The physical environment
Biotic components of the ecosystem
Productivity of the ecosystem and natural resource use
Limnological parameters: methods
Results of the limnological sampling
Discussion: floodplain habitats ecology
The Tonle Sap ecosystem is defined here as the permanent core area of the Tonle Sap lake and the surrounding natural floodplain, within the boundaries constituted by the upper flood lines. The natural upper flood levels tend to vary from year to year and are not always clear, especially in areas where flood and rainwater retention structures have been built. In most places, human activity is determined by the average expected flood levels and the boundaries of the ecosystem are conspicuous. The ecosystem as defined here also includes the Tonle Sap channel, i.e. the channel between the permanent core area of the Tonle Sap lake to the west and the Mekong river to the east, including its branches and floodplain (Fig. 1.2).
i. Climate
ii. Geography
iii. Hydrology
The atmospheric conditions of the Tonle Sap ecosystem are determined by the dry and wet-monsoon climate of the lowland plains of Cambodia. The climate is characterized by a long rainy season determined by the southwest monsoon, beginning in May-July and reaching a peak in precipitation in October. This is followed by the dry period of the northeast monsoon from November to April, which brings generally dry weather over Cambodia and makes December and January the coolest months of the year. The average annual rainfall in Phnom Penh is 1 407 mm. The rainfall across the Cambodian lowland plains increases in a southwest-northeast direction (Mekong Committee 1992). Precipitation for the Tonle Sap lake and an area that stretches about 100 km to the north and the northeast from the lake averages between 1 000 and 1 250 mm per year.
Figure 1.1 Average monthly rainfall and minimum/maximum temperatures for Phnom Penh
Figure 1.2 Map of Cambodia. The Tonle Sap ecosystem lies within the boundaries of the upper flood area of the Tonle Sap lake and channel.
Continental Southeast Asia is situated on the Sunda shelf, which is currently exposed as the Indo-Chinese Peninsula and a series of large islands including Sumatra, Java and Borneo. Current sea levels are up to 120 m above the lower parts of the shelf. Several cycles of regression and transgression have repeatedly exposed nearly the entire shelf and again covered large parts of it, including land that is up to 7 m higher than the present sea level. During the periods of exposure, extensive systems of rivers existed which have become disconnected following the transgressions.
The extensive and complex tectonic and volcanic activity of the region has shaped the course of the Mekong, although the present configuration of the river is very recent. Previously, the Mekong debouched into the Gulf of Thailand in Kampot. Many rivers in the region that are discontinuous today were connected in the past, not once but several times. Rainboth (1996) provides an overview of the recent changes of the river basins of the region. Thus, before becoming integrated in the Mekong system, the Tonle Sap channel may have been part of another large river which has disappeared since.
The Tonle Sap lake is situated in the centre of the Cambodian central plain, which has an elevation of 10-30 m above sea level and covers 75 percent of the country. The lake was formed only less than 6 000 years ago when the most recent subsidence of the Cambodian platform took place (Carbonnel, 1963).
The hydrology of the Tonle Sap ecosystem is mostly determined by the over 4 000 km long Mekong river. The lake is connected to the Mekong river through the 100 km long Tonle Sap channel. The channel and the Mekong join at the Quattre Bras near Phnom Penh, after which the river immediately branches into two arms, the larger main Mekong and the smaller Bassac river. Farther downstream these two arms are reconnected to form two equal channels as they start to fan out to form the delta, discharging into the South China Sea.
The Mekong originates in the same area in southeastern Tibet where four other Asia’s major rivers (Brahmaputra, Irriwaddy, Salween and the Yangtse) have their origin. The total drainage area of the Mekong is 795 000 km2 (Mekong Committee, 1992).
The part of the Mekong in Cambodia is 486 km long and drains a total of 156 000 km2 or 86 percent of the country (Hak and Piseth, 1999). There are two different sections. From the Khone falls at the border with Laos down to Kratie, the river type is that of a fast streaming upland form, with relatively high turbidity and characterized by a succession of rapids, deep pools and sand banks. Gradients are high, currents strong and the inundation zone is localized. The flow ratio between high and low water is very high. In the section downstream from Kratie, the Mekong shows the characteristics of a broad lowland river. The gradients are shallow and seasonal flooding by the river waters covers vast areas.
The discharge of the Mekong reflects the pattern of the rainfall distribution throughout the year and follows the monsoon seasons. Consequently, the flow patterns display predictable cyclical changes. The water level varies accordingly and fluctuations between years in the maximum water levels and discharge volumes are large. However, minimum levels are stable with little variation and do not seem related to the preceding maximum levels (Hak and Piseth, 1999). The maximum water level in the Mekong in Kratie is reached in September-October. The river flow decreases rapidly until December, and then slowly during the dry period, reaching minimum levels in late April. As the rising water level in the Mekong during the following months reaches 7 metres above mean sea level, overland flow begins and the flow in many tributary rivers is reversed.
The high velocity of the water at the beginning of the flood increases the carrying capacity and consequently reduces the clarity of the water by five to ten times compared with the dry season. The sediment load of the Mekong is low compared with other major rivers, and decreases downstream (Mekong Committee, 1992).
