ipa project

MONTENEGRIN ACADEMY OF SCIENCES AND ARTS

IPA PROJECT

„DEVELOPMENT OF HYDROLOGICAL AND HYDRAULIC STUDY OF REGULATION OF SKADAR LAKE AND BOJANA RIVER WATER REGIME” VOLUME II

This project is implemented by the Montenegrin Academy of Sciences and Arts and Albanian Academy of Sciences and managed by the Delegation of the European Union to Montenegro and The Delegation of the European Union to Albania.

The manuscript presents results of the first year research.

Academy of Sciences of Albania

Montenegrin Academy of Sciences and Arts

This project is funded by the European Union

DEVELOPMENT OF HYDROLOGICAL AND HYDRAULIC STUDY OF REGULATION OF SKADAR LAKE AND BOJANA RIVER WATER REGIME VOLUME II

Reports given in this book were presented and approved at the Conference on IPA Project “Development of hydrological and hydraulic study of regulation of Skadar Lake and Bojana river water regime”, held on 21 June 2014 in Montenegrin Academy of Sciences and Arts, and present results of the second year work on the Project. Editor: Academician Momir Djurović, MASA

CONTENTS INTRODUCTION............................................................................................................... 9

I MODELING OF HYDROLOGICAL PROCESSES IN THE BASIN OF LAKE SKADAR AND THE BOJANA RIVER ABOUT HYDROLOGICAL MODELS – INTRODUCTION................................... 11 1. CATCHMENT PROPERTIES AND HYDROMETEOROLOGICAL DATA OF SKADAR LAKE BASIN..........................................................................12 2. MIKE SHE MODEL.................................................................................................... 17 3. WATER BALANCE ANALYSIS...............................................................................28 4. CONCLUSION............................................................................................................35

II FORMATION OF THE HYDRAULIC MODEL IN THE SYSTEM THE MORAČA RIVER – LAKE SKADAR – THE BOJANA RIVER INTRODUCTION.............................................................................................................37 1. BASIC DATA ON THE HYDROLOGICAL REGIME OF THE RIVERS MORAČA AND BOJANA........................................................ 37 2. COMPUTATIONAL MODEL OF UNSTEADY FLOW – THE MORAČA RIVER SKADAR LAKE AND THE RIVER BOJANA................................................................................................49 3. NUMERIC SIMULATIONS RESULTS..................................................................56 4. SHORT MANUAL FOR USING MIKE 11 COMPUTER PROGRAM ON THE MORAČA –SKADARSKO JEZERO–BOJANA MODEL................... 71 REFERENCES....................................................................................................................78

STUDY ON THE MODELLING RESULTS OF THE HYDROLOGICAL AND HYDRAULIC PROCESSES IN SKADAR LAKE AND RIVER BOJANA BASIN

Area:

SKADAR LAKE BASIN

Study Title:

STUDY ON THE MODELLING RESULTS OF THE HYDROLOGICAL AND HYDRAULIC PROCESSES IN THE SKADAR LAKE AND RIVER BOJANA BASIN

Project Title: Development of hydrological and hydraulic study of regulation of Skadar Lake and the Bojana river water regime

Institution:

The Montenegrin Academy of Sciences and Arts

Authors:

dr Goran Sekulić, graduate civil engineer Marija Ivković, graduate civil engineer Nikola Rosić, graduate civil engineer

Coordinator of the study development: prof. dr Goran Sekulić, graduate civil engineer

Study development period: May, 2013 – April, 2014.

INTRODUCTION The project “Development of hydrological and hydraulic study of regulation of Skadar Lake and the Bojana river water regime” is aimed at forming the aforementioned models which would become the basis for all further analysis and interventions in the surveyed area. Based on the previously prepared datebases necessary for developing the required models, a specific model has been chosen, that would best adapt both to the available data and to the desired results. After analyzing the capabilities and capacities of the available models and their references, in the application worldwide, it was decided to try to establish a model of the hydrological condition and the behaviour of the Skadar Lake basin and hydraulic condition in the system river Morača – Skadar Lake – Bojana river, applying the proven MIKE models developed by the DHI (Danish Hydraulic Institute). the models MIKE SHE and MIKE 11 were used in the specific examples and their characteristics are provided in detail in the text describing the manner of their application. The success of the modeling process must certainly be proven through a calibration process of the obtained results. This process was somewhat hampered by the specificity of the very problem, space and above all, the lack of a sufficient number of measured data, but the authors believe that the quality of the results is quite satisfactory at this level of the model. The models have confirmed certain assumptions and previous findings obtained by other methods and are a sound basis for further upgrading and testing. The formation of the structure of the model presented here is the first step towards improving the knowledge of the system and presents a challenge for future users to upgrade the model, both by new inputs and new requirements for the simulation of certain conditions, which at this point might not have been performed. In order to facilitate understanding ,the text has been divided into two sections : – Development of the hydrological model MIKE SHE on the Skadar Lake basin and – Development of the hydraulic model MIKE 11 on the Morača river – Skadar Lake – the Bojana river system In addition to clarifying the basic input data, hydrologic and hydraulic phenomena that were the basis of the study and the results obtained, the authors have attempted to give future users certain instructions and hints for using the very models through text, so that this study is also a set of specific guidelines for implementation and further development of the models presented.

I MODELING OF HYDROLOGICAL PROCESSES IN THE BASIN OF LAKE SKADAR AND THE BOJANA RIVER

ABOUT HYDROLOGICAL MODELS – INTRODUCTION Hydrological models are important for a wide range of applications, including water resources planning, development and management, flood prediction and design, and coupled systems modeling including, for example, water quality, hydro-ecology and climate. However, due to resource constraints and the limited range of available measurement techniques, there are limitations to the availability of spatial-temporal data; hence a need exists to extrapolate information from the available measurements in space and time; in addition there is a need to assess the likely hydrological impact of future system response, for example to climate and land management change. In this project hydrological modeling is applied to verify the hydrological process understanding developed during the study period. Several problems arise when worked with hydrological models, such as data availability (especially when working with physically based models), spatial diversity between observed parameters and model parameters, differences between hydrological process scales and modeling scales. The choice of a model is determined by the purpose of the model and the availability of data. There are numerous watershed-scale hydrologic models. The choice of models should be based on the objectives of use. Literature reviews suggest that the MIKE SHE⁄MIKE 11 modeling package (DHI, Hørsholm, Denmark) has several advantages over other hydrologic models for estimating watershed runoff: (1) it is a distributed model and most of the algorithms in describing the water movements are based a physical processes, (2) it simulates the overland ﬂow processes commonly found in dry regions, and (3) it has been commercialized and a GIS user interface was built in the system that can directly use spatial GIS databases for model inputs. Also, the model has a strong visualization utility that makes interpretation of modeling outputs much easier MIKE SHE covers the major processes in the hydrologic cycle and includes process models for evapotranspiration, overland flow, unsaturated flow, groundwater flow, and channel flow and their interactions. Each of these processes can be represented at different levels of spatial distribution and complexity, according to the goals of the modeling study, the availability of field data and the modeler’s choices. The MIKE SHE user interface allows the user to intuitively build the model description based on the user’s conceptual model of the watershed. The model data is specified in a variety of formats independent of the model

12 domain and grid, including native GIS formats. At run time, the spatial data is mapped onto the numerical grid, which makes it easy to change the spatial discretization. MIKE SHE, in its original formulation, could be characterized as a deterministic, physics-based, distributed model code. It was developed as a fully integrated alternative to the more traditional lumped, conceptual rainfall-runoff models. A physics-based code is one that solves the partial differential equations describing mass flow and momentum transfer. The parameters in these equations can be obtained from measurements and used in the model. For example, the St. Venant equations (open channel flow) and the Darcy equation (saturated flow in porous media) are physics-based equations. There are, however, important limitations to the applicability of such physics based models. For example, – it is widely recognized that such models require a significant amount of data and the cost of data acquisition may be high; – the relative complexity of the physics-based solution requires substantial execution time; – the relative complexity may lead to over-parameterized descriptions for simple applications; and – a physics-based model attempts to represent flow processes at the grid scale with mathematical descriptions that, at best, are valid for small-scale experimental conditions. Catchments typically have a high level of spatial heterogeneity which can be prohibitively expensive to observe or comprehensively represent in the model. This is most obvious in the representation of subsurface processes because of the difficulty of observation and the high degree of soil/aquifer heterogeneity which often exists. In principle the parameters of physics-based models are measurable, but in practice this cannot be achieved at the scale of modeling application, because such measurements are essentially made at a point. Therefore, these models use averaged variables and parameters at grid or element scales which are greater than the scale of variation of the processes. Even under a “full” physics representation, parameterization (including spatial variability of parameters) of material properties does not represent catchment heterogeneity.

