Transactions on Ecology and the Environment vol 14, © 1997 WIT Press, www.witpress.com, ISSN 1743-3541

Contribution of hydrodynamic and limnological modelling to the sanitation of Lake Bled Mitja Rismal, Boris Kompare, Rudi Rajar Faculty of Civil and Geodetic Engineering, University of Ljubljana, Hajdrihova 28, 57-1001 Ljubljana, Slovenia, P.O.Box 3422 E-mail: [email protected]

Abstract Lake Bled is an alpine pearl in the north-western part of Slovenia. In this century eutrophication has progressed rapidly, endangering this previously beautiful oligotrophic lake and the tourist economy of Lake Bled region. From the 1950's several sanitation measures have been proposed and undertaken. This paper presents basic facts about the lake and modelling approaches undertaken to define the proper restoration measures. Several models were used, beginning with the simplest Vollenweider model, continuing with Imboden's two box steady-state model, progressing to a dynamic three box model, then modelling of circulation by a 2D and later by a 3D hydrodynamic (HD) model. Together with measurements of tracer dispersion the two HD models showed the basic pattern of the circulation and the mixing of inflowing water. Finally the whole problem was tackled with the aid of artificial intelligence tools (the latest approach by Kompare et al. [4] is shown elsewhere in these Proceedings). All the three eutrophication models shoed that at present we cannot expect any amelioration of the trophic state of the lake unless we drastically cut the input of nutrients to the lake. The last two models, and the dynamic one in particular, show the benefits and drawbacks of the introduction of artificial flushing of the lake with water from a nearby river, and the positive effects of the outflow of hypolimnetic water through a syphon pipe. The needed decrease of point and non-point sources of pollution with phosphorus to achieve an economically feasible mesotrophic state of the lake is also calculated. The possible negative effects of the syphonic outflow on the environment were foreseen and predicted to be negligible, which was also demonstrated after the construction of the syphon. 1 Introduction Lake Bled is a typical alpine lake. In the present century it has severely suffered from natural but above all from anthropogenic eutrophication. The inflows are very small (the exchange time being about 3 years), the encompassing watershed is urbanized, and there is marked flushing of nutrients from agricultural land and the forested part of the watershed. Besides external

126

Transactions on Ecology and the Environment vol 14, © 1997 WIT Press, www.witpress.com, ISSN 1743-3541

Water Pollution

factors, accelerated eutrophication is caused by internal ones, i.e. the morphological characteristics of the lake. This relatively shallow lake with a small hypolimnion is also protected from strong winds by a rim of hills Therefore, most of the year the lake is strongly stratified, showing anoxic conditions at the bottom and also in a great part of the hypolimnion. Anthropogenic influences are the main causes of eutrophication. Urbanisation and intensive agriculture in the watershed have increased flushing of nutrients from agricultural and urban land, compounded by outflows of cess-pits, leakage of sewers, and combined sewer overflows (CSO's) during heavy rainfalls. In 1955 a commission was established for sanitation of the lake. It elaborated a priority action plan which contained the following measures: 1. Sanitation of the Bled sewerage system All users should be connected to the sewerage system, cess-pits should be abandoned, inadequate sewers should be improved (present problems: exfiltration of sewerage into the lake and infiltration of the lake water into the sewers). 2. The mean residence time of the lake should be shortened by artificial surface flushing of the lake. For this purpose the introduction of cca. 3.0 mVs of fresh water from another watershed was proposed. Cold (cca 6 °C), oxygen rich, and relatively unpolluted water from the river Radovna was planned to be conducted into the lake during the spring and autumn turnover, to promote better mixing and oxygen uptake in the hypolimnion. 3. Oxygenating of hypolimnetic water by pumping and spraying to the surface was rejected as it does not guarantee a permanent solution. 4. The construction of a hypolimnetic syphon was also rejected, although this measure could have completely removed the totally anaerobic monimolimnetic layer at the bottom. It was the opinion of the commission that this solution does not remove the causes of the problem, but merely the consequences - in contrast to the flushing of the lake with Radovna. A diversion aqueduct for cca 3.0 mVs offreshwater from the Radovna river, (10-times the natural inflow to the lake), was first constructed. Due to constraints (loss of energy) at a hydroelectric power plant on the Radovna river downstream, the foreseen water diversion was reduced later to some 300-500 1/s. At the same time nothing was done to reduce the input of nutrients into the lake from the hinterland. For the reconstruction of (the already well developed) sewer system, more design work was needed and also alternative solutions were proposed. This is the main reason that the reconstruction of sewers is still not completed. The described measures, which were never fully realised, have not produced the desired effects. In the beginning, the oxygen balance was improved by the greater inflow of the Radovna, but after the reduction of inflow (due to the constraints of the hydroelectric power plant) the anaerobic layer returned to its original position. These newly developed conditions required another analysis of possible restoration measures taking into account the changed conditions and new knowledge in limnology.