Flow reversal also occurs in the Tonle Sap channel. By the end of May, the water level of the Mekong at Quattre Bras reaches sufficient height (9 m) to start pushing water from the Tonle Sap drainage area back in the channel, thus reversing the flow. This results in an accumulation of water from the Mekong and from its proper drainage area in the Tonle Sap lake. Extensive flooding is the result, increasing the surface area of the lake from about 2 500 km2 to over 10 000 km2 and, according to some sources, even to almost 16 000 km2 (Rainboth, 1996). The size of the lake varies between 160 km long and 35 km wide at its widest point during the low water time, and 300 km in length and over 100 km in width in some places at the height of the flooding. Maximum depth is reported to vary between less than 2 m in the dry season and up to 14 m at the peak (Rainboth, 1996; Mekong Committee, 1992; Hak and Piseth, 1999). After about five months, in October, the water level in the Mekong has sufficiently subsided for the flow in the Tonle Sap channel to reverse and the lake and the thousands of square kilometres of floodplain to be drained. Hak and Piseth (1999) estimate the volume of water yearly stored in the Tonle Sap lake at 72 Gm3. The Tonle Sap drains an area of 85 065 km2, which is 10.7 percent of the total drainage area for the Mekong. It contributes 6.4 percent of the average annual flow of the Mekong (Mekong Committee, 1992).
Farther downstream, at the border with Viet Nam, the Mekong begins to experience tidal influences, even though the water remains entirely fresh.
The primary production of the Tonle Sap ecosystem depends on the phytoplankton and the macrophyte vegetation of the floodplain. A Vietnamese study in the late 1980s (Nguyen and Nguyen, 1991) identified 197 species of phytoplankton in the Mekong, the Tonle Sap channel and the floodplains. Green algae and diatoms make up almost two thirds (64 percent) of the plankton species, one quarter are blue-green algae and the remainder is composed of Euglenophyta and some Xantophyta, Pyrrophyta and Chrysophyta. Among the diatoms, there are 17 marine species and only about 10 percent of the species are specific to the river or to the floodplain. The largest number of species was found in the main stream of the Mekong. The phytoplankton density varies considerably in the rivers and is about 40 times higher in the dry season than during the high-water period. The density is in general higher in the flooded areas than in the rivers.
The macrophyte vegetation is one of the most conspicuous elements determining the structure and nature of the habitats in the ecosystem. McDonald and Veasna (1996) and McDonald et al. (1997) provide a detailed study of the macrophyte vegetation of the Tonle Sap ecosystem. The present seasonally inundated vegetation is composed of several distinctive vegetation types that are located exclusively within the boundaries of the fluctuating shorelines. The original macrophyte vegetation of the floodplains of the ecosystem has assumed a variety of forms as a result of a thousand years of human impact and is at different stages of degradation and regeneration.
A number of cultured crops make up part of the macrophyte vegetation of the ecosystem nowadays. These include maize, mung bean, lotus, rice and a variety of vegetables. In places, growth of grasses is facilitated for grazing of herded animals (cattle, pigs, ducks).
The zooplankton of the Tonle Sap ecosystem is partly described in the study by Nguyen and Nguyen (1991). Half of the 46 species identified are rotifers, about one third are cladocerans and the remaining seven species are copepods. The number of species in the flooded areas is higher than in the river, and the composition of the zooplankton communities varies throughout the year. In the Tonle Sap channel, the zooplankton density increases by a factor of 107 during the dry season. This is attributed to the impact of the water of the Tonle Sap lake but no explanation is given on how this effect is brought about. In the flood season, zooplankton density in the flooded areas is almost one hundred times higher than in the river. The dynamics of plankton communities are complex and considerable variation in species composition and density is likely.
The biomass of the zoobenthos does not vary much with the seasons. Nguyen and Nguyen identified 57 species: 15 insects, 8 oligochaete worms, 29 molluscs and 5 crustaceans. Molluscs make up as much as 85 percent of the zoobenthos by weight. Rainboth (1996) terms the rich diversity of molluscs in the Mekong as “striking”.
The neuston has been little studied in tropical floodplains but usually makes up a non-negligible part of the aquatic ecosystem. Mosquito larvae, insects (hemiptera and coleoptera) and other invertebrates are common in floodplains, particularly in the sheltered water among the stems of floating or submerged vegetation (Welcomme, 1985).
Of all the vertebrates of the Tonle Sap ecosystem, fishes are undoubtedly the largest group, both in number of species as well as in biomass. The precise number of species in the Mekong is not known; the total number recorded or inferred from the known zoogeography of the region includes about 1 200 species (Rainboth, 1996). Welcomme (1985) ranks the Mekong in terms of fish species diversity as the third richest river in the world, after the Amazon and the Zaire, with a total number of known species of fewer than 1 000. Based on the relationship presented by the same author between the basin area of 47 of the world’s largest rivers and the number of fish species they hold (N = 0.297A0.477, with N the number of species and A the basin area in km2), as many as 1 938 species would be expected from the Mekong.
Even though the total number of species from large systems tends to be high, groups of species often are confined to very different parts of the river. The three different parts of the Mekong system in Cambodia (the Tonle Sap lake, the main river and the high estuary) support rather distinct fish assemblages (Rainboth, 1996). These differences are based on the physical characteristics and present ecological parameters of the sections and the differences in historical configurations. The complex geological and climatic history of continental Southeast Asia is largely responsible for the present-day fish diversity. The repetitive physical separation and mixing of species in the extensive river systems of the Sunda shelf has contributed to the high number of species in the Mekong system.
The ecological parameters that determine the different habitats in the Mekong have locally given rise to isolated species, although the degree of endemism is not known. The Tonle Sap ecosystem is expected to lack localized endemism (Rainboth, 1996).
