Télécharger le fichier original (Mémoire de fin d’études)

Fish migration system

Mekong fishes migrate longitudinally and laterally among critical habitats of the Mekong mainstream and its tributaries or between the floodplains and deeper areas of lakes or permanent water bodies. Migration usually takes place for all life stages of fish and is associated with dry-season refuging, flood-season feeding and rearing, and migrations for spawning as well as escaping from adverse environmental conditions (Welcome 2001, Poulsen et al. 2002). Generally, three different fish migration systems have been identified in the LMB (Valbo-Jorgensen and Poulsen 2000, Poulsen et al. 2002, 2004). The first migration system takes place in the lower part of the Mekong system between deep pools of the Mekong mainstream in Kratie-Stung Treng reach (dry-season refuge habitats) and the floodplain of Tonle Sap Lake, area South of Phnom Penh and the Mekong delta of Viet Nam together known as flood-season feeding and rearing habitats. The second occurs in the middle part of the LMB (between Khone Falls and Loei Province) and is characterized by the migration between the rapids and deep pools of the Mekong mainstream and the floodplain habitats which are connected with the Mekong’s major tributaries. The third migration system occurs in the areas of upper part of the LMB in the downstream stretch of Loei River in Thailand to Luang Prabang in Lao PDR. This last migration reach is represented by rapids with deep pools and restricted floodplain habitats.
In the three migration systems, hydrology plays a central role in structuring up- and downstream fish community dynamics such as triggering fish to migrate among critical habitats during their life cycles (Poulsen et al. 2002, Baran 2006). General seasonal migration patterns of the Mekong fishes particularly those with white and grey ecological charateristics are reflected in seasonal hydrological patterns. For instance, fishes migrate for spawning in early wet season in May and June when the Mekong’s water levels start rising. Afterwards, between July and November, both adult fish and larvae move to floodplains for feeding and growth. When water levels are falling particularly in December and January, these fishes migrate to permanent water bodies such as deep pools in the Mekong mainstream or lakes, and then remain sedentary in the permanent water bodies during the dry season (February – April). Fig. 3 gives a generalized life cycle of a Mekong fish species. Changes in hydrological patterns caused by anthropogenic activities such as infrastructure development are highly likely to distrupt the river biological system i.e. fish migration and reproduction success, which in effect alters fish community structure and reduce the overall fisheries productity in the Mekong system.

Socio-economic importance of fisheries in the Lower Mekong Basin

In 2015, the total population of the LMB was estimated at 68.9 million (So et al. 2015). Some 80% of the LMB’s dwellers is rural, and the economy highly depends on farming, fishing and aquaculture (Hortle 2009). About 66% of the LMB population was engaged in capture fisheries either part-time or seasonally (MRC 2010). At country level, ~80% of rural households in Cambodia, Lao  PDR and Thailand and 60-95% of households in Viet Nam delta were involved in capture fisheries (Hortle 2007). In large water bodies such as the Tonle Sap, commercial fishing appears to represent more than 40% of household (Ahmed et al. 1998).
Inland fish and other aquatic animals make up of more than half the animal protein consumed by people in the LMB which is more than three times the world average of 16% (Baran et al. 2013b), and which range from ~50% in Lao PDR and Thailand to ~60% in Viet Nam and ~80% in Cambodia (Hortle 2007). The average consumption of aquatic animals in the basin is 46 kg per capita per year, similar to the Southeast Asian rate of 51 kg/person/year but significantly higher than the world average of 24 kg/person/year (Baran et al. 2013b). Other inland aquatic animals such as frogs, insects, clams, shrimps, snails and snakes contribute ~6% to the total animal protein consumption (Hortle 2007).
A recent estimate indicates that, based on the first sale landing prices, the LMB capture fisheries is worth about US$11 billion annually in 2015 (So et al. 2015). The largest single fishery in the basin is the century-old dai or stationary trawl bagnet fishery on the Tonle Sap River. The fishery operates between October through March and targets mainly white and grey fishes that migrate out of the floodplains surrounding the Tonle Sap Lake to the main river channels for dry season refuge. Based on first-sale prices, the value of the fishery, on average, is estimated at around US$10 million seasonally (Ngor et al. 2015b). First sale fish prices recorded at the dai fishery indicate that there have been increasing fish prices observed particularly since the fishing season of 2006-2007 at the time when there was also global food crisis. Fish prices of small mud carps (Henicorhynchus spp.), recorded over 20-year period at the dai fishery are shown in Fig. 4. These are ecological keystone species which are the most abundant with their critical role in food security throughout the LMB and important prey species for many predatory fishes and Irrawaddy dolphins (Roberts and Baird 1995, Hurwood et al. 2008, Baird 2011, Fukushima et al. 2014, Ngor et al. 2015a).