1. CATCHMENT PROPERTIES AND HYDROMETEOROLOGICAL DATA OF SKADAR LAKE BASIN 1.1. CATCHMENT CHARACTERISTICS The Morača river basin is located in the central part of Montenegro. The river originates in northern Montenegro, under Rzac mountain at an altitude of 975 masl. In the upper and middle part of its flow, the Morača river is highly mountainous river that has cut a canyon with steep slopes. Its length is 113.4 km, and the area of the river basin to the hydrological station Podgorica is 2628 km2. The Zeta is a right tributary of the river Morača. It appears as a stream Surduk, which in Nikšićko polje, after having received two small tributaries, becomes the Sušica

13

Figure 1.1.1. Flow Accumulation derived from DEM with Arc Hydro Tools

river. Downstream of the river Rastovac confluence, the river changes its name to Zeta. During the low water period the river delves and continues to flow underground in a straight line, for about 5 km, and reemerges from the cave on the southern slope of the Planinica, at an altitude of 100 masl. At the village Tvorila the river creates the Slap waterfall. Through Bjelopavlići plain the Zeta meanders quietly. At the debris of the ancient city of Duklja, the river flows into the Morača river as its main tributary. Its length is 85 km, and the river basin area to the most downstream hydrological station Danilovgrad is 1216 km2. After the merge with its largest tributary, the river Zeta, the Morača enters the Zeta plain and flows to its confluence to Lake Skadar. The Morača river is the largest tributary of Skadar Lake (Figure 1.1.1). The steepest hillside slopes are located in the upstream and middle part of the Morača river flow. The inclinations are up to 69 degrees. The slopes of the Zeta river basin are significantly milder and they vary between 0 and 30 degrees. Some smaller areas of the basin have also high terrain slopes exceeding 45 degrees (Figure 1.1.2).

14

Figure 1.1.2. Hillside slopes for catchment area

The spatial distribution of the height on the Morača catchment as given by the Digital Elevation Map (DEM) is shown on the Figure 1.1.3. The watershed elevation ranges from 22 to 2217 masl. The mean catchment elevation is 981.53 masl.

Figure 1.1.3. Distribution of the altitude on the catchment

15

1.2. CLIMATE DATA The precipitation data on the Morača river basin are available on Danilovgrad, Kolašin, Nikšić, Podgorica, Šavnik, Andrijevo, Bioče, Bogetići, Dragovića Polje, Gornje Polje, Jasenovo Polje, Kupine, Lijeva Rijeka, Lukovo, Manastir Morača, Presjeka and Vasiljevići stations. Temperature time series exist for Danilovgrad, Kolašin, Nikšić, Podgorica and Šavnik stations (Figure 1.2.4).

Figure 1.2.4. Review of stations with available meteorological data

Precipitation stations are evenly distributed with altitude on the catchment area, but the temperature data are available only for the stations under 1000 masl. Since the catchment has a complex relief, it is necessary to have more precipitation and temperature data at all altitudes for the adequate simulation of the hydrological processes. The precipitation distribution by months and seasons indicates that here a long rainy period with a humid climate and a shorter summer period with an extremely arid climate can be separated (Figure 1.2.5). In the period from October till December the mean monthly precipitation is 873.08 mm, which is 40.16% of the annual total and from January to April the average precipitation is 796.43mm or 36.56% of the total annual precipitation. The largest amounts of precipitation are measured at Andrijevo station, followed by Danilovgrad and Kolašin stations (Figure 1.2.6).

16

Fig. 1.2.5. Mean multiannual monthly precipitation amounts

Fig. 1.2.6. Mean annual precipitation amounts for stations

17

1.3. HYDROLOGICAL DATA The Morača river catchment area for the purpose of this study is determined by hydrological profile of Podgorica. Monthly mean discharge values available for this study are those recorded from January 1948 till December 2012. Mean multiannual discharge for Podgorica profile are 158.05 m3/s and 189.66 mm respectively

Figure 1.3.7. Interannual discharge distribution

The seasonal cycle of monthly mean discharge at Podgorica station exhibits a high discharge of 184 m3/s to 284 m3/s from November till May and a relatively low discharge of 26 m3/s to 56 m3/s from June to October with maximum discharge usually in December and in August. The interannual variation of monthly mean discharge is generally small in the dry season and large in the wet season (Figure 1.3.7).

2. MIKE SHE MODEL 2.1. MODEL DESCRIPTION MIKE SHE is a deterministic - distributed hydrologic model that has been used in a wide range of applications (http://www.mikebydhi.com/Products/WaterResources/ MIKESHE.aspx). MIKE SHE model integrates the entire land phase of the hydrologic cycle and can model interception, actual evapotranspiration, overland flow, channel flow, flow in the unsaturated zone, flow in the saturated zone and exchange between aquifers and rivers. MIKE SHE applied at a catchment scale implies the assumption that

18 smaller scale equations are valid also at the larger scale; thus, it performs an upscaling operation using effective parameter values. MIKE SHE consists of the following model components: – Precipitation (rain or snow), – Evapotranspiration, including canopy interception, which is calculated according to the principles of Kristensen and Jensen, – Overland flow, which is calculated with a 2D finite difference diffusive wave approximation of the Saint-Venant equations, using the same 2D mesh as the groundwater component. Overland flow interacts with rivers, the unsaturated zone, and the saturated (groundwater) zone, – Channel flow, which is described through the river modeling component, MIKE 11, which is a modeling system for river hydraulics. MIKE 11 is a dynamic, 1D modeling tool for the design, management and operation of river and channel systems. MIKE 11 supports any level of complexity and offers simulation tools that cover the entire range from simple Muskingum routing to high-order dynamic wave formulations of the Saint-Venant equations, – Unsaturated water flow, which in MIKE SHE is described as a vertical soil profile model that interacts with both the overland flow (through ponding) and the groundwater model (the groundwater table is the lower boundary condition for the unsaturated zone). MIKE SHE offers three different modeling approaches, including a simple 2-layer root-zone mass balance approach, a gravity flow model, and a full Richards’s equation model, – Saturated (groundwater) flow, which allows for 3D flow in a heterogeneous aquifer, with conditions shifting between unconfined and confined. The spatial and temporal variations of the dependent variable (the hydraulic head) are described mathematically by the 3D Darcy equation and solved numerically by an iterative implicit finite difference technique. For integrated process-oriented catchment modeling MIKE SHE has different numerical methods from simple conceptual approaches to advanced, physics-based methods. That possibility enables modeling in situations when some model parameters are not available, therefore some processes could be conceptualized. All modules for hydrologic processes can be linked or unlinked with each other easily. For the best simulation of the hydrological processes of the Morača river basin all available spatial and temporal data are included in the MIKE SHE Flow model. The model for Morača river basin was built with modules for Overland Flow, Unsaturated Flow and Evapotranspiration. Modeling of the Saturated Flow needed information and data that were not at the disposal at the moment. Simulated data were then used as input to the Water Balance Module, where the components of the water balance were calculated.