Transactions on Ecology and the Environment vol 14, © 1997 WIT Press, www.witpress.com, ISSN 1743-3541

Water Pollution

127

For this reason we proposed to tackle the abatement of the lake eutrophication with modern modelling tools which enable cost- and timeeffective simulations of different engineering measures in lake management. 2 Basic data on Lake Bled The lake lies on the eastern edge of the Julian Alps and is surrounded by hills. The average elevation of the surface is 475.88 m a.s.l. Extensive hydro-geomorphological measurements of the lake werefirstmade by Sketelj & Rejic [16, 17]. Its main characteristics are summarized in Table 1.

parameter surface area volume original watershed watershed with Misca watershed with Radovna maximal depth of western pool maximal depth of eastern pool mean depth depth of epilimnion depth of hypolimnion annual mean inflow/outflow mean annual precipitation Table 1.

units km% 106 m3 km? knf km2 m m m m m 1/s mm

value 1. 438 25. 69 4 .87 8. 97 107. 03 30. 20 24 .80 17. 90 6 - 10 14 - 9 272/ 330 1614

Hydromorphological characteristics of Lake Bled

The detention time of the lake is approx. 3 years. The volume of the hypolimnion is too small compared to the epilimnion, i.e. the hypolimnion is only 25-48 % of the lake's total volume, which is not enough for it to maintain oxic conditions through the whole year. The size of the contributing area has changed significantly several times. From the original watershed of 487 ha (147 ha lake, 170 ha urbanised and agricultural, and 170 ha forest) the diversion of the Misca, a stream passing nearby, has added an additional 410 ha, making the watershed grow to 897 ha. This diversion was made in the previous century for energy purposes (mills) In 1964, another diversion was introduced as a restoration measure, this time from the river Radovna, which has increased the contributing area 12-times compared to the former watershed (with the Misca), and 22-times compared to the original natural watershed (see Fig. 1). The increase of watershed has caused also increased nutrients' loads and thus acceleration of eutrophication processes. The transparency varies significantly during the year. In March/April algal blooms aggregate mainly on the surface and transparency is of the order of 1.0 -

Transactions on Ecology and the Environment vol 14, © 1997 WIT Press, www.witpress.com, ISSN 1743-3541

128

Water Pollution

2.5 m; it increases to cca 4.0 m in June, and falls to its minimum in August (0.5 - 1.5 m). During autumn and winter it can be as high as 9.5 m [2, 20].

410 ha artificially added watershed by the diversion of M/SCA creek

diversion of RECICA into MISCA creek

340 ha natural contributing area to the lake

Figure 1. Growth of the watershed of Lake Bled

contribution from Misca river Zaka river immediate watershed sewer M /part of Bled lake surface Radovna river total

min. load 61 17 19 8 37 94 236 kg P/year

max. load 290 64 97 20 265 94 830 kg P/year

taken into account 290 40 58 14 152 94 648 kg P/year

Table 2. Estimation of phosphorus loads to the lake in

Transactions on Ecology and the Environment vol 14, © 1997 WIT Press, www.witpress.com, ISSN 1743-3541

