CHAPTER 8

MANAGEMENT OF RIVER FISHERIES

EFFECTS OF OTHER USES OF RIVERS AND THEIR BASINS ON FISHERIES

Because the river is only one component of a larger system, the basin, fish communities in rivers are affected not only by events occurring within the channel and its associated waters but are also subject to a range of external influences. Precipitation falling within the basin eventually finds its way into the river by surface and sub-surface flow carrying with it a variety of materials including the topsoil and any contaminants it might contain. Changing conditions in the basin can produce differences in water quality and quantity, as well as in loading with silt and other material which can directly affect channel form. Many such effects are the result of natural variability, particularly climatic shifts but more frequently such changes are the result of some human intervention. In river reaches with floodplains the situation is more complex as the floodplain has a dry season phase during which time terrestrial use is made of space which is covered with water during the flood. Because of the fertility of their soils, and their proximity to water courses for irrigation and transport, floodplains are much in demand for a wide range of activities. They are also regarded as having high development potential and efforts to ‘reclaim’ them by flood control are widespread. This chapter examines some of these effects and also describes the observed impacts on river fish communities of uses of the river and its basin other than fisheries.

CHANGES IN FLOW

Rivers respond to changes in mean annual flow by a period of adjustment to the new regime after which they stabilize in a form adapted to the altered conditions. Clearly reductions in flow will result in the progressive restriction of the stream to a smaller bed within the original channel with a concomitant loss of habitats for fish and other aquatic organisms. Conversely, overall increases in flow will lead to the enlargement of the river channel by erosion of former banks and other features; these, in time, should lead to an increase in habitat diversity through extension of the aquatic system. Biotic sensitivity to changes in water regimes are not limited to alterations in absolute quantity but also to the distribution of flow in time as food peaks can be moved to different times of the year or even suppressed altogether.

The living aquatic organisms in rivers are usually adapted to the particular patterns of flow found there, consequently changes in flow will produce changes in the biotic components of the ecosystem quite apart from those arising from the contraction or extension of the aquatic habitat. Fish communities tend to be either limnophilic or rheophilic, depending on the type of water regime prevalent in the river reach in which they live. Changes in flow will, therefore, tend to favour one or other of these communities, increasing flows leading to a larger presence of rheophilic species, decreasing flows encouraging colonization by limnophilic species. Adequate flows are also essential for the breeding and migration of many fishes. Physiologically fish respond to flood conditions3 by becoming sexually ripe and by movement to breeding grounds. Conversely, should the appropriate flow regime not occur, eggs may be resorbed, as in the case of the Rutilus rutilus, Abramis brama and Cyprinus carpio of the Arakum lakes described by Shikhshabekov (1979). Certain critical levels of flow are also needed to maintain certain types of breeding substrate in a suitable condition for spawning. For this reason, there has been much concern over the determination of minimum flow requirements for fish, particularly in rhithronic areas where high value angling resources coincide with high demands for water. Examples of such research are discussed by Stalnaker (1980) who has identified three main approaches to the general problem of evaluating instream flow requirements: (a) rule of thumb and hydrographic analyses, (b) physical habitat analyses and (c) crop-flow analyses. An example of the rule of thumb approach is the ‘Montana method’ described by Tennant (1976) (Table 8.1) whereby the effect of the various flow regimes relative to the original regime are described.

The above figures have been found to describe the effects of flow reductions on a wide range of North American rivers. Similar series have possible applications elsewhere in the world but, unfortunately, there is insufficient data in the tropics upon which to base such indices.

In Potamon reaches the problem of maintenance of flow is especially critical as the yield of systems with floodplains is closely linked to the extent of flooding. Thus, if at any time the floodplain fails to be inundated recruitment fails in that year for all those species which spawn on the plain. Furthermore, the fertility of the plain during its dry phase is conditional on its being flooded for at least part of the year.

Even if average flows within the system remain unaltered changes in the timing or form of the flood may have grave consequences. In many fish species breeding success depends on a coincidence of characters of which flow is but one. Consequently races or strains of species have adapted to a particular timing in their breeding and displacement of the floods to a different time of year may not permit them to reproduce. There are, however, signs that in some systems fish may adapt to the altered flow regimes following dam construction. In the Volga River, for example, changes in the timing of floods and in temperature of the water led to shifts in the spawning time and to an increase in the duration of the reproductive period in many cyprinid and percid species, even to the extent that a twin-peaked spawning period has appeared following reservoir construction (Kutznetzov and Fedeyev, 1979). Equally, abrupt changes in flow characteristics can influence breeding success adversely. Overly rapid rises and falls in water level can leave nests or spawning grounds dry at critical periods or can result in eggs or fry being washed away. The precipitous decline of the flood can result in fish being trapped in temporary water bodies for lack of time to find passage to the main channel of the river.

CHANGES IN SILT LOAD

Increases in silt load resulting from changes in land or water use accelerate the natural evolutionary processes of the river system, but in doing so cause a number of problems. In the rhithron deposition of fine particles of silt on what is normally a coarse substrate suffocates the rheophilic organisms that normally inhabit such reaches, cutting down on the availability of food. Such choking of the substrate may also render it unsuitable for spawning by those species requiring swift, well aerated flows and clear pebble or gravel bottoms. The silt provides an anchorage for vegetation, blocking low order streams and even diverting them into new courses. Further downstream deposition of silt on levees and on the river bottom may lead to progressive elevation of the whole channel until it stands above the level of the surrounding plain. An extreme flood in a channel so encumbered may cause the river to jump its bed changing its course by some kilometres. Excessive silting of floodplains chokes the standing waters, which disappear faster than new ones can be generated by erosion. Similarly, channels and dead arms are filled and new channels are cut to such an extent that the whole delta of a river may shift along the coastline. At the same time coastal deltaic floodplains grow rapidly, especially at their seaward end, where new land continuously appears.

Heavy silt loads also directly affect living organisms. The poor light penetration into silt laden waters reduces the depth to which phytoplankton can develop and shade out submersed vegetation. Choking of bottom substrates has also been implicated in the disappearance of benthic organisms in the potamon as well as being identified as the cause of mortalities of fish scattering eggs in such areas.

CHANGES IN WATER QUALITY

The succession of physical and chemical conditions in rivers from the headwaters to the mouth may be regarded as a natural eutrophication process. Most downstream reaches are normally enriched, and further loading with nitrogen, phosphorus and organic compounds from agricultural, industrial and urban sources appears acceptable up to a point, and may even be beneficial in initially impoverished systems. If the capacity of the ecosystem to satisfy the BOD is exceeded, however, conditions can deteriorate rapidly. Fish communities inhabiting potamon reaches of rivers are usually adapted to eutrophicated conditions and can support a measure of deoxygenation, although some of the more active species can disappear, Godoy (1975) for instance, traced the disappearance of Triurobrycon lundii from the Mogi Guassu river to deoxygenated conditions produced by such eutrophication. General pollution with toxic substances has not so far been widely reported from tropical systems, although the situation in the rivers of India described by Patil (1977) leaves little room for complacency. Here several rivers are suffering from severe contamination with industrial effluents which have adversely affected their fish stocks. Severe local pollution can occur, however, and is commonly associated with mining where seepage and direct discharge of toxic wastes can create fishless zones in the rivers affected.

In temperate rivers the situation has been more serious and the widespread degradation of the quality of water in many European and North American rivers has been widely described. Even large rivers such as the Rhine or the Vistula have deteriorated to the point where many elements of their fish fauna have been lost. Liebman and ReichenbachKlinke (1967) also recorded bad conditions in certain reaches of the Danube caused by domestic and industrial pollution, but the self-purifying capacity of this large and swift flowing river has been sufficient to keep the main stream at an acceptable quality. Studies on the floodplain, however, indicated that conditions in some standing waters can deteriorate to a point where they damage the fish stock, although in others moderate eutrophication by sewage can raise the productivity.

The effects of pollution on the aquatic life of the system may be summarized as:

(i)    lethal toxicity which kills fish at some stage of its life history. In the case of floodplain rivers this may be indirectly in that the reduction of dissolved oxygen in standing waters of the floodplain and river channel may make them unsuitable for fish that normally live there;

(ii)    sub-lethal effects which are usually difficult to detect or prove but which alter the fish's behaviour in such a manner as to prevent it completing its normal life cycle, or simply to reduce its growth or increase its susceptibility to disease;

(iii)    cumulative effects which can render fish either unsafe or unpalatable for consumption. Most pollution effects tend to be very broad affecting many different species. Whatever their immediate effects, the response at the community level is a reduction in diversity and a shift in species composition towards relatively smaller, shorter-lived forms. In other words they tend to mimic the changes expected from heavy fishing and are therefore apt to reduce the amount of fish available to the fishery. Moderate enrichment with organic substances, on the other hand, can increase the amount of fish (ichthyomass) supported by the system.

INTERACTION WITH OTHER USES

Terrestrial Animals

Wildlife

Many species of game animals move on to the floodplain during the dry season in search of the rich grazing to be found there. Certain species are more specialized to this habit than are others, and some are virtually limited to the floodplains for their distribution. Such species as the Lechwe (Kobus leche kafuensis) migrate from the centre of the plain toward its periphery as the flood rises, and return in the wake of the falling water (Fig. 8.1). Their distribution is, therefore, a mirror image of that of the fish and their dynamics are somewhat similar with a maximum in lambing as the water leaves the plain exposing new pasture (Sayer and Van Lavieran, 1975). Before the construction of the new dams on the Kafue river there was a stable lechwe population of about 94 000 individuals, but changes in the regime following the closing of the Kafue Gorge dam has reduced the amount of grazing available and most probably will cause a drop in the number of individuals supported by this plain. According to Gonzalez-Jimenez (1977) the capybara (Hydrochoerus hyrochaeris) exists in huge numbers on the floodplains of Latin America where it feeds on grasses, and also eats aquatic plants. Unlike the lechwe it swims well and inhabits permanently swampy areas.

Figure 8.1  Distribution of the Kafue lechwe: (A) at maximum flood, April 1971; (B) at low water, 2–14 September 1971. (After Sayer and Van Lavieren, 1975)

Floodplain wildlife is apt to be affected adversely by alterations to their environment in the same way as the fish. As a further example of this, Attwell (1970) noted some effects of the Kariba dam. Here the Mana floodplain is of great importance for the conservation of wildlife in the valley. Since the closure of the dam a lesser area of the plain has been flooded and the remaining portion is under severe pressure of utilization by the larger species of mammals which has produced changes in the vegetation. The impact of overpopulation is reinforced by changes in the type of grasses favoured by the new flood regime, which are tougher. Growth of plants is also less lush as the rich alluvium which used to fertilize the plain is now removed from the Zambezi water before it reaches the plain. By stabilizing the flow the dam has reduced the ecological dynamism of the river whose discharges are wrongly timed, disrupting reproductive patterns of the mammals. A similar theory to this appears in many systems where mammals and waterfowl have suffered by the drying up of the wetlands (Smart, 1976) and will also emerge when we examine the consequences of dam building to the fish.

The presence of wild ungulates and hippopotami on the floodplain has been considered by Kapetsky (1974) to be beneficial to the fish and we have already seen the importance attached by Fittkau (1973) to the crocodile populations of Amazonian rivers for the maintenance of balanced communities. In fact, the amount of nutrients recycled by the terrestrial components of the system are probably very large and must have some effect on its overall productivity. It has been suggested that the tendency for wildlife to disappear in favour of cattle early in the development of the plains, lowers this productivity as the cattle deposits less of their wastes directly into the aquatic system.

