mean ±95% CI) of the assemblage sampled during phase 2 of the lunar cycle of each month and season ..........................................................................................84

Figure 41 Estimates of mean (±95% CI) (a) species richness and (b) the Shannon Diversity Index (H) of the assemblage sampled during each quarter (phase) of the lunar cycle in January of each season ....................................................................................................85

Figure 42 MDS ordinations illustrating rank similarities in the species assemblage sampled each month during lunar phase 2, 1997–98 to 2002–03 . ....................................................86

Figure 43 MDS ordinations illustrating rank similarities in the species assemblage sampled each month during lunar phase 2, 2003–04 to 2008–09 ......................................................87


Page xii

Figure 44 MDS ordinations illustrating rank similarities between the species assemblage sampled during each lunar phase of January, 1997–98 to 2002–03 ....................................89

Figure 45 MDS ordinations illustrating rank similarities between the species assemblage sampled during each lunar phase of January, 2003–04 to 2008–09 ...................................................90

Figure 46 Changes in mean fish weight (all species combined) through the fishing season (1997–08 to 2008–09) .........................................................................................................91

Figure 47 Mean weight changes by month for the six most abundant species (1997–08 to 2008–09) ..........................................................................................................92

Figure 48 (a) Total Catch, (b) effort and (c) CPUE (1997–8 to 2008–9). ............................................93

Figure 49 The species composition of the dai fishery catch, 1997–98 to 2008–09. .............................95

Figure 50 (a) The Species Richness (S) (Mean ±95% CI) and (b) Shannon Diversity Indices (H) (Mean ±95% CI) of fish species caught by a dai on a day for the seasons

1997–98 to 2008–09 averaged across all sampling months in each season. ........................97

Figure 51 MDS ordinations based on Bray-Curtis indices of similarity derived from fourth-root transformed mean CPUE for individual dais for the 1997–98 to 2008–09 seasons

for (a) October and (b) November. .......................................................................................99


for (a) December and (b) January. ......................................................................................100


for (a) February and (b) March. ..........................................................................................101

Figure 54 Mean weight of fish landed (all species combined) 1997–2009 ........................................102

Figure 55 Trends in the mean weight of the six most abundant species in the fishery estimated for December each year (1995–06 to 2008–09). ................................................102

Figure 56 The location of the Kampong Luong gauging station in the TS-GL. .................................104

Figure 57 Daily water level (m) measured at Kampong Luong gauging station and estimates of the flooded area of the TS-GL System, 1997–2009. ......................................................105

Page xiii

List of figures

Figure 58 Illustration of the estimation of hydrological indices. .......................................................105

Figure 59 The GLM observed and predicted mean catch rates 1997–2009 ......................................108

Figure 60 Model Residuals .................................................................................................................108

Figure 61 The relationship between the mean predicted daily catch rate of a dai unit during the fishing season and the flood index (FI) for the TS-GL System. .......................109

Figure 62 The relationship between mean sampled fish weight (all species combined) and the flood index with fitted exponential mode .............................................................109

Figure 63 Mean weight estimates for the six most abundant species in the fishery in December each year .......................................................................................................110

Figure 64 Changes in mean fish weight and the flood index from 1997–98 to 2008–09. ..................110

Figure 65 Fish abundance plotted as a function of the flood index. ...................................................111.

Page xiv

List of plates

Plate 1 Arrow fence traps constructed in the flooded forest margins of the Lake . ............................ 11

Plate 2 Rows of dai nets anchored across the Tonle Sap ...................................................................15

Plate 3 Cirrhinus lobatus (trey angkam) ...........................................................................................19

Plate 4 Henicorhynchus cryptopogon (trey riel angkam) ...................................................................19

Plate 5 Paralaubuca barroni (trey slak russey) .................................................................................20

Plate 6 Labiobarbus lineatus (trey khnawng veng) ...........................................................................20

Plate 7 Henicorhynchus siamensis (trey riel tob) ..............................................................................21

Plate 8 Labiobarbus siamensis (trey ach kuk) ...................................................................................21

Plate 9 Labeo chrysophekadion (trey kaek) .......................................................................................22

Plate 10 Pangasius pleurotaenia (try chhwiet) ....................................................................................22

Plate 11 Puntioplites proctozysron (trey chrakeng) .............................................................................23

Plate 12 Thynnichthys thynnoides (trey linh) ......................................................................................23

Plate 13 Emptying the codend into sorting compartments ..................................................................28

Plate 14 The codend being emptied directly into traders boats ...........................................................28

Plate 15 Diesel engines used to haul the dai net .................................................................................28

Page xv

Acknowledgements

This paper is an output of the Fisheries Ecology Valuation and Mitigation (FEVM) Component of the MRC’s Fisheries Programme. The paper aims to improve stakeholder capacity to monitor and evaluate the status and trends of fisheries resources in the Lower Mekong Basin (LMB) in the context of fisheries management and basin development activities (FEVM Logframe Outputs 1, 3 & 5).

The paper also forms an output of the ACIAR-funded project: Analyses of three databases of fisheries data from the Mekong River (FIS/2006/137) - a collaborative research project between the MRC, IFReDI, LARReC and Murdoch University, funded by ACIAR.

The routine fishery monitoring activities described in Section 3.4 were initiated by Niek van Zalinge and later managed by Kent G. Hortle, Niklas Mattson and then Ashley Halls. The field work was supervised by Ngor Peng Bun, Deap Loueng, Yim Chea, Heng Kong, Chhoun Chamnan, Souen Sotthia and Lieng Sopha. Ngor Pengby and Tan Phalla have managed and improved the database for the past five years. Their outstanding efforts, including those of the many field enumerators, and the cooperation of the dai operators, are gratefully acknowledged.

We are grateful to Kent G. Hortle and Nao Thuok for their reviews and comments which significantly improved an earlier draft of the paper.

The preparation of this paper was facilitated by the MRC Fisheries Programme with funding from DANIDA, SIDA and ACIAR.

Page xvi

Abbreviations and acronyms

TAMCF Assessment of the Mekong Capture Fisheries (Programme)AMSL Above Mean Sea LevelANOVA Analysis of VarianceARTFISH Approaches, Rules and Techniques for Fisheries Statistical Monitoring (FAO Database)CPUE Catch Per Unit of EffortDoF Department of FisheriesDOS Disk Operating SystemFEVM Fisheries Ecology Valuation and MitigationFI Flood IndexFiA Fisheries AdministrationIFReDI Inland Fisheries Research and Development InstituteLMB Lower Mekong BasinLVF Low Value FishMDS Multi-dimensional ScalingMFCF Management of the Freshwater Capture Fisheries (Programme)MFD Mekong Fish DatabaseMRC Mekong River CommissionMRCS Mekong River Commission SecretariatPRIMER Plymouth Routines in Multivariate Ecological Research (software)SD Standard DeviationSE Standard ErrorSIMPER Similarity Percentage (sub-routine in PRIMER)TSBR Tonle Sap Biosphere ReserveTS-GL Tonle Sap-Great LakeTSS Total Suspended SolidsUNESCO United Nations Educational Scientific and Cultural Organisation

Page xvii

Glossary

Analysis of variance Analysis of variance (ANOVA) is a collection of statistical models, and their associated procedures, in which the observed variance in a particular variable is partitioned into components attributable to different sources of variation. In its simplest form ANOVA provides a statistical test of whether or not the means of several groups are all equal, and therefore generalizes t-test to more than two groups. Doing multiple two-sample t-tests would result in an increased chance of committing a type I error. For this reason, ANOVAs are useful in comparing two, three or more means.*

ARTFISH A standardized tool adaptable to most fisheries in the developing countries. Its design was driven by the need to provide users with robust, user-friendly and error-free approaches with computer software, and achieve the implementation of cost-effective fishery statistical systems with minimal external assistance.**

Blackfish Species that possess morphological and physiological adaptations to extreme environmental conditions including low dissolved oxygen concentrations, and desiccation.

Bray-Curtis (dis)similarity Index The Bray–Curtis dissimilarity is a statistic used to quantify the compositional dissimilarity between two different sites. It is equivalent to the total number of species that are unique to any one of the two sites divided by the total number of species over the two sites. In other words, it is the ratio between the turnover of species between the two sites and the total species richness over the two sites.*

Catchability coefficient The proportion of the population removed by one unit of effort.

Coefficient A multiplicative factor in some term of an expression.


Page xviii

Coefficient of determination The coefficient of determination R2 is used in the context of statistical models whose main purpose is the prediction of future outcomes on the basis of other related information. It is the proportion of variability in a data set that is accounted for by the statistical model. It provides a measure of how well future outcomes are likely to be predicted by the model.*

Correlation coefficient A measure of the strength of the linear relationship between two variables.*

Delury Depletion Model A method to estimate animal abundance by monitoring how indices of abundance (e.g. catch rates) decline in response to cumulative fishing effort.*****

Flood Index A quantitative description of the extent and duration of flooding corresponding to the area beneath and the area-duration curve above mean flood levels.

General Linear Model (GLM) The general linear model incorporates a number of different statistical models. * In this document GLM provides a general version of multiple linear regression where explanatory variables take the form of factors and covariates.

Lunar phase or (lunar quarter) Lunar quarters relate to four consecutive seven day periods starting from the new (dark phase) moon. Quarter 2, when catch rates in the dai fishery are observed to peak, corresponds to the period of approximately 7–14 days after the new moon when between approximately 50–100 % of the moon is visible. This period between what are commonly termed the first quarter and full moon phases is also known as the ‘Waxing Gibbous’ phase.

Multi-dimensional scaling (MDS) Multidimensional scaling (MDS) is a set of related statistical techniques often used in information visualization for exploring similarities or dissimilarities in data. MDS is a special case of ordination. An MDS algorithm starts with a matrix of item–item similarities, then assigns a location to each item in N-dimensional space, where N is specified a priori. For sufficiently small N, the resulting locations may be displayed in a graph or 3D visualisation.*

Page xix

Glossary

Population dynamics Population dynamics is the branch of life sciences that studies short-term and long-term changes in the size and age composition of populations, and the biological and environmental processes influencing those changes. Population dynamics deals with the way populations are affected by birth and death rates, and by immigration and emigration, and studies topics such as ageing populations or population decline.*

Primary production Primary production is the production of organic compounds from atmospheric or aquatic carbon dioxide, principally through the process of photosynthesis, with chemosynthesis being much less important. Almost all life on earth is directly or indirectly reliant on primary production. The organisms responsible for primary production are known as primary producers or autotrophs, and form the base of the food chain.*

PRIMER (Plymouth Routines In Multivariate Ecological Research Version 6): PRIMER

6 is a collection of specialist routines for analyzing species or sample abundance (biomass). It is primarily used for ecological and environmental studies. Multivariate routines include:• grouping (CLUSTER);• sorting (MDS);• principal component identification (PCA);• hypothesis testing (ANOSIM);• sample discrimination (SIMPER);• trend correlation (BEST);• comparisons (RELATE); and• diversity, dominance, and distribution calculating. ***

Recruitment The number of fish (recruits) added to the exploitable stock, in the fishing area, each year, through a process of growth (i.e. the fish grows to a size where it becomes catchable) or migration (i.e. the fish moves into the fishing area).**


Page xx

Shannon Diversity Index The Shannon Diversity index, sometimes referred to as the Shannon-Wiener Index is one of several diversity indices that can be used to measure species diversity. The advantage of this index is that it takes into account the number of species and the evenness of the species. The index is increased either by having additional unique species, or by having a greater species evenness.*

SIMPER Identifies the species primarily providing the discrimination between two observed sample clusters.*** (See PRIMER)

Species richness Species richness is the number of different species in a given area. It is represented in equation form as S. Typically, species richness is used in conservation studies to determine the sensitivity of ecosystems and their resident species. The actual number of species calculated alone is largely an arbitrary number.*

Standard deviation Standard deviation shows how much variation or "dispersion" exists from the average. A low standard deviation indicates that the data points tend to be very close to the mean, whereas high standard deviation indicates that the data points are spread out over a large range of values. The standard deviation of a statistical population, data set, or probability distribution is the square root of its variance.*

Standard error An estimate of that standard deviation, derived from a particular sample used to compute the estimate*

Survey stratification The process of dividing members of the population into homogeneous subgroups (stratum) before sampling to reduce sample variance.

Type I Error Falsely rejecting the null hypothesis when it is true.

Whitefish Migratory species intolerant of low dissolved oxygen conditions and typically inhabit lotic (flowing water) environments.

*http://en.wikipedia.org** http://www.fao.org*** http://www.primer-e.com/index.htm***** Hilborn & Walters (1992)

Page xxi

Glossary of parameters (from formulas used in the paper)

Α, a ConstantΒ, b CoefficientAD Active Days (fishing days)AG Active Gears (dai units)C CatchCPUE Catch Per Unit of EffortD Flood durationd A given day of the yearDSI Dry Season Indexe Effort (dai units)E (Cumulative) fishing efforte Exponentialε Error (residual)F Instantaneous fishing mortality rateFA Flooded AreaFC Fixed CostsFE Flood EndFS Flood Starth HaulH Shannon Diversity IndexKAN Kandal lp Lunar Phase (1–4)m Calendar monthN Fish abundance (number of fish)p Administrative zone (Province)p Probability of committing a Type I Errorpi Relative abundance of species iPP Phnom Penhq Catchability coefficientr Dai rowr Correlation coefficientR2 Coefficient of determinationS Species richnessTL Total LengthTVC Total Variable CostsVC Variable CostsWL Water Levelwt Sampled haul weighty Yield or year

Page xxii

Abstract

The Tonle Sap-Great Lake (TS-GL) system is an integral part of the history, culture, ecology and economics of the Mekong region. Species of blackfish and whitefish are the target of industrial, artisanal and subsistence fisheries operating in the TS-GL System. Strong competition exists among the fisheries to land in excess of 200,000 tonnes of fish each year equivalent to approximately 10% of the total weight of fish consumed in the entire Lower Mekong Basin (LMB) each year. The dai fishery on the Tonle Sap, established almost 140 years ago, is an important component of the industrial fishery, landing approximately 14% of the annual catch taken from the TS-GL System. It targets the refuge migrations of a multi-species assemblage of fish as they migrate from the Great Lake to the Mekong main channel via the Tonle Sap with the receding floodwaters each year. In addition to its significant socio-economic value both locally and nationally, the dai fishery provides a valuable source of data and information to monitor trends in migratory fish populations which seasonally utilise the TS-GL System and beyond. This paper represents the first attempt to compile and analyse the available data and information about the Cambodian dai fishery in a single document. It therefore serves as an important reference document for present and future workers involved in the management, monitoring and administration of the fishery. It also contains new insights into the ecology and dynamics of target fish populations important for their management. The fishery exhibits considerable spatial and temporal variation in catch rate indicators of fish biomass and abundance. This reflects (pulsed) migratory behaviour associated with the lunar cycle, the hydrological cycle (drawdown effects), depletion effects as fish migrate through the fishery, and inter-annual hydrological effects on fish growth and biomass. Above average levels of recruitment were probably responsible for the very high catches observed during 2004–05 and 2005–06. Factors responsible for these high levels of recruitment remain uncertain. There is evidence that the timing of migrations is species and size-dependent with larger species and larger individuals of the same species migrating earlier than smaller fish. These responses are consistent with earlier studies on the system and in other tropical river systems. Whilst the dai fishery is the focus of most fisheries monitoring and evaluation efforts in Cambodia, it is not the only, nor most significant, component of the entire TS-GL fishery. Other components with which it interacts and competes with, particularly the other lot, artisanal and subsistence fisheries, are also significant and therefore must be given greater consideration in the future. In spite of the present restricted focus, the monitoring efforts directed at the dai fishery have generated the only continuous long-term data set for an inland fishery in Cambodia. Analyses of indicators estimated from this data set described here have been informative for policy and management evaluation, revealing little or no compelling evidence of changes in the abundance, biomass, size or diversity of migratory fish populations that seasonally utilise the TS-GL System and beyond often over distances of more than 600 km. Furthermore, time series of these indicators have equipped managers with an important baseline against which to monitor any impacts of management and basin development activities. A key finding of this research is that inter-annual variation in the biomass of the multispecies assemblage targeted by the fishery (and hence landings) can be largely explained by flood duration and extent effects on fish growth. Fish growth, indicated by mean fish

xxiii

weight, increases exponentially with the flood extent and duration presumably reflecting changes in feeding opportunities or competition. This response has been modelled, allowing predictions to be made of how the relative biomass of the multispecies assemblage targeted by the fishery (and hence catches) are likely to vary under different flooding conditions, whether natural or modified as a consequence of climate change and/or water management projects in the Basin. Owing to the highly migratory nature of the target fish species, these predicted hydrological responses may be observed over large distances, affecting fisheries and piscivorous fish populations beyond the immediate vicinity of the system. The unexplained variation in the model may well reflect variation in recruitment to the system each year in addition to variation in fishing effort (mortality) applied by the other important fisheries within the TS-GL System or over the migratory range of the target species, reinforcing the need for two more comprehensive monitoring programmes. These results also urge caution when monitoring mean fish size as a proxy for rates of exploitation in the TS-GL System and other highly fluctuating environments. By applying depletion model theory, this research has provided the first estimates of the proportion of fish removed over the range of the fishery, dai gear catchability (efficiency), and dai fishing mortality rates subject to a number of assumptions. These results are an important first step towards understanding, and thereby controlling if necessary, the relative sources of fishing effort (mortality) over the migratory range of populations of important species of fish. Additional studies and monitoring programmes will be necessary to determine the validity of the assumptions underlying these estimates and to quantify the spatial distribution of the remaining sources of fishing mortality in the TS-GL System and beyond. Recommendations are made for these studies and programmes, as well as to improve the existing dai fishery monitoring programme.

Abstract

1. Introduction

1.1. The Tonle Sap-Great Lake (TS-GL) System

Rising in the Himalayan highlands of China, the Mekong River flows through Myanmar, Lao PDR, Thailand and Cambodia before discharging into the South China Sea in Viet Nam. At 4,880 km, it is the 12th longest river in the world and 10th largest in terms of water volume (Gupta and Liew 2007; Kummu and Sarkkula, 2008). In Cambodia it empties onto a vast alluvial plain where its channel becomes less well-defined and where more hydraulically complex floodplain processes prevail (MRC, 2005). Roughly 4,300 km from its source at the Chaktomuk junction near the city of Phnom Penh it is met on its right bank by the Tonle Sap. The Great Lake–the largest wetland in Southeast Asia (Kummu et al., 2008)–is situated at the apex of the Tonle Sap around 130 km to the northwest of this confluence. These two water bodies–the Tonle Sap coupled to the Great Lake–form the Tonle Sap-Great Lake (TS-GL) system.

The Tonle Sap-Great Lake system is an integral part of the history, culture, ecology and economics of Southeast Asia. It plays a crucial role in mitigating floods in Cambodia and Viet Nam, particularly in the case of extreme flood events (Hai et al., 2008), provides habitats for a diversity of aquatic and terrestrial plant and animal species (Campbell et al., 2006) and an important source of raw materials, nutrition, income and livelihoods for upwards of one million people living in and around it (Keskinen, 2003; Sarkkula et al., 2003 and Lamberts, 2006). Together with rice cultivation, fishing is amongst the most important of these livelihood activities (Keskinen, 2003; Ahmed et al., 1996) and areal yields from the Lake have been estimated to be 230 kg ha-1 year-1 (Baran et al., 2001).

1.1.1. Location, origins and physiography

The TS-GL System is one of the most distinctive hydro-geomorphic features within the larger Mekong River basin. The Great Lake depression that today comprises the Lake itself was believed to have formed through the subsidence of the Cambodian platform around 5,700 years ago (Carbonel, 1963 cited in Campbell et al., 2006). Higher-than-present rainfall and sea level transgression suggest that, at this time, the depression was inundated and that there was less seasonal variation in depth and possibly tidal influence in the vicinity of the modern lake (Penny, 2006). From the mid-Holocene onwards, however, a weakening of the southwest (wet) monsoon led to drier, more seasonal climate. These factors, combined with sea level regression would have contributed to hydrological and hydrodynamic conditions in the Lake becoming similar to what they are today (Penny 2006; Penny 2008).

The catchment of the Lake basin covers an area of around 67,000 km2 (Ahmed et al., 1996), bounded to the north by the Dongrek mountain range and to the southwest by the Cardamom and


Page 2

Figure 1 The Tonle Sap–Great Lake System.

Elephant ranges. It is separated into a northern and a southern basin with a deltaic region at its southern end formed by the deposition of sediments brought in from the Mekong by the Tonle Sap (Figure 1).

1.1.2. Hydrology and hydrodynamics

During the dry season (October–May) water drains from the TS-GL System into the Mekong River via the Tonle Sap. As the wet season advances from June/July onwards, however, flooding in the delta downstream of Phnom Penh causes water levels in the Mekong River to rise higher than those in the Great Lake. This causes flow in the Tonle Sap to reverse, and instead of draining into the Mekong, the waters are pushed back upstream towards the Great Lake, inundating its floodplains. At the peak of the flood the aerial extent of the Lake will increase between three and six times (van Zalinge et al., 2004). Towards the end of the flood, backed-up waters in the Lake and concurrently subsiding water levels in the Mekong, cause the flow in the Tonle Sap to reverse once more. The waters are then carried out of the Lake, back into the Mekong River and towards the delta.

By the height of the flood season in August, discharges through the Tonle Sap into the Great Lake peak at approximately 10,000 m3 s-1 or roughly 25% of the mean mainstream flows at this time of year (MRC, 2005). After the peak of the flood, in November, discharges once more decline and the backed up waters in the Great Lake begin to drain back into the Mekong at similar rates (10,000 m3 s-1) to the inflow (MRC, 2005) (Figure 2).

Figure 2 Mean, maximum and minimum inflow (negative values) and outflow (positive values) discharges from the Tonle Sap measured in m3 s-1 at Prek Kdam Between 2004 and 2008.

Introduction


Page 4

At the height of the flood, the inundated area of the Great Lake increases from approximately 3,500 km2 during the dry season to 14,500 km2 as water levels rise over the wet season (Figure 3, Kummu et al., 2008). Over this same period the Lake volume will increase from 1.5 km3 to 60–70 km3 (MRC, 2005).

0

2

4

6

8

10

12

14

0

2000

4000

6000

8000

10000

12000

14000

May Jun

Jul

Aug

Sep Oct

Nov

Dec Jan

Feb

Mar

Apr

Wat

er le

vel [

m],

AMSL

in H

atie

n

Area

[km

2 ]

area

water level

Rising flood Receding flood

Figure 3 Monthly mean water level (amsl) and inundated lake area. Error bars represent minimum and maximum mean monthly water levels (Kummu et al., 2008).

The mean annual inflow to the Tonle Sap is 79 km3–roughly comparable with its outflow (78.6 km3) and it exhibits high inter-annual variability (44.1 km3 in 1998 and 106.5 km3 in 2000) (Kummu and Sarkkula, 2008). Minimum depths of 0.5 m in April increase to depths of around 6–9 m at the height of the flood between late September and early October (MRC, 2005). Campbell et al. (2006) noted a statistically significant decline in the maximum and minimum water levels of 0.52 m between the years 1925–35 and 1996–2002. This translates to reduction in inundated area of 623 km2. They rule out the possibility of gauging error to explain this decline and suggest instead that a reduced rainfall and a minor reduction in flows from the Tonle Sap may be responsible.

Of the flows entering the Lake, 52% originate from the Mekong River via the Tonle Sap and 5% from overland flows from the mainstream over the flood season. The remainder of the flows originates from the Lake’s tributaries (30%) or directly from rainfall (13%) (Baran et al., 2007). The left

bank tributaries of Lao PDR together with the Se Kong, Se San and Sre Pok Rivers, that combined, contribute 25% of the mean annual flow volume of the Mekong River at Kratie, are important components of the flow regime in the Lower Mekong Basin and key to driving the flow reversal of the Tonle Sap (MRC, 2005). Ecological processes in the TS-GL System are therefore vulnerable to hydrological modification in upstream locations caused by water management structures and abstractions.

1.1.3. Ecology of the Tonle Sap

Biodiversity and endemism

The status of the TS-GL System as a wetland of international ecological and conservation importance was reinforced when it was approved by UNESCO as the Tonle Sap Biosphere Reserve (TSBR) in 1997 (Davidson, 2006). Many of the plant and animal communities in the Lake itself and the surrounding floodplains are adapted to the fluctuating water levels which is both a key driver of ecosystem processes and a major factor affecting the species composition in different areas of the Lake.

Davidson (2006) estimated that roughly 200 vascular plants occur on the surrounding floodplains. One of the most distinctive plant communities on these floodplains are the flooded or swamp forests that occupy about 10% of land on the floodplain area and are situated around the edge of the central portions of the Lake that are permanently inundated (Campbell et al., 2006). They comprise woody tree species that vary in height between 7–15 m, the dominant species being Barringtonia acutangula (Lecythidaceae). Other species that are also found in this community include Diospyros cambodiana, Terminalia cambodiana, a local endemic and Samandura harmandii, a narrow endemic (Campbell et al., 2006; Davidson, 2006). The swamp forests become submerged for up to 6 months of the year between August and January during which time the majority of the deciduous tree species loose their leaves. Flowering and fruiting occurs in the late dry/early wet season after which the trees drop their seeds to be carried away in the flood (Davidson, 2006). Aside from the swamp forests, the swamp scrubland vegetation community that occupies a much greater proportion of the floodplain (up to 80%) also consists of woody species, but these do not exceed more than 4 m in height. The scrubland vegetation community is dominated by Euphorbiaceae, Fabaceae and Combretaceae (Campbell et al., 2006). These flooded forests are critical for ecological processes in the basin, providing spawning, nursery and feeding habitats for fish and wildlife (Ahmed et al., 1996).

Although mammal diversity is not high around the Great Lake, distinctive primates such as the slow loris Nycticebus coucang, long-tailed macaque Macaca fascicularis and silvered langur Semnopithicus cristatus have all been recorded in the Prek Toal Core Area of the TSBR (Campbell et al., 2006). The Prek Toal is also recognised as important site for breeding colonies of waterbirds, supporting about 100 species of which 13 are of global significance (Clements et al., 2007). Near-endemic species include giant ibis Pseudibis gigantea–the largest of the world’s ibises and the only

Introduction


Page 6

Indochinese population of the critically endangered Bengal florican bustard Houbaropsis bengalensis is restricted to the grasslands surrounding the Tonle Sap (Gray et al., 2009) and Prek Toal supports the only known Southeast Asian breeding colonies of the globally threatened spot-billed pelican Pelecanus philipensis and milky stork Mycteria cinerea (Clements et al., 2007).

The TS-GL System also supports a diversity of reptiles including the Siamese crocodile Crocodylus siamensis, although numbers are low (Cambell et al., 2006). Seven species of homalospine watersnake also inhabit the system including the endemic Tonle Sap watersnake Enhydris longicauda (Stuart et al., 2000).

Occupying the permanently inundated waters and the floodplain during the wet season are upwards of 149 fish species in 35 families, although the exact number is difficult to determine due to absence of extensive surveys, survey objectives and taxonomic uncertainties (Campbell et al., 2006; Junk et al., 2006). Although the Mekong River has a relatively high level of fish endemism (24%), none of these species are restricted to the TS-GL System itself (Campbell et al., 2006). Five of the fish species that occur in the TS-GL System are globally threatened, one of which is the Giant Mekong catfish Pangasianodon gigas (Davidson, 2006).

The flood pulse

As water levels in the main channel begin to rise each year and flow into the Tonle Sap they transform the previously dry terrestrial floodplain habitat surrounding the Lake into an ephemeral aquatic habitat. Together with the increase in the amount of habitat available to aquatic organisms, nutrients and carbon are exchanged between the land and water giving rise to a peak in primary productivity. Invertebrates, fish and plants adapted to living in river-floodplains systems exploit this increased availability of food and habitat for feeding and sheltering their young and have therefore evolved to synchronise their reproduction and recruitment to coincide with this flooding period. The central importance of the flood to these processes led (Junk et al., 1989) to conclude that this ‘flood pulse’ was “the principal driving force responsible for the existence, productivity and interactions of the major biota in river-floodplains systems”. It therefore follows that any natural or anthropogenic changes to either the magnitude, timing, duration of the flood, or speed with which the floodwaters rise (rate of change), will affect the overall productivity of the system.