The total number of fish species present in the Tonle Sap is not known either. About 500 species have been described for the Mekong system in Cambodia (including the Tonle Sap ecosystem) but the real number is certainly higher (Rainboth, 1996). Furthermore, there is not much reliable information available about the biogeography of most species inside Cambodia. Available information usually does not allow assessing which species are present in the Tonle Sap ecosystem. The data collected from fishery operations tend to underestimate the number of species considerably. This is due to the uncertainty about most field identifications. Identifications based on local names of fish species are not very useful in this respect since they usually cover more than one, and often many, biological species. The lack of a practical, comprehensive fish species identification guide for use in the field by local data collectors contributes to this uncertainty.
Some of the species found in the Tonle Sap remains there permanently, while many other species use the lake and the floodplain only temporarily and migrate back and forth to the Mekong.
Amphibians and reptiles are common in the ecosystem, even though larger species like turtles and crocodiles have become rare or have disappeared altogether as the result of excessive hunting. Aquatic snakes are very common, as are toads and frogs.
The Tonle Sap ecosystem is home to some of the world’s rarest and endangered bird species including Pelicanus philippensis (Spot-billed pelican), Leptoptilos dubius (Greater adjutant stork), Cairina scutulata (White-winged duck), Mycteria cinerea (Milky stork) and Leptoptilos javanicus (Lesser adjutant stork). It shelters some of the last viable populations of bird species thought to be extinct elsewhere. The flooded forest hosts the largest breeding water bird colony anywhere in Southeast Asia covering an estimated 875 ha in Battambang province, mostly in Fishing Lot 2 (E. Briggs, personal communication, 2000). Most of the birds species of international conservation significance known to occur in Cambodia are found in habitats that are also present in the Tonle Sap ecosystem (Ministry of Environment, 1998). Bird colonies are large in places and they can locally play an important role as fish predators and in nutrient recycling (Welcomme, 1985).
Orcaella brevirostris, the freshwater dolphin that is found in the Mekong, is occasionally also seen in the Tonle Sap ecosystem. As all larger wild animals in the Tonle Sap ecosystem, it is becoming increasingly rare.
Cultured animals are increasingly becoming a permanent part of the Tonle Sap ecosystem. Inundated forest in the floodplains is cleared using slash-and-burn techniques in order to facilitate the growth of young grasses on which cattle and ducks are grazed.
i. Ecosystem productivity
ii. Natural resource use in the Tonle Sap ecosystem
The Tonle Sap ecosystem is widely believed to be one of the most productive inland waters in the world (e.g., Mekong Committee, 1992; Rainboth, 1996; Ministry of Environment, 1998). This assessment is mainly based on the amount of harvested fish, which will be discussed in more detail later on.
The productivity of the ecosystem is generally attributed to two of its particular characteristics: the flood cycle with extensive, long-lasting floods, and the vegetation of the floodplain usually described as flood or flooded forest. Fig. 1.3 provides a simplified model for the productivity of the Tonle Sap ecosystem. The precise mechanism of the productivity is not yet fully understood but high rates of nutrient cycling and the high biodiversity are key factors (T. Hand, personal communication, 1999; Junk and Furch, 1991).
The input of silt-laden water from the Mekong and from the catchment area of the Tonle Sap ecosystem provides an annual addition of silt that is deposited in the floodplain, particularly where floods have been at their deepest, as in Battambang and Siem Reap (Mekong Committee, 1992).
The migration of many fish species between the Tonle Sap ecosystem and the Mekong is extensive and diverse. Regardless of the precise migration patterns that differ for most species, there is an annual mostly passive migration of large numbers of eggs, fry, juvenile and adult fish into the Tonle Sap as the water of the channel begins to flow towards the lake from the Mekong. The monsoon rainfall in the catchment of the ecosystem proper and the rainwater collected in the Mekong cause the rise in water levels and subsequent flooding. Mekong Committee (1992) provides a model for the hydrological systems of the lower Mekong basin which further clarifies the pathways of the water in this stage.
The productivity model (Fig. 1.3) is based on the parameters that are essential for the productivity of the Tonle Sap ecosystem. The level and duration of the floods are essential as they determine the area that becomes part of the aquatic phase of the ecosystem and thereby the amount of terrestrial primary products that can contribute to the aquatic productivity.
Figure 1.3. Tonle Sap ecosystem productivity model. The dotted line indicates the boundaries of the Tonle Sap ecosystem. Products are considered final as they leave the ecosystem
As substantiated further on in this chapter, the aquatic habitats go through different ecologically significant phases which depend on the duration of flooding. The flooded habitats in the floodplain need time to develop and to perform the role they have in the aquatic production.
The migration of fish at different stages of development into the ecosystem at the beginning of the flooding is for many species very important, as it determines the initial stock levels and therefore the potential final production. Migration out of the Tonle Sap ecosystem is essential for many species, e.g. for mere survival away from deteriorating water quality, or for reproduction. Minimum levels of migration of fish in and out of the ecosystem are therefore another essential parameter for its productivity.
With silt, a number of nutrients enter the production system of the Tonle Sap. It is not known how important this input of nutrients is for productivity. The levels of silt load and subsequent siltation in the Tonle Sap ecosystem are reportedly increasing, reaching levels where the productivity of the lake could be affected (Mekong Committee, 1992). Increased siltation could alter water levels and quality and influence the migration of fish in and out of the ecosystem.