Most unmodified floodplains are used as ranges for cattle during the dry season. In certain areas seasonal migrations of the cattle herding peoples dominate the demography. Rzoska (1974 and 1976) has described the way in which the pastoral Nuer and Dinka of the Nile Sudd migrate away from their permanent villages on the higher ground of the swamps following the receding water. As they progress they burn the dead and drying aquatic vegetation to obtain the fresh shoots upon which the cattle feed. At the beginning of the flood they return to their island villages after migrating distances of up to 80 km. Similar movements have been noted from the Central Delta of the Niger (Gallais, 1967; Fig. 8.2) and the Okavango swamps (Stannard, pers. comm.), and are a prominent feature of all African wetlands. Patterns do differ in some plains, for instance, the Oueme, where the characteristic “lagunair” cattle are confined to the levees which are ditched or fenced off from the back swamp depressions. During the flood the cattle are corralled on artificial islands and fed with aquatic vegetation cut from the floating mats on the plain. Extensive cattle ranching is also a feature of Latin America, and at present the vast “llanos” of Venezuela and Colombia, which are drained by the Apure, Arauca and Meta rivers, are used mainly for this purpose. Because the llanos are either submerged by sheet flooding or extremely arid, the Venezuelan ranchers are erecting large crescent shaped dykes called “modulos” across the plain. These trap and retain the water as it retreats towards the main river channels and by evaporation and filtration they slowly dry out leaving a well-watered fringe of vegetation available for the cattle. The “modulos” also tend to trap and retain fish although their potential as fish collecting and rearing devices needs to be explored.

Large numbers of cattle are present on the plains, Rzoska (1974) recorded some 625 000 head as being present in the Sudd. On the Barotse plain there were 310 000 head (FAO/UN, 1969), on the Kafue 250 000 cattle grazed the plain from May to October (FAO/UN, 1968a), on the Shire floodplains 148 000 head are present in the Elephant and Ndinde marshes, and in West Africa on the Central Delta of the Niger, Gallais (1967) estimated some 200 000 head, whereas on the Gambia floodplain there were 300 000 head. In general cattle and fisheries are compatible, even complementary utilizations of the plain. The dung dropped by the cattle, estimated at about 500 kg/ha/yr (Shepherd, 1976) converts much of the dry season primary production into readily dissolved organic and mineral nutrients which have an important impact on the chemistry of the flood waters. Several workers have noted the enrichment of some standing waters of the floodplain by cattle who use the lakes for drinking. The exceptional fertility of the Bangula lagoon on the Shire has been attributed to this cause, and according to Gilmore (1976) lagoons of the Okavango frequented by cattle had more than three times the standing crop of more normal pools, 700 kg/ha as against 200 kg/ha. Some adverse effects of cattle have also been noted. Extensive use of portions of the Oueme floodplain for pasture have resulted in trampling and breaking down of the banks of the drain-in ponds found there, with their later abandonments and filling through siltation (Hurault, 1965). On the flooded banks of small channels and streams, clearance of vegetation for intensive grazing may result in a lessening of cover with a consequent reduction of fish population. Gunderson (1968) cited the case of a Montana stream where ungrazed reaches had 76 percent more cover, and the brown trout living there were more numerous and had about 44 percent more ichthyomass than those then in the reaches where grazing was common. The literature on the interactions between grazing in riparian environments, particularly those associated with smaller water courses, and the quality and quantity of water in United States streams has been surveyed by Platts (1981). It is clear that improper grazing practices, involving high densities of animals over prolonged periods, degrades bottom land and associated water courses. Changes include loss of vegetation cover, instability and erosion of stream banks, increase in silt load and consequently in siltation of channels and a decrease in water quantity. Because of these changing fish populations, in these cases consisting of salmonids, are much higher (by 200–400%) in ungrazed than in grazed reaches of lower order streams (order 6 or less).

Figure 8.2  The seasonal migrations of the pastoral peoples of the Central Delta of the Niger prior to the Sahelian drought. The numbers refer to separate ethnic groupings. (Adapted from Gallais, 1967)

Forestry

Because forests tend to conserve water, topsoil and nutrients, they exert a conservative effect on the aquatic systems draining forested landscapes. The removal of the trees has several readily observable effects, noteably on the water yields in streams, the timing and nature of runoff and the production of silt. Other effects are less immediately discernable but may nevertheless be important and include increased concentrations of nutrients, changes in pH and water temperature and more recently introduction of pollutants into the system.

Deforestation of the catchment area of rivers leads to changes in the flood characteristics whereby flood peaks tend to become higher and shorter as run-off is decanted straight into the channels. In forested slopes much water is retained by the vegetation and also in the top soil. As the top soil disappears there is nothing to delay the water in its move down slope. The faster rise and fall, the more unpredictable spiky flood regimes and the lower dry season flows are detrimental to many species of fish which require a smoother transition from one water phase to another. The lack of top soil and the exposure of the bedrock also tends to lower the amount of nutrients entering solution. The conductivity of the water drops leading eventually to impoverished conditions in the river. Such changes are summarized in Table 8.2.

* Rise in level in mm caused by increased flow of 1 1/m²

The process of logging itself contributes to the degradation of stream quality. With bad logging techniques excessive amounts of waste timber and soil enter the stream causing increases in stream bed load, suspended sediments and dissolved solid concentrations (Graynoth, 1979).

Clearing of the forests allows sunlight to fall on the aquatic system. Thus in rivers in the North Western United States total salmonid standing crops were significantly greater (some 1.5x) in deforested sites than in forested areas, although there were also associated charges in year class strength and species composition of the community. Similarly the more rapid recycling of nutrients through the repeated growth and death of floodplain grasses on floodplains free of trees would suggest higher levels of productivity on savanna plains.

It is now apparent that many of the great open savanna floodplains of the present day were once lined with gallery forests and the plains themselves were covered with scrub forests of the bush savanna type. Clearance of trees for agriculture, grazing and firewood have denuded these and this process is still continuing in some of the Latin American plains. As the process of denudation is historically slow it is difficult to assess the impact of these changes on the fish populations although these have undoubtaedly occurred. It seems probably that the rivers have become less stable with more frequent changes of channel. The flood regimes would also have changed as extensively forested plains tend to retain the flood waters longer. There would be a reduction too in the amount of allochthonous food available to the fish. Nevertheless, most studies seem to indicate that most savanna plains are indeed more productive than those that are forested possibly because they are at a younger stage of their succession. However, deforestation may lead to lowered productivity in some cases such as that which took place in the Grand Lac of the Mekong. Here the surrounding forest was cleared for agriculture and for firewood, and was accompanied by a decline of the fish catch by about half over twenty five years. Sao Leang and Dom Saveun (1955) attributed this to erosion and siltation in the basin and the lessened availability of allochthonous food arising from the reduced area of forests. The silting led to increased turbidity in the lake through resuspension of particles by wave action in the increasingly shallow water body. This in turn led to a drop in primary productivity which was reflected in a lowering of the fish populations. According to the information now available this process was halted and reversed during the late 1970s when political troubles brought about the collapse of population in the Khmer republic. The area around the Grand Lac became reafforested and the fish populations increased again.

Much of the increased silting noted in tropical rivers in recent years has been traced to the denudation of land in the upper reaches of their basins. Here the steeper slope of the land encourages erosion and the top soil is rapidly lost especially where marginal agriculture is practised (Eckholm, 1976).

Reafforestation of hill slopes has been proposed for the control of soil erosion, the protection of rhithronic streams and for the eventual diminution of siltation in the lower courses of rivers. Locally, however, reafforestation programmes may not be universally beneficial, as experience in Scotland have shown. Here plantations of sitka spruce (Picea sitchensis) decreased the pH of streams flowing over quarzite, schist or slate by collection of acid contaminants from rainfall (pH c. 4.3 – 4.5). In such streams salmon eggs die within a few weeks and trout are also absent. In adjacent unforested streams good trout populations exist (Harriman and Morrison, 1982).

Agriculture

The new deposit of alluvial material during the floods each year, a valuable input of organic nitrogen from blue-green algae and a mobilization of phosphorous and postash through the alternation of aerobic and anaerobic conditions (Bramer, 1980) make flood-plains some of the richest of agricultural lands. For this reason, they have attracted man's attention from early in his history as a farmer and some of the earliest civilizations apparently arose in response to the need for communal control of the flooding of the Nile, the Mesopotamean rivers or the rivers of the great plains of China. Where human population density is low, simple culture is common whereby small plots are cleared of the moribund aquatic vegetation as the floods subside. Cereal crops such as maize, sorghum or millet are usually sown and, aided by the high water table, grow to be harvested before the next flood. At this level of exploitation little modification of the plain is needed and the cultivated areas are dispersed among wooded plains. As pressure on the land increases, the floodplain woodland is progressively cleared and the environment further modified by drainage or irrigation. The needs of intensive dry season agriculture lead to the filling in of many floodplain depressions and reclamation of permanent swamps. To assist in this drainage canals are dug which dry out permanently wet areas, but these also hasten the run-off from the plain so as to lengthen the growing season. Irrigation systems often have to be installed in floodplain areas deficient in water or to compensate for water lost through improved drainage. There is thus an increasing tendency to control the flooding of the plain either by poldering, which keeps the water out of some areas, or artificial levee construction which keeps the river channel within its banks and stops the annual flood. To complete the control, upstream flood control dams and reservoirs are also built. Levels of cropping at different intensities of exploitation of bottom lands have been summarized by Thompson (1983) as:

In its simple state agriculture does not conflict with fisheries, and on such rivers as the Oueme in Africa a pattern of intensive use has evolved where both activities are pursued together. On the floodplain the maize fields are interspersed with drain-in ponds, the banks of which are used for various types of market culture, tomatoes, peppers or green vegetables. In addition, the higher levees which follow the river are used for grazing cattle, giving a balanced economy which supports a dense population. Fig. 8.10 illustrates the typical arrangement of these activities according to Hurault (1965) who described the land use patterns of this area in some detail.

Unfortunately as agriculture is intensified beyond this level and pursued as a sole objective, the effects on the aquatic system of the flood control works that become necessary are far reaching. One result, for instance, is that when the floods are controlled depositions of alluvial silt no longer fertilize the soil which can quickly become exhausted. One finds a decrease in productions such as that experienced downstream of the Kainji dam, where within ten years of its closure 50 percent of the land was no longer suitable for agriculture (Adeniji, 1975). To compensate for this applications of fertilizer are needed, putting up costs, and raising the risk of eutrophication and eventually pollution of the water courses. Furthermore, in regions where there is relatively little water and where there are high evaporation rates, salinization of the soil arises either through repeated flooding with salt water or though the drawing to the surface of ground waters whose evaporation leaves the salts in the surface layers of the soil. Should there be insufficient flood water to flush the accumulated salts into the river, the ground rapidly becomes unsuitable for crops and eventually totally barren. As yet little is known of the salinity tolerance of floodplain biota, but the salinization of lagoons definitely inhibits most freshwater forms and, although in regions near the coast colonization with halophytic species of plants may occur, the ecosystem is in some measure simplified.

The hydrological regime of a floodplain environment is well-suited to rice culture. This may be at the primitive level of floating rice such as is practiced in the Central Delta of the Niger, but in Asia, the adaptation of the lowlands of the river systems to intensive rice growing is one of the major features of the landscape. Intensive rice culture requires a complete control over the hydrological regime. Flood control dams, polders and networks of irrigation and drainage ditches are therefore installed for this purpose.

