NUMBER TWENTY NINE - OCTOBER 1998

The strategy has seven steps:

1. Diagnosis of the problem and establishment of the target for restoration; (b) Removal of existing or potential forward switches;
2. Reduction of nutrient loading;
3. Biomanipulation;
4. Restablishment of plants;
5. Re-establishment of an appropriate fish community;
6. Monitoring of the results

Deciding what is wanted and whether it can be achieved is the very important first step of the strategy. Essential to this is a clear view on why the restoration is being undertaken. This might be to reestablish a former biodiversity deduced from historical evidence, to create an attractive, though not necessarily diverse community for general amenity, to establish clear water for abstraction and less expensive subsequent treatment, or a robust general recreational fishery. The first target is likely to be the most expensive one to meet as it will require the greatest nutrient control.

Goals such as re-creation of a diverse charophyte community may not be attainable simply because nutrient loading cannot be sufficiently reduced. This will be difficult in an intensively farmed landscape, though it may be feasible where farming is extensive and supported by measures such as the Environmentally Sensitive Area legislation in the UK. Desk studies can be used to determine what degree of nutrient control is possible. Export models [91] can be used quite precisely for calculating external loads, though there are more uncertainties where internal loading (release from the sediments) is involved as generally it will be in shallow lakes. There are also now uncertanties about which nutrient, nitrogen or phosphorus, to control.

Historically there has been most emphasis on phosphorus and this has undoubtedly been effective for most deep lakes. Because of the likely normality of phosphorus release from sediments in shallow lakes, it may be more appropriate to concentrate on nitrogen control. This poses considerable difficulties because of the diffuse nature of nitrogen sources and the strong political lobby in favour of the intensive farming which has generated high artificial nitrogen loads. A further uncertainty is in linking the 'quality' of the restored community to water chemistry. There is a general acceptance that the lower the nutrient concentrations, the greater the diversity of aquatic plants and thence of associated animals, but it is not yet possible to be quantitative about the relationship. The lower the nutrient concentrations are, the better, but the costs of nutrient control are likely to increase rapidly the greater the stringency that is required.

A restoration programme may be blocked if forward switches are present that cannot be removed or their influence severely reduced. The effectiveness of a switch may be reduced if the nutrient loads can be lowered, but again we cannot yet be quantitative. The possibility, nonetheless, exists, for example, that boating activity might be sustainable at low nutrient concentrations, whereas it would cause a rapid switch to algal dominance at higher levels. This may be a useful trade-off, for some switches are not removable. Small and shallow lakes often figure in an interlocking nexus of uses such that removal of geese and ducks, fed by visitors, may simply be impossible without outcry from an affectionate public. Similarly we are yet some way from an agriculture that achieves parsimonious fertilisation and is independent of pesticides and herbicides, or from urban sewage effluents that are not complex mixtures of many substances potentially inhibitory to Daphnia activity.

Nonetheless some obvious switches can be removed. Deliberate management of aquatic vegetation by cutting or herbicides can be stopped, boats can be re-routed, water tables can be raised to minimise saline inflows. Such control does not always have to be all or nothing. Combined with nutrient control, there may be many points of compromise.

There is a large literature [37,64,121] on nutrient control which will only be briefly covered here. The topic can be dealt with under four combinations of control of point and diffuse sources of phosphorus and nitrogen (Box 5). Until recently it was generally understood that most phosphorus came from point sources, usually associated ultimately with the excreta of humans or farm stock. It now appears that diffuse sources from soils may be increasing [55,56] and frustrating attempts already made to treat effluents. The problems already associated with the removal of diffuse sources of nitrogen are thus becoming associated with the removal of phosphorus.

The ultimate source of these problems is in the nature of modern farming systems and also in the perception among agriculturists that because only a small proportion of the fertilisers put on the land is lost to freshwaters, there cannot be a big problem. The reality is that water is many times more sensitive to nutrient additions than soil and a small loss from the land can be a large absolute amount for freshwater ecosystems. The latter evolved in former environments where natural nutrient losses from the land were extremely small. The plan made for nutrient control in a given catchment will depend on the local circumstances and models may be helpful in targetting the major source.

There is now European legislation (Urban Wastewater Treatment Directive [91/271] EEC]) potentially capable of controlling point sources of both nitrogen and phosphorus from waste water treatment works, at least of large size. Unfortunately the ways that individual member states implement this legislation mean that its potential effectiveness is greatly reduced. It does not apply to small works, of which there are many, and it only applies to areas deemed by states to be sensitive to eutrophication. Restricted definitions of the latter have allowed some governments to avoid applying the legislation to many of their freshwaters.

