Summary1. Eutrophication constitutes a serious threat to many European lakes and many approaches have been used during the past 20-30 years to improve lake water quality. Results from the various lake restoration initiatives are diverse and the long-term effects are not well described. 2.In this study we evaluated data from more than 70 restoration projects conducted mainly in shallow, eutrophic lakes in Denmark and the Netherlands. Special focus was given to the removal of zooplanktivorous and benthivorous fish, by far the most common internal lake measure. 3. In more than half of the biomanipulation projects, Secchi depth increased and chlorophyll a decreased to less than 50% within the first few years. In some of the shallow lakes, total phosphorus and total nitrogen levels decreased considerably, indicating an increased retention or loss by denitrification. The strongest effects seemed to be obtained 4-6 years after the start of fish removal. 4. The long-term effect of restoration initiatives can only be described for a few lakes, but data from biomanipulated lakes indicate a return to a turbid state within 10 years or less in most cases. One of reasons for the lack of long-term effects may be internal phosphorus loading from a mobile pool accumulated in the sediment. 5. Synthesis and applications . Lake restoration, and in particular fish removal in shallow eutrophic lakes, has been widely used in Denmark and the Netherlands, where it has had marked effects on lake water quality in many lakes. Long-term effects ( > 8-10 years) are less obvious and a return to turbid conditions is often seen unless fish removal is repeated. Insufficient external loading reduction, internal phosphorus loading and absence of stable submerged macrophyte communities to stabilize the clear-water state are the most probable causes for this relapse to earlier conditions.

Climate change will alter freshwater ecosystems but specific effects will vary among regions and the type of water body. Here, we give an integrative review of the observed and predicted impacts of climate change on shallow lakes in the Netherlands and put these impacts in an international perspective. Most of these lakes are man-made and have preset water levels and poorly developed littoral zones. Relevant climatic factors for these ecosystems are temperature, ice-cover and wind. Secondary factors affected by climate include nutrient loading, residence time and water levels. We reviewed the relevant literature in order to assess the impact of climate change on these lakes. We focussed on six management objectives as bioindicators for the functioning of these ecosystems: target species, nuisance species, invading species, transparency, carrying capacity and biodiversity. We conclude that climate change will likely (i) reduce the numbers of several target species of birds; (ii) favour and stabilize cyanobacterial dominance in phytoplankton communities; (iii) cause more serious incidents of botulism among waterfowl and enhance the spreading of mosquito borne diseases; (iv) benefit invaders originating from the Ponto-Caspian region; (v) stabilize turbid, phytoplankton-dominated systems, thus counteracting restoration measures; (vi) destabilize macrophyte-dominated clear-water lakes; (vii) increase the carrying capacity of primary producers, especially phytoplankton, thus mimicking eutrophication; (viii) affect higher trophic levels as a result of enhanced primary production; (ix) have a negative impact on biodiversity which is linked to the clear water state; (x) affect biodiversity by changing the disturbance regime. Water managers can counteract these developments by reduction of nutrient loading, development of the littoral zone, compartmentalization of lakes and fisheries management

1. The effect of benthivorous bream and carp on sediment resuspension and the concentrations of nutrients and chlorophyll a were studied in sixteen experimental ponds (mean depth lm, mean area 0.1 ha, sandy clay/clay sediment), stocked with bream or carp at densities varying from 0 to 500kgha~\ Planktivorous perch {Perca fluviatilis L.) were added to some ponds to suppress zooplankton. 2. Suspended sediment concentrations increased linearly with biomass of benthivorous fish. Bream caused an increase of 46g sediment m"^ day"' per 100kg bream ha~' and a reduction of 0.38 m~' in reciprocal Secchi disc depth, corresponding to an increase in the extinction coefficient of 0.34m~V 3. No relationship was found between size of fish and amount of resuspension, but the effect of bream was twice as great as that of carp. Benthivorous feeding was reduced in May because alternative food (zooplankton) was available. 4. Assuming a linear relationship, chlorophyll a level increased by 9.0 [igl"', total P by 0.03mgr' and Kjeldahl-N by 0.48mgr' per 100kg bream ha^V Silicate, chlorophyll a, total P and total N were all positively correlated with fish biomass, but orthophosphate showed no correlation.

In this paper we analyze a long-term dataset on the recovery from eutrophication of Lake Veluwe (The Netherlands). Clear hysteresis was observed in a number of ecosystem variables: the route to recovery differed significantly from the route that led to loss of clear water. The macrophyte dominated state disappeared in the late 1960s at TP above 0.20 mg l 1 , whereas its return occurred at less than 0.10 mg TP l 1. Several regime shifts resulting in the occurrence of three alternative stable states were observed over a period of 30 years. The turbid state showed resistance to change, despite a strong and prompt reduction in Chl-a following reduction of external P-loading. The most important component that determined hysteresis in the return to clear water was not internal P-loading, but a high level of nonalgal light attenuation (through sediment resuspension) maintained by the interaction between wind and benthivorous fish. Although Chara was able to re-colonize the most shallow parts of the lake, recovery stalled and for a number of years clear (above charophyte beds) and turbid (deeper parts of the lake) water co-existed, as a separate alternative state on route to full recovery. Lake-wide clear water was re-established after bream density had been reduced substantially. This allowed a return of zebra mussels to the lake, whose high filtration capacity helped in maintaining clear water. In this study, we were able to identify the main drivers of hysteresis and regime shifts, although formal demonstration of cause and effect was not possible on the basis of field data alone. We argue that resilience of the present clear water state of Lake Veluwe very much depends on sizable populations of a few keystone species, especially Chara (stoneworts) and Dreissena (zebra mussels), and that careful management of these species is equally important as control of nutrients. Lake management should strive to maintain and strengthen resilience of the ecosystem, and this should offer protection against a renewed collapse of the clear state.

In October-December 1990, the large barbs (Barbus) that contribute more than 35% of the catch in lake Tana (northern Ethiopia) were studied. Previous authors (Riippelll837, Boulenger 1902,1911, Bini 1940 described from 6 to 23 (sub)species for the lake. Banister (1973) lumped all of these into one subspecies: Barbus intermedius intermedius Riippell, 1837. We found that the Lake Tana Barbus could be readily categorized in at least 13 discrete morphotypes, some of which were already distinguished by local fishermen. None of the known descriptions are adequate to distinguish the barbs unambiguously, which is important for monitoring and management of developing fisheries. Intermediates between morphotypes were rare (< 10%). By applying canonical discriminant analysis on a set of 17 morphometric characters (including some directly associated with feeding) our initial morphotype-distinction was confirmed. Also, differences between the morphotypes in distribution, related to depth and substratum were found, as well as differences in intestinal contents, a key to the food-niche. The high number of piscivorous morphotypes (8 out of 13) was striking as piscivory is relatively rare among cyprinids. Piscivory was found to be highly correlated with morphological (feeding related) characters. The presence of discrete morphotypes, that also differ in food-niche and distribution, strongly suggests that several distinct populations exist, that may be (partly or completely) reproductively segregated. Knowledge about these populations, that may represent separate units of fish stock, is of crucial importance for the management of sustainable fisheries and protection of the biodiversity in Lake Tana. It is possible that several species or even a unique cyprinid species flock are present, that urgently need protection.