Heckman (1979) has examined in considerable detail the ecology of rice fields in Northern Thailand which become naturally colonized by numerous fish species which contribute to the crops of the area. In fact the capture of fish in rice paddies in Thailand amounts to some 65-80 kg/ha/yr (figures similar to those estimated for natural flood-plains), which means that for the 8 million ha of paddies, the total production could be as much as 580 000 t/yr, most of which is for home consumption. His findings on general ecology of rice fields (Fig. 8.3) may be taken as representative not only of cultivated semi-aquatic ecosystems but also, to a certain degree, of natural floodplains as well. The nutrient flow within the system emphasizes the enormous proliferation in the biomass of rooted aquatic plants relative to other elements of the community. Evidently in the natural system the offtake toward man and domestic animals would be absent and this block of nutrients would be returned to the floodplain by the natural processes of decay.

Figure 8.3  General overview of ecosystem nutrient flow in rice culture: (A) early rainly season (1 May); (B) late rainy season; (C) cool season (16 October); (D) dry monsoon (5 December). (After Heckman, 1979)

Because rice growing and fisheries use the same phase of the hydrological cycle there are sometimes considerable conflicts between them. Fish are frequently accused of destroying troying rice and Matthes (1977) investigated the ways in which the fishes of the Niger river interact with this type of culture. He found that many species were present in the rice fields most of which were juveniles. Certain species were, however, present as adults and these could be classed in three groups:

(i)    fish feeding predominantly on rice, of which four emerged as the most destructive, Alestes dentex, A. baremoze, Distichodus brevipinnis and Tilapia zillii of which the last two nibble through the stalks at mid-height thus cutting down the whole plant;

(ii)   fish feeding occasionally on rice, particularly Brycinus nurse which bites at the leaves of young plants, and Oreochromis niloticus which nibbles epyphytic algae from the stems of the rice plants but also tears away at the stalk;

(iii)   species causing occasional damage through other activities such as Heterotis and Gymnarchus which construct nests of the stems, or Clarias, Heterobranchus and Protopterus which uproot small plants when probing the mud in search of their benthic food. Some species probably benefit the rice by seeking out and eating the stem borer and other insect pests or by cleaning the stems of epiphytic vegetation.

In the Central Delta of the Niger fishermen construct low dykes, 50 cm high, which protect the young rice during the early part of the rising flood when it is most vulnerable to attack and these work reasonably well, as by the time fish can penetrate the field the rice is past its most tender stage. For complete protection, permanent dams with screens are needed to exclude fish from the field entirely, but these all too often succeed only in sealing fish in rather than out.

Rice culture may affect fisheries adversely in two ways. First, the need to control insect pests has encouraged the use of insecticides leading to possible pollution of the water in the rice field and downstream of it. Second, the modifications of the environment associated with intensive rice culture are detrimental to the fish. The fact that rice and fish are not necessarily mutually exclusive is shown by the widespread practice of fish culture in rice fields.

Urbanization

In most parts of the world early human communities were established along the courses of rivers and streams and about 5 000 years b.p. the systematic colonization of the flood-plains of the Nile, Indus, Yangtse, Tigris and Euphrates rivers lead to the growth of man's earliest civilizations. With heightened demand for land brought about by population growth, the pressures for physical occupation of the lower part of water courses has increased despite the risks inherent in the season flooding of such areas. This breeds the need for flood control which in turn encourages further occupancy of the plain to the point where a considerable proportion of the population is located on the 100 year flood-plain. For example, about 16 percent of the urban areas of the United States lie within such areas, and of this over 50 percent is developed for industry or human habitation (Sabol, 1974). Small wonder, therefore, at a certain preoccupation with flooding in these areas as flood losses mount annually despite the best efforts to contain them. An analysis of the world occupancy of floodplains (UN, 1969) summed up as follows: In Western Europe new flood losses are mounting slowly as a result of the intensified use of areas in major cities which are subject to large but unfrequent overflows. In Eastern Europe there is also an increase in flood losses though somewhat slower than in North America. In South and Southeast Asia the intense use of long settled land and the development of new floodplains has more than offset the effects of major river control works. In South America, there is widespread encroachment of urban growth on floodplains. The situation in Africa is similar for, although the older settlements were sited away from the main flood zones, new cities are expanding into these areas. The result of this gradual invasion is a growing tendency to control floods.

The construction of cities tends to produce local disturbances in discharge as the existence of large areas of impermeable surface accelerate run-off in the vicinity of the city, but such effects are still slight in the basin as a whole. Some changes to flood-plain morphology also follow from the communication systems which support the city and its surrounds. Navigation by commercial craft on the river often requires some measure of regularization of flow including the installation of wiers and locks, and the canalization of exceptionally tortuous stretches to make passage easier. Wash from boats erodes banks and accelerates siltation and the destruction of marginal habitats. Roads and railways cut across the floodplains, usually on raised embankments which act as polders or dams to seal off large areas of the plain restricting movements of fish and, even more important, containing the flood waters within a smaller area. Urbanization has two additional effects which also bear on the aquatic environment. One of these, the growing need for power, has similar results to the flood control measures in that it is associated with upstream dam and reservoir construction. The second, loading of the waters with organic and inorganic substances produces pollution and eutrophication in the waters at the level of the city and downstream of it.

Hydraulic Engineering

Control of the flood in rivers is thought necessary for many of the foregoing uses, and is increasingly pursued as a major objective in the development of river basins. Three main types of structure are used either separately or together:

  (i)  dams and impoundments;
 (ii)   levees;
(iii)   canals;

These all have impacts on the riverine and associated ecosystems upstream and downstream of the structures which have caused considerable concern in a number of fora and have found expression in such publications as Ward and Stanford (1979) or Petts (1984). The main effects on fisheries of such structures are briefly reviewed in Table 8.3 and in the following sections.

Dams and impoundments

The storing of water behind a dam so that it may be released more slowly throughout the year is probably the most popular of flood control devices. It has the additional advantage that the water can do work for the generation of power and that it can be used for irrigation or cattle watering. Furthermore, fish can be grown in the sometimes extensive water bodies that are retained. In size, the reservoir can range from vast lakes such as that backed up behind the Akosombo dam on the Volta river, which has a total flooded area of 8 500 km², to small dams which desiccate completely in the dry season. Some major flood rivers such as the Volga, the Missouri and the Columbia have already been converted into a chain of reservoirs, and others such as the Danube, the Indus or the Mekong seem destined for that fate in the near future. The durability of such barrages is questionable as silting is proceding at a considerable rate in many rivers and the smaller dams have a predicted useful life of only a few years before their ability to control floods diminishes progressively. Eventually the reservoir will fail in its purpose, whether it be water storage or flood control. Larger reservoirs, of course, may take much longer to silt up, and may last for over a century but the process is progressing faster than predicted in many areas. As major sites are something of a non-renewable resource the success of a policy of hyraulic control through the use of dams seems somewhat dubious in the long-term. Meanwhile, dams produce their changes on the floodplain and the fish. These are brought about by alternations in the flood regime and by changes in silt loading which in turn alter the dynamics determining the channel shape. This in turn affects the distribution and persistence of vegetation in the downstream stretches of river. For example, below the Volta dam the more stable hydrological conditions and lack of scouring during the flood favoured the rapid development of extensive stands of submersed vegetation such as Potomogeton octandrus and Vallisneria aethiopica (Hall and Pople, 1968).

The flood regime may be suppressed completely, or at least altered in magnitude so that it does not flood such extensive areas. Very large amounts of floodplain have been lost in this manner as a few documented examples will show. In the Missouri, the amount of floodable wetlands over a 145 km sample reach of river dropped from 15 167 ha in 1879 to 7 414 ha in 1967, a loss of 67 percent of the area (Whitley, 1974). The Illinois river, too, lost about half of its floodplain as 80 939 ha of the original 161 874 ha were drained between 1903 and 1920. Subsequently, 3 238 ha have been restored as lakes. Following the closure of a dam on the Peace river, the 2 560 km² of the Peace-Athabaska delta were transformed from a thriving floodplain environment into a series of isolated mud flats (Blench, 1972). Although drainage is not so advanced in other parts of the world the intentions are there. Liepolt (1972) stated that all of the 26 450 km² flood-plain of the Danube will eventually be drained for irrigated agriculture and the original floodable area of the river is already much restricted. In the Mekong the intention is to eliminate flooding from the 49 560 km² delta area and 1 480 km² have already been lost below the Pa Mong dam site. Similarly, in the Senegal river most of the 5 000km² valley floodplain will become dry after the dams which are at present being planned in the head-waters are built. This presumably is going to be the fate of most of the world's great rivers, at least temporarily. Many smaller water courses are being modified in a similar manner, and although this is happening with less attention being drawn to individual cases, the effects of the accumulation of environmental modifications brought about by many small dams may well be impressive. A typical example of such an intervention is the Strydom dam on the Pongolo river, South Africa, which has resulted in the disappearance of 100 km² floodplain and its associated lakes, although in this case there has been an attempt to keep the wetlands intact through controlled discharges (Coke, 1970). In the case of the Mogi Guassu-Rio Grande system in Brazil described by Godoy (1975) there has been a progressive loss of feeding areas on the floodplains of the Rio Grande, by the construction of a series of dams for industrial use.

The loss of floodplain area for feeding and breeding has serious effects on the fish populations, and the reaction of the fish communities of the Chari, Niger and Senegal rivers to flood failures provoked by natural climatic variations such as has occurred in the Sahel also confirm the highly detrimental effects of suppressing the flood. The significance of floodable plains and backwaters as breeding and nursery areas in North American rivers is documented by Guillory (1979) and was early appreciated by Richardson (1921) who found a mean fish yield of 199 kg/ha where 90 percent of the water areas of the Illinois river was in backwaters, 146 kg/ha where 83 percent was backwater and 78 kg/ha where 63 percent was backwater.

In the Missouri river, Whitley (1974) traced the steady decline of catch from 680 t in 1894 to 122 t in 1963 mainly to the loss of fish habitats following the construction of the reservoirs. On this river, too, there have been modifications of the main channel by dykes on side channels and backwaters have been blocked. The river has been trained by reveted banks into a series of bends. Prior to 1900 Ictalurus punctatus and Ictiobus spp. made up the major part of the catch, but since then the introduced common carp Cyprinus carpio has become dominant.

Widespread changes occurred in the Colorado River following its conversion into a cascade system (Holden and Stalmaker, 1975; Holden, 1979). A combination of hindrance to migration, reduction in river flow and lowered mean water temperatures seriously reduced populations of endemic cyprinid species such as the Colorado squawfish (Ptychocheilus lucius) hump back chub (Gila cypha), bonytail chub (Gila elegans) and the razorback sucker (Xyrauchen tecarus) all of which are obligate riverine species. These have to a large measure been replaced by species exotic to the Colorado basin such as Micropterus salmoides, Lepomis cyanellus and Notropsis lutrensis which are more tolerant of lentic conditions. A similar replacement of endemic fish species by introduced species which are better adapted to the changed hydrological conditions has also been described from Soviet rivers. In the Volga Chikova (1974) has noted changes in species composition below the V.I Lenin Volga Hydroelectric station (Kuibyshev reservoir) where the phytophilous fishes, Abramis brama, A. ballerus and Rutilus rutilus decreased in number, whereas Stizostedion (=Lucioperca) lucioperca and “Sichel” (Pelecus cultratus ?) increased. Eliseev and Chikova (1968) concluded that the decline of the phytophilous species is due to the failure and unpredictability of the flood, which when it does inundate the plain, does so in a sporadic manner which arbitrarily strands young fish and spawn in isolated pools. Main stream spawners such as the various sturgeons, Stizostedion and “Sichel” are not so affected. Furthermore the flood peak of the new regime is now somewhat retarded favouring the later spawning species which generally belong to the second group. Long distance migrant species also declined in the Volga river following its conversion into a cascade reservoir system. Such species as Huso huso, Caspialosa kessleri or Stenodus leucicithys disappeared from all but the lowest reaches of the system whereas the regulated flows and reservoirs favoured such lacustrine species as Clupeonella delicatula which populated large areas (Poddubnyi, 1979). In the Volga delta the reduction of duration of flooding of the floodplain water bodies reduced the time available for their occupation by young fish from 50–70 days before the closure of the dams (1950s) to only 10–15 days afterwards (1960s–70s). This meant that the downstream migration of the young has been advanced with less time for growth and survival on the plain. As a result there has been a great reduction in year class strength (Koblitsbeya, 1985). At least 40 days of regular annual flood are judged necessary for the maintenance of an adequate fish community in the lower reaches of this river. According to Lelek and El Zarka (1973) and Adeniji (1975) the changes in the fish fauna of the Niger river below the Kainji dam are also traceable to the unpredictable nature of the flood. Catch in the reach between Jebba and Lokoja fell by about 50 percent in three years from about 4 400 t/yr (1967–69) (Otobo, 1978). This was accompanied by changes in species composition whereby the Characidae, Mormyridae and Clariidae declined but predatory species such as Lates and Bagridae increased in abundance (Sagua, 1978). Changes were also felt further downstream where the Lower Anambra basin fisheries declined by about 60 percent (ie. 4 000 t) because of desiccation of much of the floodplain following the closure of the dam (Awachie, 1978; 1979).