Challenges of inland capture fisheries in the Lower Mekong Basin

Many freshwater faunal species particularly fishes have experienced severe declines in their ranges and abundances, and they are now far more endangered than their marine or terrestrial counterparts (Jenkins 2003, Strayer and Dudgeon 2010). In the Mekong Basin, several dangers are identified as threats to the sustainability of the Mekong fish and fisheries. These threats stem from sources both outside and inside the fishery sector including population growth, hydropower dams, water extraction and diversion for agriculture, widespread habitat fragmentation and loss, water quality degradation, mining, farming expansion and intensification, land-use change, urbanization, climate change, pollution, overharvesting and introduced species etc. Among these threats, water resources infrastructure development, habitat loss and open-access nature of fisheries (overharvesting) in the region are among the great dangers threatening the region’s fishes and fisheries (van Zalinge et al. 2000, Welcome 2001, Halls and Kshatriya 2009, Valbo-Jørgensen et al. 2009, Welcomme et al. 2016, 2010, Ferguson et al. 2011, Ziv et al. 2012, Grumbine et al. 2012, Cochrane et al. 2014, Kummu et al. 2014, Winemiller et al. 2016, Sabo et al. 2017).

Water infrastructure development in the Mekong

During the last three decades or so, infrastructure development significantly poses by far the most significant threat to the Mekong River ecosystem, biodiversity and its fisheries (Arias et al. 2012, 2014b, Ziv et al. 2012, Piman et al. 2013, Cochrane et al. 2014, Winemiller et al. 2016, Sabo et al. 2017, Ngor et al. 2018b). For example, at least six large dams have been built in the upper Mekong River since mid-1990s (Fan et al. 2015, Winemiller et al. 2016) and in the LMB, two mainstream dams are under construction in Lao PDR and 10 others are planned. Among 144 tributaries dams, 42 are in operation, 27 under construction, 17 licensed and 58 planned by 2030 (Nielsen et al. 2015, Schmutz and Mielach 2015, Ngor et al. 2018b). These dams are known to disrupt river continuity, block migration routes of riverine fishes, dampen natural flood pulses, mute flow seasonality, fragment habitats, degrade water quality, and alter sediment and nutrient dynamics as well as other biogeochemical processes, which, in effect, alters the structure of aquatic faunal communities that adapt to natural seasonal flow dynamics as part of their life cycles (Collier et al. 1996, Agostinho et al. 2004, Graf 2006, Poff et al. 2007, Latrubesse et al. 2017, Sabo et al. 2017, Ngor et al. 2018b). Specifically, dams generate hydropower-related pulsed flows e.g. hydropeaking reacting to energy demands (from hourly to seasonally) which adversely affect riverine fishes and other aquatic organisms through, among other factors, stranding/extirpation, downstream displacement and spawning/rearing disruption (Young et al. 2011, Schmutz et al. 2015, Kennedy et al. 2016, Tonolla et al. 2017). In total, these pressures may lead to fish community compositional changes, fish recruitment failure and a continued diminishment of fisheries productivity in the system (Poulsen et al. 2002, ICEM 2010, Baird 2011, Grumbine et al. 2012, Ziv et al. 2012, Winemiller et al. 2016, Ngor et al. 2018b). Fig. 5 provides an overview of hydropower projects in the Mekong Basin.
For example, under the current functioning dams, the 3S’s dry seasonal flow shows an increase of 28% and the wet seasonal flows a decrease of 4%, when measured at the 3S outlet (Piman et al. 2013). Similarly, hydropower dams upstream of the Mekong have caused the most distinct changes to the Mekong’s flow, and their cascade impacts have been demonstrated from Chiang Sen in Thailand (the beginning of the LMB) as far as downstream in the Tonle Sap River in Cambodia which reduces flood pulses by 23% and 11% in rising and falling rates with observed changes taking place since 1991 (Arias et al. 2014a, Cochrane et al. 2014). These changes in natural flow dynamics and flood pulses have severe implications for fish community structure because, of an estimated 1200 fish species with 877 species recorded in the Mekong Basin (Rainboth 1996, Baran 2006, Baran et al. 2013b), about 87% are longitudinal and lateral migratory species (white and grey fishes) (MRC 2010, Baran et al. 2013b). Also, at least 89 migratory species including 14 endangered and critically endangered species characterize fish community from the 3S system (Baran et al. 2013a). In addition, of the 161 Mekong endemics, 17 species exist exclusively in the 3S Basin, and nowhere else on the planet (Baran et al. 2013a). More serious impacts are also expected for the fishes in the Tonle Sap Basin, hosting some 296 fish species (Baran 2005, Baran et al. 2013b). These fishes depend on natural seasonal-predictable flows and flood pulses as the main ecological trigger to disperse, reproduce and seek refuge (Valbo-Jorgensen and Poulsen 2000, Poulsen et al. 2002, 2004, Sverdrup-Jensen 2002, Baran 2006). Fig. 6 shows temporal change in daily water levels in the Mekong mainstream in Stung Treng Province over 95-year periods. Observably, there has been a general significant decrease in wet season flow (June-November), and an increase of dry season flow (December-May). Hydropower dams upstream in China have been attributed to cause the most ‘distinct change’ in the Mekong flow regimes as compared to other anthropogenic activities such as climate change (Cochrane et al. 2014, Winemiller et al. 2016, Sabo et al. 2017).