There is little legislation concerned with the control of diffuse sources of either nitrogen or phosphorus. The European Union Nitrates Directive (91/676/EEC) has potential in this regard but has not been widely used where drinking water has not been involved. Most schemes depend on voluntary arrangements for which there is little or no financial incentive to landowners. Most codes of practice for fertilisation or sludge or manure spreading, although attempting to minimise losses to waters, are orientated towards maintaining agricultural productivity with little compromise.

Arguably the damage wrought on freshwater systems by the activities of farming should be the responsibility of the polluter, but to some extent the very existence of agriculture, indisputably necessary to settled societies, inevitably means some leakage of nutrients. There is thus a case that society should pay for measures such as the leaving of substantial strips of land alongside streams as buffer zones to absorb nutrients, but equally that these losses should be those only from a sustainable agriculture and not from one that is heavily subsidised and profligate of nutrients.

In nutrient control, new problems seem continually to arise. In upland areas it now seems likely that atmospheric nitrogen loading from manure decomposition and vehicle exhausts is a very significant source [11]. Although the nitrate concentrations in receiving waters are still small compared with those running off arable land and ploughed grassland in the lowlands, they may be capable, together with a small intensification of upland farming, of damaging existent plant-dominated systems. Control of such atmospheric sources is not in the purview of lake restorers but depends on considerations fundamental to the nature of our society.

Box 5 shows various means of controlling external loads - those coming from the catchment area. Because release of phosphorus, in particular, from sediments has now been shown to be a major contributor to the phosphorus budgets of shallow lakes, there has been much pressure to remove surface, phosphorus-enriched sediment by dredging [20]. In theory this removes the source. In practice, it removes only part of the source for the underlying sediments, though lower in their phosphorus concentrations, may still be potent (and normal) sources. There are no unequivocal cases where sediment removal has been shown to have been beneficial in restoring plant dominance to shallow lakes and several where it has been definitely ineffective [2]. It may be necessary where water has become so shallow that boat navigation is impeded, but its value in ecological restoration is doubtful. The key to understanding this may simply be that it is nitrogen (which is not released in quantity from sediments) that is the key nutrient in the range relevant to the existence of alternative states.

In theory, biomanipulation is relatively easy. On a local basis, removal or modification of fish communities must be handled sensitively for anglers must be persuaded of the ultimate value of the exercise and there may be navigation problems to be solved. There is no point in removing fish if replacements can swim back in from an interconnecting river, so the biomanipulated lake must be isolated for at least a year or two in most cases. Allowing boats, but not even the smallest fish, to move in, may pose technical problems that are not easy to solve.

If carried out efficiently, the process almost always results in an increase in Daphnia populations and clearing of the water (Fig 18). Failures are usually because insufficient fish have been removed or because invertebrate predators on the Daphnia, also the potential food of the fish, have increased in number. There have been occasional instances of this with a mysid shrimp (Neomysis integer) in slightly saline waters [114], but they are unusual.

Complete removal of fish, however, is a relatively crude ecological tactic. It prevents angling for some time and much effort may be needed to convince the general public of the necessity of such a drastic move in the interests of conservation! Selective removal of the zooplanktivorous fish would be a better approach. However, the difficulty is that most European fish are zooplanktivores when young, and it is the large populations of young-of-the year fish that do most of the feeding on the Daphnia. Later stages of some species remain zooplanktivorous but most move their diet to larger invertebrates associated with the bottom or the plant beds.

Some species, such as tench (Tinca tinca), move early to bottom-feeding and characteristically do not produce large populations of young. Predators like eels (Anguilla anguilla), brown trout (Salmo trutta) and pike (Esox lucius) also have small populations and might be left when the main zooplanktivores are removed. Such selective fishing is difficult however. Netting methods are generally unselective if they are to recover the fish alive and electrofishing is inefficient at removing the smaller (and therefore more likely zooplanktivorous fish). Gill netting can be made highly selective for size, but kills the fish and is ethically unacceptable. The less drastic the fishing, the greater is the risk of failure of the biomanipulation and hence wastage of the whole effort. Further knowledge of individual fish biology is needed together with development of more subtle fishing methods to improve the sophistication of this exercise.