A localized increase in the abundance of fish, particularly of predatory species, immediately below dams has been noted from many rivers. In the Nile below the Owen falls dam populations of Barbus altianalis and the recently introduced Lates niloticus were particularly abundant. In the Niger, as we have seen, Lates also appeared in quantity especially in the reach immediately below the Kainji dam (Kainji to Jebba), although more recently it declined in abundance (Sagua, 1978) in favour of various species of Mochokidae, Characidae and Cyprinidae. Chikova also commented on the maximum accumulation of fish in the Volga as occurring in the same regions downstream of the dams. Whitley (1974) explained this locally increased production in terms of the enriched water from the reservoir which, in passing through the sluices, carries with it zooplankton, insects and fish. This enrichment does not persist for any great distance, however, and in the Missouri was only detectable for about 2 km downstream. Such concentrations are probably associated primarily with feeding, but fish also accumulate below dams where they are interrupted in their migrations and form the basis of rich tailrace fisheries such as that described by Otobo (1978) for the Kainji dam.

Fish may be entrained by the flow out of dams either through the sluices or through the associated power stations. This usually results in their death by gas disease caused by the explosive decompression from high pressure upstream to low pressure downstream of the dam.

A history of diminishing catches in the Columbia river from a peak production of 22 440 t in 1911 to only 6 800 t at present, is mainly due to the blocking of passage to migrating fish and changes in the flow characteristics of the river, rather than to the loss of floodplain area (Trefethen, 1972). In the Murray river, Australia, the 800 km migrations of the golden perch (Plectroplites ambiguus) have also been stopped by a combination of water management practices, including flood control weirs, which have left few of the original characteristics of the uncontrolled river system unchanged (Butcher, 1967). The Sakkur dam of the Indus did not affect populations of the migratory Hilsa ilisha, but the construction of the Gulam Mahommed dam further downstream of it deprived the fish of 60 percent of their previous spawning areas. Further upstream in the same river the Tarbela dam poses a threat to Tor putitora, a large cyprinid which migrates from the floodplain to the foothills of the Himalayas to breed. Hilsa stocks have become depleted in several other rivers of the Indian sub-continent. In the Godavari river a combination of upstream dams and silting through erosion are restricting the species increasingly to the estuarine delta (Nagaraja Rao and Rajalakshmi, 1976); in the Cauvery H. ilisha has ceased to ascend the river following the regulation of flow by the Stanley reservoir (Sreenivasan, 1977). Following the construction of the reservoir, Puntius dubius has disappeared from the Cauvery and Tor khudree is now much more restricted in its distribution. Species have also declined in abundance in Brazilian rivers where dam construction has been intensive in recent years. Many of the “Piracema” species have largely disappeared from the Rio Grande and its affluents in the Sao Paulo state, and Pseudoplaytstoma coruscans was eliminated from the Tiete river soon after the closure of a series of barrages there. In Venezuela the construction of the Guri dam on the Caroni River impeded the upstream migration of many species of Pimelodid catfishes as well as the prochilodontid, Prochilodus mariae and Semaprochilodus spp.

The importance of changes in silt and nutrient discharge from rivers on larger aquatic systems following damming has been noted from several areas. Ryder (1978) traces the decline of the East Mediterranean pelagic fishery to the withholding of nutrients in the Nile within Lake Nasser which in turn has led to some increase in the yield of the lake (Fig. 8.4). Increases also occurred in Lake Manyala in the Nile Delta where the reduction in flow caused by the Aswan High Dam led to the silting of the channels connecting the lake to the sea. The freshwater environment then created with the eutrophicating conditions originating from organic pollution has in recent years (1962–1973) supported a greater fish population than previously (1908–1935) (Shaheen and Youssef, 1979). A disastrous decline has been noted in fish production in the Black Sea and Sea of Azov since the main inflowing rivers, the Danube, Dnieper and Dniester have been subject to increasing control in the form of cascade systems (Tolmazin, 1979). This has led to a reduction in the inflow of nutrients which used to support large stocks of fish and shellfish in the estuarine seas. It has also prevented the discharge of freshwater which formed a well oxygenated layer over the deeper, saline and largely anoxic waters. A similar decline in water quality and fisheries of the Caspian followed the construction of the dams of the Volga River. Berdicevsky (1975) (in Carre, 1978) showed that while the overall catch remained the same a change occurred in community structure away from higher valued species to small pelagics of less commercial interest.

Figure 8.4  Regression of the annual fish catch from Lake Nasser on the annual yield of the Eastern Mediterranean fishery four years earlier. (After Ryder and Henderson, 1975)

The importance of river inputs into larger aquatic systems is also demonstrated by the influence of river flow on pelagic fish catches in Lake Kariba. Here highly significant relationships between the wet season flow of the nutrient rich tributaries (Gwaai and Sanyati rivers) and the catch in the subsequent year were established by Marshall (1982). Variations in the flow of the larger but nutrient poor Zambezi did not affect catches to the same extent. Although these relationships originated in a reservoir, the great flow to volume ratio makes nutrient conditions more closely resemble a river. Thus by extension fluctuations in discharge of nutrient rich tributaries are also liable to influence the productivity of larger but nutrient poor major rivers. This indicates the possible dangers of even comparatively small tributary dams for lowering the overall productivity of river, lake or even marine ecosystems.

Levees

The man-made levee is a type of linear dam which heightens the natural levee upwards to prevent water spreading laterally on to the plain. The main impact of this is of course to deny fish access to the feeding and breeding grounds that are necessary for their survival. Catch and specific diversity drop in much the same way as they do for dams. As the water can no longer spread over such large areas, sediment and pollution are concentrated into the river and such standing waters as remain, much as described for the Illinois river by Starret (1972) (Fig. 8.5). Here the lack of spawning grounds caused species such as pike (Esox lucius), large-mouth bass (Micropterus salmoides) and yellow perch (Perca flavescens) to diminish in abundance compared to new species such as Cyprinus carpio which were introduced into the system. The commercial catch declined from 2 613 t in 1950 to 182 in 1973, and angling success showed a similar reduction; effects which were attributed directly to loss of habitat and siltation following the leveeing of the river and the draining of the bottom lands (Sparks and Starrett, 1975).

Figure 8.5  Schematic drawing demonstrating the impact man has had during the past century on the ecology of the Illinois river and two of its adjoining bottom-land lakes near Havana. (After Starret, 1972)

The flow during the floods is also increased and as a result fish can be swept out of protected positions into unfavourable environments. The construction of levees in low lying areas of the Danube delta has increased the rate of descent of the larvae of many of the migratory fishes. Instead of spreading out over the floodplain, to grow in the standing waters, these are ejected into the sea. As a result catches of such species as Alosa pontica, Aspius aspius, Tinca tinca, Blicca bjoerkna and Abramis brama, are much reduced relative to other arms of the river where no levees have yet been constructed (Zambriborsch and Nguen Tan Chin, 1973). In the Middle Danube the Yugoslavian fishery yields have declined from 20 000 t/yr to 1 200 t/yr following the enclosure of the flood-plain zones by levees. A similar loss in fish production from Rumanian tributaries of the Danube has also been noted by Bacalbasa following the reclamation of floodable bottom land for agriculture. Here not only has the catch declined but the remaining population has come to be dominated by Carassius auratus gibelio which is useless to the fishery (Bacalbasa and Popa, 1978). Similar losses are described by Butcher (1967) from the Murray river, and pose a potential source of concern when levees are to be constructed on floodplains supporting “Subienda” or “Piracema” types of migrants. Effects comparable to those produced by levees can be anticipated from polders which essentially enclose areas of floodplain within dykes to prevent them from uncontrolled flooding. In this way many potential breeding areas may be removed from the ecosystem especially when the poldered area is of any great extent.

Standing stocks of fish in the poldered Northern portion of the Atchafalaya basin, a distributory of the Mississippi, has a lower standing stock at 555 kg/hr than the less controlled southern basin with 860 kg/hr. Land clearance and agricultural runoff leading to eutrophication are also implicated in the lowered yield from the North whereas, in the South, the original allochthonous detrital based food chains in the annually inundated habitats have been maintained (Bryan and Sabins, 1979).

Gaygalas and Blatneve (1971), traced the poor recruitment of bream, Abramis brama to the poor spawning conditions in the 23 000 ha² of summer polders, which were constructed to reclaim the floodplain depressions of the Nyamunas river delta. Summer polders are normally inundated during the winter floods but are emptied during the spring. Overly rapid drainage does not allow the fish to breed successfully, whereas with correct management, a slight prolongation of the flooded phase improved recruitment in other years, and large fish stocks survived in the canals within the enclosed area. Polders can also be flooded in a controlled manner for rice culture, forms of irrigated agriculture or, as in Venezuela, for grazing reserves for cattle. Such practices can be combined with extensive aquaculture. Where polders occupy a large proportion of the plain, the reduction in area available for the expansion of the floodwaters can result in the acceleration of flow which can alter conditions on the remaining flooded areas in a way unfavourable to fish.