Habitat loss

Wetlands and river habitat degradation and losses in freshwater ecosystems are widespread worldwide. These habitats are critical for fish spawning, rearing, feeding, or for dry reason refuge. In the Mekong system, dry season refuge are usually situated in perminant water bodies or in the Mekong mainstream (with deep pools) such as in Kratie and Stung Treng Provinces in Cambodia and Champasack Province in southern Lao PDR. The critical habitats are also found either in the main river channel of the major tributaries or floodplains such as the 3S system, the Tonle Sap system and areas south of Phnom Penh and the Mekong delta. Natural flow dynamics ensure the lateral and longitudinal connectivity among these habitats. Many Mekong riverine fishes are known to migrate longitudinally up- and downstream and laterally between tributary rivers and floodplain areas to access the crtical habitats to complete their lifecycles. Therefore, dams physically block migrating fishes from accessing the critical habitats to complete their life cycle. Also, critical habitats such as deep pools that serve as dry season refuge in the main river channel are filled up with particles, sediments released by erosions triggered by hydropower related pulsed flows. As a result, fish is disabled to access these critical habitats which reduces feeding, rearing, spawinng and recruitment success, and thereby, diminishing the system’s overall productivity.
Habitat loss is also linked to cumulative effects of flow regulation which is caused by water infrastructure development. Various models indicate that effects of hydropower dams distinctly reduce wet season water levels and increase dry season water levels (Piman et al. 2013, Arias et al. 2014a).The reduction in water levels in the flood season means that seasonally flooded habiats (spawning, rearing and feeding habitats) are less available for fish. In the Tonle Sap, seasonally flooded habitats and gallery forest are estimated to have been reduced by 13 to 22% and 75 to 83%, respectively, whereas the increase in water levels in dry season (i.e. 18 to 21% in the open area of Tonle Sap) is causing permanent submersion of existing vegetation and forests (Arias et al. 2012) triggering a permanent dieback situation of the plants in the submerged area. Thus, these type of changes in the Mekong’s natural flow patterns ultimately lead to habitat fragmentation and destruction.
Moreover, other habitat losses are caused by the expansion of agriculture land, gathering of fuelwood, as well as enlargement of settlements in the LMB floodplains as a result of increasing population and government policies. Agriculture policies often focus more on the expansion and intensification of rice farming and industrial crop cultivation. The conversion of flooded forests into farmland and settlements have been accerlated during the last two decades (van Zalinge and Nao 1999, Hortle et al. 2004). These flooded forests are imortant for fishes as shelter, sources of food supply and breeding areas.