The flooded macrophyte vegetation of the floodplain is another essential parameter of the productivity model. The ratio between primary products of the system that originate from phytoplankton and those coming from the macrophyte vegetation is not known. However, aquatic primary productivity is thought to be not as high as is the case in similar tropical lakes (Nguyen and Nguyen, 1991; Boyd, 1990; J. P. Descy, personal communication, 1997). Apart from this energy input, the flooded forest vegetation is important as a spatial structure with multiple functions, particularly that of a substrate, and which defines many different habitat types in the floodplain. The present-day vegetation of the floodplain is becoming more diverse under the impact of human activities but in most cases, this entails a loss of primary products available for the aquatic phase and paradoxically results in a loss of habitat diversity. The ecological functions of the flooded vegetation effectively result in an amplification of the primary productivity by enabling the development of subsequent associated communities of microorganisms, epiphytic algae and their associated communities and so on, all characterised by high turn-over rates of nutrients.
The high degree of biodiversity among the fish makes that a maximum of niches are created, making intensive use of the potential of the ecosystem at a maximum time, and that the available resources are used to a high degree.
These are the parameters that determine largely the productivity of the Tonle Sap ecosystem. They are therefore prime points of attention and concern for the management and conservation of the system and the livelihood of the people who are dependent on its rich production.
All elements of the Tonle Sap ecosystem that can be exploited with the mostly low-tech means locally available are being exploited. Most of the collection activities are labour-intensive. Harvesting of products from the ecosystem for subsistence use or for local trade at community or village level is done for nearly all the products; the more valuable items are also exploited on a larger scale, and some even in very large, labour and capital-intensive operations.
The natural floodplain vegetation is used for the collection of a variety of wood and non- wood forest products. This exploitation has been going on to the extent that in large parts of the floodplain the original trees have almost all disappeared and only pruned or regenerated morphs remain. The level of degradation appears to be linked with the accessibility of the area (McDonald and Veasna, 1996). In many places, the original vegetation is entirely cleared.
Wood is collected from the flooded forest for a wide variety of purposes: for domestic use as fuel wood or as charcoal, as construction material, for use in brick kilns, for fish processing (smoking and drying), for the construction of fishing gear, for use as artificial shelter for fish and shrimp in order to facilitate their capture (so-called brush parks), etc.
The flooded forest is also a rich source of non-wood forest products. These include a wide range of plants used as food and for medicinal purposes for man and husbanded animals. Lianas are collected for furniture and fishing gear production. Other plant products include fruits, seeds, resins, tubers, bark and mushrooms. Forest animals and their products that are collected include beeswax and honey, while some larger animals are collected for use as pets (macaques, iguanas, birds) or trade but they often end up as food.
Aquatic plants are collected for human consumption, as feed for farmed animals (pigs, fish) or for further cultivation, as is the case with lotus. Dried water hyacinth (Eichchornia crassipes) is sometimes used as fuel, especially in villages where after the recession of the floodwaters, large amounts of the dry plants have accumulated under the houses and create at the same time a considerable fire hazard.
Most macroscopic aquatic animals are used by humans. Several aquatic insects (mainly coleoptera) are collected for consumption. Shrimps are caught and molluscs (mainly snails) collected for food. Frogs are collected from the wild to be further raised in pens for the local food market. All reptiles are sought after. Snakes are hunted for consumption of body parts or for trade, alive or for their skins. Tortoises are collected for exquisite food and for the trade in tortoiseshell. Birds are hunted for food, for use as pets and for trade. Eggs and chicks are collected for consumption, especially in the floating villages where they are one of the few local sources of animal protein other than fish. Suntra (1995) mentions kingfisher feathers as one of the main non-wood forest products during the height of the Angkor era. The largest volume of products of the Tonle Sap ecosystem is fish. Fish are harvested in many ways. They are either consumed directly as such, or processed for added value (smoking) or for preservation (drying, smoking, prahoc and fish sauce production). Certain species are traded, some of them alive. Fish, mostly juvenile specimens, collected from the wild form the majority of seed for local aquaculture in the Tonle Sap, mostly in large floating cages and pens. Considerable volumes of the fish harvested are used as a cheap feed source for husbanded animals (crocodiles, pigs, fish).
The water of the Tonle Sap ecosystem forms per se one of the important natural resources. Floodwater is retained in the floodplain for irrigation and for cultivation of flood-recession rice. The rising water levels in the tributaries and in some places the reversed currents are used for irrigation. The Tonle Sap lake and channel form one of the important transportation ways between the centre of the country and the northwest. Thousands of tourists travel between Siem Reap and Phnom Penh each month and large quantities of freight and passengers are transported. A variety of vessels are used for these purposes, and they all have in common to be causing high levels of noise pollution, possibly impacting on (migrating) fish. The water of the lake also provides a living ground for floating villages.
Figure 1.4. Land cover of part of the floodplain of the Tonle Sap lake, 1992
Part of the land in the floodplain is used for agriculture. For this purpose, the original vegetation is removed, usually with the use of fire. Agriculture activities range from herding of cattle, pigs or ducks to the growing of corn and other cash crops like tobacco, watermelon and mung bean. Rice and a variety of vegetables are grown in the floodplain and on the riverbanks after the water has receded. Based on the model presented here, it can be expected that most types of agriculture practised in the floodplain tend to reduce the aquatic productivity of the lake.