Channelization

The channel is frequently used as a system for simplifying the natural complexity of a river. The banks are smoothed and straightened in the interest of the more rapid and contracted evacuation of water. Whilst channelized reaches of river may protect a particular area their net effect is to increase the energy content of the water making it potentially more troublesome downstream. Flow patterns are concentrated in time, producing shorter, higher spates. The rivers themselves may be shortened to some degree by the evening out of meander bends; about 120 km have been subtracted from the Missouri river between 1890 and 1947 in this way and the surface area of water has been reduced from 49 250 ha in 1879 to 24 645 ha in 1972. Trees have largely been eliminated and the flood-plain forest coverage has declined from 76% (1826) to 13% (1972). By contrast the percentage of land under cultivation has increased from 18 to 83 over the same period (Burke and Robinson, 1979). Channelization has much the same effects on species composition and abundance of fish and other organisms as other flood control measures. In the United States workers such as Congdon (1973) and Adkins and Bowman (1976) have found that the channelization of streams considerably reduces both species diversity and biomass when compared with unchannelized stretches of the same river, and Gorman and Karr (1978) also found that the general reduction in habitat diversity following disturbances to the ecosystem such as channelization produced a corresponding reduction in both diversity and stability of the fish community. As examples of this effect Groen and Schmulbach (1978) reported that fish catch per mile was 2 to 2.5 times greater in the unchannelized reaches than in channelized reaches of the Missouri. In channelized reaches many of the larger species such as lake sturgeon (Acipenser fulvescens) paddlefish, (Polyodon spathula) and blue catfish (Ictalurus furcatus) have almost been eliminated (Burke and Robinson, 1979) and catch has fallen by 80% between 1897 and 1963. Tarplee et al. (1971) found an average ichthyomass of 174.0 kg/ha in unchannelized and 55.3 kg/ha in channelized streams in N. Carolina. Jahn and Trefethen (1973) found 632.8 kg/ha of fish in unchannelized, 502.9 kg/ha in lightly channelized and 146.7 kg/ha in heavily channelized reaches of the Blackwater river, Missouri where the mean size of the fish was also much reduced in channelized reaches. Many organisms other than fish are affected by channelization and Bulkley et al. (1976) and Griswold et al. (1978) among others, have shown the bearths and the drift organisms which constitute the bulk of fish food are also mcuh reduced by the siltation of the bottom in channelized reaches. Further references describing the reduction of aquatic fauna in U.S. rivers due to this type of management were listed by Schneberger and Funk (1971) and Schoof (1980). Schoof found that in addition to the effects on the biotic elements of the system channelization can affect the physical environment by draining wetlands, cutting off oxbow lakes and meanders, reducing ground water levels and the recharge of ground waters from stream flow and by increasing sedimentation rates. Channels can also contribute to the diminution of flooded area by deflecting the flow from one river to another, as in the case of water transfers, or by bypassing major floodplain areas. The greatest projected work of this nature is probably the Jonglei Canal, which if completed will lead much of the Nile water past the Sudd to decant it back into the main river downstream of the swamp at Malakal. This will make much more water available for irrigation in the arid Nile basin, but the estimated effects of the altered regime within the Sudd will result in a reduction of maximum flooded area of between 26 and 36% depending on the water regime adopted (Nefit-Babtie, 1984).

An early phase in the channelization process is the clearance of streams of snags, logs, boulders and other large debris. This effectively reduces habitat diversity and has been responsible for a great loss of fish. Even in torrential streams from the North-West United States floodplain habitats resulting from such debris represented 6 – 29% of the summer rearing habitat and 55 – 70% of the standing crop of coho salmon juveniles. As a result large log-jammed side channels had eight times the density of coho than side channels lacking such features (Sedell, Yuska and Speaker, 1983).

Not all changes associated with channelization are detrimental to fish stocks. For example, Hesse et al. (1982) noted that the rock rip-rap, wing dykes and stone revetments used to contain the flow enormously increase the area available for periphyton. This offsets the lack of benthic food organisms present in the accelerated flow of the river.

DEVELOPMENT AND MANAGEMENT OF RIVER SYSTEMS FOR FISHERIES

OBJECTIVES AND STRATEGY

Development and management are here considered as those intentional activities which are directed at the manipulation of living aquatic resources, the environment in which they live and the human communities associated with them. The nature of these activities depends on a variety of factors among which are the purpose for which the resource is to be utilised, the objectives of management, the type of fish community and the developmental context within which management is pursued. Strategies for effective implementation of management plans also have to be conceived at different levels. At the level of the basin, priorities must be set among a number of competing uses of which fisheries is but one. At the level of the fishery the resource may have to be allocated among several interested groups and direct management techniques applied to control the type and amount of fish being caught. At the level of the fish community certain management measures can be employed to improve the productivity of the fishery. Less directly fisheries may be managed by inducing social changes or by the development of new types of fishery.

Objectives

Objectives, as opposed to uses, are here considered as those socio-economic criteria which underlie the planning process. The first level of decision is the assignment of the resource among several possible alternative uses. Such decisions usually arise from social or economic pressures. For example, whereas production of fish for food may have high priority throughout the developing world, the earning of foreign currency through export of luxury items may be judged more important. Also in industrialised countries the commercial fisheries of the past have largely been supplanted by sporting interests. Questions of access dominate recreational fisheries, and whether the resource is open to all-comers or restricted to a certain group is a matter of deep concern to the recreational fishery manager. Several main objectives consistantly recur in development and management plans for food fisheries; first the production of the maximum quantity of food at an appropriate price, second the maintenance of the quality of the fish caught, third the provision of employment and fourth the improvement of the standard of living of the fishing community. Such objectives are frequently in conflict and the attainment of one in its entirety will impede the realization of the others. For example, production of the maximum amount of cheap fish protein can only be at the cost of the quality of the fish caught and usually of the well being of the fishing community. Alternatively, regulating access to the fishery reduces employment opportunities but increases economic returns from the fishery. Decision makers should, therefore, give same weighting to the various objectives to allow for a more realistic orientation of management plans.

One of the commonest concepts for the regulation of fisheries is that of maximum sustainable yield (MSY), which predicts that any fish stock has a constant surplus production which may be removed by a fishery each year. The concept is applicable mainly to single stocks whose abundance is relatively unaffected by changes in the environment. Few river fisheries are typical of this situation, and because the concept introduces severe economic weaknesses in that to reach MSY costs incurred exceed the revenues added by increasing effort. MSY has been also somewhat discredited in recent years from the population dynamics standpoint. The hunt for a successor concept has proved unrewarding and MSY is still cited in a loose way as representing the general productive capacity of a fishery. The concept is particularly inappropriate to those river fisheries which are based on multi-species communities and here the type of model described is more applicable. As an illustration of the interaction between the objectives selected and planning, three points may be considered as critical on the yield curve in this figure.

A.   Represents a strategy based upon the management of the fishery for a few high valued species only. It implies some loss in total potential catch and a great limitation in employment opportunities but is often coincident with the maximum value to individual fishermen who maintain an elevated catch per unit effort of a highly priced product. The fishery must be highly regulated to remain in this state for an extended period.

B.   Represents a compromise at which catch is maximal while conserving acceptable quality. Individual welfare is high but access to the fishery must be restrained. In practise a fishery generally fails to remain long at this point as social pressures tend to encourage drift towards the right hand end of the curve.

C.   Represents the other extreme. It provides maximum employment and a near maximum yield of protein but a poor quality product and low level of individual welfare. Given the nature of the catch per unit effort curve, however, where decreases in catch per unit effort decline and effort intensifies, the welfare of the individual fisherman may not fall excessively when compared with point D for instance. This strategy accepts the risks inherent in pushing the fish community near to its limits of tolerance whereby any further pressure, environmental disturbance or climatic change can readily destroy the fragile equilibrium and bring about the collapse of the community.

Because it is apparently more difficult for the fishery to move from the right-hand side of the model to the left for socio-economic and biological reasons, the early setting of objectives is highly desirable to avoid arriving inadvertently at a position where options are limited. Furthermore, because impacts of human induced changes in the ecosystem tend to accelerate the fishing-up process, anticipated changes in water quality or quantity should also be taken into account when setting objectives.

Uses of Fish Resources

The various species of fish that are found in rivers can be used for a variety of purposes.

(a) Food:   This is by far the most widespread and important use of fish resources, especially in tropical and subtropical areas far removed from the coast where freshwater fish often contribute a significant proportion of the total animal protein in the diet. Fish for food may be either captured as adults for direct consumption, or may be caught while still in the juvenile or fry stages for stocking into other bodies of water or for rearing in ponds or cages through a variety of aquaculture techniques. Most of the world's rivers contain stocks of fish that are used to a greater or lesser extent for food but such fisheries are usually diffuse with relatively low yields at any one point. However, most of the great lateral expansion areas, where extensive floodplains exist, support intensive fisheries which are among the most productive in the world. The location of most of these are indicated in Figs. 1.13–1.19. It is difficult to separate the production obtained from rivers from that of lakes and intensive aquaculture in ponds in any given geographical region. It seems fair, however, to assume that about half of the world's inland fisheries production of 8.6 million tons (1983 estimate) comes from running waters, their floodplains and from aquaculture associated with lands inundated for rice culture or other purposes. Certainly in Africa just under half of the 1.4 million tons is attributed to these sources (Welcomme, 1979) and in S. America most of the 323 000 t comes from rivers as there are very few lakes there. By far the greatest proportion of inland fish is derived from Asia, and in particular China, where various types of extensive aquaculture are common, but river fisheries are also pursued. Fish rearing in floodplain rice paddies is widespread throughout Southeast Asia and the production from major rivers such as the Ganges, Brahmaputra, Mekong and Indus alone combine to give some 1.25 million tons.

(b) Sport:   Recreational fishing is particularly common in the industrialised countries of the temperate zones although this type of utilization is now spreading to other areas of the world where there is heavy industrialization. It is often difficult to define the limits between recreational and food fisheries as most fish caught for sport is also eaten, thus providing a valuable addition to diet. Furthermore there is sometimes a recreational element in subsistence fisheries even in non-industrial economies. Sport and commercial fisheries often compete for the same resource and, because the value of the same fish is considerably greater to the sport fishery than to the commercial fishery, there is a trend to favour the recreational interest with subsequent declines in the commercial catch.

(c) Ornament:   The beauty and the interesting behaviour of many of the species of riverine fishes, particularly those of the tropics, together with their small size has favoured their use for ornament or hobby. Several hundred species are regularly captured or cultured for export providing a small but significant income to some river communities. The magnitude of the world trade in aquarium fish has been described by Conroy (1975) and I.T.C.(1979). From these it emerges that large numbers are caught to compensate for wastage through disease and transport mortalities. Colombia, for instance, exported some 10 million fish in 1974, Peru 15 million, Brazil 3.5 million, Venezuela 10 million - all from the Amazon and Orinoco basins. As mortalities between capture and export range between 50 and 70 percent, this represents an offtake of at least 70 million fish, most of which are withdrawn from such restricted environments as floodplain pools or small feeder streams. Local disappearances of species due to overexploitation have already been reported from some countries. To compensate for this, and to reduce the costs of capture and collection from the wild, many of these species are now cultured in suitable warm waters. The transfer and spread of species by escape from such farms has resulted in several of the commoner forms becoming pan-tropical in distribution.

(d) Other uses:   There are a number of minor uses for riverine fish species including disease vector control, where fish such as Gambusia or Lebistes have been introduced to control mosquito larvae. Other species noteably Ctenopharyngodon idella have been transferred widely for the control of undesirable aquatic vegetation.

Developmental Stage

Four main stages can be identified in the modification of river systems and their fisheries which call for different approaches to research and management (Table 8.4) and three cutting points seem significant between one developmental stage and the next. The first of these occurs with the introduction of systematic agriculture to the riparian area which marks the end of the unmodified stage. Agriculture demands at least some attempt at drainage and regulation of floodplains and also may involve some abstraction or diversion of water courses. As the support capacity of the basin rises population increases and fishing pressure rises to the point where it produces a noticeable effect on the fish stock. The second cutting point seems to arrive with the installation of flood control structures which begin to modify the flood, impound water in lower order streams and restrict the flood area available to the fishery. At this stage exploitation of the wild stock is very intensive and its effects, added to the stresses induced by environmental modifications, can bring about rapid changes in species composition and abundance. The last cutting point is reached with the installation of one or more barrages along the course of the river. These change the nature of the downstream reaches from flood rivers to reservoir rivers with the consequent loss of much of the previously flooded area. The four stages listed do not necessarily represent an evolutionary sequence although most rivers in Europe and North America have passed successively through these steps to reach their presently highly controlled state. Similar fates are projected for many tropical rivers, but other rivers will inevitably stay at a slightly or extensively modified stage without ever becoming completely controlled. Nor does the developmental process necessarily proceed at an even pace along the whole length of the river. Indeed in many large river systems, such as the Danube or the Mekong, different reaches and tributaries may be at different developmental stages.

The four stages proposed correspond approximately to those presented by Yon and Tendron (1981) (Figure 3.7) for European rivers. A similar series of categories has been proposed by Scudder and Conelly (1984) on the basis of socio-economic criteria which to a certain extent parallels the above classification. Their four categories are:

(a)    Primarily subsistence; in which fishing methods are limited to traditional, locally manufactured gears and both men and women participate in the fishery. The catch is consumed entirely within the boundaries of the fishing community.