Water Pollution

129

It was shown ([6, 13, 14, 18]) that phosphorus is the nutrient controlling eutrophication. The ratio between N^ and P^ in Lake Bled is 80, far beyond 17, where phosphorus becomes the limiting nutrient. In recent decades the concentration of total phosphorus P^ was mostly over 50 j^g/1, i.e. the lake was in a hypereutrophic state. A summary of measured and estimated phosphorus loads is given in Table 2. 3 Determination of circulation by measurements and hydrodynamic modelling The basic circulation in alpine lakes of this type is well known: in spring thermal stratification begins to form, which becomes stronger in summer. In late autumn the cooling of the epilimnion causes destratification and subsequently the turnover of the lake water, which can last for the whole winter if there is no ice cover. Through the whole year winds are relatively weak in the lake region and experience, as well as several simulations, have shown that in summer the winds cannot overcome the strong stratification and can never cause mixing to the bottom. A more detailed knowledge of the hydrodynamic (HD) circulation helps to determine: (a) the pattern of transport-dispersion of nutrients in the lake; (b) the circulation of surface inflow water from the Radovna river and of the syphonoutflow, which also helps to understand the efficiency of the two pipeline systems, and (c) the optimal locations for measurements. A combined methodology of measurement and HD modelling was used to determine the circulation. As both the measurements and the 2D modelling have already been described by Rajar and Cetina [7], only a short description is given below. Measurements. In greatest part of the lake the circulation velocities are very small, below Icm/s (only at the surface can the wind cause local velocities up to about 30 cm/s). It would be very difficult to measure the flow velocities directly. Therefore two measurement with tracers (Rhodamine B and Uranine) were carried out. The tracer was introduced into the inflow pipeline from the Radovnariver,and the concentration of the tracer was measured over the whole lake for the next few days ([5]). The results were used as an indirect method of verification of the hydrodynamic model. Modelling. As the measurements have shown that the water flowing to the lake from the Radovna river is stabilised in a horizontal layer, 2D modelling was initially performed in 1986 [7]. From 1995 on research is going on with the goal of a complete 3D simulation of the circulation. A fully 3D, baroclinic hydrodynamic model (described in Rajar and Cetina [8]) is being applied. Description of the Circulation. Measurements with tracer were carried out with two discharges of the Radovna inflow: Case A with 200 1/s, and Case B with 600 1/s. Both cases were carried out in June with strong stratification of the lake water and with negligible winds.

Transactions on Ecology and the Environment vol 14, © 1997 WIT Press, www.witpress.com, ISSN 1743-3541

130

Water Pollution

Fig. 2 shows the results of measurements for Case A. The temperature of the inflowing water was 8-10 °C, and the temperature distribution of the lake water in Fig. 2b makes clear that the Radovna water finds its equilibrium density in the depths between 10 and 14 metres. The velocity of horizontal spreading of the tracer is about Icm/s during thefirstday, but it decreases to about 2 mm/s by the fourth day. A vertical cross-section (Fig. 2b) shows that during the following days there is almost no mixing of the Radovna water with the lake water. The effect of this facility is indeed limited to the layer of the same density/temperature as the inflowing water. The mixing is probably stronger during the winter and early spring circulation, when there is almost no stratification. But during that time of the year there is a turnover of the lake water, and oxygenation of the bottom water is enabled even without the extra inflow of the Radovna.

0 SMOmg/m* D O.S-Smg/m*

Location of the front at the time f —*- Flow direction

Figure 2: Results of tracer measurements in Lake Bled a) Spreading of the tracer; b) Vertical cross section During Case B, with a discharge of 600 1/s, the mixing of the Radovna water with the lake water was somewhat stronger. The thickness of the layer with the

Transactions on Ecology and the Environment vol 14, © 1997 WIT Press, www.witpress.com, ISSN 1743-3541

Water Pollution

131

tracer was about 8 metres. But another disadvantage of the surface inflow appeared: the jet hit the bottom causing some resuspension of nutrient-rich sediments. Subsequently, this would enrich the lake waters with phosphorus and its compounds, causing increased primary production, with eventual algae blooms and increased eutrophication. For this reason it was decided that the inflow discharge should always be kept below 500 1/s. The measurements were not able to show the effectiveness of the syphonoutflow directly, but a simple analysis of the phenomenon, based on the mass conservation law, has shown that the combined effect of the surface inflow and the bottom outflow is positive: the water of the worst quality at the bottom of the lake is drained out of the lake, thus contributing to its sanitation. Fig. 3 shows the simulated circulation for Case B (inflow 600 1/s) in the horizontal layer at the depth of 10 to 11.5 m The outlets of both inflow pipelines are at a depth of 18 m, but the 3D model properly shows that the Radovna water stabilises at depths of about 10 to 18 m. A vortex above the outlets is formed because the discharge through the eastern outlet is about 3 times greater than through the western one. The study on 3D hydrodynamic modelling is not yet fully completed. The present results show too much dispersion in the vertical direction. It has been found that a finer numerical grid and a more accurate turbulence model should be used. This will be carried out in the continuation of the research.