Biomanipulation may clear the waters and plants may spontaneously recover from seedbanks left in the sediment [95], but this is not inevitable. First the plants may have been absent for so long that no viable seeds are left, whilst many aquatic plants overwinter as vegetative propagules rather than seeds, and the longevity of these is very limited. Even if seeds or propagules are present or can enter naturally in winter floods or by the perhaps apochryphal route attached to the feet of birds, they may not readily establish. The sediments laid down under phytoplankton communities are often soft and amorphous and may not provide strong enough rooting conditions. They may shift in the wind and water currents or provide insufficient anchorage for developing rooting systems. The establishing plants may also be vulnerable to grazing by water birds such as coots (Fulica atra).

If they have to be introduced, care must be taken not to use horticulturally developed varieties bought from commercial suppliers. They will generally not survive in wild conditions and frustrate the intention to restore natural communities of conservation value.

1 Because of predation on epiphyte eating snails 2. Post-larval; pelagial means open water, middle of the lake. 3 Not native in the UK

* Scoring system: N = 5, I = -5; for breeding, - = 5, + = 3, + = 0, ++ = -5; for bottom disturbance, ++ = -5, + = 0, - = 5; for zooplanktivory, - = 5, + = -1, + = -3, ++ = -5; for piscivory, ++ = 5, + = 3, - = -5; for angling intrusion, - = 5, + = -3, ++ = -5; for abundance, + = 0, + = -3, ++ = -5; for plant destruction, - = 5, + = -1,+ = -3, ++ = -5.

Piscivores are always desirable, though the species available are few in the limited British fauna. On mainland Europe, the pikeperch (zander) (Stizostedion lucioperca) is a useful predator for it hunts small zooplanktivorous fish in open water. It is alien to Britain, however, and although already introduced, and spreading, further deliberate movement is illegal.

There is again uncertainty - this time about the outcome of particular combinations of reintroduced fish. We know quite a lot about individual fish biology but much less about how different species interact with each other, especially in an environment where fish reproductive success is very variable and dependent very much on temperature, which varies greatly from year to year. Controlled experiments in enclosures (Box 2) can be carried out on these interactions, but the scope where, say, ten main species must be investigated in all possible combinations of 1 to 10, and at different biomass levels, is daunting.

Few restoration attempts have been permanently stable or have achieved the high diversity of plant and invertebrate communities that occur naturally. Some of the reasons for this are understood, including failure to remove or sufficiently reduce forward switch mechanisms, to lower nutrient loadings effectively or to achieve sufficient fish removal. These can, in principle, easily be dealt with. Others may be more systematic. We tend to view freshwaters as isolated systems which we can treat or manage per se, forgetting that the key concept in understanding the functioning of freshwater systems is that they are part of a greater whole, the catchment area. This is acknowledged in concepts such as nutrient loading, but overseen where other aspects are concerned. Attempts at restoration of lakes will always be potentially unstable if they are not accompanied by sustainable management of the catchment also. Our objectives are usually forced to be too limited.

This concept (Fig 20) is best dealt with by example. One classic case of restoration is Zwemlust in the Netherlands, where a small lake was restored to reasonable stability by biomanipulation alone, without nutrient control. It was soon the only plant-dominated lake in the local area, with the result that many coot were attracted by the rich source of plant food. This grazing was sufficient to threaten the plant-dominated state and the lake has shown tendencies to revert to algal dominance as a result [147,149].

With the drainage and development of floodplains, this greater system has been lost, and with it the associated mechanisms which supported the within-lake buffers and preserved the plant- dominated state. It is not possible to guarantee the future state of such lakes without reinstating this greater system. Opinion and action in some countries is now beginning to favour reinstatement of floodplain systems, costly though it is, for they provide flood control for downstream sites and recharge of ground water tables, as well as amenity, angling and other values. Such ambitious restoration should be our goal and is essential to provision of a sustainable environment. The same principle should apply to other lake systems, where stable restoration of diversity and function must also depend on reversion to agriculture which minimises the use of toxic chemicals, the risks of soil erosion and other mechanisms of nutrient loss. The only guarantee of this is to ensure that the forms of agriculture used are constrained by the natural features - geological, topographical and meteorological- of the catchment and not by artificial intervention with subsidies for inappropriate technology, and that urban waste-water management is as stringent as possible.

Fig 21 Restoration is most likely to be stable where it takes place in a complete interconnected lake, river and wetland system. Isolated lakes in engineered floodplains lack the connections that can buffer them against external impacts.