Aarhus Universitets segl

No. 636: Restoration of lakes in Denmark

Liboriussen L., Søndergaard M. & Jeppesen E. (editors), 2007. NERI technical report no. 636. Part I, 88 pp. Part II 312 pp.

 

Summary


During the past 20 years several restoration projects have been conducted in Danish lakes to improve water quality. Particularly during the 1990s restoration measures were applied to 5-25 lakes annually. Today, more than 80 lakes have been subjected to restoration.

 

Most restoration initiatives have involved stocking of predatory fish (mostly pike) and removal of planktivores (particularly roach and bream). These two types or combinations of these constitute 75% of all conducted restoration projects. The remaining measures include removal of nutrient-rich sediment, oxidation of bottom water in deep lakes, addition of aluminium and protection of submerged macrophytes. Moreover, a number of interventions of a more experimental nature have been made; however, as the number of these is limited any general conclusions cannot be drawn.

 

The results from lakes where more than 200 kg planktivorous fish per hectare have been removed within maximum three years show a typical reduction of 20-50% in the concentrations of chlorophyll, total phosphorus, total nitrogen and suspended matter within the first 8-10 years following the intervention. The most enduring effects were observed for Secchi depth and the level of suspended matter. The effect on the water chemical parameters varied with the season during the year, but the most significant changes were observed in summer and in the most nutrient-rich lakes. In <st1:PlaceType w:st="on">Lake</st1:PlaceType>  <st1:PlaceName w:st="on">Væng</st1:PlaceName> and <st1:place w:st="on"><st1:PlaceType w:st="on">Lake</st1:PlaceType> <st1:PlaceName w:st="on">Arreskov</st1:PlaceName></st1:place> , where the internal phosphorus loading declined following fish stock manipulation, a return to a turbid state was observed and, with it, increased internal loading and enhanced summer concentrations of phosphorus. This indicates that a mobile phosphorus pool in the sediment may pose a risk for a return to the turbid state for a long time after the restoration event.

 

In lakes where more than 200 kg planktivororus fish were removed per hectare, a decline was observed in total phytoplankton biomass. The summer biomass was reduced by ca. 50%, and especially the share of bluegreen algae declined markedly. In contrast, the number of cryptophytes increased – also in absolute figures, which may be interpreted as an enhanced grazing pressure by the zooplankton. The relative abundances of other algae groups did not show any variations. The zooplankton biomass did not change significantly following the fish removal, except from an increase in the number of large-sized Daphnia at the cost of smaller cladocerans. This indicates a reduced predation pressure from fish as does the observed increased ratio between the biomass of zooplankton and phytoplankton.

 

The number of fish investigations conducted as a follow-up to fish removal is limited, but the existing results suggest that the fish stock does not exhibit any general changes in either total number or fish biomass. However, the species composition changes markedly, and it seems that bream is the species most substantially influenced by the interventions. Also the number of roach declines during the first year following the intervention after which it rises again; this may be one of the reasons why the effect on zooplankton and subsequently on chlorophyll a does not last. The number and the biomass of perch increase at first, only to decline again in most lakes.

 

Only few examples describe the effects on other biological variables. In many lakes, the abundance of submerged macrophytes increases after restoration, but absence of changes, i.e. lack of response by plants to the increased Secchi depth, also occurs. In some cases this may be due to waterfowl grazing. In the lakes exhibiting changes in macrophyte abundance, the density and number of various waterfowl species increase. Data on bottom animals is scarce, but the information available depicts an increase.

 

In general, fish removal experience shows that a number of positive effects are to be gained from massive removal, but that the effects may be difficult to maintain unless removal is repeated. The effects are most marked on Secchi depth and the level of suspended matter, as these seem to last 10 years or more in most lakes. In contrast, in most cases and typically 8-10 years following the intervention, no statistically significant effects of removal can be traced for chlorophyll a and nutrients. The number of lakes with data available for >10 years is limited, however. The more substantial impacts on suspended matter and Secchi depth probably relate to the fact that particularly the stock of bream declines following removal and remains low, entailing reduced stirring of bottom material. The relapse observed in most lake indicates that removal should be repeated to maintain the improvement. However, the follow-up removals may be of more moderate nature, as the bream stock, on weight basis, seems to be long term affected and as the number of piscivores and the size of perch and roach increase after manipulation. In lakes with fish removal <200 kg ha-1 within three years, no or only negligible effects were observed on the water chemical variables.