In order to describe the different habitats of the Tonle Sap ecosystem, a number of basic physico-chemical water parameters were measured, in particular those parameters that have a determining importance in sustaining aquatic life.
· Dissolved oxygen
Dissolved oxygen levels were measured using a Yellow Springs Instrument® electronic dissolved oxygen meter. The probe was calibrated against moist air, calibration that had proved satisfactory by comparing it with the Winkler method. Saturation levels were calculated using the reference tables in Boyd (1990).
· pH
The pH of samples was measured using an electronic pH meter after calibration against a neutral and an acid buffer.
· Temperature
The temperature sensor in the probe of the dissolved oxygen meter was used to measure the temperature of the samples. This was calibrated against a precision mercury thermometer.
· Secchi depth
The Secchi depth was measured by lowering a weighted Secchi disk of 20 cm diameter. The Secchi disk depth was calculated as the average of the depth where the disk disappeared during descent and the depth at which it reappeared during ascent. The rope it was attached to was marked with 10 cm intervals.
· Specific conductivity
Specific conductivity was measured using an electronic conductivity meter, correcting automatically the measured effective conductivity for that at 25°C.
Water samples were collected using a horizontal Van Dorn bottle of 2.5 litres at regular depth intervals of 1 metre, starting from just below the water surface to, as conditions would allow, near the bottom. The water quality parameters were measured from samples immediately after collection.
The sampling procedure outlined in the introduction was followed for the collection of data to assess seasonal variations in water quality. In addition, another sampling and data collection procedure was followed to determine diurnal fluctuations or patterns.
For the former, sampling and data collection were done, where and when possible, at monthly intervals, together with experimental gillnet fishing. This was repeated for three or four positions per sampling site, coinciding with the number and location of the gillnets. For the diurnal sampling, at two dates, dissolved oxygen concentrations, pH, temperature and specific conductivity of the water in the different habitats were recorded according to depth three times a day. Measurements were made at regular one-metre depth intervals from the surface to near the bottom. The first data were collected at sunrise, between 0500 and 0600 hours, then at noon and the final ones late in the afternoon, between 1700 and 1800 hours. The first series of diurnal change data was collected for all habitats late December 1996/early January 1997, and a second series of only the permanently flooded habitats was recorded at the end of May 1997, towards the end of the dry season.
i. Diurnal changes in water quality
ii. Seasonal changes in water quality
The first data were collected for all the habitats at the time when flood levels begin to recede, at the beginning of January 1997. A second series, only of the permanently flooded habitats, was recorded at the end of May 1997, towards the end of the dry season.
Figure 1.5. Diurnal change of dissolved oxygen concentrations, pH, temperature and specific conductivity at different depths for the temporarily aquatic habitats of the floodplain. (The dotted horizontal lines indicate maximum water depth at the time of sampling. Data recorded between 30 December 1996 and 10 January 1997, at the beginning of flood recession).
|
Habitat |
Dissolved oxygen concentration |
pH |
Temperature |
Specific conductivity |
|
scrubland |
|
|
|
|
|
grassland |
|
|
|
|
|
forest |
|
|
|
|
|
lotus field |
|
|
|
|
|
rice field |
|
|
|
|
|
|
¨ sunrise
|
|||
|
Habitat |
Dissolved oxygen concentration |
pH |
Temperature |
Specific conductivity |
|
floodplain pool |
|
|
|
|
|
lake, coastal |
|
|
|
|
|
|
¨ sunrise
|
|||
|
Habitat |
Dissolved oxygen concentration |
pH |
Temperature |
Specific conductivity |
|
floodplain pool |
|
|
|
|
|
lake, coastal |
|
|
|
|
|
lake, pelagic |
|
|
|
|
|
|
¨ sunrise
|
|||
The general pattern of diurnal thermal variation is one of isothermal conditions throughout the water column at dawn, in some cases with a slightly cooler surface region. During the day a temperature gradient builds up, which leads to higher, isothermal conditions in the late afternoon, or to pronounced thermal stratification in places with deeper water. The temperature differences between the surface layer and deeper water are limited, and are less than 4°C at any time in all habitats. The maximum temperatures reached in the shallow habitats at the end of the dry season are considerably higher than those of January. The temperature of water below 3.5 m of depth does not vary much at all. The maximum water temperatures recorded in December-January and May-June were 29.6°C and 33.6°C respectively, the lowest 25.0°C and 26.0°C.
All water samples analysed are neutral to slightly acid. The pH varies usually within half a unit throughout the day. This variation is higher for the two lake habitats (forest and coastal lake), which are also the habitats where the diurnal variation does not occur homogeneously throughout the water column but where there is a considerable drop in pH at larger depths towards the end of the daylight period. The pH reaches more extreme values (between 4.87 and 7.63) in the shallow habitats of the dry season than during high-water time (between 5.55 and 7.04).
Table 1.1. Secchi depths and total water depths for the habitats in December 1996/January 1997. (Secchi disk readings done at midday).
|
Habitat |
Secchi depth (m) |
Total depth (m) |
|
scrubland |
1.20 |
2.30 |
|
grassland |
1.35 |
2.60 |
|
floodplain pool |
1.25 |
3.10 |
|
lotus field |
1.10 |
3.30 |
|
rice field |
0.70 |
1.10 |
|
forest |
0.80 |
3.00 |
|
lake, coastal |
0.90 |
4.10 |
The results of the sampling and data recording for water quality for all eight studied habitat types are presented in the figures below (Fig. 1.9-1.16). All samples were taken around 0530 hours at sunrise. When no values are presented, it means that data are not available. The values shown in the graphs are the averages for three to five measurements that were made at each site for each date and depth. Fig. 1.8 shows as an example the average dissolved oxygen concentrations for the permanent water and the pelagic lake in August 1997 with their standard deviation. This source of variation has been analysed for all data, and, as in case of the example, is in most cases small or nil.