(b)    Incipient commercialisation; where fishing methods are still diversified but more modern twines and gear are used. Division of labour is more apparent with the men fishing and the women carrying out support activities. The fishery usually produces sufficient surplus for part of the catch to be sold outside the fishing community.

(c)    Primarily commercial; gill nets and seines are now the most common gear and the traditional methods have largely disappeared. The fishery is pursued almost entirely for markets external to the fishing region.

(d)    Increasing marginalization of traditional communities; a period marked by a decrease in total catch and decline in the fishery corresponding to a depletion in the resource.

MANAGEMENT OF THE RIVERINE ENVIRONMENT

There is an increasing trend at the present time to attempt to master rivers and to control their floods so as to make their basins more productive for agriculture and safer for human occupation. Whatever the long-term benefit and success of these attempts they do modify the aquatic environment in such a manner that the fish stocks are adversely affected. If rivers are to continue to produce fish, it must be within an entirely different framework from the capture fishery as practised in unmodified rivers.

Preservation of the Natural System

The biology and ecology of the majority of fishes inhabiting rivers show them to be extremely sensitive to modifications of the flood cycle and to the environmental changes that occur in modified rivers. In order to retain something of the natural diversity of the stocks of fish certain areas have to be maintained in their wild state and it is hoped that the setting aside of land or some river systems for suitable reserves be considered in the planning of land use within a river basin. As has been discussed various types of agricultural engineering and construction works also interfere with the natural pattern of flooding. Thus management policies involve planning of land use within the river basin as a whole. Levees, polders and embankments for transport systems prevent the water reaching floodplain depressions, and the remaining water is restricted within main channels in such a way as to modify its flow. An essential part of planning for fisheries is, therefore, to make sure that any lateral expansion zones which are designated as reserve areas should be kept free of such structures. A second important feature is the provision of sufficient water to produce a flood which will have the characteristics needed for the reproduction of fish species. This is both a question of quantity and timing which have to be based on an adequate knowledge of the biology of the fish species concerned. Controlled releases of water have been tried in some systems, and on the Pongola river, a series of experimental floods have successfully filled the lagoons and induced breeding in the species inhabiting them (Coke and Pott, 1970; Phelines, Coke and Nicol, 1973). Similarly, sufficient water has been released from the Shire river dam to fill the lagoons of the Elephant and Ndinde marshes, with satisfactory results for the fisheries. Kenmuir (1976) showed that the juveniles of several species of fish, including Clarias gariepinus, Brycinus imberi, B. macrolepidotus, Labeo altivelis, L. cylindricus, Eutropius depressirostris, Synodontis zambezensis, Oreochromis mortimeri and Tilapia rendalli appeared in lagoons of the Mana floodplain of the Zambezi river following flood releases from the Kariba reservoir. Furthermore sexually active adults of these species were observed moving on to the plain during flood releases that corresponded to the time when the Zambezi would normally flood. He therefore concluded that artificial releases of water (simulated floods) can stimulate breeding in riverine fishes provided that they occur during the previously normal flood period.

The correlations between drawdown factors and catch detected in some floodplain rivers, as well as the theoretical predictions of Welcomme and Hagborg's (1977) simulation indicate that standing stock and yield do not depend solely on the flooded area, but also vary according to the amount of water retained in the system during the dry season. Because of this, the shortfall in catch following a reduction in flooded area can to a certain extent be compensated for by increasing the area of residual water, provided the reduced floods still occur at the appropriate time. Where water management actions are being planned which will result in the loss of areas of floodplain the provision of large permanent bodies of water may be considered at an early stage. In all probability fish community of systems managed in this way will undergo changes in species composition probably in favour of the blackfishes much as in ordinary reservoirs.

The management of the natural system should not be restricted to the floodplain, but should be extended upstream and downstream as part of the manipulation of the whole basin. As we have seen, activities in the headwaters of the basin can alter the siltation, runoff and flow patterns of the potamon reach, often with far reaching consequences to the fishery. Equally, pollution or damming downstream can prevent species from moving up to the floodplains to feed or breed.

Instream Improvement Structures

A variety of structures can be used to restore more natural patterns where they have been destroyed or to stabilize river channels whilst maintaining the diversity of habitats therein. Although such structures tend to diminish the effectiveness of channels modified for drainage, irrigation, floodcontrol, etc. they can be used in a controlled fashion to improve the biological productivity of the systems concerned.

According to Swales and O'Hara (1980) three main categories exist:

(a)  structures which impound or modify flow;
(b)  structures which provide direct cover;
(c)  improvement of spawning areas.

Structures which impound or modify flow

Traditionally groynes or wing deflectors have been used for many years to improve salmonid streams and have been proved effective in introducing artificial sinuosity into a straightened channel. They also create slacks in the flow shadow immediately downstream and by scouring and deposition help reestablish pool riffle sequences along the channel. As a pool riffle sequence characteristically recurrs at intervals of five to seven channel widths such devices should be installed with this spacing between them. Low dams and nets are even more effective in the restructuring of the bed of the river to create pool-riffle sequences. Both these types of device should be accompanied by bank stabilization to prevent selective erosion by the deflected current. The use of structure of this type is not limited to the rhithronic reaches of rivers. Burke and Robinson (1979) and Schick et al. (1982) describe a number of structures which have been used to restore habitat diversity to channelized potamonic sections of the Missouri and Mississippi rivers. These include (i) notched dykes, where portions of transverse barriers are lower than the rest of the structure. These allow water to flow through the barrier so as to develop or preserve side channels or to prevent further land accretion; (ii) Low elevation structures, which are below the water surface for much of the time and which prevent island formation that can support vegetation. They also lead to diversified depths particularly in that a deep pool develops downstream of the structure; (iii) rootless structures are transverse dykes which do not abut on to the back, a varient, vane dykes, are angled relative to the current. Such structures encourage multiple sub-channel formation and create more bank area; (iv) chute closure structures are placed in such a way that side arms, fringing lakes etc are maintained in contact with the main river channel. They also combat permanent land accretion and the loss of spawning areas. On the whole such structures appear successful in increasing habitat diversity and although there is no immediate evidence for increases in biomass which would in any eventuality be difficult to detect in such a large river, various species do appear to be showing distinct preferences within the new richer ecosystem.

Structures which provide direct cover

At the most primitive, such structures consist of fixed or floating objects which provide shade in simulation of missing vegetation cover in larger river channels. Submerged objects may also serve a similar purpose. Thus logs, boulders, branches etc, may be placed adjacent to the bank, a principle elaborated by many tropical fisheries into brush parks or artificially created or anchored masses of vegetation.

Because snags and other large organic debris are demonstrably important in maintaining habitat diversity in small to moderate size streams (order 1 to 4) Sedell et al. (1982) proposed that, where channels had been cleared, such structures be either artificially reintroduced or be allowed to accumulate in order to increase the productivity of such waters.

Spawning area improvements

Several systems have been attempted for the improvement of bottom texture for fish which spawn over gravel, particularly where siltation has resulted in the loss of the normal substrate. These methods often consist in mechanical or hydraulic disturbance of the gravel causing deposited sediments to be washed out, although the more costly and arduous physical replacement of the substrate has also been tried. Letichevsky (1981) for instance, quotes the success of the recreation of sandy spawning substrate of Stenodus leucichithys in the lower course of the Volga, which has now supplanted a more expensive intensive fry rearing and stocking programme.

Fish passes

One of the remedies commonly proposed for blockages to migrations caused by dams is the construction of fish ladders or passes similar to those which work so satisfactorily for the salmons in North Temperate rivers. In some cases fish passes have been successfully installed in the tropics. One such, at Cachoeira de Emas on the Mogi Guassu river, has been functioning since 1936 with little apparent detrimental effect on the stock. Here many characin species leap up a series of shallow steps. The siluroids, which do not jump, migrate past the barrage through special tunnels. The several fishways at dams on the Tigris and Euphrates rivers have also been used successfully by many of the migratory cyprinid species in these rivers (FAO/UN, 1956). Furthermore fish passes in dams blocking the Himalaya tributaries of the Ganges also appear to function adequately for Tor spp. and Indian major carps to the extent that the upper chambers of the ladders are frequently used as fish traps.

Most attempts at using fish ladders with tropical species have been less successful. Bonetto et al. (1971) claimed that the efficiency of the fish ladders which bypass a series of dams on the Caracana tributary of the Parana was very low and that Salminus maxillosus, one of the major migratory species of the basin, was unable to negotiate them. Bonetto (1980) repeated the pessimistic view in describing the anticipated performance of fish ladders to be installed in the Salto Grande and Yacyreta dams, although subsequent experience with the Borland lifts installed in the Salbo Grande dam showed that certain species including: Salminus maxillosus, Pseudoplatystoma coruscans, Colossoma mitrei, Doras spp., Oxydoras spp. and Prochilodus platensis did in fact use them to pass upstream. The fish ladders installed at the Markala barrage on the Niger also did not fulfil their intended function satisfactorily. Here enormous quantities of fish were blocked during their upstream migration at low water although some fish did get through. The fisheries above the dam declined considerably after it was closed as these reaches depended on a replenishment of their stocks from fish moving out of the Central Delta. The failure of the ladder was attributed by Daget (1960) to insufficient capacity in view of the very large numbers of fish wanting to pass over it. Furthermore, the migration of the species concerned was a simple dispersal movement rather than a breeding migration and the stimulus to surmount obstacles was possibly relatively low as a consequence. The failure of conventional salmonid ladders in many tropical rivers is hardly surprising in view of the complex behavioural factors which contribute to the functioning of such structures. For example, it has been shown that the form of the standing wave downstream of wiers is essential for salmon to be able to generate enough speed to clear the crest of the weir. Presumably other species have very different requirements which need to be studied before the installation of effective fish passes can be contemplated. In any case the provision of a fish pass or ladder for upstream movement of fish would only be justified where the migration is absolutely essential for the maintenance of the fish stocks.

MANAGEMENT OF THE FISH STOCK

In principle a fishery exploits a community which in its unexploited state exists in some kind of equilibrium with itself and with the environment. Under normal circumstances such a community may be assumed to tend to maximize its biological productivity and to continue to do so when subjected to a reasonable level of exploitation. In such a situation manipulation of the community in all probability adds little or nothing to the yield of the fishery. There are two circumstances, however, under which such interventions may be desirable. First, where the fish community is lacking some element to exploit a food resource or habitat (i.e. a ‘vacant niche’) it may be considered advisable to introduce one or more new species into the waterway. Second, where the fish community as a whole, or some preferred element of it, is overfished a policy of stocking may be adopted.

Introduction of New Species

In their unmodified condition most rivers support fish communities that are sufficiently diverse to fill most of the available trophic and spatial niches. Nevertheless some systems do exist which for zoogeographic reasons have poor or incomplete faunas. Furthermore as systems are modified there is a tendency for certain indigenous species to disappear and for others to take their place. This is most noticeable in impoundments where many of the migratory riverine species are lost soon after the stabilization of flow. In many rivers species which are adequately adapted to the new conditions are absent from the original fauna and to overcome this additional species may have to be introduced from other river basins. In some cases introductions have been performed uncritically, and such species as the carp (Cyprinus carpio) or Oreochromis mossambicus have themselves been blamed for degradation of the environment or fish community structure. These and others, for instance Salmo gairdneri, have been implicated in the elimination of sometimes delicate native resources, such as the unique faunal associations of Northeast California (Cooper, 1983) or the endemic galaxiids and Protroctes oxyrhynchus from New Zealand. Such cases often arise when there has been a contemporary disturbance of the environment which may have led to the elimination of the species in any case. The presence of introduced species may thus be considered an advantage as they are able to colonise the disturbed habitats and thereby maintain or increase species diversity (Leidy, 1983). Fishes such as these, which are among the main elements for intensive aquaculture, find their way into natural waters by accidental release from farm ponds. Any introduction of a new species into the fish fauna of a river should, therefore, be preceded by a very careful analysis of its anticipated impacts. The proceedures for such an analysis have been defined by EIFAC/CECPI (1984).