Floodwaters carry sediments and nutrients essential to supporting ecosystem processes in the Lake and on its floodplain. The combined mean annual suspended sediment flux into the lake from the Mekong River (72%) and tributaries (28%) is estimated to be 7–9 million t y-1. Most sediments are deposited in the Lake or are trapped by the floodplain vegetation around the Lake margins (Sarkkula et al., 2003; Kummu et al., 2008). However, deposition and erosion of sediments are considered to be in balance and net sedimentation is therefore considered to be at or close to zero (Sarkkula and Koponen, 2003). There is considerable intra-annual variation in total suspended solid (TSS) concentrations that vary between 3.5 × 109 kg to 9 × 109 kg. TSS increases on the rising flood

and reaches its peak in July and August (Figure 4). It begins to decline thereafter on the receding flood, reaching its lowest levels at the beginning of the dry season (Kummu et al., 2008).

Inter-annual variability in river flows are strongly reflected in the sedimentation patterns. During a dry year, very little sediment reaches the Lake, whereas during a wet year, the sediments become widely distributed through the Lake system (Kummu et al., 2005). Long-term sedimentation rates have been reported at less than 1 mm y-1. Although this is not considered a significant threat to the Lake itself (Penny et al., 2005) caution that the channel morphology at the mouth of the Tonle Sap where it enters the Lake may be more sensitive to altered sedimentation regimes.

The TS-GL System is mesotrophic and most likely phosphorous limited, particularly over the high water period. Most of this phosphorous is sediment-bound and believed to be brought by the flood (Sarkkula and Koponen, 2003; Sarkkula et al., 2003). This sediment-bound phosphorous is made available to phytoplankton through the growth and subsequent decomposition of vascular plants growing in the aquatic-terrestrial transition zone (Kummu et al., 2008).

Oxygen concentrations in the Lake differ spatially (between the Lake and its floodplain), as well as through the seasons. In the permanently inundated central portion of the Lake, wind and wave mixing

-500

0

500

1000

1500

2000

2500

May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr

TSS

flux

[10^

6 kg

]

-30

0

30

60

90

120

150

TSS

conc

entr

atio

n [m

g/l]

/ W

ater

leve

l [dm

], A

MSL

Flux in

Flux out

TSS conc.

water level

Rising flood Receding flood

Figure 4 Mean monthly TSS flux, concentration and water levels in the Lake for the seasons 1997-2003 (Kummu et al. 2008).

Introduction


Page 8

ensures well-oxygenated water is distributed throughout the water column, even during the flood. In parts of the floodplain during the high water period, particularly in the North where water velocities are lower and there is sheltering by vegetation and little wind, anoxic conditions predominate in all but the uppermost layers of the water column due to oxygen consumption by decomposing organic matter of terrestrial origin inundated by the flood (Sarkkula et al., 2003 and Sarkkula et al., 2004). Oxygen concentrations measured at the floodplain-lake interface exhibit high variability possibly because of adventive mixing. The general trend in this region is towards an increase in concentration from dry season (July) to the peak of the flood (November) (Sarkkula and Koponen, 2003).

Sediment-bound phosphorous is made available to the aquatic food web through assimilation by higher plants growing in the aquatic-terrestrial transition zone (Sarkulla et al., 2003). Phytoplankton growth occurs during the phosphorous-limited high water period (October–December) when Aulacoseira diatoms and copepods dominate the pelagic trophic structure. During the low water period when turbidity is high, phytoplankton growth is concentrated near the surface layers where the predominant algae are positively buoyant Anabaena species (Sarkulla et al., 2003; Sarkulla and Kopenen, 2003). Algal blooms are not common due to high zooplankton biomass suggesting that nutrients are taken up and processed in the ecosystem very efficiently (Sarkulla and Koponen, 2003).

1.1.4. Fisheries of the TS-GL System

The fish fauna exploited by the fisheries of the TS-GL System can be separated into two principle ecological guilds: (1) Blackfish that are tolerant of anoxic conditions and may undertake lateral migrations between the Lake and floodplain and (2) Whitefish that undertake longitudinal migrations between the Lake and the Mekong River via the Tonle Sap. Amongst the latter, the Cirrhinus and Henicorhynchus genera are amongst the most important for river fisheries, whereas the former include genera such as Channa, Trichogaster, Anabas, Oxyeleotris and Mystus (Lim et al., 1999). Both groups depend on the advance and retreat of flood waters across the floodplains. An estimated 7 million snakes or 777 tonnes are removed from the Lake annually (Brooks et al., 2007). Snakes provide an inexpensive source of protein but current rates of exploitation are believed to be unsustainable (Brooks et al., 2007).

Under Cambodian Fishery Law, the fisheries exploiting these resources are divided into two broad categories: limited and open-access fisheries. Fishing takes place in two seasons: Open (October–May) and closed (June–September).

The limited access fisheries are managed by means of ‘fishing lots’ and may operate only during the open season. Fishing lots vary from a simple anchoring position for dais in the Tonle Sap to large areas (up to 500 km2) of floodplain around the Great Lake, Tonle Sap, Mekong, Bassac and other rivers and tributaries (Figure 5). Lots in the TS-GL System contain mostly natural habitats including flooded forests, shrub forest and grasslands, but rice fields and villages are sometimes within their boundaries. Approximately 35 fishing lots exist in the System fished using large scale gear

(van Zalinge, 2002). Many lots operate large-scale seines and bamboo fence traps to target blackfish species such as Channa (see Plate 1). Barrage traps are also set in the delta region of the TS-GL System to target longitudinally migrating whitefish (Lamberts, 2001). Exploitation rights for these large-scale (industrial) fisheries are issued by the Fisheries Administration on a licence basis. Large-scale fisheries licences are issued for two years. Licence revenue exceeds millions of dollars annually (van Zalinge, 2002).

Open access fisheries comprise two main components: Middle-scale (artisanal) fisheries that can operate only during the open season and family (subsistence) fisheries that can operate year-round and in flooded rice fields (Sam et al., 2002).

Artisanal and subsistence fisheries may operate in open access areas. These fisheries employ a wide variety of gear types including gillnets, seines, bamboo fence traps, cast nets and longlines. Illegal fishing gear include the use of explosives, electro-fishing and poisoning (Lamberts, 2001). Watersnakes are gillnetted, caught in bamboo traps, noosed or speared (Stuart et al., 2000).

Competition for fish resources among sectors is intense as illustrated by artisanal and subsistence fishers operating from small boats between rows of dais in the Tonle Sap. During the 2004–05 open season, reports of high catch rates also attracted thousands of additional artisanal and subsistence fishers from throughout Cambodia’s lowlands. In January 2005, an estimated 7,000 boats were fishing on the Tonle Sap including at times between rows of dai nets and around the Mekong Junction. Most fishers were using drifting gill nets, with some using larger gear such as seines and trawls. The total catch from these boats was estimated at about 4,000–5,000 tonnes in February 2005, a similar amount to the dai catch that month (Hortle et al., 2005). Fish not caught by dai nets or other industrial gear are vulnerable to capture by these competing fisheries as demonstrated by a tag-release study (Hortle et al., 2004a). Many thousands of people also fish along the banks of the Tonle Sap using cast-nets and small gear. Floodplains along the Tonle Sap are also fished where they drain into the Tonle Sap by small-scale fishers and formerly by some commercial lot fishers (Hortle pers. Comms; Valbo-Jorgensen et al., 2001 and Dubeau et al., 2001).

With the exception of the dai fishery, most of fisheries of the TS-GL System have not been subject to detailed investigation or systematic routine monitoring, and are generally poorly documented. The implications of this limited focus for interpreting data generated by the dai monitoring programme, and in the wider context of fisheries policy and management activities are discussed in Section 6.

It has been estimated that the livelihoods of more than a million people depend on the fish resources of the TS-GL System (van Zalinge, 2002). Most of these fishers are engaged in the open-access fisheries, living at the edges of the floodplain or in floating villages or houses built on tall stilts (van Zalinge, 2002).

Hap et al. (2006) reported that the total annual fish yield landed by all fisheries operating in the TS-GL System to be in the order of between 200,000 and 218,000 tonnes with a landed value of US$ 150–250 million. Van Zalinge (2002) estimated the yield to be approximately 235,000 tonnes.

Introduction


Page 10

Figure 5 Fishing Lots in Cambodia

Plate 1 Arrow fence traps constructed in the flooded forest margins of the Lake

Introduction


Page 12

These estimates are equivalent to approximately 50% of total annual yield of inland fish for Cambodia (480,000 tonnes) estimated by (Hortle, 2007). Lamberts (2006) cautions that yield estimates for TS-GL System may be unreliable due to the absence of quantitative surveys.

Van Zalinge et al. (2004) hypothesized that three factors affect fish production in the system: (1) nutrient bearing sediment that influences rates of primary production; (2) oxygen conditions on the floodplain that effectively determines habitat availability and (3) the transport of fish eggs and larvae to the system.

Threats to fishery resources

Campbell et al. (2006); Keskinen (2003) cite human population growth and concomitant over-harvesting of resources together with potential future hydrological changes as amongst the principle threats to fisheries and wildlife in the TS-GL System. To this, Hortle et al. (2004a) add habitat destruction and fragmentation by dams. It has been estimated that up to a third of flooded forests was lost between 1973 and 1993 in response to increasing demand for rice growing land and firewood, and subsequent soil errosion (Nao and van Zalinge, 2001). Lim et al. (2004) report that the construction of rural water irrigation infrastructure has also impeded fish migrations in the system. According to anecdotal reports, large and medium sized fish have become scarcer in response (van Zalinge et al., 1998 and van Zalinge, 2002). Water management projects in upstream locations may pose the greatest threat since they are likely to alter the magnitude and timing of flows, as well as trap sediments and nutrients (Ahmed et al., 1996; Lamberts, 2008).

Conservation of this valuable resource and predicting how it will respond to future anthropogenic interventions requires understanding of the physical and biological processes that support its productivity.

1.2. The Tonle Sap dai fishery

The dai fishery on the Tonle Sap, established almost 140 years ago, is an important component of the large scale (industrial) fisheries of the TS-GL System described above, landing up to an estimated 14% (33,000 tonnes) of the total catch from the system (235,000 tonnes) and up to 7% of the estimated annual inland fish landings in Cambodia.

The dai fishery targets the refuge migrations of a multi-species assemblage of fish as they migrate from the TS-GL System to the Mekong main channel via the Tonle Sap with the receding floodwaters each year. Fish also enter the Tonle Sap from its adjacent floodplains via numerous tributaries and channels (Hortle pers comms). Quantitative estimates of the relative importance of these migratory routes are not available. Judging by the distribution of fishing lots in the Lake and neck of the Tonle

Sap, and the absence of dais in the tributaries and channels of the Tonle Sap, it would therefore not appear unreasonable to assume that most fish caught in the dai fishery migrate from the Lake itself. Further studies are required to confirm or reject this assumption.

The fishery provides important seasonal employment opportunities for more than 2,000 rural people and is a major supplier of the essential ingredient for prahok, a fermented paste which is an important protein source for many, particularly towards the end of the dry season when fish is scarce (Halls et al., 2007).

In addition to its significant socio-economic value both locally and nationally, the dai fishery of the Tonle Sap also provides a valuable source of data and information to monitor trends in the migratory fish populations that seasonally utilize the TS-GL System and beyond and forming the basis of important fisheries.

Understanding the ‘drivers’ of these trends is important for effective resource management but also to gain insights into how fish populations respond to important environmental variation such as flooding patterns. Improving our understanding of the latter is likely to become increasingly important in the face of climate change and basin development which both have the potential to significantly modify flooding patterns throughout the Mekong basin.

1.3. The purpose and scope of the paper

This paper aims to provide a comprehensive and detailed description of the Tonle Sap dai fishery including target species, management and monitoring activities (past and present), and the dynamics of the exploited fish populations. The paper includes the results of analyses of the most recently available survey data. The paper discusses the implications of these findings in the context of fisheries policy and management, and basin development. Recommendations are made to improve existing monitoring activities and priority research needs are identified. The paper therefore provides an important reference for those engaged in the management of the migratory fish stocks that seasonally inhabit the TS-GL System.

The paper contributes to Logframe Outputs 1, 3 and 5 of the Fisheries Ecology Valuation and Mitigation (FEVM) Component of the MRC’s Fisheries Programme: (1) Improved information on the status and trends of the fisheries in the LMB is available to riparian governments, national Mekong Committees and the MRCS; (3) Improved information on the ecology of the fisheries of the LMB and models for basin planning purposes are available to basin planners and development agencies, and (5) Improved stakeholder and institutional capacity to (a) monitor and evaluate the status, value and trends of fisheries in the LMB, (b) assess and mitigate environmental impacts on fisheries and (c) initiate and sustain research and development activities.

Introduction


Page 14

1.4. The structure of the report

The context and importance of the Cambodian dai fishery has been described in Section 1. Section 2 gives an account of the history of the fishery, its gears and operation, target species and the disposal of its outputs. Section 2 also provides information relating to the economics of the fishery and details of management activities undertaken by the Government of Cambodia Fisheries Administration (FiA). Ad hoc and routine survey activities are described in Section 3. Details of the content and structure of databases used to store and process the data are provided in Section 4. More detailed descriptions of queries to generate raw data for analyses are available in a companion working document. Section 5 contains a description of the spatial and temporal dynamics of the fishery based upon a detailed analysis of fish abundance, biomass, fish size, and species diversity and similarity. A sub-section is devoted to examining the effect of hydrology on fish biomass. The findings are discussed in the context of fisheries policy and management in Section 6 which includes recommendations for future monitoring and research activities.

2. The Dai Fishery of Tonle Sap-Great Lake (TS-GL) System

The stationary trawl or Loh Dai fishery was introduced into Cambodia by the French colonial authorities between 1873 and 1889 for harvesting small-size fish primarily for fish oil production (Touch, 1998). The fish oil was used to replace engine oil during World War I. The importance of the fishery for fish oil production has diminished and most of the landed catch is now for human consumption.

2.1. The Location of the fishery

The Dai fishery is located in the lower section of the Tonle Sap spanning more than 30 km across the municipality of Phnom Penh and Kandal Province (Figure 6). Dai nets are arranged in up to 15 separate rows of between one and seven nets anchored perpendicularly to the channel, with the net mouths facing upstream (Table 1 D; Plate 2). The most upstream Row (15) is located approximately 35 km from Phnom Penh. Individual dais within a row are allocated an identification letter from ‘A’ to ‘H’. This letter combined with the row number provides a unique alpha-numeric identification code for each dai (e.g. 10A).

In 2001, dai row #1 with 3 dai units, situated nearby Chhroy Changvar bridge (presently known as the Cambodian-Japanese Friendship Bridge), was decommissioned as a result of a fishery policy reform in 2000. Some additional dai units (8B, 11A’, 12G and 15E) are reported to be in operation in Kandal Province every year. According to the Department of Fisheries Affairs of the Fisheries Administration (FiA), dais 8B, 11A’ and 15E are operated with licenses issued by the Ministry of Agriculture, Forestry and Fisheries whereas 12G is operated without a license.

Plate 2 Rows of dai nets anchored across the Tonle Sap


Page 16

Table 1 D

ai locations in the Tonle Sap during the 2007–08 season

ProvinceR

ow

No.

Approxim

ate cum

ulative distance betw

een rows (km

)

Coordinates

Relative transversal positions of dai nets in the

Tonle Sap channel

Total number of dai

units forming each

rowN

orth endsE

ast ends

Kandal

Province

Row

1537.50

11º53.585’104º48.580’

BC

DE

F5

Row

1433.00

11º52.110’104º47.266’

AB

C3

Row

1331.92

11º51.618’104º47.675’

A1

Row

1228.93

11º50.349’104º48.111’

AB

CD

EG

6R

ow 11

23.0711º47.447’

104º49.383’A’

AB

CD

5R

ow 10

13.1711º42.257’

104º50.515’A

BC

DE

FG

7R

ow 9

10.7711º40.963’

104º51.026’B

CD

3R

ow 8

4.8711º40.477’

104º51.360’B

CD

EF

GH

7R

ow 7

4.2811º39.685’

104º51.969’C

DE

FG

5Sub-Total

9 rows

42

Phnom Penh

Municipality

Row

63.77

11º38.867’104º52.581’

CD

EF

G5

Row

53.28

11º38.363’104º53.328’

BC

DE

F5

Row

42.75

11º38.295’104º53.809’

AB

CD

4R

ow 3

1.4011º37.640’

104º54.705’A

BC

D4

Row

20.00

11º37.068’104º55.116’

AB

CD

4

Sub-total5 row

s22

Grand total

15 rows

64

Figure 6 The location of dai fishery in the Tonle Sap

The Dai Fishery of Tonle Sap-Great Lake (TS-GL) System


Page 18

The transversal position of individual dai rows within the channel varies along the upstream-downstream axis. Rows 2–4 and row 7 are positioned towards the right bank of the channel (facing northwards), whilst row 13 and 14 are positioned towards the left bank of the channel. Other rows are positioned more centrally in the channel (Table 1). These positions have remained largely unchanged for more than a century and may have been chosen to maximize catch rates determined by river morphology and hydrology. The locations of the dai rows as well as the dai units are now controlled under Fisheries Law (see Section 2.6).

2.2. Target species

The dai fishery primarily targets small cyprinids including Cirrhinus lobatus, Henicorhynchus cryptopogon, Paralaubuca barroni, Labiobarbus lineatus, Henicorhynchus siamensis and Labiobarbus siamensis, collectively known as trey riel in Khmer, as they migrate from the Tonle Sap Lake to the Mekong River with receding floodwaters between October and March each year. Other species making at important contribution to landings are Labeo chrysophekadion, Pangasius pleurotaenia, Puntioplites proctozystron and Thynnichthys thynnoides (Figure 7 and Plate 3 – Plate 12). During the 2009–2010 fishing season, a total of 123 individual species were reported to have been landed. Small-sized fish are often used to produce Prahok, fish meal and fish sauce (see below).

Ngor (2000) notes that species composition varies during the six months of the fishing season. He reports that large and medium size fish species such as Pangasianodon gigas, Catlocarpio siamensis, Probarbus jullieni, Cirrhinus microlepis, Pangasius spp., Cyclocheilichthys enoplos etc., are caught in larger numbers at the start of the season (October and November) compared to later in the season. Inter and intra-specific differences in the timing of migrations conditioned primarily by fish size has been described in the Mekong since the mid-1950s (Welcomme, 1985). Inter-specific differences in migration timing are described in more detail in Section 5.3.2.

Figure 7 The species composition of the dai catch in 2009–10

Cirrhinus lobatus

Lobocheilos cryptopogon

Paralaubuca barroni

Labiobarbus lineatus

Henicorhynchus siamensis

Labiobarbus siamensis

Labeo chrysophekadion

Pangasius pleurotaenia

Puntioplites proctozystron

Thynnichthys thynnoides

Clupeichthys aesarnensis

Syncrossus helodes

Pangasius larnaudii

Yasuhikotakia modesta

Cirrhinus microlepis

21%

27%

12%

12%

11%

4%

2%2%

1%1%1%1% 1%1.5%

2%

Plate 3 Cirrhinus lobatus (trey riel angkam). A migratory herbivorous species, which seems to play a key role in the food chain; It is a protogynous hermaphrodite, which spawns in June and July in the main channel and in floodplains (MFD, 2003).

Plate 4 Henicorhynchus cryptopogon. Found at midwater to bottom depths in canals, ditches and small streams of floodplains, and more commonly in larger rivers as the temporary water bodies dry up. It undertakes lateral migrations on to seasonally inundated land during the rainy season, where it feeds on algae, periphyton, and phytoplankton (Rainboth, 1996).



Page 20

Plate 5 Paralaubuca barroni (trey slak russey) Feeds on zooplankton; Found in slow flowing or standing water in the Middle Mekong Basin; Processed to Prahok (MFD, 2003).

Plate 6 Labiobarbus lineatus (trey khnawng veng). Omnivorous species found basin-wide in rivers and streams. Undertakes seasonal migrations onto floodplains to spawn. (MFD, 2003).

Plate 7 Henicorhynchus siamensis (trey riel tob). Abundant herbivorous species occurring basin-wide in large and small rivers. It is highly migratory and spawns at the beginning of the flood. The species is extremely important in the dai fisheries and is also caught with other gear. Most of the fish is processed to Prahok. (MFD, 2003).

Plate 8 Labiobarbus siamensis (trey ach kuk). An important whitefish species that feeds on aquatic animals, small water plants and algae. It is believed to migrate from Cambodia to Lao PDR in January–February to spawn on upstream floodplains in June–July. After spawning, the fish migrates down the Mekong River again (MFD, 2003).



Page 22

Plate 9 Labeo chrysophekadion (trey kaek). Predominantly herbivorous occurring in flowing and standing water including reservoirs throughout the Mekong Basin. It spawns mainly in the early flood season in a variety of habitats. Commercially important marketed fresh or dried and salted.

(MFD, 2003).

Plate 10 Pangasius pleurotaenia (Chhwiet). An omnivorous catfish with a basin wide occurrence although apparently more abundant in the Lower Mekong. It is believed to undertake significant spawning migrations during the beginning of the flood season. After spawning in the mainstream, eggs and larvae drift to the nursery areas. The drifting larvae are caught with special gear and cultured. A very important food fish in the Lower Mekong and also used in the aquarium trade (MFD, 2003).

Plate 11 Puntioplites proctozysron (trey chrakeng). Omnivorous species occurring in slowly moving and standing water including reservoirs. Moves laterally into the flooded forest during the high water season. Spawning occurs during the flood. (MFD, 2003).

Plate 12 Thynnichthys thynnoides (trey linh). Pelagic feeding in a variety of habitats basin-wide including the mainstream. It migrates in the mainstream during the dry season. Enters floodplains during the high water season and spawns in the flooded littoral zone possibly throughout the flood season. Caught with a variety of small to large scale gear; Sold fresh and processed’ (MFD, 2003).



Page 24

2.3. Fishing gear and operation

A dai is effectively a stationary trawl net anchored within the river channel to intercept migrating fish. A dai can be operated singly or joined with up to six other dais in a row. It comprises a long (100–120 m) cone-shaped net suspended from two anchored bamboo rafts or (sampans) and a floating work platform (or a movable floating house) positioned at the codend of the net (Figure 8 a). A small boat or floating platform is secured between the two anchored rafts by bamboo poles that serve as a gangway and help to keep the net mouth open. The distance between the two anchored rafts (25–27 m) and the depth of the water determines the net mouth area. In all, up to 30 anchors attached to steel cables may be required to secure the dai in position. Another raft is positioned above the entrance of the net mouth. This holds a winch for raising and lowering the chain ‘footrope’ thereby opening and closing the net mouth (Figure 8 b and c). The bag net is kept open by the force of the current and with help of anchors.

Three types of nets varying according to mesh size and shape are used by dai operators during the course of the fishing season according to the prevailing hydrological conditions and the mean size of migrating fish (Table 2). All three nets are conical in shape but the Yore net has a much smaller top-panel and longer U-shaped top-rope to maintain net buoyancy during low flow conditions towards the end of the season (January–March) when lift forces exerted on the net generated by the flow are much lower. With a much reduced top-panel area, fishers must first drive fish towards the codend before hauling it.

Table 2 Mean (and range) dimensions, mesh sizes, and periods of operation of each dai net type reported by all 64 dai unit operators during the 2007–2008 season.

Net type Length (m) Width (m) Depth (m) Min mesh (m) Max mesh (mm)Dai Chieu 157 (110 – 180) 29 (30 – 27) 21 (32 – 10) 24 (50 – 13) 182 (250 – 100)

Dai Nheuk 157 (110 – 180) 29 (30 – 27) 17 (28 – 8) 16 (25 – 15) 118 (220 – 80)Dai Yor 158 (110 – 180) 29 (30 – 27) 14 (26 – 6) 15 (20 – 15) 70 (200 – 20)

Period of operationNet type Oct Nov Dec Jan Feb March

Dai ChieuDai NheukDai Yor

During the first two months of the season (October–November), the dai chieu net is used to target medium and large size fish species which fishers believe migrate from the Great Lake first followed by smaller sized species and individuals later in the season. These fisher observations appear consistent with those derived from the analysis of monitoring programme data (see Sections 5.3.2 & 5.3.3). Strong water currents during this period apply significant drag forces to the dai net which limit the

minimum net mesh sizes that can be used. During the first two months when haul weights are low but fish size is relatively large, an open weave basket made of bamboo and rattan is often attached to the codend (Figure 8 d). The Dai nheuk net is used between December and February to target smaller species including Cirrhinus lobatus, Henicorhynchus siamensis, Paralaubuca spp. (trey slak russey) and Labiobarbus lineatus (trey khnawng veng). The mesh sizes of all three nets decrease from the net mouth to the codend. Note that gear dimensions and net mesh sizes used by dai operators often exceed the legal maximum and minimum sizes respectively as stated in the Burden Book (see Section 2.8).

(a) dai nheuk

(b) mouth of a dai



Page 26

(c) Winches used to lift the codend of the net and to close the dai mouth.

(d) Open weave-baskets attaching to the codend of the dai, used especially in October and November

Figure 8 Main structures of the dai fishery in the Tonle Sap (source: Deap et al., 2003, pp. 228–232)

for small dai only bamboo codend basket

30–50 cm

1.8–

2 m

Upstream dai operators positioned closest to the Lake tend to use Dai chieu nets with a smaller maximum and minimum mesh size compared to those operators downstream with differences of up to 100 mm (Figure 9). Conversely, Yor nets used upstream tend to have a larger maximum mesh size than those used downstream. Minimum and maximum mesh sizes of nheuk nets show no obvious trend with row number.

Figure 9 Reported mean minimum and maximum mesh sizes of each net type by dai row

05

1015202530354045

0 2 4 6 8 10 12 14 16

Min

. mes

h si

ze (m

m)

Row



4540353025201510

50

0 2 4 6 8 10 12 14 16Row

Min

. mes

h si

ze (m

m)

200

250

)


0

50

100

150

200

0 2 4 6 8 10 12 14 16

Max

. mes

h si

ze (m

m)

Ro


250

200

150

100

50

00 2 4 6 8 10 12 14 16

Row

Min

. mes

h si

ze (m

m)

To empty the net of fish, the net mouth is first closed by raising the foot rope. The codend of the net is then winched aboard the working platform and emptied into sorting compartments (Plate 13) or directly into boats belonging to fish buyers during peak landing periods (Plate 14).



Page 28

Plate 13 Emptying the codend into sorting compartments

Plate 15 Diesel engines used to haul the dai net

Plate 14 The codend being emptied directly into traders boats

A survey undertaken in 2007 revealed that after 1999, diesel engines (Plate 15) began replacing the traditional hand-powered wooden winches (Figure 8 c) used to raise the foot rope to close the net mouth and to raise the codend during net hauling. Since 2006 all dai operators except 6G are now mechanized using a total of almost 2500 HP (Figure 10).

Ninety-percent of the dai operators reported that they mechanised their operations to increase the number of hauls they could make and the individual size of each haul during peak catch periods. Almost all reported their prime motive was to reduce hauling time, increasing the effective fishing time (net soak time) each day. Only 10% of operators reported that they employed engines to reduce labour costs (Table 3).

Figure 10 Cumulative number of dai operators using diesel engines to close and haul their nets (top) and (bottom) cumulative total engine horsepower (HP) employed in the dai fishery since 1999.

Year

Tota

l eng

ine

HP

Dai

s us

ing

engi

nes

Year

0102030405060708090100

1998 2000 2002 2004 2006 2008

0

500

1000

1500

2000

2500

3000

3500

1998 2000 2002 2004 2006 2008

Table 3 Reasons reported by dai operators to mechanise hauling operations

Reason % of operators (63)Increase hauls/day? 81

Lift more fish each haul? 71Both of the above? 90Reduce time hauling? 98Reduce labour ? 10



Page 30

2.4. Fish disposal

Small-sized low value fish (LVF) fish species that form the great majority of the annual catch by weight (98% in 2007–2008) are typically landed and sold to fish traders either on the floating dai platform or on the nearest riverbank. During periods of high catch rates, the entire haul may be landed directly into traders’ boats. Typical pathways of disposal are illustrated in Figure 11.

