Clearing lakes : an ecosystem approach to the restoration and management of shallow lakes in the Netherlands

In the 1950 s and 1960 s, most shallow lakes in the Netherlands shifted from macrophyte-dominated clear water lakes, towards algae-dominated turbid water lakes. Eutrophication, i.e. increased nutrient loading, is the main cause of the deterioration of the lake ecosystems. Other perturbations, such as the loss of lake-marginal wetlands (nutrient filters, habitat for pike, Esox lucius) and chemical pollution toxic to zooplankton, will have reinforced the effects of nutrient enrichment. The lake restoration strategy has been concentrated on the reduction of the external phosphorus (P) loading. However, so far this approach did not result in the water quality (in terms of transparancy, phytoplankton species, fish stock etc.) desired. A more comprehensive approach to lake ecosystem functioning may provide additional tools for lake restoration. Algal blooms in lakes will develop when the algal production is high and the algal losses are low. Production is controlled by the supply of nutrients and light. Consumption of algae by zooplankton is a major loss process. In this thesis attention is focused on both sides of the algal balance: (1) the control of the external and internal (from the sediments) P loading, and (2) biomanipulation, the manipulation of fish communities aiming at increased consumption of algae. Reduction of the planktivore fish stock may enhance the zooplankton and thus the grazing on algae. Bream (Abramis brama) and roach ( Rutilus rutilus ) are among the major fish species in many turbid Dutch lakes. Large bream feeds on zooplankton, as well as benthic organisms. Reduction of the benthivore fish stock results in reduced sediment resuspension and P release and less disturbance of rooted macrophytes.

Development of the lake restoration strategy (chapter 1)
The strategy for lake restoration gradually evolved from solely P stripping from sewage, towards a more comprehensive and ecosystem-based approach. In the Netherlands, P removal at sewage treatment plants is common practice and polyphosphates in household detergents have been replaced by less harmful compounds. TP levels in surface waters (including the Rhine river) are going down, however, in lakes algal blooms persist. Presently, additional measures directed at the reduction of non- point sources, treatment or removal of P-rich sediments and fish stock management come into focus. Other negative influences such as the ongoing loss of lake-marginal wetlands and (locally) the chemical pollution toxic to zooplankton, should be reversed as well.

Shallow lakes in the Netherlands: searching for lake restoration objectives (chapter 2)
Shallow lakes which are not, or only slightly, influenced by man, may provide clues to define key variables and processes expressing ecological sustainability. Historical studies may therefore be useful for finding specific objectives for lake restoration. One natural peat lake (Naardermeer) and two lake areas resulting from peat mining (Reeuwijk lakes and Oude Venen) were selected for such a reference study. It is well-known that many lakes in the Netherlands showed major changes in water transparency in the 1950 s and 1960 s. The information collected refers mainly to the period 1930-1950. At that time, all three lakes were dear with abundant submerged vegetation. Naardermeer, with its extremely low nutrient loading, showed the strongest indications for a stable dear water state (a high diversity of Chara species). Nutrient loading of the other two lakes was probably higher. It is speculated that the intensive commercial fishing in the Reeuwijk lakes (by 30-50 families), and the excellent pike habitat in the Oude Venen (due to natural water level fluctuations), played a significant role in maintaining the dear water state in these two lakes.

Multi-take studies: external nutrient loading and lake response (chapter 3)`
Multi-lake studies proved to be helpful in developing management criteria for eutrophication control. In 1976-1977, the first national eutrophication survey of the shallow eutrophic lakes of the Netherlands was conducted, including ca. 65 lakes, of which 14 lakes with reliable water- and nutrient budgets. More up-to-date relationships for take management were
derived by including the results of the eutrophication surveys for 1983-1985 and 1980-1988. According to the steady-state loading-response models, the in-lake TP concentration shows a proportional or 'somewhat' less than proportional response to a decrease in TP loading. Secchi depth-chlorophyll relationships and upper limits for chlorophyll in relation 4 to TP, TN and underwater light climate could be established. Lakes with an excess of TP V611 show a 'threshold response', after P loading reduction: algal biomass only goes down if the TP concentration approaches the limitation line, CHL = -24 + 1.04 TP (CHL and TP in mg m ^-3). Oscillatoria -dominated lakes may produce more algal biomass per unit P (CHL = 1 .54 TP).
Objectives for Secchi depth were derived, (1) for triggering a collapse of Oscillatoria blooms, and (2) for creating proper light conditions for the restoration of submerged vegetation. However, due to the highly variable non-algal turbidity, only upper limits can be given for the chlorophyll and TP levels necessary.