2.2. MODEL SETUP In most watershed problems one or two hydrologic processes dominate the watershed behavior. Thus, a complete, physics-based flow description for all processes in one

19 model is rarely necessary. A sensible way forward is to use physics-based flow descriptions for only the processes that are important, and simpler, faster, less data demanding methods for the less important processes. The downside is that the parameters in the simpler methods are usually no longer physically meaningful, but must be calibrated – based on experience. Water flow on the ground surface is calculated using a finite-difference, diffusive wave approximation of the Saint - Venant equations. Due to spring snowmelt discharge it was necessary to include the Snow Melt module and have adequate annual discharge distribution. The Snow Melt module uses a Degree-day melting algorithm. The melting of snow and ice is assumed to be related to air temperature as long as air temperature is above a critical threshold. The amount of snow or ice melted at a certain place, during a certain period, is assumed proportional to the sum of positive temperatures (on the Celsius scale) at the same place and in the same period. The amount of melt is linked to this positive degree-day sum by the degree-day factor. The main purpose of the Two-Layer Water Balance ET method is to provide an estimate of the Actual evapotranspiration and the amount of water that recharges the saturated zone. The model divides the unsaturated zone into a root zone, from which evapotranspiration can be extracted, and a zone below the root zone, where evapotranspiration does not occur. Evapotranspiration is extracted first from intercepted water (based on the leaf area index), then ponded water and finally via transpiration from the root zone, based on an average water content in the root zone. However, evapotranspiration reduces the water content in the root zone, creating unsaturated zone storage. The minimum water content in the root zone is the wilting point water content, but this can only occur when the water table is below the root zone. The Two-Layer Water Balance evapotranspiration method requires time series for the root depth and the leaf area index, as well as the Reference evapotranspiration. Some components of hydrological circle were not possible to model. The river network (MIKE 11) with the chainage locations and cross-section points is necessary for the flow exchange with Groundwater Flow Module and for simulation of the lateral flow. The input data to the MIKE SHE model include data on topography, land use, geology, hydrogeology and meteorology. Preparation of the input data in suitable format was the first step needed to be done for the model setup. With MIKE Zero and MIKE SHE Tools standard file formats were converted to a format readable by the model.

2.3. MODEL DOMAIN AND GRID Global Digital Elevation Model (GDEM) released by NASA and the Ministry of Economy, Trade, and Industry (METI) of Japan were used for geomorphological analyses. The GDEM data are posted on a 1 arc - second (approximately 30 m at the equator) grid and referenced to the 1984 World Geodetic System (WGS84)/1996 Earth Gravitational Model (EGM96) geoid. For the Morača river basin, tiles ASTGM2_N42E18, ASTGM2_N42E19, ASTGM2_N43E18 and ASTGM2_N43E19 were downloaded from http://earthexplorer.usgs.gov/.

20 Generation of catchment area was carried out using the basin command inside of the hydrology toolset in Arc GIS. Polygon with the catchment area information defines the model domain. All other spatial data, such as topography, are interpolated during pre-processing to that domain.

2.4. TOPOGRAPHY In MIKE SHE, the topography defines the upper boundary of the model and the top elevation of both the Unsaturated and the Saturated Zone model. It also defines the drainage surface for overland flow. Since the MIKE SHE has limitations regarding the number of computational cells, the finest topography grid resolution that was possible

Figure 2.4.1. Topography map for Morača river basin (100x100m)

to use is 100x100m. The Bilinear Interpolation method was applied for interpolation of the gridded DEM data (Figure 2.4.1).

2.5. CLIMATE DATA The daily accumulated precipitation amounts were used for modeling of the climate data. Since the Snow Accumulation and Melting Model is included daily mean tempera-

21 tures were also imported. Most synoptic and climatic stations have complete data sets for the period from January 2000 to December 2010. For that period lack of precipitation data is quite present in time series obtained from precipitation stations. In order to fill these gaps, correlation analysis was performed (Figures 2.5.1 and 2.5.2).

Figure 2.5.1. Precipitation correlation for data from stations Jasenovo Polje and Presjeka

Figure 2.5.2. Precipitation correlation for data from stations Danilovgrad and Podgorica

Figure 2.5.3. Temperature correlation for data from Danilovgrad and Podgorica station

Figure 2.5.4. Temperature correlation for data from Šavnik and Kolašin station

Time series with serious data deficiency or poor correlation with neighboring stations were excluded from further work. As opposite to precipitations, temperatures show strong correlation; therefore missing data were replaced and used for further modeling (Figure 2.5.3 and 2.5.4). In mountainous areas, precipitation and temperature can vary significantly with altitude. However, there are rarely enough gauging stations to measure their spatial variability. Since the station-based rainfall data were used, choosing to correct the precipitation and temperature for elevation allows us to define a spatially variable correction factor (Table 2.5.1).

22 Table 2.5.1. List of precipitation stations used for modeling

Stations: Precipitation station: (masl) Podgorica Danilovgrad Manastir Morača Andrijevo Nikšić Dragovića Polje Šavnik Jasenovo Polje Kolašin

Altitude

Parameters:

49 85 270 544 647 650 824 940 944

p/t p/t r p p/t r p/t r p/t

p/t-measurement of precipitation and temperature r-rainfall measurement

When snow melt is included, various snow melt parameters should be specified to present that process: threshold melting temperature, degree-day coefficient that defines the amount of the snowmelt runoff, melting coefficient for solar radiation, minimum snow thickness that covers the entire cell with snow, maximum wet snow fraction in snow storage, initial total snow storage and initial wet snow fraction. Snow accumulates until mean daily temperature is lower than threshold melting temperature defined in the model, otherwise the snow melts. That process involves conversion from dry snow to wet snow, and then to overland flow when the parameter that represents maximum wet snow fraction in snow storage is exceeded.

Figure 2.5.5. Thiessen polygon for temperature stations

23

Figure 2.5.6. Thiessen polygon for precipitation stations

Spatial distribution of the precipitation and temperature is done using the Thiessen polygons method. Two different dispositions were constructed, one for the precipitation and the other for temperature stations (Figures 2.5.5 and 2.5.6). The Snowmelt module in the model requires definition of the reference evapotranspiration, the rate of evapotranspiration from a reference surface with an unlimited amount of water. The Reference evapotranspiration is multiplied by the Crop Coefficient

Figure 2.5.7. Vegetation distribution

24 to obtain the Crop Reference Evapotranspiration. An assumption that evapotranspiration is uniformly distributed over the catchment was made because the spatial distribution of evapotranspiration was not available. The distribution of the vegetation over the catchment area is necessary for the simulation of canopy interception and the process of evapotranspiration (Figure 2.5.7). It can be defined over the model with the Leaf Area Index (LAI) and Root Depth (RD). The spatial distribution of the forest, shrub/trees and crops/grass is obtained from Global Land Cover 2000 Project (GLC 2000). The time series of the root depth and leaf area index for defined vegetation types, for either one year or for the growing season, are taken from Vegetation Database included in the model.