2.4.1. Prahok

Prahok is a protein-rich paste made of salted and fermented low-value fish typically trey riel, trey slak russey, trey khnawng veng, and trey linh. Prahok is a traditional condiment in most Cambodian dishes, a symbol of national cuisine and an important source of protein, calcium and vitamin A for the rural poor in Cambodia, particularly towards the end of dry season (April–May) when fish is scarce. Prahok is produced within the vicinity of the dai fishery for both subsistence and commerce trade. Subsistence production is mainly by poor farmers or households from rural areas where fish is scarce or where farming activities prevail. They establish makeshift processing camps along the riverbanks each year during the peak catch period (December–January). The peak catch period is often communicated by word of mouth or via announcements by the FiA on national television. Each household camp may process between 80 and 200 kg of fresh fish to produce prahok. The fish is dressed, often beheaded

LVF sauce manufacturers Markets

Provinces/cities

LVF exporters

Viet Nam

Tonle Sap dai fishery operators Fish/animal raisers

ConsumersRice farmers Phnom Penh & Kandal market retailers

LVF wholesalersa

a

a

a

a

ab

a

LVF traders

Figure 11 Fish disposal pathways for low value fish (LVF) or small size fish species. Source: So et al. (2007), p20.

a Occasional (less common) pathway. b Pathway recommended by the MoFL. for research dais (rice farmers given priority to landings over traders).

and partly or wholly eviscerated and mixed with salt. After some days it may receive further treatment including filleting, cleaning and packaging in airtight containers. Commercial producers use machines to remove the head and de-scale the fish. Salt and either rice, grains or fruit may be added to lower the pH to increase storage life and add to the taste (Sverdrup-Jensen, 2002; Halls et al., 2007).

Commercial producers supply wholesalers, retailers and exporters in markets countrywide. Commercial producers also export prahok especially to Thailand. A large-scale producer typically processes between 60 and 150 tonnes of fresh fish to produce between 20 to 50 tonnes of prahok (So et al., 2007).

Small-size fresh fish from the dai fishery are also processed into other forms of fermented fish (pra-ork), smoked fish (trey cha-eur), sweet fish (mam), and fish sauce for human consumption (Lieng et al., 1995; Ngor, 2000 and So et al., 2007). Following the growth of aquaculture along the Tonle Sap during the last two decades, small-size fish are also used as feed for fish cultured in cages or ponds including Channa micropeltes, Pangasianodon hypophthalmus and Clarias spp., as well as feed for animals including ducks, pigs and chickens (So and Nao, 1999; Hav and Ngor 2005 and So et al., 2005). For the latter, fish may be first sun-dried. So et al. (2007) report that each drying operator may buy up to 150 tonnes of small-sized fish landed by the dai fishery. Sun-dried fish is fed to ducks, chickens and pigs, and for producing fish meal for export to neighbouring countries.

Recent reports (So et al., 2007) suggest that small-size fresh fish landed by the dai fishery are also exported to Viet Nam by boat or road for both human consumption, fish and other animal feed. This practice was prohibited by the government in 1990 to control prices and ensure adequate supply for rural farmers and households (Touch, 1993).

More valuable medium and large-sized fish species including Osteochilus melanoplerus, Pangasius larnaudii, Cyclocheilichthys enoplos and Pangasianodon hypophthalmus are often kept alive in bamboo cages suspended below the working platform of the dai. A single cage can contain up to 20 tonnes of live fish (Lieng et al., 1995). These fish are sometimes sold during the closed season (March–September) when fish supply is low and prices are high. High value species are also sold during the fishing season to fish traders to supply nearby markets or for export.

Ngor (2000); So et al. (2007) report that the supply and mean size of these medium and large-sized fish species have declined since the 1990s. Factors postulated to be responsible for these declines include increasing rates of exploitation and illegal gear use, and habitat destruction and modification. Changes in the species composition of the dai fishery landings are examined in detail in Section 5.4.

2.5. Economics

From a sample of 16 dais, Hap and Ngor (2001) estimated that the mean net profit of a dai unit is approximately US$14,000 per annum corresponding to a mean annual rate of return to investment



Page 32

of 60% (Table 4). A more recent study (So et al., 2007) estimated returns to be approximately 50% but with significant variation ranging from 3%–74%. This lower return may reflect the exclusion of revenue generated from the sale of medium and large high value species from the study.

Table 4 Dai Fishery profit estimates 1999-2000. Hap and Ngor (2001), p. 224

Items Total values ('000 Riels) USDI. Gross Return 183,857 47,755 Variable Costs (VC) Salary and wages of labours 10,426 Fuel/oil 8,306 Food expenditure 3,405 Other expenses (medical, cigarette, fruit etc) 2,100 A. Total VC 24,237 Fixed Cost (FC) Lease fees 8,389 Government tax (5% on lease fee) 443 Depreciation of: Gear/equipment 17,381 Boat (motorized and non-motorized) 1,694 Engine boat 818 Engine for electricity & dynamo 170 Floating-house and cage for stocking fish 1,632 Total depreciation 21,694 Repairs/maintenance 26,007 Interest rate on borrowed funds 50,321 B. Total FC 106,855II. Total Costs 131,092III. Gross Profit (I - A) 159,620IV. Net Profit (I-II) 52,765 13,705 Opportunity Cost of Management 521V. Rate of Return to Capital 59.75VI. Operating Profit Margin Ratio 56.03

Note: Exchange rate is 3,800 Riels per one US$ (January, 2000). Opportunity cost of management = 5% of salary/wage of labour (374,839 Riels). Total value of assets = Total values (costs) of fixed asset (i.e. Gear/equipment, boat, engine for boat and electricity and floating house) and TVC; Adjusted Net Income = Net Profit + Interest Rate + Taxes. Rate of Return to Capital = (Return to Capital/Total values of Assets)* 100 Return to Assets/Return to Capital = Adjusted Net Income - Opportunity Cost of Unpaid Labour - Opportunity Cost of Management.

The price of trey riel which forms approximately 40% of the total annual landings of the dai fishery has increased by almost a factor of 10 from 1995–6 to 2007–8 (Figure 12). So et al. (2007) report a similar trend in price for both fresh and preserved fish at dai landing sites.

Using data on costs and revenue reported by So et al. (2007), the annual profit of dai units tends to decline from the most upstream dai row 15 to row 2 (Figure 13). Equipment, labour, and fuel costs are all higher upstream and unit fish prices are lower, but these differences are more than compensated for by disproportionately larger landings made by the upstream dai units. Noteworthy, is that licence fees (both official and unofficial combined) paid by dai operators appears independent of their reported operating profit (Figure 14).

The inclusion of high-value medium and large size fish which form only 3% of the catch by weight but 11% by value in 2007–08 might alter these conclusions.

Figure 12 Actual (upper line) and inflation adjusted (lower line) unit price of trey riel. In 2004, 4,000 Riel was equivalent to about US$1.

Source: Updated from Hortle et al. (2005). Correlation between annual catch and average adjusted price (r = 0.84) unexpectedly indicating that price increases with supply.

Season

0

2000

4000

6000

8000

10000

12000

14000

0

200

400

600

800

1000

1200

1400

1600

1800

2000

1995

-6

1996

-7

1997

-8

1998

-9

1999

-0

2000

-1

2001

-2

2002

-3

2003

-4

2004

-5

2005

-6

2006

-7

2007

-8

Average weighted price (Riel/kg)

Adusted price (Riel/kg)

Annual catch (tonnes)

Pric

e (r

iel/k

g)

Ann

ual c

atch

(ton

nes)



Page 34

Figure 13 Mean values of key economic variables for the Cambodian dai fishery plotted as a function of dai row number. Error bars show mean values +/- 1 S.D. P-values < 0.05 indicate a significant trend in the variable value with row number. Data from So et al. (2007).

23456789101112131415

Row

0

2500

5000

7500

10000

Labo

ur c

osts

($)

P < 0.01 P < 0.001 P = 0.02

P = 0.09 P < 0.001 P = 0.04

P < 0.001 P < 0.01 P = 0.02

a

d

g

b

e

h

c

f

i

Figure 14 License costs for 25 dais plotted as a function of their annual profit (R2 = 0.002; p = 0.85). Data from So et al. (2007).

2.6. Management

2.6.1. Legislation

The dai fishery is legislated under Cambodian Fisheries Law (2006) (article 39). The ‘burden book’ describes management legislation including operating seasons, the position of the dais in the river, gear size restrictions, payment rules and harvest rules. The Fisheries Administration (FiA) Inspectorate and Cantonment are responsible for monitoring and enforcing the rules and regulations described in the burden book.

2.6.2. Closed season

Fishing with dais in the Tonle Sap is permitted only between 1st October and 30th of March in the following year after which all dai structures have to be completely removed from the River. Re-installation of the dai gear is permitted to begin in September.



Page 36

2.7. Effort control / Licensing

Under Cambodian Fisheries Law, a dai unit is classified as a large scale fishing gear or ‘fishing lot’ (No. 028 KOR SOR KOR Table A III). Rights to these lots in the form of a fishing licence are either auctioned by the Government to the highest bidder for exclusive exploitation for a two-year period (Deap et al., 2003) or are allocated by the FiA for research purposes (Table 5).

Table 5 Dai unit licence allocation by type for the 2007–08 season

ProvincesDai fishery licences

Auctioned dai (units) Research dai (units) TotalPhnom Penh 1 18 4 22Kandal 2 26 12 38Total 44 16 60

Source: FiA (2007)

The processes and procedures for the auction of the dai fishery are stipulated in a sub-decree of Cambodian fisheries law. This sub-decree provides rules on renting inland and marine fishery domains for exploitation comprising 29 articles (CoM, 1989). The auction is open to all Cambodians except for government officials (Article 6) and conducted at provincial levels. Articles regarding the dai fishery or fishing lot auction written in the sub-decree contain rules on the auction process, payment of fishing fees by the auction winners, rules on fish harvesting and rules on enforcement responsibilities.

Since 2001, 60 dai units in 14 rows (row 2–15) have been granted licenses by the FiA to operate each season (van Zalinge et al., 2003). Unlicensed fishing occurs. During 2007–08 one dai unit in Kandal Province operated without a licence and other three units were operating under additional licenses issued by MAFF and FiA. In 1938–39, 108 dai units were permitted to fish in 23 rows (Chevey and Le Poulain, 1940) but by 1962–63 the total was just 61 units in 15 rows (Fily and d’Aubenton, 1965). The mid-eighties saw an increase to 86 units followed by a decline to present day numbers (Figure 15).

1 Research dais in Phnom Penh include all dai units in row #2 (2A, 2B, 2C and 2D). This title of research dais in Phnom has no timeframe limitation. 2 Research dais in Kandal Province include row #14 (14A, 14B and 14C), row #13 (13A) #12 (12A, 12B, 12C, 12D, 12E) and row #9 (9B, 9C and 9D). All research dais in Kandal have a limited timeframe between 2003–09.

Figure 15 Number of licensed dais (1938–2007) reported by Chevey and Le Poulain (1940); Nguyen and Nguyen (1991); Lieng et al. (1995); Ngor (2000) and Ngor and van Zalinge (2001).

40

50

60

70

80

90

100

110

120

1930 1940 1950 1960 1970 1980 1990 2000 2010 2020

Year

Dais

Lice

nsed

2.8. Gear restrictions

The ‘Burden Book’ states the maximum gear dimensions and minimum mesh sizes for the three types of nets that can be used in the fishery during the open season (Table 6).

Table 6 Maximum gear dimensions and minimum mesh sizes for the three types of dai nets

Types and size of dai gears Period of operation

Start EndDai Chieu Net dimension: - Size: 120 m (Length) x 27 m (width) - Mesh size: 20 mm – 220 mm

01 October 01 January

Dai Nheuk Net dimension: - Size: 120 m (Length) x 27 m (width) - Mesh size: 15 mm – 220 mm

02 January 28 February

Dai Yor Net dimension: - Size: 120 m (Length) x 27 m (width) - Mesh size: 15 mm – 30 mm

01 March 15 March

Source: Dai fishery’s Burden Book 2005–07


3. Routine monitoring activities and survey methodologies

3.1. Purpose of monitoring

The primary purpose of monitoring has been to generate estimates of total aggregated monthly and seasonal estimates of landings and their monetary value. These figures are reported to the Fisheries Administration (FiA), Ministry of Agriculture, Forestry and Fisheries, in Phnom Penh. Total catch and more recently, catch rate estimates have also been used to monitor resource trends in response to environmental variability and management initiatives.

3.2. A history of survey methodologies and sampling regimes

Routine monitoring of dai fishery has been undertaken since 1994 by the Management of the Freshwater Capture Fisheries programme, the Assessment of Mekong Capture Fisheries Component of the MRC Fisheries Programme (AMCF, 2003–2006) and since 2007 by the Fisheries, Ecology Valuation and Mitigation (FEVM) component of the MRC Fisheries Programme in cooperation with Inland Fisheries Research and Development Institute (IFReDI)–the research component of the FiA. Combined, these programmes have generated the only continuous long-term data set for an inland fishery in Cambodia–and one of only two in the Mekong Basin–the other being the lee trap fishery in southern Lao PDR (MRC, 2010).

It is likely that both the dai and lee trap fisheries became foci for intensive monitoring becausethey target fish migrations through ‘bottlenecks’: the Tonle Sap and the few channels passable by fish at the Khone falls in southern Lao PDR. Fish densities and therefore catch rates are high at these locations offering opportunities to sample relatively large proportions of migrating fish populations during short periods of time giving rise to high sampling efficiency and the prospect to accurate population estimates. The ‘Fishing Lots’ (barrage and floodplain) described in Section 1.1.4 also offer opportunities to sample significant proportions of exploited fish populations efficiently owing to the often large-scale nature of the fishing gears operating within them. Attempts by the FiA were made in the past to monitor these lots, but neither lot operators nor officials were prepared to cooperate (van Zalinge, 2002). Whilst related ad hoc studies and anecdotes have been published (e.g. Hortle et al., 2005; Valbo-Jorgensen et al., 2001 and Dubeau et al., 2001) attempts to routinely monitor the other, more dispersed, sectors of the fishery (i.e. middle-scale artisanal and family fisheries) have also failed in the past owing to a lack of capacity (van Zalinge, 2002).

Therefore whilst the dai fishery continues to provide valuable information about the variability and long-term trends of migratory fish populations that seasonally utilise the TS-GL System, no equivalent programmes exist to monitor populations of valuable Blackfish species that inhabit the system year-round (van Zalinge, 2002).


Page 40

Moreover, without information about the country’s other fishery sectors, it is difficult to interpret variability or long-term trends in the dai (or any other) fishery without assuming that effort has remained relatively static over the period of interest. This might not be unreasonable if the large-scale barrier and fence traps common in TS-GL System and in other bottlenecks over the migratory range of stocks exert the greatest overall effort (fishing mortality), but the artisanal and subsistence fisheries may also be significant in this respect. This subject requires further investigation.

However, a focus on only one of several components of the fishery cannot provide all the necessary data for the estimation of national landings by species–information that is important for economic valuation and for environmental impact assessment purposes. It also makes recommendations for the control of fishing effort (mortality) to ensure the sustainability of stocks difficult.

Recommendations to complement data and information generated by the dai fishery monitoring programme for management and policy planning and evaluation purposes are made in Section 6.

The earliest ad hoc surveys of the dai fishery were documented by Chevey and Le Poulain (1940) who reported catches over the 1938–39 season. They described the location of the fishery, the structure and operation of dai fishing gear, recorded the relative abundances of individual species and estimated the total catch to be around 13,500 tonnes (Table 7). Fily and d’Aubenton (1965) monitored the dai fishery between 1962–63. They provided a general description of the fishery and monitored fish catches and species composition on selected dai units. They also tested the effects of current speed on species composition and catch volume. After the 1960s, there appears to have been no more monitoring of the fishery until 1986–88 when Nguyen and Nguyen (1991) conducted a study (written in Vietnamese) on freshwater fish productivity in Cambodia that included the dai fishery.

Fisheries statistics for the Department of Fisheries (DoF; the present-day FiA) were collected more consistently from 1980 onwards. A lack of human resources and technical facilities during this early period, however, meant that rigorous sampling protocols could not be followed. Prior to the mid-1990s, data were collected using logbooks which dai operators were required to complete under their licence agreement. However, most operators under-reported their landings fearing that reporting their true landings might raise the cost of their operating licence. These logbook-generated data are therefore regarded as unreliable.

In 1994, as part of the Mekong River Commission (MRC) Fisheries Programme and in cooperation with the Department of Fisheries, the MFCF was established to improve fisheries monitoring activities, the main focus being on mobile gear and large-scale fisheries. The software programmes ARTFISH (Stamatopoulos, 1994) and LenFreq were introduced to aid data analysis. In the same year, Lieng et al. (1995) applied a new data collection scheme–the Catch Assessment Survey (CAS)–to estimate the total catch by species for the dai fishery.

Over the 1994–95 fishing season, Lieng et al. (1995) reported 73 dais distributed among 15 rows. For their study, Lieng et al. (1995) stratified the dais into three groups of rows (rows 1–5; 6–10 and 11–15) according to their downstream location. Because peak catches were strongly associated with

the lunar phase, they also introduced a ‘peak’ and ‘low’ fishing period as an additional sampling stratum. The Peak Period was defined as occurring during the waxing moon phase (khnaet) 4–6 days before the full moon and the Low Period as those days falling outside of this period during the waning moon phase (ronouch).

Dais were randomly selected for sampling within each combination of these strata and sampling was conducted twice a month in December 1994 and in January and February 1995. Catch per haul was sampled from selected dais and the number of hauls made during 24 hours estimated from the observed hauls per hour. Lieng et al. (1995) estimated that total catch for the 1994–95 season was 18,410 tonnes. Their estimates were considerably higher than those of Fily and d’Aubenton’s (1965), comparable with DoF and Chevey and Le Poulain’s (1940) estimates, and roughly equivalent to those of Nguyen and Nguyen’s (1991) (Table 7).

Table 7 Estimates of catches prior to and including the 1994-95 seasons reported by Lieng et al. (1995)

Season Total catch (tonne) No. of dais Catch/dai (tonne) Reference1938–39 13,568 106 128 Chevey and Le Poulain (1940)1962–63 2,135 61 35 Fily and d'Aubenton (1965)1981–93 5,000–12,839 97–35 – DOF statistics1983–88 7,413–18,026 86 86–209 Nguyen and Nguyen (1991)1994–95 10,755 73 161 DOF (pers. comm.) (1995)1994–95 18,410 73 252 Project estimates (1995)1995–96 63 Auction results (1995)

Only 18 dais were sampled between December 1994 to February 1995 (Annex Table 25) hindering comparisons with data collected under subsequent surveys.

Lieng et al. (1995) recommended that future monitoring programmes should introduce a sampling stratification scheme based upon dai licence cost (value) to replace the existing row stratification on the assumption that catch rates should be correlated with licence value. Whilst this assumption appears difficult to reconcile with the results presented in Section 2.5, this stratification scheme was introduced during the following fishing season. Unfortunately, this survey did not include the dai identification code (see Section 2.1), or sample haul frequency also hindering comparisons with data collected under subsequent surveys.

During the next fishing season (1996–97), the Department of Fisheries attempted a census of dai catch rates (Ngor, 2000; Ngor and van Zalinge, 2001). However, Ngor (2000) reported that only 49% of dai units were included in the survey because of insufficient resources. Dais were ranked according to their percentage contribution to the total catch recording during the survey. Those dais that landed more than 50% of the cumulative total catch were classified as ‘High Yield’ dais and the remainder as ‘Low Yield’ dais. It remains unclear how dais that were not sampled were classified.

Routine monitoring activities and survey methodologies


Page 42

Although the design of the 1997–98 survey was intended to incorporate the new High and Low Yield stratum into the sampling regime (Table 8), the database records show no evidence of having been stratified by yield before the 1998–99 season (Annex Table 25). From the 1997–98 season however, the dai yield stratum, based on the 1996–97 census, has been consistently applied to stratify the sampling programme. Dais continued to be selected randomly within each stratum (Table 8).

Table 8 Monthly dai sampling regime planned for the 1997–98 season (Deap, 1999, Ngor and van Zalinge, 2001). Dais within each combination of strata were selected randomly. Although the sampling regime reported here appears to have been stratified by dai yield in this season, examination of records in the database suggest that dai yield stratum was only introduced from the 1998–99 season onwards.

Stratum No. of dais Peak Period Low PeriodHigh Yield 8 Dais sampled 8 Dais sampled 8

Hauls/dai/day 10 Hauls/dai/day 10Total 80 Total 80

Low Yield 60 Dais sampled 14 Dais sampled 7Hauls/dai/day 10 Hauls/dai/day 10

Total 140 Total 70Total hauls 220 Total hauls 150

In 1998–99, the Department of Fisheries introduced another major stratum based on the location of the rows in two administrative zones: Kandal Province and Phnom Penh Municipality (Ngor and van Zalinge, 2001). Further stratification was included during the 2000–01 fishing season prompted by MFCF interest in the species composition of the landings by dais in row 2. During this season only, Row 2 and ‘all the other dai units’ formed an additional stratum (Table 9).

In the 2001–02 season, dai row 1 (three individual units in Phnom Penh Municipality) was Decommissioned reducing the total number of dais sampled to 65. The survey design has remained unchanged since 2002–03 with three main sampling strata: (1) Administrative Zone (Kandal Province and Phnom Penh); (2) Dai Yield (High Yield and Low Yield); and (3) Lunar Period (Peak Period and Low Period) (Table 10).

The number of dai units sampled each season has remained relatively constant since 1999–2000 (Table 11). Prior to this season, the sampling effort tended to grow each year. Since 1998–99 almost all dais appear to have been sampled at least once each season.

Prior to 2004, dais were randomly selected within each stratum. After 2004, recommendations were made to sample the same dais each year as those selected for the 2003–04 season to maintain temporal continuity (Sopha Lieng, pers. comms.). However, judging from the temporal and spatial variation in sampling effort (see Annex Table 25), these recommendations appear not to have been adopted.

3.3. Spatial and temporal variation in sampling effort

In addition to changes to the survey design, there has been considerable inter-seasonal variation in the sampling effort as measured by the total number of hauls sampled by the data collectors over the survey period as indicated by the counts of data entry forms, or ‘DOCs’ in the database (Figure 16). At the commencement of surveys in 1994, no more than 18 hauls were sampled from 18 separate dais. Thereafter, the total number of hauls sampled increased steadily to a maximum of 2,426 over the 2001–02 season. From 2003–04 onwards, sampling effort declined to around half this value (1,261) and aside from a peak in effort during 2007–08, sampling has remained consistent at around 1,200–1,500 samples per season until present.

From 1997–98, the total number of hauls sampled in each month has varied between four and approximately 650 (Figure 17 a). Most of the sampling effort was focused on the peak fishing months of November, December and January. Within each month, sampling effort also varied with lunar phase and was most intense during the Peak Period–approximately 6 days before the full moon or second lunar phase (Figure 17 b). This is consistent with the sampling protocol which prescribes an increase in sampling effort (i.e. sampling over a 24-hour as opposed to a 12-hour period) and personnel (i.e. the employment of four as opposed to two data collectors in each administrative zone).

Sampling effort can also be seen to vary longitudinally (upstream/downstream) as reflected in the counts of the total number of hauls sampled in each row and season (Figure 18 a). The general trend is for a decline in sampling effort with distance downstream. In most seasons, but particularly 2000–01 and 2001–02, Row 2 has been sampled most frequently. Rows 9 and 13 tend to be sampled less frequently in Kandal Province, particularly in the later seasons.

In contrast to the total number of hauls sampled by row and season, the mean number of hauls sampled per dai in Kandal Province (Rows 7–15) were found to be higher than those in Phnom Penh (Rows 1–6) (Figure 18 b). This is true for at least five of the twelve seasons shown (1998–99, 1999–2000, 2001–02, 2002–03 and 2004–05), but is most pronounced in 2001–02 and 2002–03. For the former case, the mean number of hauls sampled per dai in Kandal Province was twice that of Phnom Penh.



Page 44

Table 9

Monthly dai sam

pling regime proposed for the 2000–01 fishing season (N

gor and Zalinge, 2001). In this season, the Row 2 inclusion/exclusion

stratum w

as added. For each dai yield and lunar period stratum com

bination the following w

as reported: No. of dais: the total num

ber of dais in each stratum

combination; D

ais sampled: the planned num

ber of dais to be sampled each m

onth (actual counts of the number of dais sam

pled each m

onth from the database are reported in brackets); H

auls/dai/day: the proposed number of hauls/dai/day to be sam

pled (the mean num

ber of hauls/dai/day in each m

onth estimated from

the database are reported in brackets), Samples/m

onth: the proposed total number of sam

ples/m

onth (the mean num

ber of samples collected in each m

onth estimated from

the database are reported in brackets). Within each com

bination of strata dais w

ere selected randomly.

Kandal Province (43 units)

Minor Stratum

:Peak Period

Low

PeriodH

igh Yield

No. of dais

11N

o. of dais11

Dais sam

pled7 (11)

Dais sam

pled11 (11)

Hauls/dai/day

10 (2.9)H

auls/dai/day10 (3.5)

Samples/m

onth70 (19.2)

Samples/m

onth110 (25.1)

Low

Yield

No. of dais

32N

o. of dais32

Dais sam

pled15 (18)

Dais sam

pled28 (27)

Hauls/dai/day

10 (3)H

auls/dai/day- (3.4)

Samples/m

onth150 (21.2)

Samples/m

onth280 (55)

Phnom Penh M

unicipality (25 dai units)A

ll dais excluding Row

2R

ow 2 dais

Peak PeriodL

ow Period

Peak PeriodL

ow Period

No. of dais

21N

o. of dais21*

No. of dais

4*N

o. of dais4

Dais sam

pled20 (23)

Dais sam

pled40* (24)

Dais sam

pled14* (4)

Dais sam

pled40 (4)

Hauls/dai/day

10 (4)H


Hauls/dai/day

10 (11.3)H


Samples/m

onth200 (60.6)

Samples/m

onth400 (56.6)

Samples/m

onth140 (47.1)

Samples/m

onth400 (28.6)

Note: the sam

e dai may be sam

pled on several occasions in the same season.

*There are inconsistencies between the total num

ber of dais in each stratum com

bination (No. of dais) and the proposed num

ber of dais to be sampled (D

ais sampled) since it w

as expected that the sam

e dai would be sam

pled on several occasions on the same day (M

r. Pengbun Ngor, pers com

m.).

Tabl

e 10

Mon

thly

dai

sam

plin

g re

gim

e pl

anne

d fo

r the

200

2–03

fish

ing

seas

on. T

he R

ow 2

stra

tum

was

dro

pped

from

the

sam

plin

g st

ratifi

catio

n an

d th

e st

ratifi

catio

n in

this

seas

on re

flect

s the

stra

tifica

tion

at th

e tim

e of

wri

ting.

With

in e

ach

stra

tum

dai

s wer

e ra

ndom

ly se

lect

ed fo

r sam

plin

g.

For e

ach

dai y

ield

and

luna

r per

iod

stra

tum

com

bina

tion

the

follo

win

g w

as re

port

ed: N

o. o

f dai

s: th

e to

tal n

umbe

r of d

ais i

n ea

ch st

ratu

m

com

bina

tion;

Dai

s sam

pled

: the

pla

nned

num

ber o

f dai

s to

be sa

mpl

ed e

ach

mon

th (a

ctua

l cou

nts o

f the

num

ber o

f dai

s sam

pled

eac

h m

onth

re

cord

ed in

the

data

base

are

repo

rted

in b

rack

ets)

; Hau

ls/d

ai/d

ay: t

he p

ropo

sed

num

ber o

f hau

ls/d

ai/d

ay to

be

sam

pled

(the

mea

n nu

mbe

r of

haul

s/da

i/day

in e

ach

mon

th c

alcu

late

d fro

m th

e da

taba

se a

re re

port

ed in

bra

cket

s), S

ampl

es/m

onth

: the

pro

pose

d to

tal n

umbe

r of s

ampl

es/m

onth

(th

e m

ean

num

ber o

f sam

ples

col

lect

ed in

eac

h m

onth

cal

cula

ted

from

the

data

base

are

repo

rted

in b

rack

ets)

. With

in e

ach

com

bina

tion

of st

rata

da

is w

ere

sele

cted

rand

omly

.