Open-access fisheries

Both increased fishing effort, efficiency of fishing gears and increased human population size have likely contributed to high fishing pressure and, thus, overexploitation of the fisheries resources. For example, the use of monofilament nylon gillnets in the LMB has accelerated the decline of some common and commercial species such as Cirrhinus microlepis, Boesemania microlepis, Probarbus spp. and Tenualosa thibaudeaui, Pangasianodon hypophthalmus, Wallago leeri (maxTL: 150cm) and Irrawaddy dolphins (van Zalinge and Nao 1999, Deap et al. 2003, Baird 2006). These highly efficient nets were considered as a ‘wall-of-death’ for many migrating fishes (Hortle et al. 2004 p. 33). The problems caused by these fishing techniques have likely been exacerbated by population growth in the countries sharing the LMB; statistics show that the population has increased about three folds between 1960 and 2015 with about 80-85% rural dwellers (World Bank Group 2015). Factors like free entry into fishing (open-access), affordability of fishing gears (Deap et al. 2003, Hortle et al. 2004), and the combination of rising population along with the lack of complementary and alternative livelihood options, has resulted in millions of people moving into the fishing sector. In addition, prevailing illegal fishing practices such as the use of dynamite, mosquito netting with fences and other destructive fishing methods have put high pressure on fish stocks in the region. Combined with many other streesors (i.e. hydrological alterations, pollution, invasive species and climate change), Mekong fishes and fisheries are facing severe challenges in sustaining its productivity that has for centuries supported millions of peoples’ livelihoods in the region.

Objectives

As briefly described, rapid water infrastructure development in the Mekong region (particularly hydropower dams and irregation schemes) since 1991 have changed the perception of the pristine Mekong system, one of the world’s most biodiverse river basins (Cochrane et al. 2014, Winemiller et al. 2016). The Mekong’s natural flow patterns are considered a key environmental driver which plays a main role in structuring the communities of aquatic organsims both up and dowstream (Brownell et al. 2017). Although change in the Mekong flow patterns have been documented to a certain extent, its impacts on fishes and fisheries in some critical areas such as the Mekong-3S system and the Mekong largest wetland of the Tonle Sap are largely undocumented (Arias et al. 2012, Piman et al. 2013, Cochrane et al. 2014). Further, status and trends of fisheries in the LMB during this last decade have not been documented albeit the perception that the region’s fisheries have been declining (MRC 2010). Aguably, among the tropical largest wetlands on the planet, the Mekong River and the Tonle Sap, which supports one of the world’s biggest freshwater fisheries, have received little ecologcial research and conservation attention (Dudgeon 2000, Junk et al. 2006, Vaidyanathan 2011, Allen et al. 2012, Ngor et al. 2018a). Therefore, there is an urgent need to document and update the system’s fish biodiversity, i.e. to generate reliable information about fish species diversity, species’ distribution, fish community composition and evolution through space and time. Combined with data on their ecological requirements the new insights from research can inform basin development planning as well as fisheries management and fish conservation actions.
In recognition of this important fact, the overall objective of the study is to investigate the dynamics of spatial and temporal fish community structure in the Lower Mekong system i.e. Lower Mekong River (LMR) and its major tributaries. To achieve the overall objective, the specific objectives are set out as follows:
(i) describe large-scale spatial fish diversity patterns and assemblage structure in LMR and its major tributaries.
(ii) examine spatial and temporal variation of fish assemblages in the complex Tonle Sap River and Lake system;
(iii) explore the signature of ‘indiscriminate fishing’ effects by examining the rates of temporal dynamics of the entire fish biomass composition of the Mekong’s largest, commercial-scale stationary trawl bagnet Dai fishery operating in the Tonle Sap River.
(iv) investigate spatial and temporal fish community responses to flow changes in regulated and unregulated rivers of the Lower Mekong system.
This thesis is divided into two main Parts. Part I is the Synthesis and Part II comprises the corresponding publications. In this Synthesis, Article 1-5 contribute to the overall description on broad-scale spatial and temporal variation in fish diversity patterns and assemblage structure in the LMR and its major tributaries (objective i). While Article 1 describes spatial fish distribution patterns in the LMR (objective i), Article 2 specifically investigates spatial and temporal variation of fish assemblages in the complex Tonle Sap River and Lake system (objective ii). Article 3 exclusively examines the ‘indiscriminate fishing’ effects of the Tonle Sap fisheries, by analysing temporal changes in the biomass of 116 fish species that seasonally utilize the Tonle Sap River system (objective iii). Finally, Article 4 and 5 scrutinize the spatial and temporal fish community responses to flow changes in regulated and unregulated rivers of the Lower Mekong system.