2.6. MODEL CALIBRATION AND VALIDATION Since MIKE SHE is not coupled with any river flow model it was not possible to simulate river discharge and compare it with discharge registered at Podgorica hydrological station. Thus, validation of the model was possible by comparison between the simulated annual evapotranspiration and the one obtained from similar studies. The expected value of total annual evapotranspiration, expressed as precipitation ratio should be around 30%. It is also expected that interannual distribution correspond to the one in the nature. In case of reaching accumulation of the snow from October till the end of April, that will also be an indicator that the water balance is correct. The model was initially run for a ten year period. The initial run was done to achieve the approximate expected actual evapotranspiration percentage, the proper annual snow distribution and the minimal water balance error. Ten years of data were split into two halves for calibration and validation. The period from October 2000 till October 2005 was used for calibration, and the period from October 2005 till October 2010 for validation period. The model parameters represent the physical or hydrologic characteristics of a watershed. To adequately represent the system being modeled, during calibration, an iterative process, model parameters are set within an appropriate range. During that process, each model parameter is varied following a trial and error procedure, with all other parameters being constant (Table 2.6.1). Nevertheless, it was found during the calibration process that Manning‘s M coefficient, the snowmelt constant, and the threshold melting temperature were sensitive model outputs.

25 Table 2.6.1. Model parameters subjected to calibration

Model parameters

Initial Simulation

Degree-day factor (mm snow/day/°C) Threshold melting temperature (°C) Manning‘s M coefficient

Calibrated parameters

Final Calibration

4;5;6

6

0

-1;0;1

0

10

10;20;30

30

Using parameters fine-tuned in the calibration process, the model was validated with the data from the first five years. The final parameter list is presented in Table 2.6.2. Table 2.6.2.: Final parameter list

Module: Climate Reference Evapotranspiration Temperature Lapse Rate Snow Melt Degree day Coefficient Min Snow Storage Max Wet Snow Fraction Initial Total Snow Storage Initial Wet Snow Fraction Overland Flow Detention Storage Initial Water Depth Unsaturated Flow Water content at field capacity Water content at wilting point Saturated hydraulic conductivity Soil suction at wetting front ET Surface Depth

Precipitation Lapse Rate

Melting Temperature

Manning Number

Water content at saturation

Ground Water Table for lower UZ boundary

10 %/100m 6 mm/day -0.4 °C/100m 0 °C 6 mm/°C/day 3 mm 0.5 0 0 30 m^ (1/3)/s 0 mm 0m 0.2 0.05 0.01 1.41e-4 m/s -0.1 m 0.25 m 1m

26

27

Figure 2.6.1. MIKE SHE results for cathment area of Skadar Lake

28

3. WATER BALANCE ANALYSIS The purpose of the water balance component of this assessment is to characterize the climatic, surface water, and groundwater components of the watershed. The water balance is intended for use as a screening tool to further evaluate water resources allocations within the watershed and to identify water balance components that may require further analysis during the next levels of watershed planning. The main analytical formulation of annual water balance for large areas is : P = ET + RO + ∆S where P represents precipitation, ET evapotranspiration, RO runoff and ΔS change of water resources, which can be neglected if we analyze water balance for a long time period. But if we need to analyze the water budget on catchment scale, the equation becomes more complex : P = ET + I + RO + ∆S + U ± L and includes I as plant interception, U as base flow and L as leakage to or from the catchment. Since the river flow module is not included, the validation of the model was done comparing simulated annual evapotranspiration with the value obtained in earlier studies. In case of reaching accumulation of the snow from October till the end of April, that will be also an indicator that the water balance is correct. Table 3.1. Annual water balance summaries for calibration period

Year

RO (mm)

P (mm)

ET (mm)

Change in storage ΔS (mm)

Infiltration (mm)

Error (mm)

2001 2002 2003 2004 2005

2632.52 1502.26 1920.88 2988.34 2735.23

3583 2480 2526 4194 3169

993 907 810 993 904

-17 7 -7 3 11

2614 1569 1737 3208 2259

7 8 9 10 7

Table 3.2. Annual water balance summaries for verification period

Year

RO (mm)

2006 2007 2008 2009 2010

2015.39 1814.86 2050.76 2326.55 3519.87

P (mm) 2877 2046 3211 3383 4411

ET (mm) 899 753 933 925 953

Change in storage ΔS (mm) -23 12 -7 -8 11

Infiltration (mm) 2010 1287 2295 2472 3458

Error (mm) 5 6 2 7 5

29 The annual water balance summaries for the calibration and validation periods are presented in Table 5 and Table 6. The runoff (RO) is presented as the discharge observed at Podgorica station. The year in the table represents a water year, the period of 12 months, the period between October 1st of one year and September 30th of the next. The water year is designated by the calendar year in which it ends. In the calibration period, the annual precipitation varies between 2480 and 4194 mm and from 1814 to 3519 mm in verification period. The discharge in the calibration period is in the interval from 1502 to 2988 mm, making the runoff coefficient within 60 to 86% (Tables 5 and 6). The runoff coefficient in the verification period is around 74%. Unfortunately, some of the very important rainfall stations on the catchment have significant data gaps and they are excluded from this study. The existing data point to conclusion that there is much more precipitation fell over the catchment than the one calculated by the model, so we can assume that the runoff coefficient could be lower. The total calculated error is less than 0.3%. Infiltration varies both spatially and temporally but in general it depends on soil moisture condition, soil type and type of vegetation. Earlier studies indicate that infiltration over the Morača catchment is large and can be up to 80% of falling rain. The model indicates that the infiltration accounts for 71.12% of the total precipitation amount for the period 2001-2005 (Figure 3.1.). Plant transpiration, evaporation from intercepted water, ponded water and evaporation from snow are included in the calculation of evapotranspiration. The loss of rain water through evapotranspiration is 28.88%. Those figures are 71.98% and 28.02% for the verification period (Figure 3.2.).

Figure 3.1. Share of the water balance components in total precipitation amounts (period 2000-2005)

Figure 3.2. Share of the water balance components in total precipitation amounts (period 2005-2010)

Although adequate snow accumulation needs temperature data on higher altitudes, the model indicates that the formation of the snow cover on the Morača catchment mostly starts in November, when temperatures fall under the threshold value of 0°C for snow accumulation. The model includes melting due to solar radiation and energy in rain besides melting due to air temperature. The biggest snow accumulation is in January and February. With increase in temperature during the spring the process of snowmelt starts and can last till the end of May (Figure 3.3.). Snow cover on the highest peaks melts till June and July.

30

Figure 3.3. Monthly changes in snow accumulation and snowmelt for period 2001–2005

The seasonal water balances for calibration and verification period are given in Tables 3.3. and 3.4. Seasonal distributions of the water balance components for both periods are respectively presented on Figures 3.3. and 3.4. Soil type and soil moisture condition define the percentage of precipitation that will infiltrate to the ground. As we noticed before 70% of the precipitation goes to infiltration, thereby the smallest amounts infiltrate to the ground in summer. Significant amounts of water infiltrate during the rest of the year. Seasonal distribution of the water balance components for the years 2000 to 2005 are presented on figures 3.5 to 3.9.