Kan

dal p

rovi

nce

(43

units

)Ph

nom

Pen

h M

unic

ipal

ity (2

2 un

its)

Min

or S

trat

um:

Peak

Per

iod

Low

Per

iod

Peak

Per

iod

Low

Per

iod

Hig

h Y

ield

No.

of d

ais

11N

o. o

f dai

s11

No.

of d

ais

6N

o. o

f dai

s6

Dai

s sam

pled

10 (1

1)D

ais s

ampl

ed11

(11)

Dai

s sam

pled

6 (6

)D

ais s

ampl

ed6

(7)

Hau

ls/d

ai/d

ay10

(6.4

)H

auls

/dai

/day

5 (5

.7)

Hau

ls/d

ai/d

ay16

.7 (7

.8)

Hau

ls/d

ai/d

ay10

(4.9

)Sa

mpl

es/m

onth

100

(46.

5)Sa

mpl

es/m

onth

55 (3

4.3)

Sam

ples

/mon

th10

0 (4

3.1)

Sam

ples

/mon

th60

(23.

1)Lo

w Y

ield

No.

of d

ais

32N

o. o

f dai

s32

No.

of d

ais

16N

o. o

f dai

s16

Dai

s sam

pled

25 (2

7)D

ais s

ampl

ed29

(29)

Dai

s sam

pled

16 (1

6)D

ais s

ampl

ed16

(17)

H

auls

/dai

/day

10 (5

.4)

Hau

ls/d

ai/d

ay5

(5.2

)H

auls

/dai

/day

11.3

(5.4

)H

auls

/dai

/day

8.7

(4)

Sam

ples

/mon

th25

0 (6

9.6)

Sam

ples

/mon

th14

5 (6

6.6)

Sam

ples

/mon

th18

0 (5

5)Sa

mpl

es/m

onth

140

(43)

Not

e: th

e sa

me

dai m

ay b

e sa

mpl

ed o

n se

vera

l occ

asio

ns in

the

sam

e se

ason

.



Page 46

Table 11 Summary of changes made to the dai fishery survey design (1994–2008) and numbers of dais sampled each season.

Season Database structure and features Dais sampled1994–95 Stratified by lunar period

Low sample effort16

1995–96 Stratified by lunar periodNo dai identification reported

no records

1996–97 Stratified by lunar periodCensus conducted to determine dai yield

48

1997–98 Stratified by lunar periodDai yield stratum not yet introduced

59

1998–99 Stratified by lunar periodAdministrative zone stratum introducedDai yield stratum introduced (20 October 1998)

60

1999–00 Unchanged from above 642000–01 Stratified by lunar period, administrative zone and dai yield

Row 2 dais in Phnom Penh stratum sampled separatelyRow 1 discontinued

67

2001–02 Stratified by lunar period, administrative zone and dai yieldRow 2 dai stratification not applied

64

2002–03 Unchanged from above 682003–04 Unchanged from above 662004–05 Unchanged from above 642005–06 Unchanged from above 622006–07 Unchanged from above 642007–08 Unchanged from above, but separation of the catch into ‘Large Fish’ and ‘Small

Fish’ 65

Figure 16 Changes in sampling effort between the 1994–95 and 2008–09 expressed as the total number of hauls sampled in each season.

Season

0

500

1000

1500

2000

2500

3000

94-9

596

-97

97-9

898

-99

99-0

000

-01

01-0

202

-03

03-0

404

-05

05-0

606

-07

07-0

808

-09

No.

of h

auls

sam

pled

(a)

(b)

0

200

400

600

Tota

l no.

hau

ls s

ampl

ed

97-98 98-99 99-00 00-01

01-02 02-03 03-04 04-05

05-06 06-07 07-08 08-09

0

200

400

600

Tota

l no.

hau

ls s

ampl

ed

Oct Nov Dec Jan Feb MarMonth

0

200

400

600

Tota

l no.

hau

ls s

ampl

ed




250

500

750

1000

1250

Tota

l no.

hau

ls s

ampl

ed

97-98 98-99 99-00 00-01

01-02 02-03 03-04 04-05

05-06 06-07 07-08 08-09

250

500

750

1000

1250

Tota

l no.

hau

ls s

ampl

ed

1 2 3 4Lunar Phase

250

500

750

1000

1250

Tota

l no.

hau

ls s

ampl

ed

1 2 3 4Lunar Phase

1 2 3 4Lunar Phase

1 2 3 4Lunar Phase

Figure 17 Total number of hauls sampled each (a) month and (b) lunar phase (Quarters 1–4) each season.



Page 48

(a)

100

200

300

400

500

Tota

l no.

hau

ls s

ampl

ed

97-98 98-99 99-00 00-01

01-02 02-03 03-04 04-05

05-06 06-07 07-08 08-09

100

200

300

400

500

Tota

l no.

hau

ls s

ampl

ed

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15Row number

100

200

300

400

500

Tota

l no.

hau

ls s

ampl

ed

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15Row number

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15Row number

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15Row number

(b)97-98 98-99 99-00 00-01

01-02 02-03 03-04 04-05

05-06 06-07 07-08 08-09

0

2

4

6

8

No.

of s

ampl

es/d

ai

0

2

4

6

8

No.

of s

ampl

es/d

ai

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15Row

0

2

4

6

8

No.

of s

ampl

es/d

ai

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15Row

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15Row

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15Row

Figure 18 (a) Total number and (b) mean number of hauls per dai (±SD) sampled by row each season. The broken vertical line indicates the boundary between Phnom Penh Municipality (Rows 1–6) and Kandal Province (7–15).

3.4. Current survey methodology

The following sections give further details of the current survey design and enumeration methods employed since 2005–06 designed to generate estimates of total catch weight and value by species landed by the fishery each season.

3.4.1. Sample stratification

Nine rows of dai operate in Kandal Province comprising 11 High Yield units: 10A, 10B, 10D, 10E, 11A, 12A, 12B, 12C, 13A, 14A and 14B (Table 12). The remaining 27 units are Low Yield units including three additional units which are often seen operating in this stratum. Within the Phnom Penh Municipality, there are five dai rows and six High Yield units: 2A, 2B, 2C, 2D 3C and 3D. The remainder (16) are Low Yield units (Table 12). The Peak Period commences seven days before the full moon (second quarter) and ends one day before full moon. All days that fall outside of this period are classified as the Low Period. Sampling occurs on approximately 17 days each month. During the Peak Period, dais are sampled every day and during the Low Period, every second or third day.

Table 12 The relative locations of High (shaded cells) and Low Yield (unshaded cells) dais. A’ = additional dai that came into operation subsequent to the rows having been labelled.

Province Row No. Relative transversal positions of dai nets in the Tonle Sap channel

Total number dai units

Kandal Province

Row 15 B C D E F 5Row 14 A B C 3Row 13 A 1Row 12 A B C D E G 6Row 11 A’ A B C D 5Row 10 A B C D E F G 7Row 9 B C D 3Row 8 B C D E F G H 7Row 7 C D E F G 5

Sub-Total 9 rows 42

Phnom Penh Municipality

Row 6 C D E F G 5Row 5 B C D E F 5Row 4 A B C D 4Row 3 A B C D 4Row 2 A B C D 4

Sub-total 5 rows 22Grand total 15 rows 64



Page 50

3.4.2. Data collectors and data collection process

Survey enumerators are trained and supervised by IFReDI. Typically, two enumerators are assigned to each administrative zone each sampling two dais per day. For each dai, they sample 3–4 hauls over a 12-hour period. During the seven days corresponding to the Peak Period each month, sampling occurs over a 24-hour period. Additional enumerators are employed during this period–three in Kandal Province and two in Phnom Penh Municipality. During this period, eight, instead of four dais are also sampled daily in each zone. However, day and night samples are not distinguished in the database. The duration of the ‘Peak Period’ does not appear to be fixed each month, rather it appears to be subjectively identified by survey supervisors on the basis of changes to catch rates.

3.4.3. Variables enumerated

Catch per haul

The catch per haul is the total weight (kg) of all species combined in each haul, or lift of the dai net. During the Low Period the total catch per haul is relatively small and can be weighed using a five kilogram scale. All individual fish and each species can be enumerated during this period. During Peak Periods, however, catches per haul are usually too high to be weighed on a scale. In these instances, a visual estimate of the fish catch per haul is made by the enumerator with guidance from the dai operator. Three alternative methods for visually estimating the weight of the haul are applied:

1) Visual estimation per codend: Enumerators estimate the proportion of the codend that is filled by the haul. The weight of a full codend is known (100–200 kg depending on the design) and the total weight of the hauls is reported as a fraction of this;

2) Visual estimation per basket: If the quantity of fish in a single haul exceeds the volume of the codend, the data collectors count the number of baskets sold directly to buyers waiting in boats beside the dais. The weight of a single basket is known and the total catch per haul can then be estimated as the product of the unit weight and the number of baskets landed; and

3) Visual estimation per boat: Very large hauls are loaded directly into the boats of the buyers. In these cases, enumerators report landed weight estimated by the dai operator on the basis of the capacity of the boat.

Species composition and sub-sampling

If the haul weight is less than 5 kg, the total haul is sampled. Individual fish in the haul are separated into species and the number and total weight of fish belonging to each species recorded. For haul weights greater than 5 kg, sub-samples totalling between 5 kg and 15 kg are first taken from the haul.

Total length (TL) of individual fish

Within each sample or sub-sample, ecologically and economically important species (large fish, high value or high demand) are selected for length and weight measurement. The total length (TL) of each fish is measured with a measuring board accurate to one millimetre. Large fish are measured to the nearest centimetre and small fish are measured to the nearest 0.5 cm. It is left to the discretion of the enumerators to decide which species are important.

Fish size

This variable was introduced in the 2007–08 fishing season to determine the relative proportion of large and small fish. Large fish comprise a smaller proportion of the total weight of the catch, but command a higher price. They are therefore usually kept alive by the dai operators in a cage suspended under the floating platform of the dai. Fish of the same species are first sorted into large and small categories before recording their number, weight and unit value.

The weight categories as they are entered into the data entry form are therefore:

A = Total Weight of all fish per haul B = Total Weight small fish C = Sample Weight of small fish (sub-sample of B) D = Total Weight big fish

Fishing effort

Effort is measured as the number of hauls per dai per day. This is estimated by the enumerator from the average observed soak time or interval between each successive haul over the 12 or 24 hours monitoring period. Enumerators correct the estimates for time spent repairing or cleaning nets during the unobserved period following the advice of the dai operator.

To estimate the total landings made by the fishery each month from the sampled catch rates, the following additional measures of fishing effort are reported:

1) Active fishing gear: The number of dais operating during the survey month. Enumerators record the number of dais observed operating on sampling days to provide a mean estimate for the month; and

2) Active fishing days: The number of days corresponding to the Peak and Low Period each month subjectively estimated from field observations and daily catch rates.

Initially these raising factors were informally reported but they are now required to be recorded on data collection forms.



Page 52

Fish Prices

Unit prices are recorded for each species. Fish prices typically change each year according to supply (Sopha Lieng pers comm.) and are used to estimate landing values.

3.5. Catch estimation methodology

3.5.1. Aggregated total catch

The estimation of total catch for the dai fishery is based on the principles outlined in Stamatopoulos (2002). The dai unit forms the primary sampling unit. The daily catch rate (CPUE) of a dai unit is estimated as the product of the sampled weight for haul, i and the estimated number of hauls in a day:

Where d = day, m = month, p = administrative zone (province), Kandal (KAN); Phnom Penh (PP), l = lunar period, y = dai yield category, Dai = individual dai unit, wt = weight of haul, and h = estimated number of hauls in a day. The mean daily CPUE (mean total weight dai-1 day-1) for any given stratum combination within a month is then obtained by averaging across all dai unit samples in each stratum combination.

The monthly estimate of the total catch ( C ) is given by:

The total monthly fishing effort (Effort) is:

Where AD is the estimate of active days and AG is the estimate of active gears. The total catch for a month is therefore given by:

The total catch for the season (TotalCatch) is then obtained by summing the monthly catches:

Cm,p,l,y = CPUEm,p,l,y x Effortm,p,l,y

Effortm,p,l,y =ADm,p,l,y x AGm,p,l,y

p=KAN l=2 y=h

TotalCatchm= ∑ ∑ ∑ p=PP l=1 y=l

m=Mar

TotalCatch = ∑ m=Oct

CPUEd,m,p,l,y,Dai,i =wtm,p,l,y,Dai,i .hd,m,p,l,y,Dai

Figure 19 Outline of the stratification of the dai fishery sampling regime. The mean catch weight per haul (CPUE) is calculated for a dai on a day (large shaded area) within each stratum. The total catch is obtained by multiplying the stratum-specific estimate of the mean daily CPUE by the two stratum-specific raising factors: the active dais and active days.

Month (Total Catch)

Kandal Province

Low Period

Low Yield

Dai 3

Species nameNo. of fishBody weightLength (selected species)

Mean daily CPUE per stratumMean daily Effort per stratum

Phnom Penh Municipality

High Period

High Yield

Dai 1 Dai 2

Big fish sample

Small fish sub-sample

Total haul

SEASON (TOTAL CATCH)



Page 54

3.5.2. Species-wise CPUE and catch

A sub-sample of small fish is taken from each haul that is too large for a total sample, whereas all the big fish are removed from the haul as a total sample. These two samples, i.e. the total sample of large fish and sub-sample of small fish, are then separated into species, enumerated, weighed and selected species are measured to obtain Total Lengths (TL) (Figure 19 and Section 3.3.3). The species composition of the total catch in terms of weight and abundance can therefore be estimated from the proportion of species present in the small fish sub-sample to which the big fish total sample is later added.

4. The Dai Fishery database: storage and processing

4.1. Database evolution

Between 1994 and 2006, the data recorded by the data collectors was stored in ARTFISH software (Stamatopoulos, 1994; 1995). Over this period two successive versions of ARTFISH were used: the DOS version which stored data files in Dbase IV (*.dbf) format (1994–2001) and the Windows version (2001–2006) which was stored in text file (*.txt) format. Individual data files in ARTFISH were stored in nested folders tiered by administrative province, fishing season and month (Figure 20). The data files themselves were separated into fishing effort (EF), fishes landed (LN) and estimates of total catches (ES). In total, ARTFISH generated 500 files analysable only by the software itself. Additional limitations of the earlier DOS version of ARTFISH included the fact that: (1) it only captured 38 of the numerically most dominant species (Cans and Ngor, 2006) and grouped the remainder as ‘X-OTHERS’, and (2) that a maximum of only 20 species could be reported for any given landing (Baran et al., 2001b).

Figure 20 Structure of the database files held in the DOS version ARTFISH showing folders tiered by Administrative Province, season and month. Alphanumeric file labels are coded by (1) the type of file: fishing effort (EF), fishes landed (LN) and estimates of total catches (ES) and (2) landing site (2 digits), month (2 digits) and year (2 digits) (Cans and Ngor, 2006).


Page 56

Baran et al. (2001b) identified the need to transform the database into a form that would enable analyses by means of a range of different software packages. They developed an access function that merged the individual ARTFISH files into a single database and restructured the species table using the ‘restructure’ function in SPSS Version 13. A detailed description of the steps taken to merge the DOS Version files and the Access source code used to do this can be found in Appendix A of Cans and Ngor (2006).

Cans and Ngor (2006) identified several problems with the storage of effort data in the original DOS version of ARTFISH. These included: (1) the fact that mean number of hauls per day was calculated manually prior to the data being entered into the system, (2) the active days raising factor was only available in printed reports and (3) mean number of hauls per day, active days and active gears were missing in several instances in the DOS version. To resolve this, Cans and Ngor (2006) estimated the mean number of hauls per day for a gear type and period on the basis of prior records.

The basic structure of the files and folders storage system of the Windows version of ARTFISH was different from the DOS version, with the primary tier being the fishing season rather than province and different file types being stored in separate folders rather than being differentiated by file name (Figure 21). The merging into an Access database appears to have been more straightforward than it was in the case of the DOS version.

Problems identified in the Windows version of ARTFISH included: (1) species lists and species codes appeared to be mismatched from one month to the other; (2) species lists and codes were occasionally duplicated and (3) there were inconsistencies in the spelling of Khmer common names of fish species in the two lists. The species lists were therefore corrected and linked to a standard species table used by the AMCF project which incorporates the scientific name and general information on the species which was obtained from the Mekong Fish Database (2000).

The database developed by Cans and Ngor (2006) contains three types of data tables. These contain the sampling information and primary data collected, four lookup tables that contain gear, species and sampling season information and two additional tables that contain data on water level and lunar phase (Table 13). Descriptions of the fields in each of these tables appear in Annex, Table 26 – Table 34.

Figure 21 Structure of the database files held in the Windows version ARTFISH showing folders tiered by Season, Administrative province, month and file type (effort, landings, results and look up tables).

Figure 22 Entity-relationship diagram for the dai fishery database developed by Cans and Ngor (2006) and in use between 2006–2008. Bold text indicates primary keys.

Table 13 Description of the main tables in the dai fishery database developed by Cans and Ngor (2006) and in use between 2006–2009.

Tables DescriptionData Tablestbl_MainWin&Dos Main table: Stores all the sample information: date, location, gear used,

total weight of the haul, weight, value, price and fish number of the sample, number of hauls per day.

tbl_SpeciesWin&Dos Sample species table: Stores species information caught in the sample, species name, code, weight, number, value and price.

tlb_Effort Effort table: information on the effort, date, stratum, gear code, active gears and active days.

Lookup Tablestlkp_GearCode List and code of the fishing gear usedtlkp_Species Species list in Khmer with the MFD, 2000 corresponding codetlkp_SeasonYear This table provides the season year according to the datetlkp_SpeciesStandard List of fish names both English & Khmer (translation), general species

information issued from MFD, 2000Additional Tablestbl_PhnomPenh Port Water Level Water level at Phnom Penh Port from 1994 to present.tbl_moonFace Daily moon face percentage table added for further analysis

The Dai Fishery database: storage and processing


Page 58

The primary keys in the main database tables (bold text in Figure 22) linked via one-to-one or one-to-many relationships include:

(1) DOC: the document number of the paper record; (2) Date: date the sample was collected; (3) StratumNum: administrative zone (Kandal or Phnom Penh); (4) GearCode: an alphanumeric gear code that incorporates the lunar phase and dai yield

(B0001–B004) as well as the stratifications applied in the earlier sampling years (B0005–B0010) i.e. (a) no stratification (All DAI) applied prior to the 1997–98 season and (b) the stratification scheme that included or excluded Row 2 (Row2/Excl2) that was applied between the 1998–99 and 1999–00 seasons (Appendix II, Table 10–11); and

(5) KhmerNameCode: Khmer species name linking the samples of individual species with the standard species table.

4.2. Description of the current database

Since Cans and Ngor (2006) developed the Access version of the dai fishery database, it has undergone several further modifications resulting from minor changes to the sampling regime as well as the inclusion of length frequency tables (Table 14).

4.2.1. Data tables

As in the original Access database, the Main Table (tbl_MainWin&Dos), describes sample information relating to the date of the sample, the gear used (DaiID, dai size and mesh size) and gear code (stratum combination) (Table 14). It also contains aggregated catch values including: the measured or estimated total weight of each haul (TotalWeight) and the total weight of the samples taken from the haul. Each row in the database relates to a single sampled haul which can be traced to the original data entry form via a document number recorded in the DOC field. The four primary keys that link the hauls to other tables in the database are the same as those in the original database: DOC, date and GearCode and StratumNum (Administrative zone).

In addition to the way in which information about the catch is recorded, other minor changes include the manner in which dais are identified and described. In the original database, the size of the dai was described by a single figure. This, ‘SizeDai’ field in the original database is now split into three fields that describe the depth, width and length of the dai and the minimum mesh size as measured at the codend of the bagnet. The ‘Skipper’ field has been replaced by ‘DaiID’ that contains the alphanumeric code for the row number (1–15) and the transversal position across the channel (A–G). Many of the fields remain unchanged (Annex Table 36).

The second principal data table in the database, referred to as tbl_SpeciesWin&Dos, is linked to the Main table by the four primary keys and contains information pertaining to the abundance (FishNumber) and weight (Catch) of individual species in the sample (Annex Table 37). Also in this table are fields containing information on the value and price of the individual species. The 2009 format provides for the separation of fish samples into ‘big’ and ‘small’–a stratification which was introduced in 2007 and therefore incorporates three additional fields namely: sample weight of the small fish (S_Sampled weight), total weight of the small fish (S_TotalWeight), and total weight of the big fish (B_TotalWeight). The third principal data table (tbl_Effort) in the 2009 database contains information on the active days and active gears and remains unchanged from the original database (Annex Table 38).

4.2.2. Lookup Tables

The data tables are linked to four look-up tables that contain information linking the season, date and year (tlkp_SeasonDateYear), gearcodes (tlkp_GearCode) and species information (tlkp_Species and tlk_SpeciesStandard) (Annex Table 39–Table 41).

Table 14 Description of the main tables in the dai fishery database 2009. Asterisks indicates tables added subsequent to Cans and Ngor (2006).

Tables DescriptionData Tablestbl_MainWin&Dos Main table: Stores all the sample information: date, location, gear used, total

weight of the haul, weight, value, price and fish number of the sample, number of hauls per day

tbl_SpeciesWin&Dos Sample species table: Stores species information caught in the sample, species name, code, weight, number, value and price

tlb_Effort Effort table: information on the effort, date, stratum, gear code, active gears and active days

Lookup Tablestlkp_GearCode List and code of the fishing gear usedtlkp_SeasonYear This table provides the season year according to the datetlkp_Species Species list in Khmer with the MFD, 2000 corresponding codetlkp_SpeciesStandard List of fish names both English & Khmer (translation), general species

information issued from MFD, 2000Length Frequency TablesLeng_tbllengthFreq* Species, total length and number of fish in the length intervalLeng_tblSpecies* Species, total weight of the haul from which the sample was taken, weight of the

sample, weight of the sample used to obtain length informationLeng_tblLocation* Administrative zone from which the length data was obtainedKey additional TablesAnnual Hydrological Indices Flood Indices, water levels from 1994 to present, flood start and end,

flood durationtbl_LunarAge&Phase Daily moon age in days from the new moon and phase

tblOtherInfo* Beginning and end of the bagnet soak time, total catch per haul, number of hauls

The Dai Fishery database: storage and processing


Page 60

4.2.3. Length Frequency tables

The three length frequency tables were incorporated into the database in 2007. These tables detail the length and number of individuals of each species sampled (Leng_tbllengthFreq), the weight of the species sample selected for length measurements (LengtblSpecies), and location (Leng_tblLocation) (Annex Table 42–Table 44).

4.2.4. Additional tables

The key additional tables in the current database contain information on flood magnitude, timing and duration (Annual Hydrological Indices), lunar age (in days) and phase (tbl _LunarAge&Phase). Other tables include daily records of water levels at Prek Kdam (Pkdam), Phnom Penh port (Ppport), and Kampong Luong (Kluong). The table ‘Dai units by row & Et’ lists the number of dai units by season and row (Annex Table 45–Table 48).

4.3. Current Database Query descriptions

The queries in the database have been designed to extract information on individual dai and daily catch rates, to estimate the total catch for each stratum (e.g. month and season), as well as to calculate these statistics for individual species. A number of alternative approaches (queries) are available according to different assumptions regarding the structure of the data itself and/or the reliability of certain variables such as information on fishing effort. Thus variation from the application of the different methods represents “model” or “structural” uncertainty–an additional form of uncertainty in catch estimation that exists alongside the statistical uncertainty associated with estimates of variables of interest derived from the data. Details of these queries are described in a working document held at IFReDI.

Figure 23 Entity relationship diagram for the Dai fishery database in use from 2009.

5. The Ecology of the Fishery

5.1. Longitudinal (upstream/downstream) variation

5.1.1. Aggregated catch, effort and CPUE

On average, most (76%) of the catch is landed in Kandal province, upstream of Phnom Penh, where the majority (64%) of the dai units are located (Table 15). The resulting catch rates in Phnom Penh are on average less than half the rates estimated for Kandal, indicating a substantial difference (decline) in fish abundance between Kandal and Phnom Penh.

Table 15 Estimates of aggregated catch and effort by season and municipality.

Catch EffortSeason Phnom

Penh(kg)

Kandal(kg)

Total (kg) Phnom Penh (%)

Kandal (%)

Phnom Penh

(Dai units)

Kandal(Dai units)

Total (Dai units)

Phnom Penh (%)

Kandal (%)

98–99 5,951,708 5,835,913 11,787,621 50% 50% 25 38 63 40% 60%99–00 3,389,174 8,892,105 12,281,279 28% 72% 25 38 63 40% 60%00–01 5,723,004 23,417,799 29,140,803 20% 80% 25 38 63 40% 60%01–02 6,249,828 12,327,838 18,577,667 34% 66% 22 38 60 37% 63%02–03 5,043,345 10,675,150 15,718,495 32% 68% 22 39 61 36% 64%03–04 2,529,287 8,067,296 10,596,583 24% 76% 22 39 61 36% 64%04–05 4,368,836 21,526,992 25,895,828 17% 83% 22 40 62 35% 65%05–06 5,924,983 32,801,968 38,726,952 15% 85% 22 42 64 34% 66%06–07 2,947,799 19,444,630 22,392,429 13% 87% 22 42 64 34% 66%07–08 1,939,297 11,584,989 13,524,286 14% 86% 22 42 64 34% 66%08–09 1,792,376 11,217,913 13,010,289 14% 86% 22 42 64 34% 66%Mean 4,169,058 15,072,054 19,241,112 24% 76% 23 40 63 36% 64%

A detailed examination of mean sampled catch rates plotted as a function of row number or cumulative effort measured from the most upstream row of dai units closest to the Lake is indicative of a depletion of fish (Figure 24). The depletion response is most pronounced between rows 13 and 6 reflected in the significant linear decline in catch rates.

Removals of fish by other small-scale gear such as gillnets and seines operating between the dai rows may have contributed to this apparent depletion. However, information about this component of the fishery is sparse and poorly documented (see Section 1.1.4).


Page 62

The depletion response is non-linear over the range of the fishery. Fish may become more vulnerable to capture (i.e. gear catchability increases) where catch rates rise between rows 15 and 13 and again through rows 6 to row 2, possibly as the result of changing hydrological and morphological conditions which favour fish capture. Differences in small-scale fishing effort and/or migrations of fish from adjacent floodplains between dai rows might also be responsible for this non-linear response.

Assuming that both: (i) migrations of fish from adjacent floodplains to between-dai row locations and (ii) removals of fish by any small-scale gears operating between the dai rows, are negligiable, then the observed depletion response can be described by the depletion model of Delury (Hilborn and Walters, 1992). This model can provide approximate estimates of the mean numbers of fish arriving at the dai fishery during each season, the catchability coefficient of a dai unit, and the rates of exploitation by the fishery. The catchability coefficient, (q) is a measure of gear efficiency and can be defined as the proportion of the population removed by one unit of effort. Since we are considering changes in fish abundance indicated by catch rates through rows of dais, (r) instead of through time (t), the Delury model is expressed as:

Where N15 is the number of fish arriving at row 15 and Er is the cumulative fishing effort measured in dai units from the most upstream row 15. Thus, E15 = 0.

Figure 24 Mean ln transformed catch rates (1997–2009) plotted by row.

CPUEr = qN15e – qEr

Figure 25 The Delury depletion model fitted to mean ln transformed catch rates (1997–2009) expressed as the number of fish caught per dai unit per day, and cumulative fishing effort measured in dai units from the most upstream row 15.