WEST

LENGTH Q VELOC. 0

SCALE 250 0.05

500 rr O.'O^-

Figure 3: Velocities in the layer at the depth of 10 to 11.5m, simulated by 3D HD model (The outlets of the two inflow pipes are at a depth of 18 m.) 4 Elaboration of sanitation measures with the use of limnological models Because former remedial actions for restoration of water quality were ineffective, in 1979 the Institute of Sanitary Engineering (IZH) was given the task of proposing variant sanitation measures for Lake Bled, to evaluate them in

Transactions on Ecology and the Environment vol 14, © 1997 WIT Press, www.witpress.com, ISSN 1743-3541

132

Water Pollution

the sense of their contribution to water quality, and to elaborate a list of priority tasks. In these studies ([9, 10]) several methods of lake sanitation were taken into consideration, i.e.: - Reduction of nutrient input into the lake. - Surface flushing with water from the river Radovna. - Deep-water flushing via syphoning (outflow) of hypolimnetic water. - Destratification of the lake and aeration of the anaerobic layers with air jets or with aerator shaft pumps. - Chemical precipitation of nutrients in the lake. From thefinancialpoint of view the last two methods are the least attractive also they do not offer a permanent solution of the problem So they were not advised as feasible, unless in an emergency. Thefirstthree methods were found to be feasible, each by itself and in combination. So we focused our research on these three methods [9, 10]. Each of the three methods has its advantages and disadvantages. It is not possible to totally cut-off the import of nutrients into the lake, but only to reduce them to some extent. Surface and/or deep-water flushing with water from the river Radovna has limited capacity. The only practical way to adequately evaluate the contribution and efficiency of the listed methods was mathematical modelling of the whole ecosystem, i.e. the lake and the contributing area. In 1980 we used a simple Vollenweider model [19] as afirstorientative approach. The practical value of this model is its simplicity and the need for very few data. With this model it was not possible to evaluate the differences between surface- and deep-water flushing; instead, the differences between different rates of flushing was explored, see Table 3.

Radovna

Total inflow

m^/s 0 0.228 0.728 1.228 1.728

rrP/s 0.272 0.500 1.000 1.500 2.000

Phosphorus load from Radovna kg P/year

0 107 344 581 817

Permissible p losphorus load to obtain folktwing category: oligotrophic mesotrophic kg P/year kg P/year 188 288 483 662 824

352 540 906 1 241 1 545

Table 3. Permissible load with P^f after Vollenweider's model [19] From thisfirstinsight into the behaviour of Lake Bled ecosystem it can be seen that artificial flushing cannot improve the water quality. Therefore, it is necessary to reduce the external load and maybe to combine it with artificial flushing of the lake. So we decided that a more sophisticated model is needed to

Transactions on Ecology and the Environment vol 14, © 1997 WIT Press, www.witpress.com, ISSN 1743-3541

Water Pollution

133

show the differences between the various ways and rates of flushing, and external as well as internal loading. We chose the model from Imboden [3] and extended it (Rismal [10]) to model surface and below surface inflows and outflows. This model is also among the simpler ones; we chose it intentionally so as to avoid unnecessarily complex processes, which are difficult to model even if one has perfect data, which did not hold in our case. The model is twobox and contains four equations describing the processes of photosynthesis and respiration of phyto- and zoo-plankton in the epilimnion and hypolimnion.

needed surface outflow in mVs by resuspen sion frc)m the permissible load with bottom mg P / (m2 day) Ptot 0 100 gP/(m2y) kgP/year L 10 0.20 294 2.227 0.15 1.343 220 0.10 147 0.442 0.05 73

needed deep-water outflow i.e. via syphon in mVs by resuspen sion from the bottom mg P / (m^ d; iy) 0 10 100 1.496 0.306 2.295 1.173 0.459 2.891

Table 4. Permissible load with P^ after Rismal's model [10]