31 Table 3.3.: Seasonal water balance summaries for calibration period

2001 2002 2003 2004 2005

Winter Spring Snow ET Storage Infiltr P Snow ET Storage Infiltr change change change change (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) 1495 190 199 4 1103 610 -198 300 -10 522 473 67 148 7 254 490 -209 287 -5 420 946 328 120 -9 509 333 -397 231 13 490 1229 415 159 4 653 1164 -416 383 -2 1204 870 420 103 -16 364 446 -481 293 3 635

2001 2002 2003 2004 2005

Summer Autumn P Snow ET Storage Infiltr P Snow ET Storage change change change change (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) 486 -16 203 7 294 938 23 252 -14 421 0 261 -2 163 787 134 193 -13 344 0 170 -12 187 1226 69 333 0 192 -33 152 -2 77 1458 50 270 7 500 -3 255 15 235 1473 64 267 13

P

Infiltr (mm) 678 474 826 1132 1130

Table 3.4.: Seasonal water balance summaries for verification period

2006 2007 2008 2009 2010

Winter Spring Snow ET Storage Infiltr P Snow ET Storage Infiltr change change change change (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) 1155 453 138 0 565 463 -592 338 -4 725 732 124 166 2 444 502 -168 275 -15 414 883 266 153 9 456 1015 -435 336 -8 1127 931 163 167 -16 598 527 -353 269 9 604 1913 365 185 -1 1368 779 -463 304 16 926

2006 2007 2008 2009 2010

Summer Autumn Snow ET Storage Infiltr P Snow ET Storage Infiltr change change change change (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) 408 -22 211 11 210 839 161 190 -20 509 223 0 126 12 86 530 94 172 2 266 307 -5 165 4 143 1139 174 292 -7 681 461 0 243 -5 223 1492 170 266 8 1049 350 -20 178 -16 210 1245 119 255 2 871

P

P

32

Figure 3.3. Seasonal distribution of the water balance components for period 2001-2005

Figure 3.4.: Seasonal distribution of the water balance components for period 2005-2010

33 Total Error 7

Precipitation 3583

Evapotranspiration 993 Snow -Storage change 0 OL-Storage change 0

Canopy-Storage change 0

UZ-Storage change -17

Infilt. incl. Evap 0

2614

2614 Boundary flow

SZ-Storage change 0

0 Accum ulated w aterbalance from 1.10.2000 to 1.10.2001. Data type : Storage depth [m illim eter]. Flow Result File : C:\Users\Marija Ivkovic\Desktop\Moraca\Result\Moraca.she - Result Files\Moraca Title : Text :

Figure 3.5. Seasonal distribution of the water balance components for the year 2000 - 2001

Total Error 8

Precipitation 2480 Evapotranspiration 907

Snow -Storage change Canopy-Storage change 5 OL-Storage change 0 0 UZ-Storage change 7

1569


1569 SZ-Storage change 0

Boundary flow 0

Accum ulated w aterbalance from 1.10.2001 to 1.10.2002. Data type : Storage depth [m illim eter]. Flow Result File : C:\Users\Desktop\Moraca\Result\Moraca.she - Result Files\Moraca Figure Seasonal distribution of the water balance components for the year 2001 – 2002 Title :3.6.Text :

34 Total Error 9


Snow -Storage change Canopy-Storage change -5 OL-Storage change 0 0 UZ-Storage change -7

1737



Boundary flow 0

Accum ulated w aterbalance from 1.10.2002 to 1.10.2003. Data type : Storage depth [m illim eter]. Flow Result File : C:\Users\Desktop\Moraca\Result\Moraca.she - Result Files\Moraca Figure Seasonal distribution of the water balance components for the year 2002–2003 Title :3.7.:Text :

Total Error 10



3208



Boundary flow 0

Accum ulated w aterbalance from 1.10.2003 to 1.10.2004. Data type : Storage depth [m illim eter]. Flow Result File : C:\Users\Desktop\Moraca\Result\Moraca.she - Result Files\Moraca Figure Seasonal distribution of the water balance components for the year 2003–2004 Title :3.8.:Text :

35 Total Error 7



2259



Boundary flow 0

Accum ulated w aterbalance from 1.10.2004 to 1.10.2005. Data type : Storage depth [m illim eter]. Flow Result File : C:\Users\Desktop\Moraca\Result\Moraca.she - Result Files\Moraca Figure Seasonal distribution of the water balance components for the year 2004–2005 Title3.9.: : Text :

4. CONCLUSION Uncertainties in the estimation of the water balance components depend on the accuracy of the measured input data into the model: the one with temporal variability (e.g. rainfall) and the one with spatial variability (e.g. rainfall, vegetation, groundwater levels, soil hydraulic properties etc.). Developing a water balance or water budget for the Skadar lake basin was difficult due the absence of adequate data and information. MIKE SHE model for the Skadar lake catchment area shows satisfactory results and could be improved in the future with new input data, if they become available meanwhile. It is no longer acceptable to manage groundwater and surface water independently of one another. Advances in data collection and availability, as well as computer resources, have now made distributed, physics-based watershed modeling feasible in a wide range of applications. MIKE SHE is one of the few commercially available codes that has been widely used for integrated hydrologic modeling. MIKE SHE’s process based framework allows each hydrologic process to be represented according to the problem needs at different spatial and temporal scales. This flexibility has allowed MIKE SHE to be applied at spatial scales ranging from single soil profiles, to the field scale, and up to the watershed scale. Furthermore, each process can be represented at different levels of complexity. MIKE SHE has a modern, windows-based user interface that includes advanced tools for water quality, parameter estimation and water budget analysis.

II FORMATION OF THE HYDRAULIC MODEL IN THE SYSTEM THE MORAČA RIVER – LAKE SKADAR – THE BOJANA RIVER

INTRODUCTION The data bases for the formation of the hydraulic model were collected through all the previous activities during the life of the project, both within field investigations and on the basis of the research literature on long-term observations and statistical analyses of temporal and spatial data series. Generally, the most important basis for the formation of a hydraulic flow model can be summarized as: – Topographical and morphological database – Hydrological-meteorological database – Hydraulic database – Psammological database (sediment data) – Geological and geomechanical databases Data on the hydrological regime of a watercourse, that is, on water levels and discharge are the most important basis for the formation of a hydraulic model. Without knowing the values of these hydrological parameters and how they change, it is impossible to understand the issue of watercourses or engage into any further activity on the model formation. For this reason, we shall first provide an overview of the availability of these databases for watercourses within the system, for the rivers Morača and Bojana separately.

1. BASIC DATA ON THE HYDROLOGICAL REGIME OF THE RIVERS MORAČA AND BOJANA 1.1. THE RIVER MORAČA The Morača River is the largest tributary of Lake Skadar. The river Morača basin covers highland areas of the southeast Dinarides, at an altitude ranging from 5 m.a.s.l to slightly above 2400 m. Its approximate catchment area is 2650 km2. The flow of the river Morača is formed on the absolute elevation of 975 m.a.s.l. from a range of temporary and permanent streams, of which Rzački Potok accounts for the largest share. The river’s most upstream flow, the approximate length of 3 km, flows toward the northeast to the absolute elevation of 780 m.a.s.l, where after receiving waters from Javorski Potok, it suddenly changes direction and starts running in a northwest – southeast direction to the absolute elevation of 354 m.a.s.l at Kotlišta. Having received water from its left tributary the Slatina at Kotlište, the Morača turns southwards and runs in a meridional direction to the absolute elevation of 58 m.a.s.l in the area of Bioče, where it receives the Mala Rijeka waters from the left. Leaving Bioče the Morača turns once

38 Srednji godišnji proticaji

300

HS Podgorica (1948 - 2012) Osmotreni podaci Qsr,vis Linear (Osmotreni podaci)

280 260 240

Q (m 3/s)

220 200 180 160 140 120 100 80

Qsr,viš = 157.7 m3/s

60 1948 1952 1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 2012

God

Figure 1.1.1. The mean annual discharge trend analysis at HS Podgorica

again and continues its flow in a direction south – southwest until its confluence with Lake Skadar at the absolute elevation within the limits of 4.53 – 9.84 m.a.s.l. The Morača riverbed has a total length of 97.6 km while its length in a straight line from its source to the mouth is approximately 69 km. More detailed analyses of the basin and especially hydrological parameters that were analyzed were given in previous statistical interpretations of the available measurements and in the hydrological model, that is, the project reports related to that section. What is particularly important, and what is the basic input data for the hydraulic calculation, are typical discharges of the Morača river which could be taken as input values in analyses. These discharges have been processed in the hydrological report and as such will serve to the purposes of further hydraulic analyses. Since the hydrological station Podgorica on the Morača controls the balance of the entire Morača with tributaries and has the longest and most reliable series of data, it has been chosen as the key station for the water balance analysis in hydrological processing. For this station data have been provided on mean annual discharge as well as data on maximum water. From historical data and charts it is evident that the largest recorded disharge during the period in which measurements were performed at the station in Podgorica is 2073 m3/s. Since this is a relatively long and reliable measurement period, this value can be accepted with great certainty.