Taking logarithms, we then get:

This form of the model can be fitted with linear regression methods where ln CPUEr is the dependent variable and Er is the independent variable. The regression coefficient (slope) gives the estimate of (q) (Figure 25). The average number of fish arriving at the dai fishery during a fishing season is estimated from the regression intercept a and the slope, b: N15 = ea/-b.

ln CPUEr = ln[qN15] – qEr

The regression model was found to be highly significant (p < 0.001) with slope, b = - 0.028 and intercept, a = 10.745. The average number of fish arriving at the dai fishery each day of each fishing season, N15 is estimated to be 1,657,052.

The model predicts that each dai removes approximately 2.8% of fish migrating through the fishery equivalent to an instantaneous fishing mortality rate, F (F = Eq) of 1.79 for 64 dai units. In other words approximately 83% of fish arriving at the dai fishery are caught. This compares with 95% estimated for the 1938–39 fishing when 106 dais were licensed to fish. Estimates of (F) and the proportions of migrating fish escaping and removed for a range of effort are given in Table 16.

The Ecology of the Fishery


Page 64

Table 16 Estimates of fishing mortality, (F) for the dai fishery for a range of fishing efforts.

Dai Units F (per season) Proportion Escaped Proportion Captured0 0 1.00 0.00

10 0.28 0.76 0.2420 0.56 0.57 0.4330 0.84 0.43 0.5740 1.12 0.33 0.6750 1.4 0.25 0.7560 1.68 0.19 0.8170 1.96 0.14 0.8680 2.24 0.11 0.89

90 2.52 0.08 0.92100 2.8 0.06 0.94110 3.08 0.05 0.95

5.1.2. Catch diversity and composition

The results of the Delury depletion analysis described above demonstrated a strong depletion response of fish abundance and biomass through the fishery. If some species are selectively removed by the gear, e.g. if mesh sizes select only larger species, then downstream changes in the species composition of the catch would be expected.

PRIMER software (Clarke and Warwick, 2001) was used to test for significant differences in species diversity and assemblage composition among dai rows and also between administrative zone (Phnom Penh Municipality and Kandal Province). Mean daily dai catch rates (N dai-1 day-1) were employed as the index of species abundance and dai units were treated as sample replicates. Only those species forming more than 0.1% of the total catch between 1997–98 and 2008–09 were included in the analyses.

Species diversity was indicated by the total number of species (Richness, S) and the Shannon Diversity Index (H) that accounts for both the abundance and evenness of the species present:

where (pi) is the proportion of species i relative to the total number of species present, (k). High values of (H) are indicative of high evenness or richness. Low values are indicative of low species rich or evenness i.e. the assemblage is dominated by a one or a few species. Both the counts of the number of species (S) and calculation of the Shannon Diversity Index were performed.

Non-parametric multi-dimensional scaling (MDS) was used to explore assemblage (dis)similarity among dai rows and between administrative zones on fourth-root transformed species abundance estimates. Assemblage similarity was described using the Bray-Curtis similarity

kH = –∑Piln Pi i=1

index. The similarity percent (SIMPER) routine was used to identify taxonomic groups that most contributed to sample(dis) similarity. To eliminate the confounding effects of month and lunar phase (see Section 5.3.2), only catch rate samples corresponding to the peak month (January) and lunar phase (second quarter) were selected in each season.

The number of species (S) in each season varied between a minimum of two and a maximum of 80 species (mean = 21.2). The Shannon Diversity Index (H) ranged between 1.1 and 2.5. Both (S) and (H) tended to decrease almost stepwise downstream between Kandal Province (Rows 7–15) and Phnom Penh municipality (Rows 1–6) (Figure 26). This pattern is particularly pronounced after 2001–02. In 2001–02, 2002–03 and 2006–07, approximately twice as many species were recorded in Kandal Province compared to Phnom Penh.

The MDS ordinations confirmed the existence of significant dissimilarities in species assemblages reported in the two administrative zones (Figure 27 and Figure 28). However, consistent patterns of dissimilarity among dai rows were not apparent in the ordinations.

In six of the 12 seasons, C. lobatus contributed the highest percentage dissimilarity between administrative zones except in 1997–98 and 2004–05 (Table 17). P. barroni and L. lineatus also contributed significantly to assemblage dissimilarity between the administrative zones in most years. However, these patterns were not consistently evident. For example, during the 1997–98 season, Parambassis apogonoides contributed the most (33.6%) to the assemblage dissimilarity but in 2000–01 and 2005–06, the greatest contribution to the dissimilarity was made by H. cryptopogon (21.9 and 29.2% respectively).

The distance between the most upstream and downstream dais does not exceed 30 km (Section 2.2) and the distance between rows 6 and 7 is no more than 500 m. Differences in (S) and (H) at the Phnom Penh/Kandal boundary may simply be an artefact of sampling rather than the existence of real longitudinal changes in species diversity.

Selective depletion of certain species downstream or species-specific catchability differences between dais was not detected by the multivariate analyses. The observed longitudinal differences in the assemblage may simply be an artefact of the capacity of sampling teams to identify and report species present in landings. Different dai sampling teams are responsible for sampling dais in the two administrative zones and these were never interchanged, thereby increasing the likelihood inconsistencies in species identification and/or enumeration. That these differences are not consistent among seasons may be due to changes in sampling personnel over the period of observations.



Page 66

(a)

(b)

Figure 26 Estimates of (a) species richness (S) (Mean±SE) and (b) Shannon Diversity Index (H) (Mean±SE) for each row from Row 1 (downstream) to 15 (upstream) for the seasons 1997–98 to 2008–09 averaged across months and lunar phases. The broken line indicates the boundary between Phnom Penh Municipality (Rows 1–6) and Kandal Province (Rows 7–15).


Figure 27 MDS ordinations based on Bray-Curtis index of similarity derived from fourth-root transformed mean CPUE for individual dais in January during the peak lunar phase (second quarter), 1997–98 to 2002–03 seasons. Black triangles represent Phnom Penh Municipality, circles represent Kandal Province and replicates are individual dais labelled by Row number. Stress values are reported at top right and vary from 0.11 to 0.16. Clusters are grouped by percentage similarity and reported at bottom left. The broken line indicates the split between administrative zones.

1

3

45

56

67

7

7

8

8

8

8

99

1010 11

11

11

12

12

12

13

14

14

2D Stress: 0.13

11

2

2223

3

3

4

4

5

5

5

6

6

8

8

8

9

10

11

12

12

14

14

15

15

15

15

2D Stress: 0.13

2

2

2

23

33 3

4

4

44

5

55

5

5

6

6

6 6

7 7

7

8

8

8

8

8

99

9

10

10

11

11

1212 14

14

2D Stress: 0.11

2

2

2

2

3

4

4

5

58

8

99

10

10

10

1011

11

11

1213

14

1515

2D Stress: 0.15

1

1

1

22

2

2

3

3

3344 4

5

5

5

5

6

6

8

91010

10

10 11

1212

13

1414

14

2D Stress: 0.12

22

2

2

3

3

3

3

4

4

4

5

5

6

6

7

7

77

8

8

9

9 9

10

10

10

1111

12

12

12

13

14

14

14

2D Stress: 0.11

1997-98

60%

55%

55%

55%

55%

55%

1999-00

2001-02

1998-99

2000-01

2002-03


Page 68

2

2

2

2

3

3

34

4

5

5

5

6

6

6

6

7

7

8

8

8

9

9

10

10

10

10 11

11

12

12

12

12

13

14

14

15

15

15

2D Stress: 0.15

2

2

2

2

3

3

33

4

4

4

4

5

5

5

5

6

66

6

7

7

78

8

8

8

9

9

9

10

10

10

10

10

11

11

11

11

12

12

12

12

12

13

14

14

14

1515

2D Stress: 0.18

2 2

22

33 3

3

4

44

4

5

5

5

5

5

66

6

66

3

777

7

7

8

8

8

88

8

99

9

10

10

10

10

10

10

10

1111

11

11

11

12

12

12

12

1212

13

14

1414

15

151515

15

2D Stress: 0.21

2

2

2

2

3

3

3

34

4

4

4

5

56

6

6

777

8

8

8

9

9

10

10

10

10

111112

12

12

12

12

13

14

14 14

1515

2D Stress: 0.16

2

2

2

2

3

3

3

3

4

4

4

5

5

5

5

5

6

6

6

67

77

8

8

8

8

8

9

9

10

10

10

10

10

111111 12

12

12

1314

14

14

15

15

1515

2D Stress: 0.15

2

2

2

3

3

3

3

4

4

4

5

5

5

5

6

6

7

77

7

7

8

88

9

910

10

11

11

11

11

12

1212

12

12

13

14

14

15

15

15

2D Stress: 0.17

Figure 28 MDS ordinations based on Bray-Curtis index of similarity derived from fourth-root transformed mean CPUE for individual dais in January during the peak lunar phase (second quarter), 2003–04 to 2008–09. Black triangles represent Phnom Penh Municipality, circles represent Kandal Province and replicates are individual dais labelled by Row number. Stress values are reported at top right and vary from 0.1 to 0.17. Clusters are grouped by percentage similarity and reported at bottom left. The broken line indicates the split between administrative zones.

2003-04

55%

60%

65%

60%

60%

60%

2005-06

2007-08

2004-05

2006-07

2008-09

Tabl

e 17

Spe

cies

con

trib

utin

g up

to 8

0% o

f mea

n cu

mul

ativ

e pe

rcen

tage

(Cum

%) d

issi

mila

rity

bet

wee

n Ph

nom

Pen

h M

unic

ipal

ity a

nd K

anda

l Pro

vinc

e in

ea

ch se

ason

.

Spec

ies

1997

–98

1998

-99

1999

-00

2000

-01

2001

-02

2002

-03

2003

-04

2004

-05

2005

-06

2006

-07

2007

-08

2008

-09

Mea

n%

Cum %

Cir

rhin

us lo

batu

s9.

645

.933

.49.

222

.215

.427

.44.

29.

31.

618

.97.

320

.020

.0Pa

rala

ubuc

a ba

rron

i9.

915

.38.

214

.07.

38.

312

.01.

633

.34.

614

.31.

311

.631

.6La

biob

arbu

s lin

eatu

s-

14.9

6.9

7.5

12.8

1.9

1.9

9.1

2.9

26.5

1.2

8.8

10.0

41.6

Hen

icor

hync

hus c

rypt

opog

on-

--

21.9

12.3

6.6

13.9

11.2

29.2

3.6

16.8

-9.

651

.2H

enic

orhy

nchu

s sia

men

sis

--

-2.

56.

85.

73.

911

.37.

61.

56.

219

.86.

257

.4Pu

ntio

plite

s pro

ctoz

ysro

n-

--

-0.

7-

--

-8.

59.

632

.24.

261

.7C

lupe

icht

hys a

esar

nens

is18

.2-

1.8

4.3

0.8

5.6

3.9

2.9

--

1.9

-3.

264

.8La

beo

chry

soph

ekad

ion

--

1.8

-0.

70.

71.

7-

-16

.63.

111

.83.

067

.8Pa

ram

bass

is a

pogo

noid

es33

.6-

--

-0.

7-

--

--

-2.

970

.7La

biob

arbu

s sia

men

sis

--

1.2

6.2

2.8

4.4

1.2

--

--

-2.

172

.7Th

ynni

chth

ys th

ynno

ides

-3.

93.

16.

31.

32.

23.

1-

-2.

41.

6-

2.0

74.7

Ost

eoch

ilus l

ini

6.6

3.9

-3.

70.

92.

21.

4-

0.8

--

-1.

676

.3Pa

rach

ela

siam

ensi

s-

-1.

14.

21.

88.

11.

5-

--

-2.

21.

677

.9Ya

suhi

kota

kia

mor

leti

-4.

36.

34.

14.

1-

--

--

--

1.5

79.4



Page 70

Figure 29 Estimates of mean weight of the fish assemblage sampled from the dai fishery by row and season. Estimates are weighted by numbers of fish sampled.

5.1.3. Fish size (weight)

An analysis of variance (ANOVA) indicated that the mean weight of fish caught using nets with the most common mesh size (i.e. mesh range from 100 to 15 mm) recorded in the database for both municipalities, does not vary significantly with dai row (Table 18). Mean weights were however found to vary significantly with both month and season (see Sections 5.3.3 and 5.4.3). This suggests that the trends illustrated in Figure 29 are likely to reflect the use of nets with larger meshes in the downstream rows of the fishery, particularly in the Phnom Penh Municipality (rows 1–6), rather than species or fish size-specific changes in gear catchability with dai row. However, this could not be confirmed because of the generally poor quality and paucity of mesh size records in the database but the available evidence supports this interpretation at least for chieu nets (see Section 2.3). Effort should be directed towards improving the quality and quantity of these records in the database.

Table 18 ANOVA test results for the null hypothesis that the average weight of fish (irrespective of species) landed by the fishery does not change with dai row number after accounting for the effects of month, season and mesh size. Mean weight estimates by dai row (1–15), month (October–March) and season (1998–99 to 2007–08) were weighted by the numbers of individual fish sampled.

SourceType of

III Sum of Squarres

df Mean Square F Sig.

Corrected Model .003a 32 8.839E-05 7.131 .000

Intercept .002 1 .002 130.777 .000

ROW 1.024E-05 1 1.024E-05 .826 .366

SEASON .002 8 .000 16.149 .000

MONTH .000 4 6.107E-05 4.927 .001

SEASON *MONTH .001 18 7.662E-05 6.181 .000

Error .001 100 1.239E-05

Total .019 133

Corrected Total .004 132

Test of Between-Subjects Effects, Dependent Variable: Weight (kg)a R Squared = .695 (Adjusted = .598)

5.2. Variation in catch rates among individual dais

In February 2008, IFReDI undertook a survey over a four day period (9–13 February 2008) to estimate the depth of water and water velocity below each dai unit as potentially important factors responsible for the observed variation in catch rates among individual dai units. Depths ranged from 8 to 28 m and velocity from 0.2 to 0.57 m/s (Table 19). Most dais had fish depths of between 10 and 14 m, but much deeper water is found below rows 2, 3, 4 and 14 (Figure 30). Water velocities of between 0.3 and 0.5 m/s are common throughout the fishery but below this range only around row 2 (Figure 31). As expected, water velocity declines with depth (Figure 32). Neither water depth, or velocity was found to have any detectable effect on catch rates during the survey period (Figure 33 and Figure 34)



Page 72

Table 19 Estimates of water velocity and water depth below each dai recorded in February 2008.

Dai Row Water Velocity (m/s) Depth (m)2A 2 0.22 24.02B 2 0.20 24.52C 2 0.30 26.02D 2 0.32 28.03A 3 0.21 17.03B 3 0.29 17.03C 3 0.39 18.53D 3 0.31 20.04A 4 0.34 15.04B 4 0.31 17.04C 4 0.34 19.04D 4 0.35 19.05B 5 0.37 10.05C 5 0.43 11.05D 5 0.42 12.05E 5 0.42 14.05F 5 0.41 15.06C 6 0.54 16.06D 6 0.45 15.06E 6 0.43 13.06F 6 0.44 10.06G 6 0.41 9.57C 7 0.36 13.07D 7 0.37 14.07E 7 0.37 16.07F 7 0.45 16.07G 7 0.37 13.58C 8 0.39 13.08D 8 0.39 13.58E 8 0.46 15.08F 8 0.39 12.58G 8 0.32 13.08H 8 0.28 10.09A 9 0.41 16.59B 9 0.32 15.09C 9 0.29 13.010A 10 0.37 13.010B 10 0.41 11.010C 10 0.57 10.010D 10 0.49 9.510E 10 0.44 9.510F 10 0.42 8.010G 10 0.34 8.0

11A1 11 0.33 12.511A 11 0.36 12.011B 11 0.36 14.011C 11 0.32 13.011D 11 0.34 12.012A 12 0.35 10.012B 12 0.33 11.012C 12 0.36 11.512D 12 0.35 13.012E 12 0.35 12.512F 12 0.35 11.013A 13 0.28 13.014A 14 0.34 22.014B 14 0.40 18.014C 14 0.34 16.0

Figure 30 Estimates of mean depth below each dai row.



Page 74

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0.50

0.55

0.60

5 10 15 20 25 30

Depth (m)

Wat

er v

eloc

ity (m

/s)

Figure 31 Estimates of mean water velocity at each dai row.

Figure 32 The relationship between water velocity and depth (R2 = 0.18; p < 0.001).

Figure 33 Mean loge-transformed catch rates during each lunar quarter (1–4) of the survey month (February 2008) plotted as a function of the estimated water depth below the dai unit.



Page 76

Figure 34 Mean loge-transformed catch rates during each lunar quarter (1–4) of the survey month (February 2008) plotted as a function of the estimated water velocity at each dai unit.

5.3. Intra-annual variation

5.3.1. Abundance (CPUE)

The dai fishery operates during the falling water period targeting the migrations of fish as they move from the Lake to the Mekong mainstream with the receding flood waters. During this period, fish abundance indicated by daily catch rates of sampled dai units exhibits significant variation both between and within the six months comprising the fishing season. Catch rates can vary from approximately 1 kg to 80,000 kg.dai-1.day-1 (Figure 35–Figure 38).

Typically catch rates peak (mean = 10,818 kg.dai-1.day-1; ± SD 18,708) during January (nine out of the 12 seasons examined) and are lowest (mean = 86.9 kg.dai-1.day-1; ± SD 438.7) in October (eight out of 12 seasons) (Figure 39 a).

In all the seasons examined, mean catch rates peaked during the second quarter of the lunar cycle and were lowest during the 4th quarter (Figure 39 b). Here the lunar quarters relate to four consecutive seven day periods starting from the new (dark phase) moon. Quarter 2 therefore relates to the period of approximately 7–14 days after the new moon when between approximately 50–100% of the moon is visible. This period, between what are commonly termed the first quarter and full moon phases, is also known as the ‘Waxing Gibbous’ phase.

Overall, mean fish abundance indicated by loge-transformed daily catch rates of sampled dais (1997–98 to 2008–09) varied significantly (p < 0.001) between months and between the four phases (or quarters) of the lunar cycle. In addition to significant inter-annual variation in mean fish abundance, fish migrations therefore appear to be strongly influenced by the lunar cycle, and possibly also water levels as peak migrations typically occur around January or December coinciding with the end of the flood season.



Page 78

Figure 35 Mean dai CPUE (kg) calculated for each day of the fishing season during (a) 1997–98, (b) 1998–99 and (c) 1999–2000.


Page 80



Page 82

(a)

(b)

Figure 39 loge-transformed average daily catch rates (kg/dai/day) by (a) month and (b) lunar phase for the seasons 1997–98 to 2008–09.

5.3.2 Species diversity and similarity

Sections 6.3.1 and 6.3.2 illustrated how fish migrations tend to peak with each ‘Waxing Gibbous’ phase of the lunar cycle and towards the end of the flood period (December–January). Species richness in catch samples was found to peak either during November (2002–03; 2004–05; 2006–07 and 2007–08) or December (five seasons: 1997–98; 1999–00; 2000–01; 2005–06 and 2008–09) (Figure 40 a) consistent with anecdotes (Chhun Haing Tong, pers comms). The Shannon-Weiner index of diversity peaked more frequently during November (eight of the 12 seasons) (Figure 40 b). These differences in the timing of peak values reflect the dominance of one or more species in the catch early in the season.

The species richness of the assemblage caught by the fishery does not appear to vary significantly during the lunar cycle, although there is some evidence that species richness is marginally higher during phase 1 and 2 between 2002–03 and 2008–09 (Figure 41 a). The diversity index did not display any consistent trends or patterns associated with lunar phase (Figure 41 b), suggesting that either the lunar phase has little or no species-specific effect on migration or it may be that a quarterly (lunar phase) resolution is insufficient to detect such effects.

However, ordinations illustrating the rank similarities in the assemblage composition between months in each season during the second quarter of the lunar cycle showed evidence of a change in composition during the season (Figure 42 and Figure 43). Although these trends were not consistent among seasons, there was some evidence of assemblages dissimilarity at the start (October/November) and end (January/February/March) of the season. This apparent shift in the composition of the assemblage was attributable to: (i) a decline in relative abundance of Clupeichthys aesarnensis at the start of the season (8 out of the 12 seasons); (ii) an increase in the relative abundance of most other species, particularly the Cirrhinus, Labiobarbus and Henicorhynchus genera from October to January; and (iii) a decrease in the relative abundance of these three genera thereafter (Table 20). These patterns are consistent with the results for the univariate diversity indicators described above.

Ordinations of the rank similarity of the assemblage composition among the four lunar phases in January each season revealed some evidence of assemblage dissimilarity between the first (Phases 1 and 2) and second half (Phases 3 and 4) of the lunar cycle, but not consistently among seasons (Figure 44 and Figure 45). Nonetheless, these signals do suggest that the intensity of migrations among species varies during the lunar cycle.



Page 84

(a)

(b)

Figure 40 Estimates of mean (±95% CI) (a) species richness (S) and (b) Shannon Diversity Index (H) (Mean ±95% CI) of the assemblage sampled during phase 2 of the lunar cycle of each month and season.

(a)

(b)

Figure 41 Estimates of mean (±95% CI) (a) species richness and (b) the Shannon Diversity Index (H) of the assemblage sampled during each quarter (phase) of the lunar cycle in January of each season.



Page 86

Figure 42 MDS ordinations illustrating rank similarities in the species assemblage sampled each month during lunar phase 2, 1997–98 to 2002–03. Samples from phase 1 were used in those cases when no samples were available for phase 2. Similarities between samples were described using the Bray-Curtis index using fourth-root transformed mean catch rates sampled from individual dais (O = October, N = November, D = December, J = January and F = February). Stress values range from 0.12 to 0.16. Clusters samples are grouped by percentage similarity and these values are reported at bottom left.

1997-98

40%

1998-99

30%

1999-00

30%

2001-02

30%

2000-01

40%

2002-03

40%

Figure 43 MDS ordinations illustrating rank similarities in the species assemblage sampled each month during lunar phase 2, 2003–04 to 2008–09. Samples from phase 1 were used in those cases when no samples were available for phase 2. Similarities between samples were described using the Bray-Curtis index using fourth-root transformed mean catch rates sampled from individual dais (O = October, N = November, D = December, J = January and F = February). Stress values range from 0.1 to 0.14. Clusters samples are grouped by percentage similarity and these values are reported at bottom left.


2003-04

40%

2005-06

40%

2007-08

40%

2004-05

40%

2006-07

40%

2008-09

40%


Page 88

Table 20

The seasonal frequency of upward (↑) and dow

nward (↓) changes in the abundance of com

mon species betw

een consecutive months, 1997–98 to

2008–09. The average percentage contribution made by each species to the overall sam

ple dissimilarity for each com

parison, generated using the SIM

PER routine of the the PRIMER softw

are, is also given.

October–N

ovember

Novem

ber–Decem

berD

ecember–January

January–FebruaryFebruary–M

arch

↑↓

%↑

↓%

↑↓

%↑

↓%

↑↓

%

Acantopsis sp.2

0.62

0.61

0.4Albulichthys albuloides

10.3

10.3

Amblyrhynchichthys m

icracanthus1

0.4

Barbichthys nitidus2

0.8Belodontichthys truncatus

10.5

Yasuhikotaki helodes1

0.32

0.7

Yasuhikotaki modesta

20.7

31.4

10.3

10.5

Yasuhikotaki morleti

42.3

31.4

31.5

31.4

Clupeichthys aesarnensis

82

5.52

0.71

31.8

32

2.22

1.2C

yclocheilichthys armatus

10.4

10.3

Cyclocheilichthys enoplos

11

1.2G

yrinocheilus aymonieri

10.6

10.6

Henicorhynchus cryptopogon

22

1.57

2.89

4.66

34.6

52.6

Cirrhinus lobatus

37

4.411

6.21

105.8

92

5.84

12.7

Henicorhynchus siam

ensis6

2.61

31.6

31.2

52.5

Labeo chrysophekadion1

0.5Labiobarbus lineatus

51.9

73.2

62.4

72.8

51

2.8Labiobarbus siam

ensis1

0.71

0.73

1.32

11.3

10.5

Hem

ibagrus nemurus

10.6

Osteochilus lini

10.3

31.4

11

1.52

1.01

0.6Pangasius pleurotaenia

10.4

Parachela siamensis

10.3

Paralaubuca barroni2

32.6

94.8

210

6.37

14.3

42.3

Parambassis apogonoides

10.3

10.5

Parambassis ranga

11

0.71

0.51

0.7Puntioplites proctozysron

10.6

Rasbora tornieri1

0.3Thynnichthys thynnoides

10.5

21.2

21.0

21.0

10.5

Count of num

ber of species10

172

156

1115

612

2

Figure 44 MDS ordinations illustrating rank similarities between the species assemblage sampled during each lunar phase of January, 1997–98 to January, 2002–03. Similarities between samples were described using the Bray-Curtis index using fourth-root transformed mean catch rates sampled from individual dais. Stress values range from 0.1 to 0.14. Clusters are grouped by percentage similarity reported in the bottom left corner of each ordination.


1997-98

40%

1998-99

40%

1999-00

40%

2001-02

40%

2000-01

40%

2002-03

40%


Page 90

1

1

1

1

1

1

11

1

1

1

1

1

1

1

222

22

2

2

2

2 222 22

22

22

2

2

22

22

2

2

2

2

2

2

22

2 2

2

2

2

2

2

2

2

2

3

33

3

3

3

3

3

3

3

33

3

3

3

3

4

44

4

4

4

4

44

4

4

444

4

44

4

4

2D Stress: 0.11

Figure 45 MDS ordinations illustrating rank similarities between the species assemblage sampled during each lunar phase of January, 2003–04 to January, 2008–09. Similarities between samples were described using the Bray-Curtis index using fourth-root transformed mean catch rates sampled from individual dais. Stress values range from 0.08 to 0.12. Clusters are grouped by percentage similarity reported in the bottom left corner of each ordination.

2003-04

40%

2005-06

40%

2007-08

40%

2004-05

40%

2006-07

40%

2008-09

40%

Figure 46 Changes in mean fish weight (all species combined) through the fishing season (1997–98 to 2008–09). Estimates of the mean are weighted by numbers of fish sampled.

5.3.3. Size (weight)

In spite of significant inter-annual variation, changes in the mean weight of fish caught during the fishing season follow a largely consistent pattern with a rapid rise in fish size at the start of the season which peaks in November or December and is then is followed by a decline (Figure 46). This is consistent with the monthly changes in species diversity reported in Section 5.3.2.



Page 92

These results suggested that the mean weight changes observed during the season reflect changes in the timing of migration of species of different sizes with a greater abundance of larger species migrating during the flood recession in November and December compared to other months.

Monthly changes in mean weight for the most abundant species are illustrated in Figure 47. Lobocheilos crytopogon exhibits a monotonically increasing weight through time with some slowing of growth following the end of flooding typical of the pattern of growth of species inhabiting tropical rivers (Welcomme, 1985). The remaining species, except O. lini, also exhibit evidence of declining growth rates through the season, but often with deviations above and then below the expected trajectory. This may be indicative of larger individuals migrating first during the falling water period (November and December) followed by smaller juvenile individuals.

The tendency for both larger species and larger individuals of each species to leave the Great Lake earlier than smaller fish was reported by workers during the 1950’s as well as in other tropical river systems (see Welcomme 1985 for review). Dai fishers also report this behaviour (Section 2.3). Given that large fish also often fail to leave deeper floodplain pools (Welcomme, 1985) suggests that depth is a major factor affecting the timing of refuge-seeking fish migrations along with dissolved oxygen and temperature.

Figure 47 Mean weight changes by month for the six most abundant species (1997–98 to 2008–09). Estimates of the mean are weighted by numbers of fish sampled.

5.4. Inter-annual variation

5.4.1. Catch, effort and CPUE

During the 12 year period for which effort data are believed to be reliable, estimates of total annual catch for the fishery (aggregated across species) has varied from between approximately 8,000 and 33,000 tonnes, with a mean of approximately 15,000 tonnes but with no obvious trend (Figure 48 a). Very high catches were observed in2004–05 and 2005–06. Effort has ranged from 60 to 68 dai units, again with no obvious trend (Figure 48 b). Catch per dai per season–an index of fish biomass (Figure 48 c) exhibits a similar coefficient of variation as catch per season (approximately 46%) and there is also no obvious trend through time. Catch rates ranged from approximately 120 to 530 tonnes dai-1 season-1 with a mean of approximately 240 tonnes dai-1 season-1.