Materials and methods

Study area

This study covers the Lower Mekong system: the LMR and its major tributaries. LMR extends from the Golden Triangle which marks the borders of Thailand, Lao PDR, China and Burma, and which consists of Cambodia, Lao PDR, Thailand and Viet Nam. Key largest tributaries of the LMB include the TSRL and the Sekong, Sesan and Srepok Rivers known as the 3S Rivers (Fig. 8).

Data collection

This study uses data from the long-term routine daily artisanal fish monitoring (2007-2014) in the LMB and a standardized catch assessment of the stationary trawl Dai fishery (2000-2015), the largest commercial fishery in the Mekong Basin. Data were made available by the Fisheries Program of the Mekong River Commission (MRC) that technically and financially supported the monitoring and catchment assessment programs.
For the daily artisanal fish monitoring, standard sampling procedures of the MRC (MRC 2007) were applied. Fishers were trained on sampling procedures, fish identification and the use of data recording forms. They were supervised by the fishery researchers from the fisheries line agencies and research institutes of the MRC member countries with technical support from the MRC fisheries monitoring specialist. Fish photo books containing more than 200 fish species were also made available for all fishers to assist them in fish identification. Fish captured were identified to the species level and counted. Unidentified species were kept in formalin and taken to laboratory in the central office in each of the respective countries for further identification by professional taxonomists. At the end of each sampling quarter, the fishery researchers collected all recorded forms and data from all fishers. The recorded data were cross-checked with fishers for its accuracy and completeness before being brought to the national central offices for transfer into the national fish monitoring databases. The databases were quarterly cleaned and synchronized into a regional database with the help of an MRC database expert and capture fisheries specialist prior to the analyses.
For the Dai fishery, time series data of the fishery’s standardized catch assessment between 2000 and 2015 were used. The fishery operates seasonally from October through February/March in a specific location along the lower section of the Tonle Sap River, stretching about 4-30 km north of Phnom Penh. All Dai (64 units) are organized into 14 rows (row 2 to row 15) and operated individually or jointly of up to 7 units in a single row with the most upstream row 15 situated close to the Tonle Sap Lake. General concepts and formula for assessing catches and catch composition are outlined in Stamatopoulos (2002), and these concepts were used to frame the sampling protocols and assessing catches of the fishery. The sampling unit was based on Dai unit and a randomly stratified sampling method was used for the catch assessment. More specifically, Dai units were stratified based on: (i) administrative space divided into two strata (Phnom Penh Municipality and Kandal Province), (ii) time
– the lunar period (low period and peak period) and (iii) Dai types (high yield and low yield Dai units). Random sampling on catches per haul or catches per unit of effort (CPUE; including CPUE for species in catch composition) and daily number of hauls of a Dai unit were conducted in each stratum, lunar period and Dai type within each month for monthly catch estimate. Likewise, fishing effort (number of active Dai units and active days) were recorded according to the stratification framework throughout each fishing month over the whole fishing season. Apart from sampling data on total catch for each species in each season, data were also obtained for the number, weight and length of some common and commercial individual fish specimens caught per day of each fishing season. These species (i.e. Henicorhynchus lobatus, Labiobarbus lineatus, Pangasianodon hypophthalmus, Cyclocheilichthys enoplos, Cirrhinus microlepis, Osteochilus melanopleurus) are among the most ecologically, socioculturally (food nutrition and security) and economically important species in the region (Rainboth 1996, Poulsen et al. 2004, Sabo et al. 2017). Therefore, they were used to examine the temporal changes in body weight and length for this study (Article 3).
In addition, this study uses a fish species list (about 900 species and their ecological attributes) that was obtained from the Mekong Fish Database (MFD 2003); the species list was updated by cross-checking with FishBase (Froese and Pauly, 2017), the Catalogue of Fishes Online Database and other literature sources i.e. (Rainboth 1996, Rainboth et al. 2012, Kottelat 2013). Moreover, other fish datasets i.e. maximum total length (maxTL), trophic level and habitats in the water column were consulted from FishBase.
Article 1 uses daily fish monitoring datasets from 38 sites along the Lower Mekong River collected from November 2000 to December 2001. Article 2 uses 4-year daily time-series datasets from artisanal fishers (stationary gillnets and cylinder traps) in six sites: first site located on the Tonle Sap River and the other five sites situated in each of the five provinces around the Tonle Sap Lake from 2012 to 2015, whereas Article 3 uses the 15-year standardized seasonal catch assessment data of 116 fish species from the commercial-scale Dai fishery in the Tonle Sap River from 2000 to 2015. Finally, Article 4 and 5 uses a 7-year daily stationary gillnet monitoring data (riverine habitat) from six sites in the complex Mekong-3S system and Tonle Sap River.