39 Maksimalni godišnji proticaji HS Podgorica (1948-2012) 2500

Osmotreni podaci Series3 Linear (Osmotreni podaci)

2300 2100

Q (m 3/s)

1900 1700 1500 1300 1100 900 700 500

Qmax,viš = 1259 m 3/s

300 1948 1952 1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 2012

God

Figure 1.1.2. The maximum annual discharge trend analysis at HS Podgorica

In the existing project documentation in which the regulation of the Morača river was previously discussed and which was available during the development of the model, the data on the maximum water discharge recorded on the hydrological station with the value of Q max = 2040.0 m3/s, recorded in 1968, was used. Also, for the same section there is information that the watercourse at times dries up (minimum flow is Q min = = 0 m3/s and that the mean perennial discharge at the hydrological station “Botun” is Qsr = 136.0 m3/s. Since the “Botun” station is located downstream of the Podgorica station it would be interesting if there was information on the maximum value of the discharge at Botun when the already mentioned maximum discharge at the Podgorica station amounts to 2073 m3/s. The reconstruction of these data on the basis of available data lies on numerous assumptions, because the lower riparian zone of the Morača River, downstream from the confluence of the Sitnica at Botun is exposed to many problems of heterogeneous nature (urban, municipal services, environmental, social ...), which disrupt the natural flow and its continuity. The gravest problems are hydro-technical ones conditioned by the watercourse irregularity on the one hand, and unplanned exploitation of the waterway and the riparian areas on the other hand. To further verify the adopted values, thanks to modern measuring equipment which was an integral part of the project procurement activities, control measurements were carried out on the river Morača at the two previously mentioned sections. To be more specific, the measurements were performed at the section Podgorica, at the location of Moskovski most (Moscow bridge) and the section Botun, the location of Golubovci-

40

3800

Vodotok: Moraca HS: Podgorica

Povratni period (godina)

Skala vjerovatnoće max godišnjih proticaja 2

10

20

100

1000

10000

3600 3400 3200

F-je raspodjele:

3000 2800

― ― ― x

2600 2400

Q(m3/s)

2200 2000

Pirson 3 Log Pirson 3 Kricki Menkelj Gumbel

1800 1600 1400 1200 1000 800 600 400 200 0

Vjerovatno

99.9

99

95

90

80

50

30

10

5

1

ća pojave (%)

0.1

0.01

Figure 1.1.3. The probability scale of maximum annual discharge at HS Podgorica 1948-2012 ( watercourse: The Morača, return period (year), distribution functions , probabilty of occurence)

Vukovački most. Simultaneous measurements were performed on these two profiles primarily to better determine the abovementioned relation of their discharges. The values of 152.9 m3/s and 167.6 m3/s were measured at the section Podgorica and the section Vukovački most respectively. As can be seen, these values are very close to the mean perennial discharge values. The downstream section at Vukovački most exhibits some 10% higher discharge value. If we translate these values into high flow values this would mean that the value of 2073 m3/s at the Podgorica section corresponds to that of 2280 m3/s at Vukovački most. It has been unequivocally proven that the existing data and hydrological analyzes show that the high flow values on the river Morača are greater than those taken into account in the existing technical documents used for the planned regulation structures in this section of the river. The most current project of the regulation of the river Morača includes the area from confluence of the Sitnica into the Morača to Skadar Lake, that is, the area which is the subject of this hydraulic model. That project, or the solutions for the regulated section offered by it, are adopted as input data for the respective hydraulic model, because its solutions guarantee complete protection for this part of the river flow which is the most vulnerable to flooding. The project took into account that in assessing high flow levels, one should take into account that the hydropower plants, planned to be built in the upper course of the Morača river flow, will affect the hydrological regime of the watercourse so

41

Figure1.1.4. Measurements of discharge on the river Morača at the sections Moskovski most in Podgorica and Vukovački most in Golubovci using the “SonTek” Doppler ultrasound device

42 that it is partially changed. In its middle and lower course the high flow discharge will be reduced while the low flow discharge will increase. The effects of hydropower plants will also manifest by reduced sediment transport through the river bed, and therefore less sediment deposition on the section from the mouth of Sitnica to Ponari, which will, again, increase the river’s channel capacity. The abovementioned current regulation project summarizes all the previous/precedent data of earlier projects and measurements of the Hydrometeorological Institute and in its conclusion of the entire reviewed documentation it proposes and adopts the following values given in the table below:

Section(from– to) (km) The Sitnica confluence, From the Sitnica confluence to the Cijevna confluence

Up to the Cijevna confluence

Low water level

Discharge Q(m3/s) Discharge ranking High water level Medium Probability of occurrence water level 2% 1%

Qsr 165.02

v.v.Q2% 2200.03

v.v.Q1% 2600.03

10.31

175.02

2380.03

2800.03

10.31

197.92

2830.03

3200.03

m.v.Q95% 10.31

1 Hydrometeorological Institute Podgorica,2007. 2 “The main project of the regulation of the Morača River on the Vrbišće – Mišurica section” in 1987. “The main project of the regulation of the Morača River on the Vrbišće – Cijevna confluence section” in 1988. “A preliminary design of the regulation of the Morača River on the Mišurica – Ponari section”, 2007. 3 “The main project of the regulation of the Morača River on the Vrbišće – Cijevna confluence section” in 1988. “A preliminary design of the regulation of the Morača River on the Mišurica – Ponari section”, 2007.

At the lowest downstream section, downstream of Ponari, Morača is constrained by Skadar Lake backwater, which is why large amounts of water overflow from its relatively small channel and flow through Žabaljsko Polje and Ceklinsko Polje. Regulation of the bed on this section at this stage is not necessary, and in view of its disposition in relation to Lake Skadar and the river Bojana, it is estimated that its channel regulation should be defined within the framework of regulating the most downstream section of the river Bojana and Skadar Lake regulation in the confluence areas of both watercourses. One of the analyses of the hydraulic model, which is the subject of this paper and in which as input data the results obtained in the previously described hydrological model, MIKE SHE, were used, served to test the specified values. If as input data of the hydraulic simulation the time series corresponding to the time series for which a simulation in the hydrological model MIKE SHE was carried out (i.e., the period from 17.11.2010 to 09.12.2010), given that the initial discharge rate is constant and equal to the average discharge of the Morača, a runoff hydrograph is obtained which also points to similar values of maximum discharge as offered in the aforementioned sources. The hydrograph generated on the basis of the results of the hydrological model is shown in Figure1.1.5.