 

The effect of pike stocking has been estimated in a report elaborated by the Danish Fisheries and Research Institute, and based on data material on the impact of stocking in 47 lakes it was concluded that pike stocking almost always is of no significance for the water quality. Stocking of pike fry as a singular restoration method is therefore not recommended.

 

Oxidation has substantial effects on the accumulation of phosphorus in the bottom water, probably due to the redox-sensitive binding of phosphorus in the sediment. In most cases the accumulation in bottom water during summer was more than halved and also the accumulation of ammonium declined. Two examples involving cessation of oxidation for one year show increasing phosphorus concentrations in the bottom water. Thus, oxidation is a long-term project if reduced release of phosphorus is to be maintained, and whether the method can be used to create permanent effects remains to be elucidated. Increased temperatures and turnover of organic matter in connection with the oxidation may furthermore enhance the risk for the creation of a larger pool of mobile phosphorus in the sediment.

 

Experience with aluminium treatment of Danish lakes includes only 6 lakes, 5 interventions have been conducted within the last 6 years and 3 within the last 2 years. The potential of the method is therefore not yet clear. In all projects the addition of aluminium has had an immediate and significant clearing-up effect, leading to markedly reduced nutrient concentrations. However, the long-term effects are still to be elucidated, indications show signs of a relapse to the turbid state in Lake Sønderby Sø, the first aluminium-treated lake in autumn 2001. There is still doubt, though, whether the external loading to Lake Sønderby Sø was adequately reduced.

 

<st1:place w:st="on"><st1:PlaceType w:st="on">Lake</st1:PlaceType>  <st1:PlaceName w:st="on">Brabrand</st1:PlaceName></st1:place> is still the only large-sized lake from which nutrient-rich sediment has been removed with the aim to diminish internal phosphorus loading. The impact on lake water clarity has remained insubstantial due to a still too high external loading of nutrients. Sediment removal has been successfully used in a number of smaller lakes where growth conditions for large isoetids have been improved.

 

The cost of the various interventions varies considerably, but fish removal is the cheapest with an average cost of 13,000 DKK per hectare lake. The next cheapest method is oxidation with an average cost of 20,000 DKK per hectare lake, followed by aluminium treatment with an average cost of 33,000 DKK per hectare lake. However, comparison of costs between the different restoration types is rendered difficult by the fact that the estimations do not necessarily cover identical parameters, and that the effects and period of an intervention or repeated interventions are not considered.

 

Our conclusion is that in most cases an improvement of lake water quality can be obtained by restoration; however, various factors must be taken into consideration such as, for instance, how long time the effects can be maintained. The nutrient content still plays a decisive role for the water quality, and restoration intervention cannot be used as an alternative to a reduction of the external nutrient loading – particularly not if long-lasting effects are to be achieved.

 

The results from the many restoration interventions are based on a unique collection of data, but should be interpreted with caution. The data quality varied extensively, both as to type, coverage and duration of the intervention. This has rendered treatment and cross-comparisons of data difficult; however, data treatment has been standardised to allow comparison of general tendencies. This implies that the individual lakes may exhibit a different pattern.

 

The interpretation of obtained results is further complicated by the fact that most restoration initiatives have been undertaken simultaneously with a reduction of the external nutrient loading, and also a possibly declining internal phosphorus loading should be taken into account. The effects of the restoration intervention alone may therefore be difficult to differentiate from those of the reduced nutrient loading. However, for the fish stocking method this effect has been considered by correcting for the generally declining nutrient levels observed during the past 20 years. Furthermore, in a number of lakes various interventions have been conducted simultaneously.

 

Most restoration projects have been conducted in fairly nutrient-rich lakes. Half of the 40 lakes in which fish have been removed showed initial phosphorus concentrations above 0.16 mg P l-1 and a total nitrogen concentration of >1.9 mg N l-1. Many restoration measures have been applied to lakes too nutrient-rich to allow assumption of long-lasting effects according to earlier recommendations. The effects on nutrient-poor lakes cannot be estimated due to the lack of interventions, which implies that determination cannot be made of the nutrient conditions at which restoration may be expected to entail permanent results.

 

Full report in pdf-format:

Part I (6,604 kB)

Part II (10,371 kB)