Different flushing and resuspension rates of P from the sediment were calculated with an extended model of Imboden, to evaluate the possible behaviour of the lake within the real limits of controlling mechanisms. The results of the model show clearly the regions of prevalence for surface- and for deep-water flushing; see Table 4. N.B.: Values are only entered for feasible combinations, i.e. surface flushing is only effective with zero resuspension from the bottom (resuspension of 100 mg P/(nfday) was estimated to be the highest expected value). From Table 4 it can be seen that deep-water flushing has a great advantage over surface flushing, e.g. for obtaining the load of 220 kgP/year four times more water is needed for surface flushing. Also surface temperature would be decreased by intensive surfaceflushing,which renders the lake less attractive for tourists. It can also be seen from the table that flushing alone does not significantly increase water quality in the lake. The reason for this is the transport of nutrients along with the inflows. The only way to improve water quality is then simultaneous use of deep-water flushing and reduction of the external load. In the next step we decided to evaluate the predictions of the extended Imboden model with a more accurate and also time dependent (dynamic) model. In the years 1990-92 we developed a dynamic model according to Griffin and Ferrara [1]. The model simulates time-dependent balances of oxygen and

Transactions on Ecology and the Environment vol 14, © 1997 WIT Press, www.witpress.com, ISSN 1743-3541

134

Water Pollution

phosphorus in the lake in accordance with external and internal forcing functions. The model could not be quite properly calibrated and verified because of an insufficient data base. We had to use literature data for some parameters. Nevertheless, the model qualitatively correctly displays the expected behaviour. Therefore we took the results of the model to be decisive for the definitive choice of the optimal sanitation procedure. A series of simulations was conducted with variant combinations of locations and rates of inflows and outflows over a period of several years (Rismal & Cvikl [15]). Here we would like to describe only the extreme cases, the worst and the best: (1). The worst case is the situation before the construction of the syphon (before 1980). There is only the Misca river inflowing to the lake (no diversion of the Radovna river), and there is only surface outflow (the syphon is not in operation). From Fig. 4 two peaks in organic phosphorus, due to algal growth, can be clearly seen in early spring and in autumn, and an extremely low concentration of oxygen in the whole hypolimnion in the autumn. Anoxic conditions are developed at the bottom over almost the whole year (in particular if we take into account that biologically anoxic conditions take place at concentrations less than 1.0 mg Oi/l). (2). The best situation is when the only inflow (besides the immediate watershed) into the lake is the deep inflow of the Radovna river (200 1/s) and the syphon is working at full capacity, see Fig. 5. The model shows that the water quality significantly improves already in two years; in five years a great reduction of biomass concentration can be expected as well. This indicates that primary productivity as a measure of trophic state is decreasing. At the same time the concentration of oxygen in the hypolimnion is increasing and the periods of anoxia at the bottom are getting shorter. The model also shows the drop in photo synthetic production of oxygen in epilimnion due to decrease of biomass. Oxic conditions at the bottom at the same time prevent resuspension of phosphorus, thus drastically reduce internal load and speed-up the processes of the lake restoration.

5 Proposed and accomplished sanitation measures According to the results of ourfirstmathematical models of the lake, and based on a solid background in limnology and related sciences, a priority list of important conclusions and reasonable restoration measures was proposed [9, 10, 11, 12], as follows: 1. Successful and rational restoration of the lake (regarding operation, investment and operational costs) is only possible by the use of adequately sophisticated and calibrated mathematical models. All the data needed,

Transactions on Ecology and the Environment vol 14, © 1997 WIT Press, www.witpress.com, ISSN 1743-3541

Water Pollution

135

(mg/m3)

10.00

_Porg,Epi

9.00

-Porg.Hypo

8.00 7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00

days 1 61 121 181 241 301 361 421 481 541 601 661 721 a) Concentration of organic phosphorus (mg/m3)

8000 7000 6000 5000

2 O u

30002000

nSTHpi— O.Hypo 0,d

days 1 61 121 181 241 301 361 421 481 541 601 661 721 b) Concentration of oxygene Fig. 4: Results of the dynamic model for the worst case especially chemical and bio-chemical analyses, their sites and frequencies of sampling etc., must be strictly subordinated to the needs of the model. 2. First step in the approach to sanitation of the lake is to state the limiting factor of eutrophication. In our case it is phosphorus. This limiting factor determines all other planned activities, i.e. the search for the main sources of pollution, the postulation of feasible restoration techniques and the evaluation of possible short- and long-term effects.