Whole-lake study of Veluwemeer: lake flushing for control of Oscillatoria blooms and internal phosphorus loading (chapter 4)
Lake Veluwemeer (3,356 ha, mean depth 1.25 m) has suffered from a year-round Oscillatoria bloom from 1971 onwards. Early in 1979, the P loading of the lake was reduced from 2.7 to 1.5 g P m ^-2y ^-1. Monthly TP budgets, however, showed a substantial net P release from the sediments during summer and therefore the prospects for recovery were poor. Summer peaks in TP coincided-with extremely high pH (pH 9-10) and it was hypothesized that the Oscillatoria bloom supported a self-perpetuating process of algal activity, high pH, P release and even more algal activity. Winter flushing with water that was low in TP and high in Ca ²⁺and HC0 _3- could interrupt this vicious cycle. The effects were spectacular. After the first winter flushing in 1979-80, summer pH dropped by one whole unit, TP and chlorophyll more than halved and summer algal growth has been P-limited ever since. However, Secchi depth only increased from 0.20 to 0.30 m. This disappointing transparency could be explained by reduced chlorophyll content per unit algal biovolume. Oscillatoria , which prefers dim light conditions, lost its dominant position after the cold winter of 1985. After the species shift, summer transparency was still limited to 0.40-0.50 m. Model calculations showed a decrease of summer internal P loading from 1.0-8.4 before to 0.0-1.7 mg m ^-3d ^-1
after the measures. Low summer pH values could be explained by the enhanced CaC0 ₃ precipitation during the spring. Apparently, the low pH resulted in reduced P release from the sediments and P-limited algal growth during summer. The reduced algal biomass and consequently lower sediment oxygen demand, and maybe also increased N0 ₃^-levels during early spring, will have reinforced the better binding of P to the sediments. Manipulation of the carbonate system proved to be an effective tool in controlling internal P loading. During winter, conservative behavior (net growth = 0) of the Oscillatoria population may be assumed. Therefore, Oscillatoria blooms can be effectively (>95% ) removed from wellmixed lakes by flushing in November-February, with three times the lake volume.

Biomanipulation in shallow lakes: concepts, case studies and perspectives (chapter 5)
High fish stocks in algae-dominated lakes tend to impose a homeostasis on the lake ecosystem, which then resists the recovery of the lake. In today's turbid lakes, the common phenomenon of the 'spring clear water phase', induced by zooplankton grazers, fails to appear. Large numbers of planktivorous bream and roach, throughout the year, as well as high densities of 'inedible' filamentous cyanobacteria, prevent the peaking in population of the efficiently grazing large-bodied Daphnia. Submerged macrophytes play a key role in maintaining the clear water state throughout the summer. For the reestablishment of the submerged vegetation, it is necessary to restore the spring dear water phase. Biomanipulation, i.e. a substantial fish stock reduction, could trigger a shift from a stable turbid water state to an alternative stable clear water state. Nine case studies were evaluated for testing the applicability and perspectives of biomanipulation. It was concluded that a single substantial fish stock reduction (>75%) during winter, offers good chances for achieving clear water.in the next spring. The filamentous cyanobacteria and the possible development of invertebrate predators (Neomysis, Leptodora) on Daphnia are uncertain factors for successful biomanipulation. Rapid colonization of submerged macrophytes, stabilizing the dear water state, has been demonstrated in small lakes (< 30 ha). High nutrient levels, ultimately (in two cases, after five and seven years), lead to a shift back to the turbid water state. Top-down control by the predatory fish (pike and perch, Percafluviatilis) seems restricted to small lakes, with a high degree of 'patchiness' (patches of macrophytes and open water). An alternative stable clear water state may be expected in the TP range of 50-100 mg m ^-3(or higher TP levels for very small lakes). More and more lakes in the Netherlands approach these TP levels, so the chances for biomanipulation are improving. Fish control is more difficult in large lakes, and particularly in networks of interconnected lakes. Additionally, in large lake areas the reestablishment of vegetation takes more time. Winter fishing on a regular basis (rather than a single fish stock reduction) may then be promising, but case studies are needed for further evaluation.