43

Fig. 1.1.5. A runoff hydrograph on the Morača section at the entrance of Lake Skadar, according to the input data from the hydrological model MIKE SHE

It should be emphasized that, despite the limitations highlighted in the description of the application of the hydrological model MIKE SHE, the results of the hydraulic model in this case exhibit a fairly good agreement with reality and can be taken as confirmation of the above preliminary results. The maximum discharge value is 2650 m3/s, which is approximately equal to the values that have so far been measured on that section. The Morača River on the surveyed section (from the mouth of the Sitnica to Botun in Ponari) is characterized by large sedimentation of sand and gravel. Significant deposits of alluvial material are also characteristic of the upstream section of the river Morača, but the Botun – Ponari section is characterized by a very shallow channel overflooding not only at high-level but also at medium-level flow, so that the processes of sedimentation are more frequent. Due to the rich natural deposits of sandy and gravel soils as well as growing needs of them, the downstream Morača sections received great economic importance. In this context, their floodplains have have seen heavy exploitation of sand and gravel over the recent years, although without any established standards, given that research, recordings, observations and measurements have still not been performed on which to establish annual production of suspended and bed load. Since the instability of the cross-section of the Morača on the surveyed section is evident, the technical solution of the Morača channel regulation from the current regulation project has been opted for in the model analysis, since it provides a uniform discharge section - channel, which consists of the basic trapezoidal channel and a highflow channel. The channel is structured so as to be able to pass the medium water flow Qsr = (165.0-197.9) m3/s with the required banking in the range of (0.66-1.00) m (average 0.75 m).The high flow channel has been designed to be able to pass a 50-year high flow v.v.Q2% = (2200.0-2830.0) m3/s with the top of embankment of ΔZv.v.Q2% = (0.55-1.64) m (1.04 m on average), while the solution has additionally been secured by the requirement to

44

Fig 1.1.6. The adopted cross section of the future regulated Morača river channel

ensure and secure the passage of a 100-year high flow v.v.Q1% = (2600.0-3200.0) m3/s, whereas the term “safe” means “with the required top of embankment” and “with necessary reserve capacity” in the amount of ΔZv.v.Q1% = (0.01-1.19) m (0.57 m on average).

1.2. THE RIVER BOJANA After the river Po in Italy the Bojana river is the second most important tributary of the Adriatic Sea. Despite its importance in this respect, the existing data on the hydrological characteristics of the river are very modest. For various reasons (economic and political ones, among others) the cooperation between the hydrometeorological services of the former Yugoslavia, that is, Montenegro and Albania, was poor, which affected the necessary measurements on the interstate river. Although the flow of the river is relatively short, about 40 km, the river’s hydrological regime is extremely complex because it depends on a variety of natural and anthropogenic factors. By the mid-nineteenth century, the hydrological regime of the river Bojana depended solely on natural factors, that is, the hydrological regime of Skadar Lake, from which it outflows. However, during the catastrophic floods of 1859 the Drim cut a new bed and since then it has been flowing into the river Bojana about 4.5 km downstream from the exit of Bojana from Lake Skadar. Figure 6 schematically shows a wider area of Skadar Lake and river Bojana, and the old and new beds of the Drim. As most of the waters of the river Drim flowed through the new river bed to the meeting point with the Bojana, the old Drim river bed that flowed into the Adriatic Sea 25 km east of the mouth of the river Bojana, gradually became clogged with sediment, so that today it serves as a drainage channel only. Cutting of the new bed of the Drim had a profound impact on the hydrological regime of Lake Skadar and river Bojana, and the stability of the banks in the confluence area of the old Drim river bed. The Drim transported large amounts of sediment which the river Bojana, due to the limited

45

Fig 1.2.1. A schematic representation of Skadar Lake, the Bojana and the new and the old channel of the Drim

sediment transport capacity could not carry to the Adriatic Sea. Due to this reason, significant amounts of sediment from the river Drim were deposited in the river Bojana. As a result of sedimentation the Bojana channel capacity decreased significantly, which resulted in difficulties in discharge from Lake Skadar. Flooding in the Skadar Lake riparian area became much more frequent than in the period before the new Drim river bed had been cut. Cutting of the new Drim bed had not only influenced the hydrological regime of Lake Skadar and the river Bojana but also the stability of the banks in the confluence area of the old Drim river bed. As the sediment transport from the old bed of the river Drim almost entirely stopped, the process of intensive erosion of the Adriatic Sea coastal area started in the wider confluence area of the old Drim bed, which was located about 25 kilometers east of the mouth of the river Bojana.

46 Namely, the natural balance was disrupted between the amount of sediment being transported to the coast and the amount of sediment being eroded and transported into the sea by waves . At the end of the 1960s Albania saw the construction of several large hydropower plants on the Drim River and its tributaries. The creation of several large reservoirs completely changed the natural regime of the river Drim and consequently, the sediment regime was also changed. Since commissioning the first hydroelectric power plant on the Drim River, the hydrological regime of the river Bojana besides natural factors, has begun to be significantly influenced by anthropogenic factors as well. Therefore, the analysis of the hydrological regime of the river Bojana presented in this paper refers to the situation after the construction of hydropower plants on the Drim River . The first measurements of the cross section of the river Bojana were carried out in 1928. It was only in 1965 that the following recordings were made which allowed the impact of the sediment from the Drim to be observed on filling the bed of the river Bojana. It was noted that between these two recordings as much as 2.5 million cubic meters of sediment were deposited in the bed of the river Bojana. That is why it is extremely important for the modeling needs to perform new recordings of the cross sections of the river Bojana in order to identify changes in the cross-sections up to the present date and their correct areas. Based on the results of measuring the level and discharge till 1965, it was noted that the mean annual discharge of the river Bojana on the way out of the Lake was about 320 m3/s. The results of previous measurements on the Drim indicate that the mean annual discharge at the mouth of Bojana amounted to 280 m3/s. Results of subsequent studies have shown that the mean annual discharge of the river Drim was 360 m3/s, so that the mean annual discharge at the mouth of Bojana river into the Adriatic Sea was 680 m3/s . According to the Albanian data on mean monthly discharge, as shown in Table 1.2.1, the mean annual discharge at the mouth of Bojana river into the sea is Q = 682 m3/s. The maximum average monthly discharge occurs in January, and amounts to Q = 1067 m3/s. Table 1.2.1 Mean monthly and mean annual discharge of the river Bojana (according to Albanian sources) /14, 19/

River I II III IV V VI VII VIII IX The Bojanathe Adriatic 1067 899 850 879 861 588 322 195 210 Sea

X

XI

XII Qsr.g

382 770 1030 682

The Drim-the confluence 493 459 446 507 490 293 155 104 414 228 396 501 351 in the Bojana

Data on the Bojana discharge on the Montenegrin side are virtually non-existent. The list of measurement stations of the RHMI (Republic Hydr-Meteorological Institute) of Montenegro (Table 1.2.2) shows that for the sections that are established on the river Bojana there are no data on the average, maximum and minimum discharge.