Transactions on Ecology and the Environment vol 14, © 1997 WIT Press, www.witpress.com, ISSN 1743-3541

Water Pollution

(mg/m3)

10.00

Porg.Epi

9.00

Porg.Hypo

8.00 7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00

1

181 361 541 721 901 1081 1261 1441 1621 1801 a) Concentration of organic phosphorus d*Y*

(mg/m3) 8000 7000 6000 5000 4000 3000 2000 1000

1

181 361 541 721 b) Concentration of oxygene

901

1081 1261 1441 1621 1801 days

Fig. 5: Results of the dynamic model for the best case (foreseen to be implemented) 3. The total annual phosphorus load to the lake is soundly estimated to be 648 kg, of which the Misca river alone brings almost one halt i.e. 290 kg The part which cannot be reduced, i.e. input on the surface and from the immediate contributing watershed, is 250 kg. Modelling and other analyses showed that the optimal priority tasks list is as follows:

Transactions on Ecology and the Environment vol 14, © 1997 WIT Press, www.witpress.com, ISSN 1743-3541

Water Pollution

137

3.1 The largest pollutant, the Misca, can be relatively easily excluded by diversion to its original course by passing the lake (and not flowing into the lake). This is also not questionable because the Misca no longer serves as a power source for mills. The other major sources of nutrients cannot be so easily managed. So the next feasible step is: 3.2 Construction of the hypolimnetic siphon, as the model showed its prevalent advantage among the other possible sanitation measures. 3.3 Sanitation of sewers, which partly leak sewerage to the lake, and partly drain water from the lake; connection of all urbanised areas to the sewerage system. 3.4 Restoration of combined sewer overflows (CSO's) to decrease the quantities of mixed sewerage and rainfall-runoff diverted to the lake during rainy periods down to an acceptable 8 kg Ptt for an averagely wet year. For this purpose retention basins were designed (Rismal, [12]) to avoid overflows at rain intensities smaller than 40 l/(s*ha). 4. The results of the models and practical experience with functioning of the syphon since 1980 have shown: 4.1 Permanent flushing of concentrated nutrients and products of decomposition via the syphon decreases the rate of aging of the lake and can even to some extent shift the lake water from eutrophic to mesotrophic or even oligotrophic state. 4.2 The syphon positively changes the thermal balance of the lake, whereas the introduction of cold water from the Radovna river into the hypolimnion along with surface flushing has the opposite effects. 4.3 Surface flushing with as much as 3.0 mVs of Radovna water did not flush out algae, as was expected. The residence time of approx. 3 months with this flushing is still far too long compared to the time of growth of algae of approx. 7-14 days. 4.4 The net positive effect of the Radovna river is limited to flows between 0.2 and 0.3 nf/s, or slightly more if the Misca river is not flowing into the lake. Namely, greater inflows also mean a greater input of nutrients, which cannot be balanced out by increasedflushingrate. Unfortunately, for several reasons, not all of the proposed sanitation measures have been accomplished. The syphon pipeline has been successfully constructed (first pipe in 1980, second pipe in 1982). The main pollutant, the Miscariver,is still flowing into the lake. The sewer system was renovated to some extent, but the retention basins were not constructed as was proposed. 6 Conclusions With several mathematical models of Lake Bled ecosystem and with a synthesised knowledge of limnology and sanitary engineering we succeeded in elaborating step by step (from simple to complex) a sound priority task plan for the lake's sanitation. According to this plan, which proposed a combination of sanitation measures to be the most optimal, only one measure was completely

Transactions on Ecology and the Environment vol 14, © 1997 WIT Press, www.witpress.com, ISSN 1743-3541

implemented, i.e. the construction of a syphon to divert anaerobic and nutrientrich hypolimnetic water. This single measure has given significant and easily observed results even in itsfirstyear of operation. It was concluded from the results that since the beginning of the operation of the syphon (begun in 1980) the concentration of phosphorus has dropped by more than a factor of two. The transparency also increased and concentration of oxygen in the hypolimnion increased too. In the past fifteen years blooms of Oscilatoria rubescens, which is the sign of developed eutrophication, were seldom seen, which is a great contrast to their previous abundant appearance each year [2]. The effectiveness of the syphonis also evident to the naked eye of the inhabitants of Bled. The predictions of models, based on feasible reductions of external loads of phosphorus, have shown that our expectations were correct. Because of the lack of quality data, the dynamic model could not have been properly calibrated, so the quantitative predictions might not be completely valid; but the results of the syphonoperation over more than 15 years have given us confidence in the model. The potential negative effects of the syphon outlet on the environment (smell, toxic compounds of anaerobic digestion, eutrophication and oxygen depletion of the receivingriver,etc.) were foreseen [11] to be negligible, which has also been proven through experience. We hope that the other proposed sanitation measures will also be implemented in the near future, which will make it possible for the lake to retain its mesotrophic status and maybe even to regain oligotrophic conditions.