5.4.2. Catch diversity and species composition

Historical records of the species composition of the dai catches date from surveys undertaken in the 1930s (Chevey and Le Poulain, 1940) and 1960s (Fily and D’Aubenton, 1965), but as pointed out in Section 3.2, consistent and regular sampling of the fishery dates from the 1994–95 fishing season. Comparing catches from the early surveys with those of the 1994–95 period, Lieng et al. (1995) found that the dominant species had remained largely unchanged with catches being dominated by the Cirrhinus, Lobochelius and Henicorhynchus genera (collectively classified as Henicorhynchus species) that had the highest relative abundance in the 1938–39 season and comprised 25.4% and

05,00010,00015,00020,00025,00030,00035,00040,000

97-9898-9999-0000-0101-0202-0303-0404-0505-0606-0707-0808-09

Cat

ch (t

onne

s)

50

55

60

65

70

75

80

97-9898-9999-0000-0101-0202-0303-0404-0505-0606-0707-0808-09

Effo

rt (d

ai u

nits

)

0

100

200

300

400

500

600

97-9898-9999-0000-0101-0202-0303-0404-0505-0606-0707-0808-09

Season

CPU

E (to

nnes

/dai

)

Figure 48 (a) Total Catch, (b) effort and (c) CPUE (1997–98 to 2008–09).



Page 94

67.5% of the catch in the 1962–63 and 1994–95 seasons respectively (Table 21). The 1962–63 season differed from the other two seasons in that Cirrhinus microlepis (18.6%) and Thynnichthys thynnoides (14.9%) were second and third most dominant respectively. Paralaubuca barroni (listed as typus) comprised only 0.3% of the catch in 1962–63 whereas it comprised 13.4% in the 1994–95 season.

Table 21 Estimated catches of the most abundant species reported for the years 1938–39 (Chevey and Le Poulain, 1940), 1962–63 (Fily and D’Aubenton, 1965), 1994–95 (Lieng et al., 1995) – the latter estimated as a sum of the catches for December, January and

February 1994 (adapted from Lieng et al., 1995).

1938–39 1962–63 1994–95

Species Rel. abundance

% Total Catch

Total Catch (tonne)

% Total Catch

Henicorhynchus spp. 25.4 12,432 67.5Paralaubuca typus 1 0.30 2,460 13.4Osteochilus hasseltii2 Dangila spp.3 - 979 5.3Thynnichthys thynnoides 14.9 515 2.8Morulius chrysophekadion4 2.0 471 2.6Cirrhinus microlepis 18.6 398 2.2Botia spp.5 0.2 270 1.5

1 Paralaubuca typus (Bleeker, 1864) (Trey slak russey) is currently listed in the Dai Database as Paralaubuca barroni (Fowler, 1934).2 Lieng et al. (1995) and Baran et al. (2001) lists Osteochilus hasseltii (Valenciennes, 1842) as (Trey kros). In the current database, (Trey kros) is listed as Osteochilus lini.3 Dangila sp. is a synonym for a Labiobarbus sp. Lieng et al. (1995) list this species as (Trey khnawng veng). In the current database (Trey khnawng veng) is listed as L. lineata (Sauvage, 1881). The correct spelling for this species, however is lineatus.4 Morulius chrysophekadion (Bleeker, 1850) (Trey kaek) is a synonym for Labeo chrysophekadion (Bleeker, 1850) as it is currently listed in

the database.5 For the purposes of comparison, the three Botia (now named Yasuhikotakia) spp. were combined in the 2008–09 catch estimation.

Baran et al. (2001b) analysed the species composition of the dai catches between 1995 and 2000 by means of Principle Components Analysis, focussing on the 35 most significant species. They concluded that although the relative abundance of some species changed between years, there appeared to be no significant inter-annual variability in catch composition for the bulk of the species. Subsequent to Baran et al. (2001b) report, there have been no further analyses of species composition.

In this section, data from the 1997–98 season (the first reliable catch estimates for reasons outlined in Section 3.3) to present (2008–09) are examined for evidence of changes in species composition following the methods of Clark and Warwick (2001).

Consistent with the historical record, the Cirrhinus, Lobochelius and Henicorhynchus genera (previously referred to as the Henicorhynchus group) were found to be the most important genera in the catch, contributing between 2,715 to 12,712 tonnes, or 20% and 50% of the annual catch, respectively. The other dominant species consistent with the historical record include: P. typus, L. lineatus, T. thynnoides and L. chrysophekadion (formerly listed as M. chrysophekadion). There is evidence of an increasing contribution to the total catch from the Cirrhinus, Lobochelius,

Henicorhynchus and Paralaubuca genera from 1997–98. There is also considerable inter-annual variation in the contribution made by other species (Other spp.), particularly during the earlier seasons (1997–2003) (Figure 49). However, as with the diversity indices, preliminary investigations appear to support the conclusion that these trends may be an artefact of changes to the survey techniques, i.e. sample stratification and/or species identification, and therefore may not be biologically significant.

Figure 49 The species composition of the dai fishery catch, 1997–98 to 2008–09.

4,270

1,345

671592400

4,874

97- 98 12,624 t

3,333

1,179207336

140

98 - 99 9,373 t

2,759

775222369

535

7,155

99 - 00 11,954 t

6,184

2,741

807937

440

9,423

00 - 01 21,346 t

3,642

605

1,486

290597

11,886

01- 02 18,505 t

2,863

814

1,090

475417

9,230

02- 03 14,890 t

2,715

462594

177181

1,695

03 - 04 5,869 t

9,386

2,508

859224400

5,397

04 - 05 18,733 t

12,712

6,754

1,814455769

10,663

05 - 06 33,640 t

9,321

3,386

2,336

670774

5,481

06 - 07 21,969 t

4,844

1,426

1,279

271293

2,707

07- 08 10,823 t

5,126

1,081

1,327

246235

2,174

08 - 09 10,190 t

3,824

Heicoryhnchus spp. Paralaubuca barroni Labiobarbus lineatus

Thynnichthys thynnoides Labeo chrysophekadion Other spp.



Page 96

The greatest number of species reported caught by a single dai on a day was 80 by dai 12B on 22 January 2002. The second and third highest number of species reported to have been caught on a dai on a day (75 and 77 species respectively) occurred on the two days prior to this date (20–21 January 2002). The mean number of species caught by a dai on a day across all seasons was 22.8. The highest diversity (H) was recorded on these same three days: 20, 21 and 22 January 2002: 4.01, 3.99 and 3.97 respectively.

The mean species richness and diversity of the dai catches across all months and lunar periods show evidence of considerable interannual variability (Figure 50 a and b). After a decline from 1997–98, estimates of the mean number and diversity of species caught by a dai on a day show a steady increase between 1998–99 and 2002–03. Thereafter, aside from a peak in 2006–07, the values decline and then stabilise to a mean of approximately 26 species/dai/day. Thus the trends suggest an overall increase in the number and diversity of species has occurred over the twelve year period. However, strong parallels exist with the number of hauls sampled in each season (Section 3.3, Figure 17) suggesting that these trends merely reflect changes in sampling effort and/or the skill of enumerators rather than changes in the fish community itself. Sampling effort in terms of the total number of hauls sampled each season has varied considerably since the start of the monitoring period. A total of 435 hauls were sampled in 1997–98, with this number increasing to 2,426 hauls in 2001–02. Sampling effort then declined somewhat after 2003–04 and this is paralleled by the sharp drop in species numbers and diversity over this season. The only inconsistencies in the parallels between species diversity and effort is that the former peaked in 2006–07, whereas the latter peaked in 2007–08.

Multivariate analyses revealed significant differences (p < 0.01) in relative species abundance among fishing seasons (Figure 51–Figure 53). Whilst these differences were not significant for every pair of seasons tested using the ANOSIM routine, differences in the assemblage before and after the 2000–01 season were evident in the MDS plots.

Catch records from the earliest surveys (1930s), suggest that the catch composition has remained relatively unchanged through the seasons with the smaller, short-lived annual cyprinids belonging to the Cirrhinus, Lobocheilos, Henicorhynchus, Paralaubuca and Labiobarbus genera remaining the most dominant in terms of weight and number. Other important non-cyprinid fish important in the catch include the larger Pangasiid catfish such as Pangasius larnaudii as well as the Yasuhikotakia genus (e.g. Y. modesta) belonging the Cobitidae family (Table 22). Whilst there is some evidence that these species have increased in relative abundance through time, this trend may simply reflect sampling error. Efforts to minimise sampling error or species mis-identification will be required to determine if the assemblage utilising the TS-GL System is changing with time.

Figure 50 (a) The Species Richness (S) (Mean ±95% CI) and (b) Shannon Diversity Indices (H) (Mean ±95% CI) of fish species caught by a dai on a day for the seasons 1997–98 to 2008–09 averaged across all sampling months in each season.

(b)

(a)



Page 98

Table 22 Principal families and species making up 99% of the dai catches listed in descending order of their total percentage contribution. The number of species in each family is reported together with the three dominant species by weight in each family, also in descending order of their percentage contribution.

Family No. of species SpeciesCyprinidae 53 Cirrhinus lobatus

Paralaubuca typusHenicorhynchus cryptopogon

Pangasiidae 9 Pangasius larnaudiiPangasius pleurotaenia

Pangasianodon hypophthalmusCobitidae 5 Yasuhikotaki modesta

Yasuhikotaki morletiYasuhikotaki helodes

Siluridae 8 Belodontichthys truncatusPhalacronotus micronemusWallago attu

Gyrinocheilidae 1 Gyrinocheilus aymonieriSoleidae 1 Synaptura marginataBagridae 8 Hemibagrus nemurus

Mystus singaringanHemibagrus wyckii

Clupeidae 4 Clupeichthys aesarnensisTenualosa thibaudeauiAnodontostoma chacunda

Engraulidae 3 Coilia lindmaniSetipinna melanochirLycothrissa crocodilus

Cynoglossidae 1 Cynoglossus feldmanniSciaenidae 1 Boesemania microlepisBelonidae 1 Xenentodon cancilaNotopteridae 2 Chitala ornata

Notopterus notopterusPolynemidae 3 Polynemus multifilis

Polynemus dubiusPolynemus borneensis

Bagriichthidae 1 Bagrichthys majusculus

2D Stress: 0.25

2D Stress: 0.22

97-98

98-99

99-00

00-01

01-02

02-03

03-04

04-05

05-06

06-07

07-08

08-09

Figure 51 MDS ordinations based on Bray-Curtis indices of similarity derived from fourth-root transformed mean CPUE for individual dais for the 1997–98 to 2008–09 seasons for (a) October and (b) November.



Page 100

2D Stress: 0.19

2D Stress: 0.17

97-98

98-99

99-00

00-01

01-02

02-03

03-04

04-05

05-06

06-07

07-08

08-09

Figure 52 MDS ordinations based on Bray-Curtis indices of similarity derived from fourth-root transformed mean CPUE for individual dais for the 1997–98 to 2008–09 seasons for (a) December and (b) January.

2D Stress: 0.21

2D Stress: 0.19

97-98

98-99

99-00

00-01

01-02

02-03

03-04

04-05

05-06

06-07

07-08

08-09

Figure 53 MDS ordinations based on Bray-Curtis indices of similarity derived from fourth-root transformed mean CPUE for individual dais for the 1997–98 to 2008–09 seasons for (a) February and (b) March.


5.4.3. Size (weight)

Assuming that recruitment is constant, a decline in the mean weight of individuals in a population through time is indicative of increasing rates of exploitation as fewer large (older) individuals survive with time (Sparre and Venema, 1992).


Page 102

Figure 54 Mean weight of fish landed (all species combined) 1997 – 2009. Estimates of the mean are weighted by numbers of fish sampled.

Figure 55 Trends in the mean weight of the six most abundant species in the fishery estimated for December each year (1995–96 to 2008–09). Estimates of the mean are weighted by numbers of fish sampled.

Estimates of mean weight for all species combined as well as those species that contribute to the majority of the catch by weight have shown considerable variation during the fifteen year monitoring period, but with no evidence of a continuous monotonic decline (Figure 54 and Figure 55).

5.5. Hydrological effects on fish biomass

The hydrology of the TS-GL System and the basin as a whole may change in the future as a consequence of basin development activities such as hydropower dam construction and water abstractions or diversions for irrigation, as well as climate change. Improving our understanding of hydrological influences on the fisheries resources of the TS-GL System potentially offers a means to predict how the fisheries resources in this system and in other parts of the LMB are likely to respond to future hydrological conditions under different scenarios of basin development and climate change.

Hydrological effects on fish populations have been described for several tropical and European river systems from as early as 1910. Typically, these have been described in the form of correlations or linear regressions between indices of flood extent and duration and annual catch or some other index of fish biomass such as catch per unit of effort, CPUE. Other workers (e.g. Dudley 1972; Kapetsky, 1974) have demonstrated correlations between hydrological indices and annual growth increments of fish (see Welcomme, 1985 for review).

Several workers including Lieng et al. (1995); Baran et al. (2001 b) and van Zalinge et al. (2004) have described similar effects for the TS-GL System using regressions between annual catch estimates for the dai fishery and maximum annual water level measured at Phnom Penh as a proxy of flood extent. However, these earlier models took no account of changes to fishing effort in the fishery and the selected flood indices accounted only for the extent of flooding over a relatively short period of time 1–31 days and therefore poorly described the inter-annual changes in flood duration.

Below, we describe a model that employs the mean daily catch rate of a dai unit as the biomass index. A flood index (FI) is used to quantify both the extent and duration of the flood each year, (y):

Where (FAy,d) is the flooded area of the TS-GL System in year (y) on day (d), measured above the mean flooded area for the model period 1 January 1997 to 31 March 2009.

Estimates of (FAy,d) were derived from daily observations of water level (WLy,d) at Kampong Luong gauging station in the Great Lake (Figure 56 and Figure 57) and the following second-order polynomial (John Forsius pers comms):

FAy,d = 716.64 + 1094.19WLy,d + 30.05WL2y,d

FIy = ∑ FAy,d d



Page 104

Figure 56 The location of the Kampong Luong gauging station in the TS-GL.

Figure 57 Daily water level (m) measured at Kampong Luong gauging station and estimates of the flooded area of the TS-GL System, 1997 to 2009.

-1.5

-1

-0.5

0

0.5

1

1.5

1

Water Level / Flooded Area

Mean WL / FA

Flood duration

Max WL / FA

Min WL / FA

Flood amplitude

Rising waterduration

falling waterduration

Time(Days)

Floodstart

Floodend

Dra

wdo

wn

Floo

d ris

e

Flood index

Dry seasonindex

Dry season duration

Figure 58 Illustration of the estimation of hydrological indices. WL–Water level; FA–Flooded area.



Page 106

A general linear model (GLM) was used to test whether the flood index had a significant effect on fish biomass indicated by sampled loge-transformed daily dai catch rates accounting for the effects of inter-annual differences in fishing effort (Section 5.4.1), and the spatial (dai row) and intra-annual (month and lunar phase) variation in catch rates associated with fish migrations described in Sections 5.1.1 and 5.3.1.

The flood index, (fi) and the fishing effort, e (number of dais) were therefore treated as covariates in the model and lunar phase, (lp) (1–4); month, (m) (October–March); and dai row, (r) (1–15) were treated as fixed factors:

ln CPUEijkl = α + βfii + βei + lpj + mk + rl + εijkl

i = 1,2,..n; j = 1,2,3,4; k = 10,11,12,1,2,3; l = 1,2,…15.

Table 23 Summary of the indices used to describe the various attributes of the hydrograph hypothesised to affect fish biomass.

Index Description Calculation Hypothesised effect ReferencesFlood Index, FI

Measure of the extent and dura-tion of flooding.

Sum of flooded area for each day of the flood.

Potentially influences growth, survival and recruitment and therefore biomass.

Welcomme (1979; 1985), Welcomme and Halls. (2001); Kapetsky, 1974 and Dudley (1972).Dry Season

Index, DSIMeasure of the severity and duration of the dry season.

Sum of dry area for each day of the dry season.

Potentially influences fish survival and therefore spawning stock biomass and recruitment.

Annual Flood Index

Measure of dry season and flood season habitat availability.

FSI - DSI Combined effect of FSI and previous DSI.

Flood Start, FS

Relative start time of flood.

Number of days since January, 01 to start of flood

Potentially influences survival of drifting larvae and feeding opportunities.

Flood End, FE

Relative end time of flood.

Number of days before or after January, 01 corresponding to end of flood.

Determines flood duration and therefore feeding opportunities.

Flood Dura-tion, D

Duration of the flood.

Number of days between flood start and flood end date.

Potentially influences feed-ing and growth opportunities. Growth effects may also influ-ence size-dependent mortality and subsequent recruitment.

Welcomme and Hagborg (1977); Halls and Welcomme (2004).

Flood Rise Rate, FRR

Rate of flooding. Flood rise / rising water duration

Influences recruitment success and colonisable area of floodplain.

Welcomme and Hagborg (1977); Halls et al. (2001); Halls and Welcomme (2004).

Drawdown Rate, DDR

Rate at which waters recede.

Drawdown / falling water duration

Influences survival rates and therefore subsequent recruitment.

Welcomme and Hagborg (1977); Halls et al. (2001); Halls and Welcomme (2004).

Where εijkl is the residual (unexplained) component and (n) is the number of observations.

Dai row, (r) was treated as a fixed factor in the model rather than a covariate because of the apparent non-linear decline in catch rates with dai row or cumulative effort, (Er) illustrated in Section 5.1.1.

Other hydrological indices were also considered during model fitting including indices to account for variation arising from potential dry season (survival) related, and rate of flood rise and fall related effects (Figure 58 and Table 23). However, the (FI) alone was found to provide the best-fitting and parsimonious model (Table 24). Catch rates were predicted to increase with the (FI) and, as expected, catch rates were predicted to decline with increasing fishing effort. Overall the model explains almost 70% of the variation in the observed catch rates (Figure 59) and the model residuals are reasonably well behaved (Figure 60). However, the model predictions overestimate some of the very low catch rates and underestimate the very high ones (Figure 59).

Table 24 ANOVA table for the GLM model

Source Type III Sum of Squares df Mean Square F Sig.Corrected Model 25427.520(a) 349 72.858 38.867 .000Intercept 901.877 1 901.877 481.121 .000Effort 320.501 1 320.501 170.976 .000FI 740.492 1 740.492 395.028 .000MONTH 3039.725 5 607.945 324.319 .000Lunar Phase 3700.012 3 1233.337 657.945 .000Row 462.154 14 33.011 17.610 .000Month * Lunar phase 1914.482 15 127.632 68.088 .000Month * Row 248.902 70 3.556 1.897 .000Lunar phase * Row 105.415 42 2.510 1.339 .071Month * Lunar phase * Row 592.624 198 2.993 1.597 .000Error 11307.169 6032 1.875 Total 223390.183 6382 Corrected Total 36734.689 6381

Dependent Variable: ln CPUE (kg/dai/day)R Squared = .692 (Adjusted R Squared = .674)



Page 108

Figure 60 Model Residuals.

0

1

2

3

4

5

6

7

8

9

1997

1998

1999

1999

2000

2001

2001

2002

2003

2003

2004

2005

2006

2006

2007

2008

Year

Ln C

PUE

(kg/

dai/d

ay)

Observed

Predicted

Figure 59 The GLM observed and predicted mean catch rates 1997–2009. For the purposes of illustration, mean monthly, instead of daily, catch rates are illustrated.

050

100150200250300350400450

0 200000 400000 600000 800000 1000000 1200000

Flood Index (km2 days)

Mea

n C

PUE

(kg/

dai/d

ay)

Figure 61 The relationship between the mean predicted daily catch rate of a dai unit during the fishing season and the flood index (FI) for the TS-GL System.

This response appears to be mediated through the effects of the (FI) on fish growth. Growth variation, indicated by mean fish weight responds in the same way having a similar exponent value and therefore the same four-fold increase in predicted values from the minimum to the maximum observed flood index (Figure 62). This same type of growth response is also evident for the six most abundant species (Figure 63).

The model predicts that fish biomass, indicated by the mean daily catch rate of a dai unit during the fishing season (October–March), increases exponentially with the (FI) (Figure 61) as follows:

CPUE = 83.88.e1.6063E–06FI


0.000

0.005

0.010

0.015

0.020

0.025

0 200000 400000 600000 800000 1000000 1200000

Wei

ght (

kg)

Flood Index (km2 days)

Figure 62 The relationship between mean sampled fish weight (all species combined) and the flood index with fitted exponential model. Weight = 0.0054e1.231E–06FI . R2 = 0.59.


Page 110

0.000

0.005

0.010

0.015

0.020

0.025

97-98

98-99

99-00

00-01

01-02

02-03

03-04

04-05

05-06

06-07

07-08

08-09

Season

Wei

ght (

kg)

0

200000

400000

600000

800000

1000000

1200000

Floo

d In

dex

(km

2 day

s)

WeightFI

Figure 64 Changes in mean fish weight and the flood index from 1997–98 to 2008–09.

Figure 63 Mean weight estimates for the six most abundant species in the fishery in December each year (corresponding to the end of the flood season) plotted as a function of the flood index.

ANOVA demonstrated that when variation due to the flood index is accounted for, there is no significant change in mean weight through time for any of the six species examined. The relatively low mean weights observed during the last six years are therefore likely to reflect growth responses to below average flood conditions, rather than the effects of increasing rates of exploitation as suggested by some workers (Figure 64).

Figure 65 Fish abundance plotted as a function of the flood index.

No such relationship was evident between fish abundance indicated by mean CPUE for the season, expressed as number of fish caught per dai per day, and the (FI) (Figure 65).

Inter-annual fish biomass variation in the TS-GL System indicated by catch rates sampled from the dai fishery appear to be growth mediated in response to feeding opportunities dictated by flood extent and duration. Mean fish weights for 2004–05 and 2005–06 were consistent with the flood index. Therefore, above average levels of recruitment were probably responsible for the very high catches observed during these two seasons described in Section 5.4.1. Factors responsible for these high levels of recruitment remain uncertain but a campaign by the FiA to confiscate illegal gear, particularly during 2003 and 2004, may have been influential (Hortle et al., 2005). The Mekong River Commission (MRC) report that sediment inputs, which can influence primary production in the TS-GL System, peaked during these two seasons (MRC, pers. comms.). These conditions may have also favoured recruitment.


Page 112

6. Summary and Conclusions

The Tonle Sap-Great Lake (TS-GL) system is an integral part of the history, culture, ecology and economics of the Mekong region. As well as mitigating floods in Cambodia and Viet Nam, it provides a source of raw materials, nutrition, income and livelihoods for upwards of one million people living in and around it. The system is of international ecological and conservation importance supporting a diverse flora and fauna including 149 species of fish, five of which are globally threatened. The annual flood pulse is believed, by some, to be the principle driving force responsible for this productivity and diversity, transporting and recycling biolimiting nurtrients and providing diverse ephemeral critical habitats.

Blackfish and whitefish species are the target of industrial, artisanal and subsistence fisheries operating in the TS-GL System. The large-scale industrial fisheries are managed through a system of fishing lots - demarcated areas of productive fish habitat of up to 500 km2, or positions along migration corridors such as the Tonle Sap and other tributaries forming the system. These limited access fisheries operate only during the open season (October–May) and typically employ large barrier, bagnet or fence type gears designed to divert and/or intercept migrating fish. Artisanal (middle-scale) and subsistence (family) fisheries operate in open access areas with a variety of gear types including gillnets, seines, bamboo fence traps, cast nets and longlines. The former may operate only during the open season, whereas subsistence fisheries operate year-round. Strong competition exists among the fisheries to land in excess of 200,000 tonnes of fish each year equivalent to approximately 10% of the total quantity of fish consumed in the entire LMB each year.

The dai fishery on the Tonle Sap, established almost 140 years ago, is an important component of the industrial fishery landing approximately 14% of the annual catch taken from the TS-GL System. It targets the refuge migrations of a multi-species assemblage of fish as they migrate from the Great Lake to the Mekong main channel via the Tonle Sap with the receding floodwaters each year. Up to 15 rows of stationary trawl nets are set over a 35 km stretch of the Tonle Sap in Kandal and Phnom Penh Municipalities, with 1–7 trawl nets or dai units forming each row. Diesel engines have recently replaced the traditional hand-powered wooden winches for hauling the nets as frequently as 4 times per hour during peak catch rate periods.

In addition to its significant socio-economic value both locally and nationally, the dai fishery provides a valuable source of data and information to monitor trends in migratory fish populations which seasonally utilise the TS-GL System and beyond. Monitoring these trends provides an important means to evaluate the performance of fisheries management measures as well as potential impacts arising from basin development activities.


Page 114

Small cyprinids belonging to the Cirrhinus, Lobocheilos, Labiobarbus and Henicorhynchus genera form the bulk of the catch landed by the dai fishery. Other species making an important contribution to landings are Labeo chrysophekadion, Pangasius pleurotaenia, Puntioplites proctozystron and Thynnichthys thynnoides.

Small species are sold to fish traders either on the floating dai platform or on the nearest riverbank for marketing, processing or export. The remaining more valuable medium-sized and large-sized fish species including Osteochilus melanoplerus, Pangasius larnaudii, Cyclocheilichthys enoplos and Pangasianodon hypophthalmus are often kept alive in bamboo cages suspended below the working platform of the dai and sold during the closed season (March–September) when the fish supply is low and prices are high.

Typical annual operating profit per dai unit is in the region of US$14,000. Annual profit tends to decline from the most upstream dai row 15 to row 2 with declining catch rates. Licence fees (both official and unofficial combined) paid by dai operators appears independent of their reported operating profit. Unexpectedly, the price of trey riel appears positively correlated with supply (landings).

Under Cambodian Law, the fishery is managed using a closed season together with effort and gear size restrictions. Ad hoc surveys of the fishery began in the late 1930s but comprehensive routine monitoring programmes did not begin until 1994–95 undertaken by the FiA/IFReDI with support from the MRC and the FAO. The survey design has evolved over the 15-year monitoring period mainly with changes to survey stratification to reduce sample variance. In addition to changes to the structure and stratification of the sampling regime, there has been considerable inter-seasonal variation in the sampling effort as measured by the total number of hauls sampled by the data collectors over the survey period.

The database used to store and process the data collected from the fishery has also evolved and been supported by different software platforms. This paper has provided a detailed description of the changes that have been made including the full details of the tables providing a useful reference document for both IFReDI and others working with the database. Database queries used to extract data from the tables have been described in a companion working document (Cambodian Dai Fishery Database–Query Reference Manual) held at IFReDI. This Manual contains a range of alternative queries that were developed to analyse the data. The alternative queries described in the manual reflect different assumptions regarding the structure of the data itself and/or the reliability of certain variables such as information on fishing effort. The conclusions drawn from the data generated by the queries may differ slightly depending on these underlying assumptions. This variation effectively represents "model" or "structural" uncertainty–an additional form of uncertainty in catch estimation that exists alongside the statistical uncertainty associated with estimates of parameters using the data. Due to ambiguity surrounding the interpretation of the data no single method has been recommended.

Catch rates sampled from the dai fishery exhibit considerable spatial and temporal variation. Catch rates in Phnom Penh (downstream) are on average less than half the rates estimated for Kandal

indicative of a depletion effect on the migrating fish. A Delury depletion model was used to describe this depletion response. It predicted that each dai unit removes approximately 2.8% of fish migrating through the fishery equivalent to an instantaneous fishing mortality rate, (F) of 1.79 for 64 dai units. In other words approximately 80% of fish arriving at the dai fishery are estimated to be caught.

However, the non-monotonic decline in catch rates with cumulative effort evident in the depletion model suggests that the required assumption of constant catchability may have been violated. This might reflect the use of different net mesh size (see Section 5.1.3) and/or non-random sampling effort through space and time (Section 3.3). The use of larger mesh sizes would be expected to result in lower catch rates and vice-versa.