Dissolved oxygen concentrations vary between 0.20 and 7.00 mg.l-1, corresponding at relevant temperatures with 3 percent and 94 percent saturation respectively. All cases in which surface levels of dissolved oxygen of less than 1 mg.l-1 were recorded occurred in August at the onset of the flooding. All but three of the 161 samples from any depth that had oxygen concentrations of less than 1 mg.l-1 were collected in August. All samples with oxygen concentrations of more than 5.5 mg.l-1 were collected in the three permanently aquatic habitats (floodplain pool, coastal and pelagic areas of the lake). Whereas oxygen levels tend to go down towards the end of the dry season in both areas of the lake, the opposite is true in the floodplain pool. When the pool gets flooded with lake water in August, there is a sharp drop in dissolved oxygen levels.
Except for the strong variations in August, the pH levels follow roughly the same pattern as do the concentrations of dissolved oxygen. The extreme values measured were 4.65 and 7.90. The largest difference in pH recorded at any given time between depths in one habitat was 3.07, in the floodplain pool; the smallest (0.76) was found in the rice field. The pH is usually homogenous throughout the water column, except in cases where there is a pronounced gradient or stratification in the dissolved oxygen concentration.
Figure 1.9 Seasonal variations of water quality for the scrubland habitat
Figure 1.10 Seasonal variations of water quality for the grassland habitat
Figure 1.11 Seasonal variations of water quality for the floodplain pool habitat
Figure 1.12 Seasonal variations of water quality for the lotus field habitat
Figure 1.13 Seasonal variations of water quality for the rice field habitat
Figure 1.14 Seasonal variations of water quality for the forest habitat
Figure 1.15 Seasonal variations of water quality for the coastal lake habitat
Figure 1.16 Seasonal variations of water quality for the pelagic lake habitat
There are two peaks in the temperature of the water. The first one coincides with the warm-season months of April and May, the second one with the onset of the flooding in July-August. Except for the floodplain pool and the pelagic lake where the highest temperature is reached in April, maximum water temperatures occur in July and August for all other habitats. The temperature range recorded for all samples is from 25.0°C to 32.0°C. The highest variation throughout the year within any given habitat is 7.0°C (floodplain pool), the smallest 1.5°C, in the rice field. The temperature is almost always homogenous throughout the water column at dawn, the time when the samples were taken.
The specific conductivity (25°C) for all samples varies between 32.0 and 132.2 µS.cm-1. The largest variation throughout the year was registered in the floodplain pool with a difference of 100 µS.cm-1, the smallest variation (50.9 µS.cm-1) was found in the coastal areas of the lake. However, for this habitat, the data for August 1997 are not available; they were probably considerably higher than those recorded in July, as is the case with all other habitats. Like for temperature, there are two peaks in specific conductivity throughout the year. In Fig. 1.17, the data of all the specific conductivity measurements of all habitats and depths are plotted in function of the time.
The Secchi depth in the shallow murky waters of the dry-season habitats was only a couple of centimetres and could therefore not be accurately measured in these conditions. The maximum Secchi depth measured is 1.55 m (Table 1.1).
i. Diurnal changes and patterns
ii. Seasonal changes and patterns - phases of flooding
The results of the limnological sampling show the variation and dynamic character of the parameters that characterize the different habitats. The annually recurring event of flooding creates a very large, dynamic inter-phase between the water and the solid fractions of the ecosystem. The quality of the water of the lake and the floodplain is determined by the processes in the inter-phase and consequently varies considerably, reaching extreme values in transition situations. Water quality is one of the most directly determining factors for the temporal and spatial distribution of aquatic organisms in general, and for fish in particular.
The diurnal sampling demonstrates the influence of the water level on water quality of the different habitats. Places with high water levels show a temperature-driven diurnal stratification like the one that is typical for shallow tropical lakes. Dissolved oxygen and pH levels vary during the day in a way suggesting that photosynthetic activity plays an important part. Where the water depth is not high, the diurnal variation is not sufficient for clear strata to be formed but there are pronounced gradients throughout the water column. In the shallowest habitats, the changes in water quality occur in a homogenous manner throughout the water column. Regardless of the water level, conditions are homogenous in all floodplain habitats at the end of the night. From the limited number of data, it appears that the stratification in the lake itself does not last but is diurnal as well, at least in the volume of water at the time of the sampling (January). The sampling did not take into consideration gradients that might exist between the very edge of the flood zone and the cooler water offshore. Temperature stratification is likely to occur in any of the habitat types, provided that the water depth is sufficient.
The diurnal variations in water quality parameters are larger, and the extreme values more pronounced in the dry season than during the high-water period. Temperatures below 3.5 m do not appear to be affected by the diurnal changes higher up in the water column. This is the normal situation for stratification due to surface heating (Welcomme, 1985). The subsequent mixing is likely mainly due to density differences and wind action. This diurnal cycle has an effect on the availability of nutrients. These may be recycled at high rates due to the high temperature, and are redistributed daily over the entire water column. In such conditions, nutrient limitation of planktonic production is unlikely (Boyd, 1990).