7 References 1. 2.

3. 4.

5.

Griffin, T.T. & Ferrara, KA. A Multicompartment Model of Phosphorus Dynamics in Reservoirs, Water Res. Bull., 1984, pp. 777-788. HMZ. Research on water quality in Slovenia, Hydrometeorological Institute of the Republic of Slovenia (HMZ). Research publications for years 1976 to 1991 (in Slovenian). Imboden, DM Phosphorus Model of Lake Eutrophication, Limnology and Oceanography, March 1974, Vol. 19 (2). Kompare, B , Dzeroski, S. & Karalic, A Identification of the Lake Bled ecosystem with artificial intelligence tools M5 and FORS, Proceedings of the Water Pollution 97. Bled, Slovenia, 18-20 June, 1997. Leibundgut, C, Moeri, Th., Peschel, H. Stroemungsuntersuchungen Mitt els Tracerversuchen im Bledsee, Report, Geographisches Institut der Universitaet Bern, 1983.

Transactions on Ecology and the Environment vol 14, © 1997 WIT Press, www.witpress.com, ISSN 1743-3541

Water Pollution

139

6. Loffler, H, Sampl, A. Stellungnahme zu den Restauration - Massnahmen Blejsko jezero, Osterreichische Akademie der Wissenschaften, 1982. 7. Rajar, R. and Cetina, M. Mathematical simualtion of two-dimensional lake circulation, Proceedings, HYDROSOFT, Southampton, 1986, Elsevier Publ. 8. Rajar, R., and Cetina, M. Hydrodynamic models as a basis for water quality modelling: An experience. Ecological Modelling, accepted for publication, 1997. 9. Rismal, M. A study of the syphon for the sanitation of Lake Bled, report, FAGG - IZH, Ljubljana, 1979. (in Slovenian) 10. Rismal, M. An assessment of diverse sanitation methods for Lake Bled, Gradbeni vestnik, Ljubljana 1980 (29), 2-3, pp. 34-43. (in Slovenian) 11. Rismal, M. An assessment of negative influences of the syphon outflow from Lake Bled on the environment, Gradbeni vestnik, Ljubljana, 198la (30), pp. 51-54. 12. Rismal, M. The influence of the sewerage of Bled on the pollution of the lake with nutrients, Vodoprivreda 13, 73 (1981/5), 1981b, pp. 383-393. (in Croatian) 13. Rismal, M., An assessment of sanitation measures for Lake Bled and the obtained results, Vodoprivreda 14, 75-76 (1982/1-2), 1982a, pp. 9-14. (in Croatian) 14. Rismal, M., The use of limnological models for the analysis and mitigation of eutrophication in lakes and impoundments, Vodoprivreda 14, 78-79 (1982/4-5), 1982b, pp. 391-393. (in Croatian) 15. Rismal, M. & Cvikl, M. Dynamic limnological model of Lake Bled, Report, University of Ljubljana, FAGG - IZH, Ljubljana, 1991. 16. Sketelj, J. in Rejic, M. Preliminary report on research on Lake Bled, Gradbeni vestnik 61-64, 1958-59. (in Slovenian) 17. Sketelj, J in Rejic, M. Physical, chemical and biological research on changes in Lake Bled (study). Report on the integral monitoring during Dec. 14-16, 1966, FAGG - IZH, Ljubljana, 1967. 18. Stauffer, R. Summary of Discussions on Lake Bled, Report, Ms, 1982. 19. Vollenweider, R.A. Advances in defining critical loading levels for phosphorus in lake eutrophication. Mem. Inst. Ital. Idrobiol, 1976, 33: 53-86. (in Italian) 20. Vrhovsek, V. Primary productivity of eutrophic systems (Lake Bled), Annual reports for 1976-1991. (in Slovenian)