While widespread or impetuous transfer of new species are to be decried, the great success of these same species under other circumstances points to the role of properly considered introductions. Thus most of modern aquaculture is founded on a comparatively small number of introduced species, the tilapine cichlids have been invaluable in increasing the yield from Asian reservoirs (Fernando, 1976) and the common carp supports fisheries in many highly eutrophicated streams. Typical of such successful transfers is that of the chinese carps into the Amu Darya river. A combination of the macrophyte eating Ctenopharyngodon idella with the planktonophage Mylopharyngodon piceus, Hypophthalmichthys molitrix, Aristichthys nobilis and Parabramis pekingensis colonised the system including the adjoining Kara Kum canal thereby laying the basis for a rich fishery.

Stocking

Three main motives for stocking exist:

(a)  Stocking to maintain production in the face of intensive exploitation

(b)  Stocking to mitigate or compensate for adverse effects of some activity within the river basin

(c)  Stocking to increase production from individual components of the aquatic system, usually through extensive aquaculture or ranching;

With the exception of trout streams and some sport fisheries in the temperate zones systematic stocking of rivers to maintain or improve their fish stocks is not yet a widespread practice. In fact the tendency has been more for rivers to act as a source of supply of juvenile fish seed for stocking into reservoirs and fish culture installations. The requirements for successful stocking of salmonids have been extensively studied and rules of thumb for numbers of fish and age of fish to be stocked have been derived for various waters (EIFAC, 1982). The requirements and indeed the rationale for stocking with other species is less well understood and in many cases current practice appears arbitrary and counterproductive. There is considerable room for the expansion of stocking as a tool for the management of river systems.

MANAGEMENT OF THE FISHERY

Traditionally a number of practices exist for the control of the fishery. These usually involve direct or indirect control on the fishermen by limiting access to the fishery, or on the practise of fishing through control of season, location or type of gear (Fig. 8.6). Usually attempts at direct control through legislation result in costly and sometimes oppressive enforcement programmes, whereas less direct manipulation of the economic forces regulating the fishery may be more effective. Whichever the approach adopted, if controls do not correspond to biological realities regulating the fish community attempts at regulating the fishery will be counterproductive.

Regulation of Access

As shown in Fig. 7.13 the total catch is related to the number of fishermen operating on the river. Furthermore the individual artisanal fisherman seems to have a limited fishing power in that he is physically capable of removing only a certain quantity of fish from the system in any one year. Because of this the solution that seems most appealing in this type of fishery is a simple restriction on the number of fishermen operating in a certain region. Such control of access through licensing remains one of the most important managerial tools, particularly in commercial or recreational fisheries. An allied technique, the fixing of catch quotas for individual fishermen poses enforcement problems in artisanal fisheries but may be practical where more intensive industrialized practices are adopted. Many traditional systems also practice forms of access control by allowing only certain groups within the community to fish the waters. In many rivers, especially in Asia, such control is achieved by dividing the river and the plain into lots, for which individuals or groups of fishermen compete at auction each year. The job of policing the lot rented then devolves upon the fishermen groups. The renting of stretches of river to fishermen for restricted periods has been criticized on the grounds that any group with only temporary tenure of a fishery will attempt to gain the maximum profit within the time available to it. This leads to ruinous fishing practices where the stock is exploited to the point of exhaustion. On the other hand, where areas are held by tradition or contract for a number of years as in many parts of Africa, the fishermen are free to husband the resources and develop the fishery. A final form of access control is usually less direct by manipulating the economic conditions of the fishery so as to dissuade fishermen from entering the fishery. This may take the form of taxation, price control on fish landed or regulating the supply of fishing gear all of which are aimed at lowering net revenues to the fishermen.

Unfortunately the regulation of access to the fishery by traditional or by planned management tends to break down under the pressure of unstable socio-economic conditions. This occurs particularly where there is a high rate of rural unemployment and landlessness and fishing is viewed as a relief occupation producing an influx of numerous new and uneducated persons into the fishery.

Figure 8.6  Diagram of the relationships between the main techniques for fishery management

Increasing the Catch Capacity of Fishermen

Improvements in the catch capacity of the individual fisherman can be brought about in several ways. Introduction of new types of gear, or the amelioration of existing ones enables the individual to capture larger numbers of fish of a greater range of species. Better materials mean that the gear is less liable to damage, consequently less time has to be spent in its maintenance and more on the fishing grounds. Proper boat design and motorization of craft can get the fishermen to and from the fishing grounds faster. Historically these processes have already occurred in many rivers, first with the introduction of perfected gill nets, seines and cast nets, second by the introduction of nylon twine and netting, and third by the adoption of the outboard motor. In some fisheries too, traditional methods, especially those centred around cross river barrier traps which capture migrating fish, are as efficient as the more modern techniques that seek to replace them. The effectiveness of the fishery may also be conditioned by geographical accessibility or by the lack of adequate facilities for treatment which limit the amount of fish that can be handled. Because of this some fisheries do operate below their full capacity. However, the efficiency of much of the traditional and improved modern gear is such that individual stocks have collapsed in some river systems and other whole communities are near maximum exploitation, if not actually overfished. Improving the efficiency of the individual fisherman does not, therefore, appear to be appropriate in many cases and should be regarded more as a means of fine tuning the fishery as development of other sectors occurs. In other words, in areas where there is a great demand for rural employment it is probably advisable to restrain the introduction of more efficient methods and to concentrate on better marketing and other support facilities. In such circumstances, where better gear is introduced, the improved performance is either dissipated in a diminished catch per unit effort as the stock declines, or the less prosperous and efficient fishermen, who cannot use or adapt to the new gear, are forced out of the fishery. As the general economy of a country increases on the other hand, there is a tendency for labour to be withdrawn from the rural sector including fishing. In such cases, fisheries have declined due to fishermen leaving their trade, which in any case becomes less and less profitable. Here a rapid increase in the individual efficiency of a few fishermen can to a certain extent halt the decline and preserve a reasonable way of life for those who choose to remain in their previous profession.

Closed Seasons

By the nature of their biology, most river fisheries have built-in closed seasons. In the floodplain this lasts from just after bankfull on the rising flood to peak floods when the fish population is too dispersed and individual fish are too small in size for them to be readily available to the majority of methods of capture. In the torrential upper reaches flows are frequently too high for effective fishing during some of the year. This generally discourages fishing throughout the period as yields are low, effort needed to capture fish high, and there are additional dangers of loss of gear in the currents and floating vegetation masses. This effective closed season makes biological sense in that it allows the fish to reproduce relatively undisturbed and for the young to grow to a reasonable size before they are exposed to the fishery. Closed seasons outside this time are of limited value, although restrictions on the overly heavy fishing of migrating adult fish to the spawning grounds may be necessary in some places, as experience with Labeo in Africa have shown.

Reserved Waters

As fishing of the floodplain and river channel in the dry season becomes more intensive there is a risk of local over-exploitation of the stock. For this reason, traditional fisheries have long been based on the designation of certain floodplain depression lakes and reaches of the river as reserves which remained unfished. In larger systems, there are usually inaccessible areas which form reserves as they are infrequently exploited and it is probably from such areas that the Niger and Senegal rivers were recolonized during temporary reversals of the arid conditions of the Sahelian drought. As a management measure the conservation of certain areas is probably a wise move and gains force when other pressures are being applied to the system.

Mesh Regulations

The use of mesh selective gear almost always entails a consideration of the mesh sizes to be adopted which can only be viewed relative to the characteristics of the stock to be exploited. As has been described, the fishing-up process almost always involves a drift downwards in mesh size which needs considerable enforcement of legislation to stop. Any lower limit on mesh size, therefore, has to be imposed in the face of a natural trend to disregard it. It also poses the classic dilemma of how to manage multi-species stocks consisting of species with a range of sizes. If the objective of the fishery is to exploit only the large species of the community, the imposition of mesh limits which protect the immature fish is probably the only way to do it effectively. However, one almost certainly neglects a considerable proportion of the potential ichthyomass in this type of fishery as is shown by the number of species caught by progressively lower mesh sizes (see Fig. 8.7). If the mesh size is lowered to take advantage of the smaller species, then almost automatically the larger ones will disappear. One possible solution to the dilemma is the limitation of mesh size in major gears such as seine nets, gill nets, etc., coupled with a use of a variety of minor gears which are aimed at particular smaller elements of the stock. In other words, in a fish community which is highly diverse it is as well to maintain an equally diversified fishery.

Banning of Certain Gears

The restriction or complete outlawing of more destructive fishing practices is most important. However, even such methods may be appropriate in some circumstances. Poisoning of water courses is of course liable to damage the stocks of fish when carried out in the main channel of the river, whereas its use for removing fish from temporary floodplain pools or for eradicating undesirable species may be quite permissible. Unfortunately were the use of poisons allowed in one habitat it would rapidly extend to others. Gear is often prohibited for reasons other than those bearing directly on the fish stock. Long lines, for instance, are regarded with disfavour by users of cast nets which may become entangled in the hooks. Barriers which completely block the river channel, thus stopping fish migrations, are as likely to be removed for reasons of navigation as for fisheries.

Figure 8.7  The number of species captured by gill nets of different mesh sizes (after Reizer, 1974)

Accessibility

One of the main factors determining the degree to which the fisheries of the world's rivers are exploited is their accessibility and proximity to major centres of population. Many of the tributaries of major rivers such as the Ganges or the Amazon are still relatively unexploited due to the absence of access roads, and some of the greatest floodplains are still relatively little utilized by reason of their distance from suitable markets and lack of living space for the fishermen. The Okavango delta, the Gran Pantanal of the Paraguay river or the Central flood regions of the Zaire river are typical examples of areas where development is conditional on the provision of suitable access routes and centres for the marketing of catch, as well as artificial mounds or other appropriate locations for the fishermen to settle near the fishing grounds during the flood.

DEVELOPMENT OF NEW OR ALTERNATIVE FISHERIES

Aquaculture

Natural production of rivers can be exceeded by various forms of husbandry, intensive or extensive aquaculture. Such methods are widespread particularly on floodplains, and various techniques are available for the different developmental stages of rivers.

Blocking of natural floodplain depressions

Complete blocking of small channels draining water from depressions is common in many systems. Fish are conserved in the pools thus created and are fished later in the seasonal cycle at times when fish from other sources are becoming scarce. This method is often the first systematic attempt at husbandry or extensive aquaculture. Experiments with more permanent dams have been carried out by Reed (FAO/UN, 1969a) in the Niger and by Reizer (1974) in the Senegal. Reed's work showed that the area of standing water on the plain was increased by such installations (Fig. 8.8), and Reizer showed how the level and area of such pools differ before and after damming (Fig. 8.9). Harvest from such pools may be quite respectable - about 185 kg/ha/yr in the case of otherwise unmanaged pools in the Niger and up to 500 kg/ha/yr in depressions which are stocked with fry and fed.