The mineral content, as approached by pH and specific conductivity, does not vary much throughout the day. Nevertheless there is a significant change of specific conductivity between samples from the floodwater recession period (December-January) and the samples taken at the end of the dry season (May-June). The first period corresponds to a low mineral content, with specific conductivity (25°C) as low as 32 µS.cm-1, and the second one to higher dissolved solids levels with specific conductivity up to 92 µS.cm-1. This may be the result of intense mineralizing processes when the terrestrial vegetation gets flooded. The pH variations seem to be mostly the result of short-term biological processes, such as phytoplankton photosynthesis and activity of heterotrophic bacteria in the stratified water column.
The diurnal patterns of water quality variation indicate that dissolved oxygen levels are an important factor in determining the possible distribution of fish in a habitat. Of the parameters studied, dissolved oxygen is the only one to reach values that exclude large groups of fish from these waters. High temperatures also have an excluding role but for a much smaller group of species than the dissolved oxygen.
Apart from the distinctive diurnal variation of the limnological parameters throughout the water column, all the habitats also exhibit equally important changes in water quality in the course of the year. These variations are closely linked to the process of flooding and the associated mechanisms of decomposition and aquatic primary production.
For the habitats in the floodplain that exist as aquatic habitats only during a part of the year, these variations are extreme; for the permanent waters there are seasonal changes that are linked to the flooding but the amplitude of the variations is somewhat limited. At the beginning of the flooding in August, the levels of dissolved oxygen in the pelagic area of the lake are comparable with those at other times of the year and much higher than any of the levels measured in or near the floodplain at that time. This indicates the presence of processes in the floodplain that increase the oxygen consumption and that reduce the aquatic oxygen production, resulting in very low dissolved oxygen concentrations.
The low levels, and for the permanent aquatic habitats the sharp drop in dissolved oxygen levels during the early flooding, even in habitats with little decomposable matter, indicate that the decomposition process is not localized to the origin of the organic matter. The elevated levels of dissolved oxygen in the floodplain pool before the flooding, and the subsequent sharp drop when the pool gets flooded with lake water, are probably due to increased primary production in the pool, linked to the fact that the pool is somewhat protected from wind and wave action and run-off and therefore has a lower turbidity and less aquatic respiration.
In general, and except for the floodplain pools that are surrounded by scrubland, habitats with little organic matter on or near the soil (grassland, rice field, lotus field), show minimum dissolved oxygen levels that are higher than those of the scrubland. The forest is also one of these habitats as most of the degradable biomass of the trees is situated above the level of flooding at such time.
The seasonal variation in specific conductivity shows the result of a combination of two processes: the decomposition and mineralization of organic matter and the influx of silt-laden water from the Mekong at the onset of the flooding (August) resulting in a sharp increase in conductivity. Later in the flood season, the decomposition process slows down but the evaporation during the dry season and subsequent concentration of ions will result in an increase in conductivity. The first rains of June and July cause a slight dilution of the ions, thus reducing conductivity. The effective conductivity is influenced by temperature, and reaches more extreme values.
Even though the flooding itself, i.e. the extension of the lake water over the land, is a rather steady process, its consequences on the quality of the water, both above the freshly inundated parts as in the permanent lake itself, are far from steady and continuous. The progression of the water onto the land causes important and fast changes in the composition and quality of the lake water, having important consequences for the animal life that can potentially survive in it.
From an ecological point of view, several different phases can be distinguished in the flooding process, each showing important differences in potential for sustaining animal life. To illustrate this, the limnology data for the surface samples from the grassland habitat are plotted together on one relative scale showing the variation of these parameters between their extremes during the flood period (July-February) (Fig. 1.18). A better understanding of the processes associated with the flooding and of their consequences is important for a better understanding of the relationships between the new aquatic habitats that are being formed in the flooded area and the animal populations in them. The characteristics of these habitats and their variations are major factors determining the composition and size of the fish populations and their changes over time.
In the case of the Tonle Sap ecosystem, dissolved oxygen concentration is the water quality parameter that has the greatest impact on accessibility of the newly inundated habitats for fish and many other aquatic organisms as benthos and zooplankton.
In general, the flooding process shows four phases, based on the impact of the flooding on the aquatic habitat formation. In so far as the available data allow to conclude, the processes are similar for all the different types of habitat in the floodplain but the extent of the consequences of the flooding varies somewhat for each habitat. There is, however, a notable difference for the floodplain pools, where a phenomenon occurs which is hybrid between what is typical for the floodplain habitats and what takes place in the permanent core area of the lake. The timing of the onset and the duration of the different phases also depend on the position of the habitat in the floodplain and the time of flooding.
The following phases can be distinguished for the floodplain habitats:
I. Initial phase
The water depths in the newly flooded area are rising but still low. Turbidity is very high due to suspended deposits and to the high turbidity of the inundating lake water. This causes the temperature of the shallow water to rise fast and high during the day. Dissolved oxygen concentrations are very low or decreasing rapidly. Conductivity also rises and reaches maximum levels, indicating high levels of minerals.