Statistical analysis

Seasonal partitioning

In the Tonle Sap system (Article 2), the unique tropical flood pulse with flow reversal system i.e. rising water levels with flow direction to the Tonle Sap Lake (inflow) and falling water levels with reverse flow direction to the Mekong River (outflow) plays a pivotal role in influencing the intra-annual variation in fish community structure. For this reason, three seasons are defined to reflect the importance of the TSRL flood pulse system, using the 10-year mean intra-annual variation of daily water levels measured at the Tonle Sap Lake (Kampong Loung in Pursat [PS]): inflow or high flow period (July-October), outflow (November-February) and low-flow (March-June). In the Cambodian Mekong and 3S systems, seasonality is defined by a general wet and dry season of the tropical zone for the investigation of the intra-annual variation of fish communities (Article 1, 4, 5). The seasonal partitioning was based on 9-year mean daily water levels of the Mekong River, when entering Cambodia (at Stung Treng [ST]), with wet season covering the period from June to November and dry season from December to May.

Data preparation

For Article 1, all fish catches are transformed into relative abundance to reduce the effect of varying fishing efforts between sites and averaged to annual mean relative abundance prior to analysis. For Article 2, 4, 5, daily abundance data on stationary gillnet (and cylinder traps for Article 2 only) are computed as mean daily samples and then aggregated into weekly species abundance data. Article 3 is based on seasonal catch assessment data from all 64 units of the stationary trawl bagnet (Dai) fishery operating in the Tonle Sap River.

Flow seasonality and predictability

To quantify the strength of seasonality, Colwell’s seasonality index (Colwell 1974) on site daily water levels (Mekong, Sesan [3S], Tonle Sap) is computed using Colwells function of hydrostats package. The seasonality index M/P which is the Colwell’s measure of contingency (M) standardized by Colwell’s within-season predictability (P) (Colwell 1974, Tonkin et al. 2017) is used. In addition, modern wavelet analysis is applied to quantify the strength of predictability of site hydrology, using analyze.wavelet function, from WaveletComp package of the ‘mother’ Morlet wavelet (Roesch and Schmidbauer 2014).