Whole-lake study of Wolderwijd: biomanipulation for promoting the clear water state (chapter 6)
Lake Wolderwijd (2,555 ha, mean depth 1.60 m) has suffered from Oscillatoria blooms, turbid water and a poor submerged vegetation as a result of eutrophication since the early 1970s . From 1981-1984 the lake was flushed (via Veluwemeer) during winter, with water low in TP and high in Ca ²⁺and HCO ₃^-. TP and chlorophyll a in the lake more than halved, but Secchi depth in summer only increased from 0.20 to 0.30 m. In the hope of triggering a shift from the algae- dominated turbid water state to a macrophyte-dominated dear water state, the lake was biomanipulated during winter 1990-91. The fish stock, mainly bream and roach, was reduced from 205 to 45 kg ha ^-1. In May 1991, 575,000 (217 ind ha ^-1) pike fingerlings were introduced. From 1989 onwards, lake flushing was intensified in order to reduce TP and the Oscillatoria bloom. In spring 1991, the lake water cleared as a result of grazing by Daphnia galeata. The clear water phase lasted for only six weeks. Macrophytes did not respond as strongly as was expected on the basis of the results from the small-scale case studies (ch. 5). Most of the young pike died. However, from 1991 to 1993, the submerged vegetation has gradually changed. Characeae began to spread over the lake (from 28 ha in 1991 to 438 ha in 1993). The water over the Chara meadows was clear, probably as a result of increased net sedimentation within these areas and reduced mixing between (clear) water from vegetated areas and (turbid) water from non-vegetated areas. It is hypothesized that expansion of the Chara meadows might ultimately result in a shift of the whole lake to a long-lasting clear water state. In order to promote the Chara, the fish stock reductions, which aimed at a spring dear water phase, should be continued.

Guiding lake restoration and management (chapter 7)
The relationships between nutrient loading and shallow lake response are complex and tend to be different for the process of eutrophication and the reverse process of oligotrophication. The bottom line is that both clear water lakes and turbid water lakes resist changes in nutrient loading, showing the phenomenon of hysteresis. In shallow lakes of moderate productivity, the submerged vegetation plays a key role in stabilizing the dear water state. In turbid water lakes, algal blooms and particularly Oscillatoria blooms (resistant to low P and low light, reduced edibility for Daphnia ), are self- reinforcing and therefore resistant to restoration efforts. Additionally, resuspension of sediments by wind and benthivorous fish may contribute to the stability of the turbid water state. Exceptional weather conditions or special actions ('switches') may trigger a shift from the one state into the other. The external nutrient loading should first of all be reduced in lake restoration. Furthermore, negative trends such as pollution with chemicals, toxic to Daphnia or macrophytes, should be reversed. The rehabilitation of lake-marginal wetlands will also contribute to a sustainable clear water state. However, as the turbid water state is extremely stable, additional measures may be necessary. Certain 'blockages' have to be removed to get the recovery process started:

(1) the Oscillatoria bloom

After external loading reduction, eliminating the Oscillatoria bloom should be the first priority in lake restoration. For many lakes it will be difficult to achieve TP < 20-50 mg m ^-3, which is needed for a collapse of the bloom. Therefore, additional switches are needed. Washout by flushing during winter is a promising tool (flushing with at least three times the lake volume, during November- February).

(2) the bloom-mediated P release from the sediments

Manipulation of the carbonate system through flushing with water low in TP, but rich in Ca ²⁺and HC0 ₃^-, was successful in Veluwemeer. Sediment removal in the Dutch peat lake Geerplas failed to produce the desired results. Obviously, dredging is no panacea for solving problems of P release in shallow Dutch lakes. It seems wise to focus on external loading reduction and more appropriate in-lake measures.

(3) the abundance offish, preventing Daphnia and macrophytes from developing

After fish stock reduction (>75%) and clearing of the lake water, rapid colonization of submerged macrophytes has been demonstrated in small lakes (< 30 ha). In large lakes and particularly in networks of interconnected lakes the fish stock is more difficult to control. Additionally, in large lake areas the reestablishment of submerged vegetation takes more time. Winter fishing on a regular basis, rather than a single fishing operation, may be promising then. However, case studies are needed for further evaluation of the efficacy of repeated fishing for promoting Daphnia. Prerequisites for successful biomanipulation are low TP levels (TP < 100 mg m ^-3, or higher TP levels for very small lakes) and low numbers of inedible algal species, such as 0scillatoria.