Furthermore, the corresponding estimates of fishing mortality generated by this analysis will be biased if removals of fish between dai rows by small-scale (artisanal and subsistence) gear were significant. Significant lateral migrations of fish from or to adjacent floodplains between dai rows would also bias the estimates. Current knowledge of these small-scale fisheries and fish migrations routes in the TS-GL System is sparse and poorly documented. This will need to be addressed in order to reliably interpret these results.

If lateral migrations can be ignored, then the estimate of the proportion of the fish removed over the range of the dai fishery (regardless of gear type present) would remain unchanged i.e. approximately 80%. However, the catchability coefficient, (q) for the dai gear and thereby the estimate of the instantaneous fishing mortality rate, (F) for the dai fishery would be upwardly biased if removals by small-scale gear were significant.

Most importantly, the estimate of the proportion of fish removed by the dai fishery, and the corresponding estimate of the average instantaneous fishing mortality rate (F) during the dai fishing season, relate only to the population of fish remaining after exploitation (removals) by other gear operating upstream in the Lake including those used by other lot, artisanal and subsistence fisheries. They are no estimates for the entire population of fish inhabiting the TS-GL System during a period of one year.

Water depth and velocity effects on individual dai catch rates could not be detected. In addition to depletion effects on the population, the observed spatial variability in dai catch rates may also reflect differences in catchability influenced by factors such as bathymetry or other hydrological variables not considered here.

Whilst no strong evidence was found to suggest that the species assemblage landed by the dai fishery varies significantly through the dai rows, some evidence was found to suggest that the mean weight of fish landed increases downstream. This may reflect the use of nets with larger mesh size.

Fish abundance, indicated by the mean daily catch rate of sampled dai units, exhibits significant variation both between and within the six months comprising the fishing season. Catch rates typically

Summary and Conclusions


Page 116

peak in December or January and are lowest in October. Catch rates also peak during the second quarter of the lunar cycle known as the ‘Waxing Gibbous’ phase. These results indicate that fish migrations are strongly influenced by the lunar cycle, and possibly water levels as peak migrations typically occur around January or December coinciding with the end of the flood season as also suggested by Baird et al. (2003). This pulsed, rather than continuous, form of migratory behaviour from the Lake peaking during the second quarter of the lunar cycle was reported as early as the 1950’s (Welcomme, 1985).

However, it remains uncertain why migrations from the Lake should peak during this period particularly when levels of illumination are high potentially increasing the threat of predation by sight-feeding predators. For this reason, migrations of European eels (Anguilla anguilla) and lampreys (Petromyzonidae) tend to be weaker driving full-moon or moonlit conditions (Wootton, 1990). Seaward migrations of coho salmon (Onchorhynchus kisutch) smolts also tend to peak during the (dark) new moon phase. In the Tonle Sap however, illumination may not matter as the water is very turbid, and predators would be swamped (Hortle, pers comms).

Lunar phase-related migrations of amphidromous or anadromous species returning to rivers are generally interpreted in the context of variation in tidal height than as a consequence of variable light intensity (Lucas and Baras, 2001). Historically, tidal influences on the TS-GL System may have been significant (see Section 1.1.1). Baird et al. (2003) therefore proposed that stenohaline species including small cyprininds may have evolved migratory behaviour to avoid encountering saline intrusions in the Mekong mainstream associated with the highest (spring) tides. However, this is hard to reconcile with the fact that spring tides occur every 14 days (twice per month), whilst migrations clearly peak only once per month.

There was also evidence that different species migrate from the Lake at different times with a general trend toward increasing species richness and diversity from the beginning of the season (October), to the middle (November and December), before tailing off toward the end of the season in March. Corresponding changes in mean fish weight were also observed peaking in November or December and followed by a decline. These patterns may be indicative of larger species migrating during the flood recession in November and December compared to other months. There is also some evidence that larger individuals of the same species migrate first during the falling water period (November and December) followed by smaller juvenile individuals.

The tendency for both larger species and larger individuals of the same species to leave the Great Lake earlier than smaller fish was reported by workers during the 1950s as well as in other tropical river systems (see Welcomme 1985 for review). This behaviour is also reported by the dai operators. Given that large fish also often fail to leave deeper floodplain pools, Welcomme (1985) suggests that depth is a major factor affecting the timing of refuge-seeking fish migrations along with dissolved oxygen and temperature.

The fishery experiences significant spatial, intra- and inter-annual variation in catch rates and landings in catch rates and therefore total landings. During the 12 year period for which effort data

are believed to be reliable, total catch (aggregated across species) by season has varied from between approximately 8,000 and 33,000 tonnes, with a mean of approximately 15,000 tonnes but with no obvious trend. Very high catches were observed in 2004–05 and 2005–06. Effort has ranged from 60 to 68 dai units, again with no obvious trend. The mean weight of fish has shown considerable variation during the fifteen year monitoring period, but with no evidence of a continuous monotonic decline. Multivariate analyses provided no compelling evidence to indicate that species composition has changed significantly through time. Rather changes in species richness appear to reflect changes in sampling effects/error.

The extent and duration of flooding in the TS-GL System each year has a significant effect on fish biomass migrating from the Lake indicated by dai catch rates after accounting for variation in fishing effort. Relative fish biomass increases exponentially with the index of flood extent and duration exhibiting a four-fold increase across the observed range. This response appears to be mediated through the effects of the flood index on fish growth, rather than on recruitment.

Observed mean fish weight for 2004–05 and 2005–06 were consistent with those predicted (expected) for the flooding conditions. Therefore, above average levels of recruitment were probably responsible for the very high catches observed during these two seasons.

6.1. Management Implications

This paper represents the first attempt to compile and analyse the available data and information about the Cambodian dai fishery in a single document. It therefore serves as an important reference document for present and future workers involved in the management, monitoring and administration of the fishery. It also contains new insights into the ecology and dynamics of target fish populations important for their management.

The paper has attempted to place the dai fishery in the wider context of the fisheries of the TS-GL System, Cambodia and regionally. This has served to illustrate that whilst the dai fishery is the focus of most fisheries monitoring and evaluation efforts in Cambodia, it is not the only, nor most significant, component of the entire TS-GL fishery. Other components with which it interacts and competes with, particularly the other lot, artisanal and subsistence fisheries, are also significant and therefore should be considered.

Like the lee trap fishery in southern Lao PDR, it is likely that the dai fishery became a focus for intensive monitoring because it targets fish migrations through a ‘bottleneck’ giving rise to high sampling efficiency and the prospect of accurate population estimates owing to the relatively large proportion of the population that can be sampled over short periods. Adding to this, attempts to monitor other lot fisheries have failed in the past because neither lot operators nor officials were prepared to cooperate. Attempts to monitor the other, more dispersed, sectors of the fishery (i.e. middle-scale artisanal and family fisheries) also failed owing to a lack of institutional capacity.



Page 118

In spite of this restricted focus, the monitoring efforts directed at the dai fishery have generated the only continuous long-term data set for an inland fishery in Cambodia–and one of the only two in the Mekong Basin–the other being for the Lee trap fishery.

Analyses of indicators estimated from this data set have been informative for policy and management evaluation, revealing little or no compelling evidence of changes in the abundance, biomass, size or diversity of migratory fish populations that seasonally utilise the TS-GL System and beyond often over distances of more than 600 km (Baird and Flaherty, 2004; Poulsen et al., 2004 and Adamson et al., 2009). Furthermore, the time series of these indicators have also equipped managers with an important baseline against which to monitor any impacts of management and basin development activities.

A key finding of this research is that inter-annual variation in the biomass of the multispecies assemblage targeted by the fishery (and hence landings) can be largely explained by flood duration and extent effects on fish growth. Fish growth, indicated by mean fish weight, increases exponentially with the flood extent and duration described here by the ‘flood index’. Presumably food resources and available feeding time increases with the flood index, or competition for food resources is less intense during larger, longer floods (Halls et al., 2008).

This response has been modelled, allowing predictions to be made of how the relative biomass of the multispecies assemblage targeted by the fishery (and hence catches) are likely to vary under different flooding conditions whether natural or modified as a consequence of climate change and/or water management projects in the Basin. Owing to the highly migratory nature of the target fish species, these predicted hydrological responses may be observed over large distances, affecting fisheries and piscivorous fish populations beyond the immediate vicinity of the system.

The unexplained variation in the model may well reflect variation in recruitment to the system each year in addition to variation in fishing effort (mortality) applied by the other important fisheries within the TS-GL System or over the migratory range of the target species, reinforcing the need for more comprehensive monitoring programmes (see below). These results also urge caution when monitoring mean fish size as a proxy for rates of exploitation in the TS-GL System and other highly fluctuating environments.

By applying depletion model theory, this research has provided the first estimates of the proportion of fish removed over the range of the fishery, dai gear catchability (efficiency), and dai fishing mortality rates subject to a number of assumptions described above. These results are an important first step towards understanding, and thereby controlling if necessary, the relative sources of fishing effort (mortality) over the migratory range of populations of important species of fish. Additional studies and monitoring programmes will be necessary to determine the validity of the assumptions underlying these estimates and to quantify the spatial distribution of the remaining sources of fishing mortality in the TS-GL System and beyond (see below).

6.2. Recommendations

6.2.1. Monitoring fisheries resources in the TS-GL and beyond

The dai fishery monitoring programme has provided a relatively efficient means of obtaining valuable data and information about the ecology, variability and long-term trends in migratory fish populations that seasonally utilise the TS-GL System for fisheries policy and management purposes. However, it is not a panacea for all information needs because of the characteristics of the dai fishery including its interaction with other important fisheries.

Characteristically, the dai fishery targets only migratory whitefish species and therefore does not provide any information about valuable blackfish species that inhabit the TS-GL System year-round. Moreover, whilst located in a migration corridor or bottleneck, the dai fishery effectively operates only over an area spanning approximately 35 km of the Tonle Sap. Without detailed knowledge of the migratory range of the species targeted by the fishery and their population structure it is uncertain over what geographical scale information generated by monitoring the dai fishery applies. Monitoring other fisheries that target these stocks is also necessary to understand, and thereby control, if necessary, sources of fishing effort (mortality) over the migratory range of populations of important species of fish, as well as to provide other important catch-related data.

Its interaction with other fisheries exploiting the TS-GL System also makes it difficult to interpret variability or long-term trends in the dai (or any other) fishery without assuming that effort has remained relatively static over the period of interest. This might not be unreasonable if the large-scale barrier and fence traps common in TS-GL System and in other bottlenecks over the migratory range of stocks exert the greatest overall effort (fishing mortality) since their effort remains largely unchanging. However, effort exerted by the artisanal and subsistence fisheries may be significant and variable in response to fish abundance or may have increased with population growth. At the same time, many fish populations will never enter the TS-GL System and therefore never be vulnerable to capture by the dai fishery or others operating in the system.

These characteristics combine to make it impossible to rely solely on the dai fishery monitoring programme to meet all the information needs of managers and policy makers. Consideration should be therefore given to establishing additional monitoring programmes to supplement the data and information currently generated by the dai fishery monitoring programme.

Detailed recommendations for these additional monitoring programmes including the statistical aspects of their design lie outside the scope of this paper and will depend largely on the information needs of the FiA and other major stakeholders reflecting the country’s and regional fisheries policy and development priorities. A consultative process with these stakeholders will therefore be necessary to develop programmes that meet their data and information needs. Useful guidelines to support this process are described among others by FAO (1999) and Halls et al. (2005).



Page 120

However, on the basis of the findings described in this paper, future programmes should attempt to sample basic statistics (catch by species and effort by gear type) from the major sectors of the fishery (lot fisheries, artisanal and subsistence fisheries) stratified by major habitat type. Estimates of the total number and size of gear within each sampling stratum could be used to raise catch rate samples and effort census to total estimates, and to standardise effort across gears to account for differences in their catchability (Gulland, 1983). Mapping the results would help to identify where management efforts to control fishing mortality might best be targeted. Data generated by this type of survey would also help to improve understanding of the migratory behaviour of fish in the TS-GL System and elsewhere and provide estimates of total catch weight and value by species for economic valuation and for environmental impact assessment purposes.

These efforts might begin in the TS-GL to complement the existing data and models for the dai fishery and to test the validity of the assumptions described above before extending to other locations in the country depending on available resources and capacity.

6.2.2. Recommendations to improve the existing dai fishery monitoring programme

In addition to these supplementary monitoring activities, the following recommendations are made to improve the existing dai fishery monitoring programme.

There is evidence of considerable variation in sampling effort between the dai rows. This is significant from a survey design (and total catch estimation) perspective because of the apparent decline in catch rates from the most upstream row 15 to row 2 which is believed to reflect the depletion of fish as they pass through the fishery. Disproportionate sampling effort directed towards the most upstream dai rows in any given year could therefore generate unrepresentatively high sample catch rates for the entire fishery and corresponding total catch estimates. The current programme should therefore be reviewed by a qualified statistician specialising in survey design to maximise the accuracy and precision of estimates of catch and effort, and to allow for valid inter-annual comparisons of estimates of total catch, effort, and fish biomass indicated by dai catch rates, given the available resources. Reasons for apparent longitudinal differences in species size, diversity and composition between upstream and downstream locations remain uncertain but differences in dai net mesh size and the capacity of survey teams to correctly identify species between upstream and downstream locations may be important. These differences might account for the unexplained variation in the depletion model and observations of species composition through space and time. It is therefore recommended that greater effort be given to accurately recording net mesh size for each sampled haul. The ability of enumerators to correctly identify species should also be checked and training provided where necessary.

Whilst numerous database queries have been written for the purposes of the analyses described in this paper, it is recommended that ‘reports’ be programmed using the existing software platform to automatically generate basic information in the format that is routinely required by the FiA and the MRC for monitoring and evaluation purposes.

6.2.3. Further research

The relative contribution of the upstream versus adjacent floodplain sources of fish to dai catches needs to be investigated to determine the reliability of the estimates of the proportion of migrating fish removed by the dai fishery and the corresponding fishing mortality rate estimates. An expansion of survey activities to gear targeting fish migrations through channels connecting adjacent floodplains to the Tonle Sap (downstream of Kampong Tralach) during the open season would provide useful data for this purpose.

Efforts to establish the migratory range of the target species targeted by dai fishery would determine the geographic extent over which any management measures or modification to hydrological conditions in the TS-GL System are likely to propagate to add to recent work (see Halls et al. in press).

Analyses of the length frequency data sampled from the fishery might also provide further insights into the population dynamics of the species targeted by the fishery including growth, mortality and recruitment.



Page 122

7. References

Adamson, E.A.S.; Hurwood, D.A.; Baker, A.M and Mather, P.B (2009) Population subdivision in the Siamese mud carp Henicorhynchus siamensis in the Mekong River basin: implications for management. Journal of Fish Biology, 75: 1371-1392.

Ahmed, M.; Navy, H.; Vuthy, L. & Tiongco, M (1998) Socio economic assessment of freshwater capture fisheries of Cambodia. Report on a household survey, MRC/DANIDA/DoF. Management of the freshwater capture fisheries of Cambodia, 186pp.

Baird, I.G and Flaherty, M.S (2004) Beyond National Borders: Important Mekong River Medium Sized Migratory Carps (Cyprinidae) and Fisheries in Laos and Cambodia. Asian Science Fisheries, 17: 279-298.

Baird, I.G.; Flaherty, M.S and Phylavanh, B (2003) Rhythms Of The River: Lunar Phases and Migrations of Small Carps (Cyprinidae) In The Mekong River. Nat.Hist.Bul.Siam Soc, 51(1): 5-36.

Baran, E.; Starr, P and Y. Kura (2007) Influence of built structures on the Tonle Sap fisheries. Cambodia National Mekong Committee and Worldfish Center. Phnom Penh, Cambodia, 44pp.

Baran, E.; Van Zalinge, N and Ngor, P. B (2001a) Analysis of the Cambodian Bagnet ("Dai") fishery data. ICLARM, Penang, Malaysia, Mekong River Commission Secretariat and Department of Fisheries, Phnom Penh, Cambodia, Penang, Malaysia, 50pp.

Baran, E.; van Zalinge, N.; Ngor, P. B.; Baird, I. G and D. Coates (2001 b) Fish resource and hydrobiological modelling approaches in the Mekong basin. ICLARM, Penang, Malaysia and the Mekong River Commission Secretariat, Phnom Penh, Cambodia, 60pp.

Brooks, S. E.; Allison, E. H and J.D. Reynolds (2007) Vulnerability of Cambodian water snakes: initial assessment of the impact of hunting at Tonle Sap Lake. Biological Conservation, 139: 401-414.

Campbell, I. C.; Poole, C.; Giesen, W and J. Valbo-Jorgensen (2006) Species diversity and ecology of Tonle Sap Great Lake, Cambodia. Aquatic Science, 68: 355-373.

Cans, G and Ngor, P. B (2006) Cambodian Bagnet ("Dai") fishery database: overview of the conversion from ARTFISH to MS Acces database description. Mekong River Commission, Fisheries Ecology, Valuation and Mitigation Component, Phnom Penh, Cambodia, 26pp.


Page 124

Chevey, P and F. Le Poulain (1940) La pêche dans les eaux douces du Cambodge 5e mémoire. Travaux de L'Institut Océanographique de L'Indochine, 195pp.

Clarke, K.R and R.M. Warwick (2001) Change in marine communities: an approach to statistical analysis and interpretation, 2nd edition. PRIMER-e Ltd. Plymouth.

Clements, T.; O'Kelly, H and Visal, S (2007) Monitoring of large waterbirds at Prek Toal, Tonle Sap Great Lake 2001–2007. UNDP/GEF Tonle Sap Conservation Project, 60pp.

CoM (1989) Sub-decree on renting inland and marine fishery domains for exploitation (in Khmer Version), Council of Ministers, Phnom Penh.

Davidson, P (2006) The biodiversity of the Tonle Sap Biosphere Reserve: 2005 status review. Phnom Penh: UNDP/Wildlife Conservation Society.

Deap, L (1999) The bagnet [dai] fishery in the Tonle Sap River. In Annual Meeting of the Department of Fisheries of the Ministry of Agriculture, Forestry and Fisheries, eds. N. Van Zalinge and T. Nao. Mekong River Commission and Department of Fisheries, Phnom Penh.

Deap, L.; Degen, P and van Zalinge, N (2003) Fishing Gears of the Cambodian Mekong. Phnom Penh, Cambodia Fisheries Technical Paper Series IV, 269pp.

Dubeau, P.; Ouch, P and Sjorslev, J. G (2001) Estimating fish and aquatic animal productivity/yield per area in Kampong Tralach: an integrated approach. Cambodia Fisheries Technical Paper Series, 3: 143-164.

Dudley, R.G (1972). Growth of Tilapia of the Kafue floodplain, Zambia: predicted effects of the kafue Gorge Dam. Ph.D. Dissertation, University of Idaho, Moscow, USA, 50pp.

FAO (1999) Guidelines for the routine collection of capture fishery data. FAO Fish.Tech. Pap. 382, FAO, Rome, 113pp.

Fily, M and F. d'Aubenton (1965) Cambodge - Grand Lac Tonle Sap - Technologie des Pêches 1962–1963. République Française, Minisţère des Affaires Etrangères, service de coopération technique.

Gray, T. N. E.; Collar, N. J.; Davidson, P. J. A.; Dolman, P. M.; Evans, T. D.; Fox, H. N.; Chamnan, H.; Borey, R.; Kim Hout, S. and R.N. van Zalinge. (2009) Distribution, status and conservation of the Bengal Florican Houbaropsis bengalensis in Cambodia. Bird Conservation International, 19: 1-14.

Gulland, J.A (1983) Fish Stock Assessment: A Manual of Basic Methods, Chichester: John Wiley and Sons, 223pp.

Gupta, A and S.C. Liew (2007) The Mekong from satellite imagery: a quick look at a large river. Geomorphology, 85: 259-274.

Hai, N.T., Sunada, Y.K., Oishi, S., and Ikejima, K. Tonle Sap Ecosystem Fish Species Biological Groups and Hydroecological Index. International Journal of Ecological Economics and Statistics, 8: 6-26.

Halls, A.S.; Paxton, B.R.; Hall, N.; Hortle, K.G.; So, N.; Chea, T.; Chheng, P.; Putrea, S.; Lieng, S.; Peng Bun, N.; Pengby, N.; Chan, S.; Vu, V.A.; Nguyen Nguyen, D.; Doan, V.T., Sinthavong, V.; Douangkham, S.; Vannaxay, S.; Renu, S.; Suntornratana, U.; Tiwarat, T. and Boonsong, S. (2013). Integrated Analysis of Data from MRC Fisheries Monitoring Programmes in the Lower Mekong Basin. MRC Technical Paper No. 33, Mekong River Commission, Phnom Penh, Cambodia, 130pp.

Halls, A.; Lieng, S.; Ngor, P. B and Tun, P (2008) New research reveals ecological insights into Dai fishery. Catch and Culture, 14(1): 8-12.

Halls, A. S., Arthur, R., Bartley, D., Felsing, M., Grainger, R., Hartmann, W., Lamberts, D., Purvis, J; Sultana, P., Thompson, P and Walmsley, S (2005) Guidelines for Designing Data Collection and Sharing Systems for Co-Managed Fisheries. Part 1 & 2: FAO Fisheries Technical Paper. No. 494/1&2. Rome, FAO. 2005, Part 1, 42pp and Part 2, 108pp.

Halls, A.S and Welcomme, R.L (2004) Dynamics of river fish populations in response to hydrological conditions: A simulation study. River Research and Applications, 20: 985-1000.

Halls, A.S.; Kirkwood, G.P and Payne, A.I (2001) A dynamic pool model for floodplain-river fisheries. Ecohydrology and Hydrobiology, 1(3): 323-339.

Halls, A.S., Sopha, L and Ngor, P (2007) Landings from Tonle Sap dai fishery in 2006–07 above the 12-year average. Catch & Culture, 13(1): 7-11.

Hap, N and Ngor, P (2001) ‘An economic analysis of fish production in the Dai Fisheries in Phnom Penh/Kandal Province, Cambodia’, in N. van Zalinge, S. Nuov, R. Ounsted and L. Sopha (eds.), Cambodia fisheries technical paper series, Volume III, The Management of the Freshwater Capture Fisheries of Cambodia Component of the Mekong River Commission’s Program for Fisheries Management and Development and the Department of Fisheries of Cambodia, Phnom Penh.

Hap, N., Chuenpagdee, R and J. Kurien (2006) Livelihood importance and values of the Tonle Sap Lake fisheries. Inland Fisheries Research and Development Institute, Phnom Penh, Cambodia, 8pp.

References


Page 126

Hav, V and Ngor, P (2005) An overview of aquaculture in Cambodia, Working Paper 2, An output from the DFID-funded Post-Harvest Fisheries Livelihood Project, Department of Fisheries, Phnom Penh.

Hilborn, R and Walters, C.J (1992) Quantitative Fisheries Stock Assessment. Choice, Dynamics and Uncertainty. London, Chapman Hall, 570pp.

Hortle, K.G (2007) Consumption and the yield of fish and other aquatic animals from the Lower Mekong Basin. Mekong River Commission, Vientiane. MRC Technical Paper No. 16, 87pp.

Hortle, K.G.; Lieng, S and Valbo-Jorgensen, J (2004) An introduction to Cambodia's inland fisheries. Mekong Development Series No. 4. Mekong River Commission, Phnom Penh, Cambodia, 41pp.

Hortle, K.G.; Ngor, P., Hem, R and Lieng, S (2004) Trends in the Cambodian dai fishery: floods and fishing pressure. Catch & Culture, 10(1): 7-9.

Hortle, K.G.; Ngor, P.; Hem, R and Lieng, S (2005) Tonle Sap yields record haul. Catch & Culture, 11(1): 3-7.

Junk, W J; Bayley, P B and Sparks, R E (1989) The flood pulse concept in river-floodplain systems. In. Dodge, D P (ed). Proceedings of the International Large River Symposium. Canadian Special Publication of Fisheries and Aquatic Science, 106: 110-127.

Junk, W. J.; Brown, M.; Cambell, I. C.; Finlayson, M.; Gopal, B.; Ramberg, L. and B.G, Warner (2006) The comparative biodiversity of seven globally important wetlands: a synthesis. Aquatic Science, 68: 400-414.

Kapetsky, J.M (1974) Growth Mortality and Production of Five Fish Species of the Kafue River Floodplain, Zambia. PhD Thesis. University of Michigan, 194pp

Keskinen, M (2003) Socio-economic survey of the Tonle Sap, Cambodia. A Master of science thesis submitted for inspection in Espoo, 2003.

Kummu, M and J. Sarkkula (2008) Impact of the Mekong River flow alteration on the Tonle Sap flood pulse. Ambio, 37: 185-192.

Kummu, M.; Koponen, J and J. Sarkkulan (2005) Modelling sediment transportation in Tonle Sap Lake for impact assessment. In: SIMMOD '05 International Conference on Simulation and Modelling 2005. Simulation and Modelling: Integrating Sciences and Technology for Effective Resource Management. Asian Institute of Technology, Bangkok, Thailand. (http://www.mssanz.org.au/simmod05/papers/C3-02.pdf).

Kummu, M.; Penny, D.; Sarkkula, J and J. Koponen (2008) Sediment: curse or blessing for Tonle Sap Lake? Ambio, 37(7): 158-163.

Lamberts, D (2008) Little impact, much damage: the consequences of Mekong River flow alterations for the Tonle Sap ecosystem. In: Modern Myths of the Mekong, eds. M. Kummu, M. Keskinen & O. Varis: 3-18.

Lamberts, D (2001) Tonle Sap fisheries: a case study on floodplain gillnet fisheries in Siem Reap, Cambodia. RAP Publication. FAO Regional Office for Asia and the Pacific, Bangkok, Thailand, 133pp.

Lamberts, D (2006) The Tonle Sap Lake as a productive ecosystem. Water Resources Development, 22: 481-495.

Lieng, S.; Yim, C and van Zalinge, N. P (1995) Freshwater fisheries of Cambodia, I: the bagnet (Dai) fishery in the Tonle Sap River. Asian Fisheries Science, 8: 255-262.

Lim Song, S.; Lieng, S.; Ing, T and S Heng (2004) The unsustainable exploitation of inland fisheries resources in Cambodia. Discussion Paper 14 In: Overcoming factors of unsustainablility and overexploitation in fisheries: selected papers on issues and approaches. International workshop on the Implementation of International Fisheries Instruments and Factors of Unsustainability and Overexploitation in Fisheries , Siem Reap, Cambodia, 19-16 September 2004. FAO/Japan Government Cooperative Programme. (available google book).

Lim, P.; Lek, S.; Touch, S.T.; Mao, S.-O and Chhouk, B (1999) Diversity and spatial distribution of freshwater fish in Great Lake and Tonle Sap river (Cambodia, Southeast Asia). Aquatic Living Resources, 12: 379-386.

Lucas, M.C and Baras, E (2001) Migration of Freshwater Fishes. Oxford, Blackwell Science Ltd, 420pp.

MRC (2005) Overview of the hydrology of the Mekong Basin. Mekong River Commission, Vientiane, 73pp.

MFD (2003) Mekong Fish Database. Mekong River Commission 2003.

Nao, T and N, van Zalinge (2001) Challenges in managing Cambodia fisheries. How we can meet them. Mekong River Commission and Department of Fisheries, Phnom Penh.

Ngor, P and N, van Zalinge (2001). Dai (Bagnet) fishery: 1994/95-2000/01: catch assessment methodology and results. Project for Management of the Freshwater Capture Fisheries of Cambodia. Mekong River Commission/DoF/DANIDA, 47pp.

References


Page 128

Ngor, P. B (2000) Dai fisheries in the Tonle Sap River of Phnom Penh and Kandal Province (including a review of the census data of 1996-97). In: van Zalinge, N.P.; Nao, T and S. Lieng (eds.) Management aspects of Cambodia’s Freshwater Capture Fisheries. Eleven presentations given at the annual meeting of the Department of Fisheries of the Ministry of Agriculture, Forestry and Fisheries. Pp 30-47. Mekong River Commission and Department of Fisheries Phnom Penh, Cambodia, 27-28 January 2000, 169pp.

Nguyen, X.T and V.H, Nguyen (1991) Investigation of the freshwater fish productivity of Cambodia (1986-88) Fishery Service of Viet Nam, Hanoi, 126pp.