Spatial and temporal description of fish community

All data analyses are performed in R (R Core Team 2017). Summary statistics, cluster analyses (using hclust with Ward hierarchical, and K-means clustering methods), boxplots, scatterplots, bubble plots, violin plots, jittering plots and histograms are applied to give a descriptive overview on the spatial and temporal dynamics of fish community structure, as well as weight and length of individual fishes by site and entire species pool in relevant study locations.
Unconstrained ordination techniques, e.g. Nonmetric Multidimensional Scaling (NMDS) and Principal Components Analysis (PCA) (Borcard et al. 2011, Kassambara 2017) are used to visualize fish assemblage samples in a two-ordination plane for the description and analyses of spatial and temporal variability of fish assemblage patterns in important areas of the LMB (Article 2, 4, 5). In addition, for time-series analyses, various time-series analytical tools are applied (Article 2-5). These tools include Whittaker–Robinson periodograms (Legendre and Legendre 2012, Dray et al. 2017), cross-correlation analyses (Shumway and Stoffer 2011), wavelet and cross-wavelet analyses (Roesch and Schmidbauer 2014).
For statistical tests, Permutational Multivariate Analysis of Variance (PERMANOVA) using adonis function of vegan package (with 999 permutations and bray method) is used to test the influence of different factors (e.g. cluster, season and year) on the fish community composition. Complementary, contrast methods are applied to test the pairwise differences between different levels in each of these factors, using pairwise.adonis function in R. In addition, non-parametric Wilcoxon rank-sum and Turkey’s multiple comparison tests are performed to test the significant differences between variables i.e. survey sites or weeks/years over the study period. For correlation tests, non-parametric Spearman’s correlation tests are used. Significance at the 0.05 level is applied for all tests. Further, to identify species indicator characterizing fish communities in a study site or a cluster, multipatt function from indicspecies package is applied (Cáceres and Legendre 2009, De Cáceres and Jansen 2011).

Table des matières

I. Introduction
1.1 A brief about the Mekong system
1.1.1 The Mekong River
1.1.2 The 3S Rivers
1.1.3 The Tonle Sap system
1.2 The Mekong fisheries
1.2.1 Fish community structure
1.2.2 Fish migration system
1.2.3 Socio-economic importance of fisheries in the Lower Mekong Basin
1.3 Challenges of inland capture fisheries in the Lower Mekong Basin
1.3.1 Water infrastructure development in the Mekong
1.3.2 Habitat loss
1.3.3 Open-access fisheries
II. Objectives
III. Materials and methods
3.1 Study area
3.2 Data collection
3.3 Statistical analysis
3.3.1 Seasonal partitioning
3.3.2 Data preparation
3.3.3 Flow seasonality and predictability
3.3.4 Spatial and temporal description of fish community
3.3.5 Species diversity
3.3.6 Linear regression models
IV. Results
4.1 Summary of recorded catches in the Lower Mekong Basin
4.2 Overall fish assemblage structure and diversity
4.2.1 The Lower Mekong River
4.2.2 The complex Mekong-3S system
4.2.3 The Tonle Sap system
4.3 Spatial variation in fish abundance distribution
4.3.1 The Lower Mekong River
4.3.2 The Mekong-3S system
4.3.3 The Tonle Sap system
4.4 Temporal dynamics of fish community
4.4.1 Temporal variation of fish community in the Tonle Sap River and Lake
4.4.2 Temporal dynamics of fish communities in the Lower Mekong system
V. Discussion
5.1 Fish species richness and diversity
5.2 Spatial variation in fish community structure
5.3 Temporal variation in fish community structure
5.3.1 Flow variation in the Lower Mekong system
5.3.2 Intra-annual variation in fish community structure
5.3.3 Inter-annual variation in fish community structure
VI. Conclusion and implications for fisheries management and conservation
6.1 Conclusion
6.2 Implications for fisheries management and conservation
VII. Further research
References
Annexes
Annex 1. List of indicator species in each cluster in the Lower Mekong River
Annex 2. Important species contributing to overall beta diversity

Télécharger le rapport complet