Figure 8.8 Tracing from aerial photograph of the Niger floodwater retention dams, showing the original water area and the area flooded after the placing of the dams. (After FAO/UN, 1969a)

Drain-in ponds

In more advanced culture systems on floodplains depressions are regularized and deepended and eventually their number may be increased by digging new ponds into the surface of the plain. Three river systems particularly have benefitted from this form of management. In the Oueme about 3 percent of the 1000 km2 surface area is occupied by drain-in ponds some of which are several kilometers long (Fig. 8.10). Similar constructions on the Mekong floodplain ranged from 20 to 100 m in length and 2–3 m in depth. In Bangladesh some 30 000 ponds have been formed from the borrow-pits which were excavated during the construction of the artificial islands upon which much of the rural population lives. Drain-in ponds are usually fished by blocking a portion with a bamboo barrier, successively removing the accumulated vegetation and advancing the barrier until the fish are confined in a small space from which they are easily removed. Drain-in ponds are also used in conjunction with rice fields where, as the water is drained prior to harvesting, the fish may retreat into ponds or ditches constructed for this purpose. Yields from drain-in ponds may be high. In 1955–58, 34 ponds from the Oueme produced a mean of 2.1 t/ha and even though the catch dropped the same installations were still producing over 1.5 t/ha ten years later. In the Mekong this type of pond was suppressed following the recommendations of Chevey and Le Poulain (1940) who felt that the stagnant waters retained at the end of the dry season damaged the crops as they spread out over the plain at the beginning of the next flood. A similar system is used in Equator where the “Chameras” fisheries for Dormitator latifrons of the Chone river involve the construction of ponds on the floodplain, either by dyking or by excavation. The young of D. latifrons stocked into these are grown-on through the dry season.

Husbandry and fish culture in backwaters

Backwaters and other residual waterbodies separated from the main channel as a result of such river control works as revetting and channelization may be used for rearing fish. In Hungary, for example, such waterbodies are being developed for aquaculture (Pinter, 1983) and compare very favourably with the more traditional rearing of fish in ponds. Yearling fish are stocked into such waterbodies which are particularly suited to the raising of herbivorous species. Little additional feed is required over and above the natural production, although this is encouraged with a limited application of fertilizers. The success of this method is indicated by comparison with pond farming (Table 8.5).

Figure 8.9  Difference in changes in (A) depth and (B) area of a floodplain pool before and after damming of the main access channel. Encircled letters refer to individual illustrations in Figure 1.8

Refuge traps

In the Oueme river vegetation masses may be planted deliberately at the end of the flood, either attached to the bank (Fig. 8.10) or recessed into it at the mouth of the channels which drain the plain. They are left to collect fish for about two months after which they are fished and replaced to be emptied again towards the end of the dry season. Harvest from such “refuge traps” or “fish parks” can be quite high and in the Oueme river 15 installations of this type of mean area 440 m² gave a mean harvest equivalent to 1.88 t/ha (of park) per fishing or 3.88 t/ha/yr between 1958 and 1968.

The practice of deliberately planting vegetation or branches on the water to attract fish is in fact very widespread in rivers as well as in coastal lagoons having been recorded by Chevey and Le Poulain (1940) from the Mekong, Stauch (1966) and Reed et al. (1967) from the Niger and Benue systems and Meschkat (1972) for Equador. Welcomme, (1972a) carried out an evaluation of this method of fishing as it is practised in the lagoons associated with the delta of the Oueme river. The refuge traps in this river and its associated lakes and lagoons are of two types in addition to the masses of floating vegetation described above. There are small circular types about 22 m² in area which give up to 2.8 t/ha for each harvest. As they are harvested up to ten times during the 7 month dry season an annual yield as high as 28 t/ha can be obtained without apparently affecting the catch by other fishing methods in the area. Larger rectangular parks or “acadjas” are harvested less frequently, but also achieve very high yields. The yield of brush parks is directly related to the frequency of fishing and to the density of planting with greater yields of fish coming from parks with the greater density of branches. In the freshwater zone, a variety of species are attracted to the Oueme fish parks from which up to 32 species were recorded. By contrast, in the brackishwater zone, only two species, Sarotherodon melanotheron and Chrysichthys nigrodigitatus made up to 95 percent of the individuals present. These breed actively in the park throughout the year and as a consequence the population builds up rapidly (Fig. 8.11) thus accounting for the very high yields of up to 8 t/ha/yr in 1957–59.

Figure 8.10   Portions of the Oueme floodplain showing the distribution of different activities: (A) river channel with brush parks and drain-in ponds; (B) a dense aggregation of drain-in ponds. In both illustrations the plain is divided into fields. (By courtesy of IGN, Paris)

More recent investigations have shown that the fish stock increased in a similar manner in 1969–70 although yields were somewhat lower as a result of changes in the lagoon environment brought about by the construction of a port. The only other such installations that have been systematically investigated are the ‘Samra’ parks of the Grand Lac which according to Chevey and Le Poulain (1940) would appear to have given comparable yields to the Benin type of installation. Brush park fisheries are regarded by many as a somewhat mixed blessing. There is little doubt that they put up the overall productivity of the body of water in which they are used, but they also may shorten its life through accelerated silting.

Rice-fish culture

The widespread practice of fish culture in rice fields may be considered an almost ideal method of land use which produces both a carbohydrate and protein crops from the same piece of land. It has reached very high standards in Asia and Madagascar but is still in the experimental stage in Africa and Latin America. Coche (1967) has dealt comprehensively with the situation up to 15 years ago. The combined culture of fish and rice is widespread throughout Southeast Asia and represents a logical, integrated approach to the management of these types of wetlands. Ruddle (1980) classifies the various ways in which rice and fish can be managed as crops from the same piece of land.

Figure 8.11   Relation between yield and length of time of installation of fish parks in the Oueme Delta: (•) means for 1957–59; (x) means for 1968–70

He estimates that capture yields are about 135 kg/ha in the Malaysian Peninsula.

Two main types of exploitation can be distinguished:

(i)   capture systems where the fish that gain access through natural channels may be trapped.

(ii)   culture systems where the field is deliberately stocked with fry.

The fish may be taken as a single annual crop with a single rice crop, an intermediate crop between one rice harvest and the next planting or during the growing season of the rice. Frequently fish are trapped casually during the growing season, and are then drained into specially constructed sump ponds or canals as the fields are dried (see Fig. 8.12). Yields from trapping during the growing period have been measured at about 132 km/ha by Tang Cheng Eng et al. (1973). If the rice is cropped only once per year, the sump ponds yield about 162.8 to 262.9 kg/ha (mean 200 kg/ha) but if the field is double cropped fish yields drop to 68.2–143.0 (mean 95.7) kg/ha. At the same time between 2 277 and 2 975 kg/ha of rice were produced. Fish represents between 12–50 percent of the total income in single cropping and 6–27 percent in double cropping systems. Similar yields have been recorded in Thailand in the Mekong basin where 433 hm² of paddy fields were sampled giving fish harvests ranging between 50 and 1 710 kg/ha (mean 433 kg/ha) (Thailand, Department of Fisheries, 1974). In Sri Lanka rice-fish culture has not been practised because of the short (3–4 weeks) duration of the flood phase.

There has been a long term decline in yield from ricefield fish production. The main reasons for this are: First the increasing use of pesticides, weedicides and fungicides associated with the modern advanced agrarian practices needed to grow the high yielding varieties of paddy. The effects of pesticide treatment of rice for the control of stem borer insects, has caused concern and such workers as Kok (1972) or Saanin (1969) have attempted to define the toxic and sub-lethal effects of the common pesticides used (which include Endrin, Methyl Parathion, Dieldrin, Diazinon and -BHC). While no definite conclusions can yet be drawn as to the precise effects of such environmental contaminants, it is evident that the use of less toxic products can result in improved yields of rice whilst safeguarding the supply of fish.

Figure 8.12   Adaptations of the paddy field for concurrent rice-fish cultivation (after Ruddle, 1980)

Second the concept of multiple cropping in rice cultivation leaves little scope for fish culture because of the short duration of the flood over the rice field (Kassim, et al. 1979). The production of a plant and a protein crop from the same area of floodable land is undoubtedly of great value in maintaining a nutritionally balanced yield pattern, although one that is not necessarily maximal in terms of absolute tonnage consequently projects have been suggested for the extension of this type of culture. For instance, the management of the 32 000 ha Candaba swamp. This plain is flooded by the Pampanga and Angat rivers and is one of a similar series of floodplains in the Philippines. The plain has been enclosed in a series of dykes which include a gate for the harvesting of fish. The fields are at present used for water melon or rice growing during the dry season when the fish remain in the canals. Fish yields based solely on natural productivity range from 300–500 kg/ha in 7 months. The melon crop produces 5 000 kg/hm. Delmendo suggested ways in which this area could be improved with the inclusion of fish canals at the foot of the main dykes and an intensification of fish culture by stocking and manuring the ponds.

Intensive aquaculture in ponds

The intensive culture of salmonids in ponds associated with the valleys of rivers is very common in temperate countries, but the use of the potamon in the tropics for the purpose is more recent.

Once the flood is completely controlled the former floodplain is converted to irrigated agriculture. Aquaculture farms, with ponds or pools specially adapted to the rearing of fish, can be associated with the irrigation net-works. Such farms can either rear fish to a size where they can be sold directly for consumption, or can raise fry for stocking into such permanent lakes as remain in the plain or into fields where fish are reared as a secondary or alternative crop. Developments of this type are perhaps best seen in Thailand, although they are also common in many other parts of Southeast Asia. In the Chao Phrya river basin for instance, the control of the water regime over much of the floodplain has decreased natural fish production. To compensate for this loss many state and private fish farmers rear various species of cyprinid and siluroids as well as freshwater prawns (Macrobrachium rosenbergii). The impervious nature of the alluvial soils, the flat terrain and the ready access to water supplies through irrigation canals makes this type of modified floodplain particularly suitable for aquaculture in ponds.

Cage culture in rivers

Once flow is stabilized within the main river channels, or even before such stabilization occurs, fish may be cultured in floating cages. This type of culture usually applies to species of high commercial value, such as Pangasius sutchii which is cultured in this manner in Thailand. The fish are usually caught as fry and reared to marketable size in such structures.

Reservoir Fisheries

Reservoirs support fish populations and fisheries where yields are comparable to or higher than the rivers they submerge. The net gain is usually far higher in rapids reaches where the relatively low productivity of the main channel is easily exceeded by the new lake, but in any system where a reservoir replaced a floodplain either by submerging it or by stopping its flooding the benefit or loss to the fishery as a whole has to be carefully evaluated. In the case of the Upper Mekong, many of the impounded tributary rivers have little floodplain either in the submerged reaches or downstream of the dams (Sidthimunka, 1972). However, according to the Mekong studies the mainstream Pa Mong reservoir will eliminate flooding for some 700 km downstream of the dam, resulting in a loss of catch of about 2 150 t from the river, whereas the reservoir is expected to produce only about the same amount of fish. Similarly, at least 6 000 tons of fish have been lost below the Kainji dam in Africa which must be deducted from the 5 400 t produced by the reservoir behind the dam.

In many other reservoirs the amount of fish lost downstream of the dam surpasses the potential of the dam itself, although such considerations are usually secondary where power generation or irrigated agriculture are also taken into account.

The fish community that establishes itself in the reservoir is based entirely on that pre-existing in the river basin, unless exotic species are introduced. There are usually considerable modifications in community structure as many species are unable to adapt to the new environment. Earliest to disappear from the body of the reservoir are the migratory whitefishes although they might survive in the upper portion of the lake from which they can readily move upstream to breed. Blackfishes tend to adapt better, and the main body of the lake is often colonized by species that were relatively insignificant elements of the preimpoundment fauna. In some rivers, for example, those of the Indian subcontinent, Fernando (1976) suggested that few suitable blackfish occur. Consequently, lacustrine fish communities are not readily established with the species native to the system, and, failing the introduction of exotics, populations have to be maintained by continual stocking.