Penny, D (2006) The holocene history and development of the Tonle Sap, Cambodia. Quarternary Science Reviews, 25: 310-322.

Penny, D (2008) The Mekong at climatic crossroads: lessons from the geological past. Ambio, 37(3): 164-169.

Penny, D.; Cook, G and S.S, Im (2005) Long-term rates of sediment accumulation in the Tonle Sap, Cambodia: a threat to ecosystem health? Journal of Paleolimnology, 33: 95-103.

Poulsen, A.F.; Hortle, K.G.; Valbo-Jorgensen, J.; Chan, S.; Chhuoun, C.K.; Viravong, S.; Bouakhamvongsa, K.; Suntornratana, U.; Yoorang, N.; Nguyen, T.T and Tran, B.Q (2004) Distribution and Ecology of Some Important Riverine Fish Species of the Mekong River Basin. MRC Technical Paper, 10: 116pp.

Rainboth, W.J (1996) Fishes of the Cambodian Mekong. FAO Species Identification Field Guide for Fishery Purposes. FAO, Rome, 265pp.

Sam, N.; Lieng, S and Thor, S (2002) Improving Inland Capture Fishery Statistics in Cambodia. In New Approaches for the Improvement of Inland Capture Fishery Statistics in the Mekong Basin. RAP Publication 2003/01, FAO, Rome, 145pp.

Sarkkula, J and J, Koponen (2003) Modelling Tonle Sap for environmental impact assessment and management support. Water Utilization Program - modelling of the flow regime and water quality of the Tonle Sap MRCS/WUP-FIN Project, 110pp.

Sarkkula, J.; Baran, E.; Chheng, P.; Keskinen, M.; Koponen, J and Kummu, M (2004) Tonle Sap pulsing system and fisheries productivity. In XXIX International Congress of Limnology

(SIL 2004). Lahti, Finland.

Sarkkula, J.; Kiirikki, M.; Koponen, J. and M. Kummu (2003) Ecosystem processes of the Tonle Sap Lake. In 1st Workshop of Ecotone Phase II. Phnom Penh, Cambodia.

So, N.; Sy Vann, L and Yumiko, Y (2007) Study of the catch and market chain of low value fish along Tonle Sap River, Cambodia and implications for management of their fisheries - A preliminary study. Consultancy report for the WorldFish Center. Inland Fisheries Research and Development Institute and WorldFish Center’s Greater Mekong Region, Phnom Penh, Cambodia, 56pp.

So, N.; Eng, T.; Souen, N and Hortle, K.G (2005). Use of freshwater low value fish for aquaculture development in the Cambodia's Mekong basin. Inland Fisheries Research and Development Institute, Phnom Penh and Mekong River Commission Fisheries Program, Vientiane, Lao PDR.

So, N and Nao T (1999) Aquaculture Sector Review, Cambodia. Constancy report for APIP/World Bank- Fisheries Component, Department of Fisheries, Phnom Penh, 50pp.

Sparre, P and Venema, S.C (1992) Introduction to tropical stock assessment. FAO Fisheries Technical Paper 306/1. Rome, FAO, 376pp.

Stamatopoulos, C (1994) ARTFISH: a microcomputer system for the statistical monitoring of artisanal fisheries. Version 1. FAO Rome.

Stamatopoulos, C (1995) Statistical monitoring of the freshwater capture fisheries in Cambodia. Mission report, field document No. 3. DoF/MRC/DANIDA. Management of the freshwater capture fisheries of Cambodia project, Phnom Penh, 6pp.

Stamatopoulos, C (2002) Sample-based fishery surveys. FAO Fisheries Technical Paper 425. FAO Rome, 132pp.

Stuart, B.L.; Smith, J.; Davey, K.; Din, P and S.G. Platt (2000) Homalospine watersnakes: the harvest and trade from the Tonle Sap, Cambodia. TRAFFIC Bulletin, 18: 115-124.

Sverdrup-Jensen, S (2002) Fisheries in the Lower Mekong Basin: status and perspectives. Mekong River Commission, Phnom Penh. MRC Technical Paper.

Touch, S.T (1998) Inland fisheries in historic prospective, an afterthought of commercialization. MRC Symposium, 7-8 December 1998, 17pp.

Touch, S.T (1993) Fish supply and demand in rural Svay Rieng Province, Cambodia, A thesis submitted as a partial fulfillment of the degree of master, Asian Institute of Technology (AIT), Bangkok.

Valbo-Jorgensen, J.; Chan, S and Chhuon, K. C (2001) Lateral fish migrations between the Tonle Sap River and its floodplain. Cambodia Fisheries Technical Paper Series, 3: 120-128.

References


Page 130

van Zalinge, N (2002) Update on the status of the Cambodian inland capture fisheries sector with special reference to the Tonle Sap Great Lake. Catch and Culture, 8(2): 1-5.

van Zalinge, N.; Sarkkula, J.; Koponen, J.; Loeung, D and Ngor, P.B (2004) Mekong flood levels and Tonle Sap fish catches. In: Welcomme, R.L., Petr, T. (Eds.), Second International Symposium on the Management of Large Rivers for Fisheries. FAO, Bangkok, Phnom Penh, Cambodia.

van Zalinge, N.; Degen, P.; Pongsri, C.; Nuov, S.; Jensen, J. G.; van Hao, N. and X, Choulamany (2003) The Mekong River system. In Second International Symposium on the Management of Large Rivers for Fisheries. Phnom Penh.

van Zalinge, N.; Thouk, N and Touch, S.T (1998) Where there is water, there is fish? Fisheries issues in the Lower Mekong Basin from a Cambodian perspective. Project for Management of the Freshwater Capture Fisheries of Cambobia. Mekong River commission. Seventh Common Propoerrty Conference of the International Association for the Study of Common Property, Vancouver, Canada, 10-14 June 1998, 14pp.

Welcomme, R.L and Hagborg, D (1977) Towards a model of a floodplain fish population and its fishery. Environmental Biology of Fishes, 2: 7-24.

Welcomme, R.L (1979) Fisheries Ecology of Floodplain Rivers, London: Longman, 317pp.

Welcomme, R.L and Halls, A.S. (2001) Some considerations of the effects of differences in flood patterns on fish populations. Ecohydrology and Hydrobiology, 1(3): 313-321.

Welcomme, RL (1985) River Fisheries. FAO Fisheries Technical Paper 262: 330pp.

Wootton, R.J (1990) Ecology of Teleost Fishes. Chapman & Hall, London, 404pp.

8.

Ann

ex

Tabl

e 25

Sum

mar

y fr

om th

e ex

istin

g da

i dat

abas

e of

the

num

ber o

f dis

tinct

dai

s sam

pled

in e

ach

of th

e di

ffere

nt st

rata

bet

wee

n 19

94 a

nd 2

008

(cou

nt o

f un

ique

Dai

ID in

the

Dai

Fis

hery

dat

abas

e). D

ai Y

ield

and

Lun

ar P

erio

d st

rata

: B00

01 H

igh

Yie

ld- P

eak

Perio

d; B

0002

Hig

h Y

ield

- Low

Per

iod;

B

0003

Low

Yie

ld- P

eak

Perio

d; B

0004

Low

Yie

ld-L

ow P

erio

d; B

0005

All

DAI

- Low

Per

iod;

B00

06 A

ll D

AI- P

eak

Perio

d; B

0007

Row

2- L

ow

Perio

d; B

0008

Row

2- P

eak

Perio

d; B

0009

Exc

l2- L

ow P

erio

d; B

0001

0 Ex

cl2-

Pea

k Pe

riod.

No

effo

rt in

form

atio

n in

term

s of t

he n

umbe

r of

haul

s or d

ai ID

was

reco

rded

for t

he 1

996–

96 se

ason

.

1994

–200

0Ph

nom

Pen

h M

unic

ipal

ityK

anda

l Pro

vinc

eB

oth

Tota

lR

ow L

abel

sM

onB

0001

B00

02B

0003

B00

04B

0007

B00

08B

0009

B00

010

B00

01B

0002

B00

03B

0004

B00

07B

0008

B00

09B

0001

0B

0001

B00

02B

0003

B00

04B

0005

B00

0694

–95

Dec

27

9Ja

n3

25

Feb

44

94–9

5 To

tal

513

1895

–96

Nov

No

data

on

effo

rt or

dai

iden

tity

Dec

Jan

Feb

Mar

95–9

6 To

tal

96–9

7N

ov6

39

Dec

1424

38Ja

n16

1430

Feb

118

19M

ar13

1396

–97

Tota

l47

6210

997

–98

Nov

816

24D

ec11

3647

Jan

1130

41Fe

b6

3137

Mar

421

2597

–98

Tota

l40

134

174

98–9

9O

ct4

1310

210

39N

ov5

710

147

612

768

Dec

87

1014

417

711

78Ja

n6

73

137

711

1771

Feb

76

111

79

611

5898

–99

Tota

l30

2737

5235

4146

4631

499

–00

Oct

46

37

47

76

44N

ov5

612

135

89

1775

Dec

57

1119

79

1627

101

Jan

55

1216

913

619

85Fe

b7

711

135

108

2283

Mar

37

810

610

713

6499

–00

Tota

l29

3857

7836

5753

104

452


Page 132

2000–2008Phnom

Penh Municipality

Kandal Province

Both

TotalR

ow L

abelsM

onB

0001B

0002B

0003B

0004B

0007B

0008B

0009B

00010B

0001B

0002B

0003B

0004B

0007B

0008B

0009B

00010B

0001B

0002B

0003B

0004B

0005B

000600–01

Oct

417

717

146

Nov

42

1415

68

718

175

Dec

54

1813

310

822

21

187

Jan4

419

179

67

161

184

Feb4

420

198

86

151

85M

ar4

47

137

47

955

00–01 Total25

1895

7733

4335

973

13

2432

01–02O

ct4

413

139

74

963

Nov

44

1418

921

314

87D

ec4

418

164

176

978

Jan4

417

1610

1311

1590

Feb4

420

118

145

773

Mar

26

127

2701–02 Total

2022

8280

4084

2961

41802–03

Oct

63

137

64

1610

65N

ov6

610

158

1015

1585

Dec

66

915

87

1220

83Jan

65

1110

104

1314

73Feb

65

137

96

149

69M

ar3

34

102

57

842

02–03 Total33

2860

6443

3677

76417

03–04O

ct6

128

1945

Nov

65

1212

87

1413

77D

ec6

512

1311

818

1083

Jan5

411

108

614

1270

Feb4

17

67

211

1048

Mar

55

03–-04 Total21

2142

5334

3162

64328

04–05O

ct6

212

86

720

566

Nov

65

1313

109

1814

88D

ec6

512

108

922

1688

Jan6

412

1210

719

1383

Feb6

116

99

616

1376

04–05 Total30

1765

5243

3895

61401

05–06O

ct6

310

127

421

1174

Nov

22

1212

78

2015

78D

ec2

213

139

1021

1686

Jan6

515

1212

418

1183

Feb6

213

510

417

764

Mar

55

1005–06 Total

2214

6354

4530

10265

39506–07

Oct

35

86

53

1113

54N

ov5

79

129

813

1578

Dec

46

1113

88

1223

85Jan

66

149

105

2213

85Feb

66

126

46

1113

6406–07 Total

2430

5446

3630

6977

36607–08

Oct

15

6N

ov5

79

1212

1122

22100

Dec

65

1413

149

2216

99Jan

66

1612

1814

3016

118Feb

74

118

75

1616

7407–08 Total

2422

5045

5140

9075

397G

rand Total233

219510

52425

1895

77396

430658

7263

13

292

2094,221

8.1. Dai Fishery principal data tables (Cans and Ngor 2006)

Table 26 tbl_MainWi&Dos

Name Type DescriptionDOC Text Document number of the original paper record , number of the data page (per month)DAY Number Number of the day of the monthMONTH Number Number of the monthYEAR Number YearDate Date/Time Concatenation of Day, Month and YearGearCode Text Gear Code, linked with the Gear list tableNumber of haul Number Number of haul per daySampled weight Number Total weight of the sampleTotalWeight Number Total weight of the haulRecorder Text Recorder nameStratumNum Number Identification of the strata: 1 Phnom Penh, 2 Kandal, and 3 both of them.SKIPPER Text Name of the skipper or No of the Dai unit/ no of row ( No of row (=dai)/Letter of

unit in a row, ex: 12E = dai number 12, 5th bagnet.SizeDai Number Size of the dai in metersMeshed Text Meshed useNOSP Number Number of people working at the dai stationTOTV Number Total value of the sample ( SampledWeight*TOTP) ( * 1,000 riels )TOTP Number Weighted average of the sample price ( * 1,000 riels)TOTN Number Total number of individuals in the sample

Table 27 tbl_SpeciesWin&Dos

Name Type DescriptionDOC Text Document number , number of the data page ( per month) foreign key from Main tableDate Date/Time Date foreign key from main tableGearCode Text Gear code key from main tableStratumNum Number Stratum number, foreign key from main tableKhmer name Text Khmer species nameCatch Number Weight of this species (kg)Value Number value of the catch = catch* pricePrice Number value ( *1,000 riels) per kgFishNumber Number Number of individuals of that species

Annex


Page 134

Table 28 tbl_Effort

Name Type DescriptionStratumNum Number Stratum numberGearCode Text Gear codeRecorder Text Recorder nameDate Date/Time Datemonth Number Monthyear Number YearActive Gear Number Number of active gearActive Day Number Number of active days

Table 29 tlkp_SeasonYear

Name Type DescriptionMONTH Number MonthYEAR Number YearSeason Text Fishing season

Table 30 tlkp_GearCode

Name Type DescriptionGear Code Text Code of the gearGear name Text Name of the gear

Table 31 tlkp_Species

Name Type DescriptionCode in form Number Code of the species in the field data entry formMerge for form Text Concatenation of the Khmer name and the codeMFD CODE Number MFD, 2000 corresponding code

Table 32 lkp_SpeciesStandard

Name Type DescriptionSpeciesNo Number LEK number 1–193 was missing 65,176,192 (190 species), plus new nos. added for new

AMCF databasesSpeciesCode Number from Dai data entry formlngSpeciesCode Number MFD species code. 2–1323, total 924, numbers > r 1323 added arbritarilytxtSpeciesName Text Species Name in MFDTAXON Text Taxa used for analysis, can combine species by user depending on the dataFamily Text FamilyFish or OAA Text Fish or other aquatic animaltxt Habitat Texttxt Indigenous Text Indigenous, ExoticBlackWhite Text Black, White, Estuarine, MarineFeeding Simple Text Simplified feeding categoryCommEnglish Text English NameKhmerName Text Khmer NameVietnamese Text Vietnamese NameLao Text Lao NameLocalName Thai Text Thai NameboolInclude Yes/No Included in LEK surveyGenusMnS Text Migration survey genus nameSpeciesMnS Text Migration survey species nameMFDGenus Text MFD genus nameMFDSpecies Text MFD species nameLogbook Yes/No In drop-down list for data entryFeeding Text Feeding categoryComment Text Comment on the names usedmemFeeding Memo Notes on feeding from MFDtxtSpeciesStatus Text Indicate the status of the species (Questionable, Exclude, Expected or Confirmed)Cambodia Text Indicates if the species occurs in Cambodia, Source (Rainboth, 1996)LaoPDR Text Indicates if the species occurs in Lao PDR, Source (Kottelat)Thailand Text Indicates if the species occurs in Thailand, Source (Chavalit)VietNam Text Indicates if the species occurs in Viet Nam, Source (Kottelat and Yen)Yunnan Text Indicates if the species occurs in Yunnan (China) (Dr Chen)Rainboth Text Indicates if this species included in Rainboths field-guideOldGenusRainboth Text Generic name of the species as used in Rainboth field-guideOldSpeciesRainboth Text Specific name of the species as used in the Raimboth field-guidememGlobalOcc Memo Remarks on the global distribution of this speciesmemMekongOcc Memo Remarks on the occurrence within the Mekong basinmemHabitat Memo Information on habitattxtDryHabitat Text Typical habitat where the species can be found during the dry seasontxtFloodHabitat Text Typical habitat where the species can be found during the flood seasonMaxLenFemales Text Maximum length for females; note this was always > for malestxtMaxTypeFem Text Type of length measurementReservoirs Text Recorded in reservoirs, man-madeSwamps Text Recorded in swampsLakes Text Found in lakesFloodplain Text Recorded on floodplainsBrackish Text Recorded in brackish waterEstuaries Text Recorded in estuariesMarine Text Recorded in the seaAnadromous Text Known anadromous - ascends rivers from the sea to breedCatadromous Text Known catadromous - descends rivers to estuaries or the sea to breedPotamodromous Text Known potamodromous - migrates within rivers to breedLimnodromous Text Known limnodromous - migrates within lakes to breedOceanodromous Text Known oceanodromous - migrates in the sea to breedAmphidromous Text Known amphidromous - moderates up or downstream to breed

Annex


Page 136

Table 33 tbl _Phnom Penh Port Water level

Name Type DescriptionDate Date/Time DateLevel (m) Number Water level at Phnom Penh port

Dai Fishery additional Tables (Cans and Ngor, 2006)

Table 34 tbl _MoonFace

Name Type DescriptionDate Date/Time DateMoon Illumination Percent Number Percentage of moon face illumination

Table 35 Alphanumeric gear codes in the Access database incorporating to lunar phase and dai yield (B0001–B004), as well as the alternative stratifications applied in the earlier sampling years, i.e. no stratification (All DAI) applied prior to the 1997-98 season and the scheme that included or excluded Row 2 (Row2/Excl2) applied between the 1998–99 and 1999–00 season.e

Gear code Gear nameB0001 High Yield- Peak PeriodB0002 High Yield- Low PeriodB0003 Low Yield- Peak PeriodB0004 Low Yield- Low PeriodB0005 All DAI- Low PeriodB0006 All DAI- Peak PeriodB0007 Row2- Low PeriodB0008 Row2- Peak PeriodB0009 Excl2- Low PeriodB00010 Excl2- Peak Period

8.2. 2009 Dai Fishery database principal data tables

Table 36 tbl_MainWi&Dos

Name Type DescriptionDOC Text Document number of the original paper record , number of the data page (per month)DaiID Text Name of the skipper or number of the Dai unit/row ( number of row(=dai)/ Letter of

unit in a row, ex: 12E = dai number 12, 5th bagnet.DAY Number Number of the day of the monthMONTH Number Number of the monthYEAR Number YearDate Date/Time Concatenation of Day ,Month and YearGearCode Text Gear Code, linked with the Gear list tableNumber of haul Number Number of hauls per dayS_Sampled weight Number Total weight of the sample taken of the small speciesS_TotalWeight Number Total weight of all the small species in the haul (TotalWeight– B_TotalWeight)B_TotalWeight Number Total weight of big species in the haul (not sub-sampled)TotalWeight Number Total weight of haul the haul (small and big species combined)Recorder Text Recorder nameStratumNum Number Identification of the administrative zone: 1 Phnom Penh, 2 Kandal and 3 both zonesFisher Text Name of the Dai ownerSizeDai_Depth Number Depth of the Dai (m)SizeDai_Width Number Width of the Dai (m)Size Dai_Length Number Length of the Dai (m)Meshed_Max Number Maximum mesh size of the bagnet (mm)Meshed_Min Number Minimum mesh size of the bagnet (mm)Mesh Size_ART-FISH ONLY

Text Mesh size from ARTFISH database NB Not classified as Max or Min (Max and Min data combined)

NOSP Number Number of people working at the dai stationTOTV Number Total value of the sample ( SampledWeight*TOTP) ( * 1,000 riels )TOTP Number Weighted average of the sample price ( * 1,000 riels)TOTN Number Total number of individuals in the sample

Annex


Page 138

Table 37 tbl_SpeciesWin&Dos

Name Type DescriptionDOC Text Document number , number of the data page ( per month) foreign key from Main tableDate Date/Time Date foreign key from main tableGearCode Text Gear code key from main tableStratumNum Number Stratum number, foreign key from main tableKhmer new name Text Khmer species nameS_Catch Number Weight of the small species (kg)S_Value Number Value of the small catch=catch* priceS_Price Number Value of the small catch (/1,000 riels) per kg - ie all prices have been divided by 1,000S_FishNumber Number Number of individuals of small speciesB_Catch Number Weight of the big species (kg)B_Value Number value of the big catch= catch* priceB_Price Number value of the big catch ( /1,000 riels) per kg - ie all prices have been divided by 1,000B_FishNumber Number Number of individuals of big species

Table 38 tbl_Effort

Name Type DescriptionStratumNum Number Stratum number (Administrative Zone)GearCode Text Gear codeRecorder Text Recorder nameDate Date/Time Datemonth Number Monthyear Number YearActive Gear Number Number of active gearsActive Day Number Number of active days

Table 39 tlkp_GearCode

Name Type DescriptionGear Code Text Code of the gearGear name Text Name of the gearPeriod_numeric Number Numeric code for the period (Peak = 1; Low = 2)Yield_numeric Number Numeric code for the Dai Yield classification

(Both = 2; High = 1; Low = 0)

2009 Dai Fishery database lookup tables.

Table 40 tlkp_Species

Name Type DescriptionNewCode Number Code of the species in the field data entry formKhmerName Text Khmer species nameKhmerNameCode Text Concatenation of the khmer name and the codeMFD CODE Number Mekong Fisheries Database 2000 corresponding code

Table 41 tlkp_SpeciesStandard

Name Type DescriptionSpeciesNo Number LEK number 1–193 was missing 65,176,192 (190 species), plus new nos. added for

new AMCF databasesSpeciesCode Number From Dai data entry formlngSpeciesCode Number MFD species code. 2–1323, total 924, numbers > r 1323 added arbritarilytxtSpeciesName Text Species Name in MFDTAXON Text Taxa used for analysis, can combine species by user depending on the dataFamily Text FamilyFish or OAA Text Fish or other aquatic animaltxtHabitat Text HabitattxtIndigenous Text Indigenous, ExoticBlackWhite Text Black, White, Estuarine, MarineFeeding Simple Text Simplified feeding categoryCommEnglish Text English NameKhmerName Text Khmer NameVietnamese Text Vietnamese NameLao Text Lao NameLocalName Thai Text Thai NameboolInclude Yes/No Included in LEK surveyGenusMnS Text Migration survey genus nameSpeciesMnS Text Migration survey species nameMFDGenus Text MFD genus nameMFDSpecies Text MFD species nameLogbook Yes/No In drop-down list for data entryFeeding Text Feeding categoryComment Text Comment on the names usedmemFeeding Memo Notes on feeding from MFDtxtSpeciesStatus Text Indicate the status of the species (Questionable, Exclude, Expected or Confirmed)Cambodia Text Indicates if the species occurs in Cambodia, Source (Rainboth, 1996)LaoPDR Text Indicates if the species occurs in Lao PDR, Source (Kottelat)Thailand Text Indicates if the species occurs in Thailand, Source (Chavalit)VietNam Text Indicates if the species occurs in Viet Nam, Source (Kottelat and Yen)Yunnan Text Indicates if the species occurs in Yunnan (China) (Dr. Chen)

Annex


Page 140

Rainboth Text Indicates if this species included in Rainboths field-guideOldGenusRainboth Text Generic name of the species as used in Rainboth field-guideOldSpeciesRainboth Text Specific name of the species as used in the Raimboth field-guidememGlobalOcc Memo Remarks on the global distribution of this speciesmemMekongOcc Memo Remarks on the occurrence within the Mekong basinmemHabitat Memo Information on habitattxtDryHabitat Text Typical habitat where the species can be found during the dry seasontxtFloodHabitat Text Typical habitat where the species can be found during the flood seasonMaxLenFemales Text Maximum length for females; note this was always > for malestxtMaxTypeFem Text Type of length measurementReservoirs Text Recorded in reservoirs, man-madeSwamps Text Recorded in swampsLakes Text Found in lakesFloodplain Text Recorded on floodplainsBrackish Text Recorded in brackish waterEstuaries Text Recorded in estuariesMarine Text Recorded in the seaAnadromous Text Known anadromous - ascends rivers from the sea to breedCatadromous Text Known catadromous - descends rivers to estuaries or the sea to breedPotamodromous Text Known potamodromous - migrates within rivers to breedLimnodromous Text Known limnodromous - migrates within lakes to breedOceanodromous Text Known oceanodromous - migrates in the sea to breedAmphidromous Text Known amphidromous - moderates up or downstream to breed

Table 42 Leng_tbllengthFreq (Asterisks indicate name changes or new fields)

Name Type DescriptionDOC Text Document number and number of the data page ( per month)

foreign key from Main tableDaiID Text Dai identification numberDate Date/Time Date foreign key from main tableSpeciesName Text Species NameLength Number Total Length (cm)Num_Fish Number Number of fish of length (Length Field)

2009 Dai Fishery database Length Frequency Tables (new)

Table 43 LengtblSpecies (Asterisks indicate name changes or new fields)

Name Type DescriptionDOC Text Document number , number of the data page ( per month) foreign key from Main tableDaiID Text Dai identification numberDate Date/Time Date foreign key from main tableSpeciesName Text Gear code key from main tableTotal Number Total weight of the haul from which the sample was takenTotal_Sample Number Total weight of the sampleSubtotal Number Weight of the sample by species used to obtain the length information

Table 44 Leng_tblLocation (Asterisks indicate name changes or new fields)

Name Type DescriptionDOC Number Document number , number of the data page ( per month) foreign key from Main tableDaiID Text Dai identification numberDate Date/Time Date foreign key from main tableRecorder Text Name of the recorderLocation Text Administrative zone (Phnom Penh and Kandal)

Table 45 tbl_LunarAge&Phase* (Asterisks indicate name changes or new fields)

Name Type DescriptionDate Date/Time DateMoonAge Number Number of days following the New MoonQuarter Number Moon phase (1–First Quarter, 2–Second Quarter, 3–Third Quarter and 4–Full Moon)

2009 Dai Fishery database Additional Tables

Table 46 tblOtherInfo* (new table) (Asterisks indicate name changes or new fields)

Name Type DescriptionDOC Text Document number , number of the data page ( per month) foreign key from Main tableDate Date/Time Date foreign key from main tableGearCode Text Gear code key from main tableStratumNum Number Stratum number, foreign key from main tableTime_Start* Date/Time Beginning bagnet soak timeTime_End* Date/Time End bagnet soak timeTotalCatch/haul* Number Total catch per haul (kg)No_of_haul* Number Number of hauls

Annex


Page 142

Table 47 Annual Hydrological Indices* (Asterisks indicate name changes or new fields)

Name Type DescriptionSeason Text Fishing seasonFI (km2 days) Number Flood IndexFiy-1 (km2 days) Number Flood Index–1 year DSI (km2 days) Number Dry Season IndexAFI (km2 days) Number Annual Flood IndexFlood start Number Serial days since 1st JanuaryFlood end Number End of flood (days relative to January, 1)Flood duration 1 Number Days between start and end of floodFRR (m/day) Number Flood Rise rateDDRy-1 (m/day) Number Drawdown rate (previous year)

Table 48 tbl_LunarAge&Phase* (Asterisks indicate name changes or new fields)

Name Type DescriptionDate Date/Time DateMoonAge Number Serial days since new moonQuarter Number Lunar quarter (moon age / 7)



P.O. Box 623, Phnom Penh, CambodiaTel. (855-23) 425 353 Fax. (855-23) 425 363

Office of the Secretariat in Vientiane (OSV) Office of the Chief Executive Officer 184 Fa Ngoum Road, P.O. Box 6101,

Vientiane, Lao PDRTel. (856-21) 263 263 Fax. (856-21) 263 264

© Mekong River CommissionE-mail: [email protected]: www.mrcmekong.org

E-mail: [email protected] Website: www.mrcmekong.org


P.O. Box 623, Phnom Penh, Cambodia

Tel. (855-23) 425 353 Fax. (855-23) 425 363

Office of the Secretariat in Vientiane (OSV), Office of the Chief Executive Officer

184 Fa Ngoum Road, P.O. Box 6101, Vientiane, Lao PDR

Tel. (856-21) 263 263 Fax. (856-21) 263 264

the stationary trawl (dai) fishery of the tonle sap-great lake

Documents