There is readily decomposable organic matter available in the dead (plant) material layer on top of the soil, as well as on the standing vegetation. The fast depletion of oxygen and the sharply rising conductivity indicate that aerobic decomposition and mineralization is going on at a high rate. Aquatic primary productivity under these conditions is probably zero; terrestrial primary productivity by (partially) flooded plants diminishes or stops. Overall, during this early phase, highly unfavourable conditions for the development of a diverse fauna are formed. The freshly inundated areas are only accessible for air-breathing fish, capable of surviving almost exclusively on atmospheric oxygen (e.g. Anabas testudineus), or using the very thin surface layer that usually is well oxygenated, many of them feeding on exogenous food like flying insects. Towards the end of this first phase, with the diluting effect of the rising water levels, there is a rise in dissolved oxygen concentration and conditions slowly improve.
Figure 1.18.a Phases of flooding. Seasonally aquatic habitats
Figure 1.18.b Phases of flooding. Seasonally floodplain pool.
II. Transitional phase
Water depths are steadily rising and reach maximum levels. Turbidity is diminishing considerably, due to the settlement of larger suspended particles and to the increased water volume. Higher water levels reduce the suspension of soil particles by wave action. Dissolved oxygen levels are further rising and reach their highest concentrations. This is the result of the increased aquatic primary production. The reduced turbidity has created a deeper euphotic zone. The readily decomposable matter has disappeared and the nutrients contained therein have recycled and become available for aquatic primary production. pH reaches maximum values. The combined effect of reduced respiration and increased photosynthesis contributes to higher dissolved oxygen levels. Terrestrial primary production (in the habitats with standing macrophyte vegetation) is further reduced to a minimum due to increased flood stress. Secondary aquatic communities are developing and provide new feeding opportunities. The habitats are now accessible for most fish species but in the early stages of this phase, mostly fishes with supplementary breathing features such as diverticula of the branchial cavity (e.g., Channa striata, Trichogaster trichopterus) are found.
III. Proliferation phase
Water depths have peaked and slowly begin to recede. Turbidity is minimal but still considerable. Water temperature drops and reaches minimum values during these months with also the lowest atmospheric temperatures. Dissolved oxygen levels are still high but tend to become lower towards the end of the phase. Water conductivity drops sharply, reaching minimum levels. This could indicate that the aquatic primary productivity has become nutrient limited, which would explain the falling levels of dissolved oxygen. However, oxygen concentrations are still such that most fish species will not experience these as a stress factor and thus limit their distribution. A wide variety and large biomass of feed communities have developed, contributing to a large number of diverse ecological niches.
IV. Terminal phase
Water levels are strongly going down now, leading in the end to total desiccation of the non-permanent aquatic habitats. Whereas turbidity did hardly change during the first half of the water level drop, it now increases sharply, probably because a critical depth has been reached where wave action kicks up considerable quantities of deposited matter. This probably is also the cause for the conductivity to rise again. Dissolved oxygen levels decrease somewhat further but remain relatively elevated. The shallower waters are more easily heated and the temperature rises again considerably towards the end of the flooding. The receding water levels cause many aquatic organisms to die off, sometimes massively. This is the case with much of the floating aquatic vegetation such as Eichchornia crassipes. Aerobic decomposition levels are increasing as a consequence. The aquatic primary production is reducing due to increased turbidity and adverse conditions for photosynthesis at shallow water depths. Overall, the conditions in the flooded habitats are turning adverse for aquatic fauna and flora, ending in total desiccation and disappearance of the aquatic habitat. The latter is of course not the case for the habitats that exhibit a permanent stand of water. Less favourable water quality is urging many fish species to migrate into the permanent lake water. This is also the time of the massive migrations out of the lake but it is probably not the reduced levels of dissolved oxygen that trigger the migration from the floodplain habitats. These levels remain relatively high until the very end of the flooding. Fish with supplementary breathing features can remain somewhat longer but are eventually forced to migrate or they die off. Some fish are capable of migrating over land.
This succession of events seems to be similar in all the habitats in the inundated area. The extent of oxygen depletion varies somewhat between the habitats but not as much as could be expected based on the potential amount of aerobic decomposition taking place in the habitats. The oxygen depletion in habitats with relatively low levels of macrophyte biomass such as floodplain pools and lotus fields is similar to that in the habitats like scrubland where there are large quantities of decomposable matter. The patchiness of the inundated area and water movements probably cause a spreading of the suspended organic matter.
The last phase of the flooding is somewhat different in the case of the floodplain pool habitat, which remains aquatic (Fig. 1.18). After the sharp drop in water levels and the related phenomena such as reduced dissolved oxygen, high turbidity and rising temperatures, the pool seems to recover from this degradation in water quality during the terminal phase and to develop according to an own dynamic. Turbidity diminishes, temperatures rise high, and dissolved oxygen levels reach maximum values. This creates a microenvironment in the floodplain that offers favourable conditions for many fish species. The pH, however, reaches unfavourable low values. This microenvironment is abruptly disturbed when the pool gets flooded with the rising lake water as the annual cycle resumes.
Apart from the temporal variation in water quality, there also is a spatial one. This is mainly linked to the elevation of the site and therefore to the flood levels. Higher areas with short flood periods are expected to show similar phases in the flooding as lakeshore sites but the level and time at which the processes reach their local maximum vary.
Some of the scheduled sampling could not be carried out for a variety of reasons. Particularly, data that should have been collected together with the experimental gillnet fishing in the forest habitat are missing. The strong wind and wave action caused the experimental gillnets to become displaced and entangled in the vegetation, or made the nets collapse. Accessibility of the sampling sites varied according to the progression and receding of the floodwater. At some points during the transition, especially after the flooding, access to some sampling sites was impossible.
midday
late afternoon ---- water
depth