47 Table 1.2. 2. Values of hydrological parameters for the period up to 2002 for the major rivers of the Adriatic basin (excerpt from the RHMI table)

3 4 5 6 7 8

31,4 2628 2073 158,7 7,93 1226 1 3,03 2,1

58,3 101 2,82 79,3 228

0 298

6,2 0,46

56,6 63,7 3,07

0

PERIOD

Min H

Perennial average H

Max. H

Minim. Q

Perennial average Q

Max. Q

Catchment area

zero elevation Km from confluence

Skadar 24,6 Lake Podgorica Ribnica Morača 32,53 Brodska Rijeka Skadar 8,32 Njiva Crnojevića Lake Skadar 12,12 Orahovo Orahovštica Lake Skadar 4,86 Crkla Skadar Lake Lake Skadar 4,56 Plavnica Skadar Lake Lake Adriatic 0,09 Reč Bojana Sea Adriatic -0.07 Fraskanjel Bojana Sea

1 Podgorica Morača 2

BASIN

WATERCOURSE

STATION

THE ADRIATIC SEA BASIN

242 18 1948-2002 79

0 1949-2002

272

33,5

-1 1987-2002

216

35,2

0 1965-2002

*

*

*

*

* 496

165 -20 1950-2002

*

*

*

*

*

208 20 1948-2002

10,8 18160

*

*

* 434

96

-9 1952-2002

20 16520

*

*

* 603

183

2 1960-2002

593

The data on the high-flow levels of the river Bojana are very unreliable because due to the persistent uplift of the riverbed it was not possible to define a reliable flow curve. According to some estimates the discharge at the mouth of the river Bojana in the Adriatic Sea, with the probability of occurrece once in 100 years is Q=8500 m3/s (SchneiderJacoby, M., U. Schwarz, P. Sackl, D. Dhora, D. Saveljic and B. Stumberger (2006): Rapid assessment of the Ecological Value of the Bojana-Buna Delta (Albania / Montenegro). Euronatur, Radolfzell23/). According to this source the Bojana discharge, for the same return period, at the ouflow of the Skadar Lake is Q = 3930 m3/s, while the Drim discharge is Q = 6530 m3/s. These data are problematic because coincidental occurrence of flood waves has not been taken into account. It seems that in order to perform more reliable estimates of the high-flow levels of the Bojana we should take account only of the data on the Drim high flow levels. Namely, the high-flow levels of the Drim river significantly hamper the river Bojana outflow so it can be roughly assumed that the high-flow levels of the Bojana are the same as the high-flow levels of the Drim. The mathematical model that is the subject of this paper resorted exactly to such an assessment, so that the Drim high-flow levels have been analyzed alongside their impact on the whole upstream and downstream system from the Drim confluence in the Bojana. Data on the Drim high-flow levels vary significantly. Thus, it can be found in the literature that during the great 1896 floods the Drim discharge was about 7000 m3/s (Rapid Assessment of proposed Hydropower Plants on Drin River near Ashta (south of Shkodra). Euronatur, 2009./20/). The results of hydrological analysis in the design phase

48 of the Vau I Dejës reservoir indicated that the maximum of maximum Drim discharge, during extreme flood waves, could amount to as high as 9000 m3/s. 1.2.1. THE HYDROLOGICAL REGIME OF THE BOJANA RIVER UPON THE CONSTRUCTION OF HYDROPOWER PLANTS Three large reservoirs built on the river Drim in Albania are, among other things, of great importance to the hydrological regime of the river Bojana. These are reservoirs Vau I Dejës, Komani and Fierze whose positions are given in Figure 6. At the beginning of the 1970s the most downstream reservoir Vau I Dejës was built. This reservoir is located in the immediate vicinity of Skadar and the confluence of the Drim into the Bojana. The reservoir was formed after the construction of the three dams: Qyrsad, Zadeja and Ragam. The total volume of the reservoir Vau I Dejës is 0,623·109 m3. In terms of discharge of water from the reservoir, the Qyrsad and Zadeja dams are the only important, because the Ragam dam serves only for formation of the reservoir. As mentioned above, it was determined that the maximum discharge of the Drim, during flood waves could amount to 9000 m3/s. Therefore, spilways with great channel

Figure 7. Reservoirs Vau I Dejës, Komani and Fierze on the Drim

49 capacity were built on the dams Qyrsaq and Zadeja. Three spillway bays were built on the Qyrsaq with the channel capacity of 1400 m3/s each, that is, 4200 m3/s total, whereas two spillway bays were built on the Zadeja dam with the channel capacity of 1200 m3/s each, that is, 2400 m3/s total This means that if great flood waves approached the accumulation of Vau I Dejës, 6600 m3/s could be discharged through the Qyrsaq and Zadeja dams spillways. Five Francis turbines were built on the Qyrsaq and Zadeja dams, each of which with the installed flow of 113 m3/s, which makes the total discharge through the turbine 565 m3/s. Therefore, the maximum discharge through the spillways and the turbines of the Vau I Dejës reservoir is 7165 m3/s. So this is the maximum flow of the Drim which flows into the Bojana River directly downstream of the dam and as such will be used in one of the model iterations. In all the available documentation on the mean annual discharge of the river Bojana, after the construction of reservoirs Vau I Dejës, Komani and Fierze, the discharge data of Qsr = 680 m3/s is still valid. It is interesting that during the construction of the dams it was planned to build the spillway on the third dam Ragam on Vau I Dejës reservoir. However, as at that time the decision on the construction of the most upstream reservoir Fierze had already been made, the idea was abandoned. The Fierze reservoir, whose construction was completed in 1979 has a huge volume of as much as 2.7 ·109 m3. The Komeni reservoir, which is the middle step on the Drim River, was built in 1988 and has a volume 0,43 ·109 m3. The total volume of all three reservoirs on the Drim River equals 3.76 ·109 m3 and is greater than the average volume of Skadar Lake, which is 2.6 ·109 m3.

2. COMPUTATIONAL MODEL OF UNSTEADY FLOW – THE MORAČA RIVER SKADAR LAKE AND THE RIVER BOJANA INTRODUCTION The hydraulic regime of the river Morača, Skadar Lake and the river Bojana has been analyzed by applying the linear model of unsteady flow using the module of the software package MIKE11 to solve the complete Saint Venant equations [1]. The subject of the analysis is the section of the Morača river from the confluence of the river Zeta to the point where it flows into Skadar Lake and the whole course of the river Bojana (Figure 1). The computational model includes the influence of Skadar Lake, the River Drim and the Adriatic Sea on the flow of water through the channels of the Morača and the Bojana. The text below presents the results of numerical simulations designed so as to analyze the mentioned impacts separately or jointly. It is important to note that all the parameters of the model described could not be fully calibrated and, therefore, some caution should be taken with the interpretation of the results.

50

Figure 2.1. The elements of the hydraulic model MIKE11: the Morača – Skadar Lake – the Bojana

Open-channel flow is often modeled by a linear (1D) model. Such an approach to the problem implies that the complex flow pattern in open channel-flow is simplified, and by calibration of the model parameters the accurate simulation of the natural process

51 is achieved. Although in most cases open-channel flow is three-dimensional, (flow rate has three components) for solving most practical problems the aforementioned onedimensional analysis is used. This means that the average velocity over the cross section is adopted and that the streamlines of the flow of water are quasi-parallel as well as that the vertical distribution of load is hydrostatic. A computer program MIKE11 enables an analysis of the flow of water with free surface based on a mathematical model of the linear flow consisted of Saint-Venant equations. These equations represent the law of conservation of momentum and mass in a differential form and were obtained by averaging the spatial equations of flow over the depth and width of the bed. The dependent variables in the Saint-Venant equations are discharge and elevation levels in cross section and the independent variables are one spatial coordinate of the section and time. Given that it is a system of differential equations, in order to perform flow calculation, in addition to data on the geometry of the river section and parameters which define the flow resistance, it is necessary to set initial and boundary conditions and define the scheme of numerical calculations. Numerical solution of Saint-Venant equations in MIKE11 is based on the AbbottIonescu implicit scheme. Calculation, according to this scheme, is performed using the method on staggered grid which means that changes of flow and elevation levels over time are not calculated at the same computational points [1]. Information on the geometry of the flow field as well as the initial and boundary conditions used in the calculations are described below.

2.1. THE MODEL OF THE RIVER NETWORK The line, which schematically represents the flow direction in the considered analysis, is shown in Figure 2.1. The shown line, which actually represents a computational network, is defined in a separate editor of the software package MIKE (river network editor) and has a role in the topological sorting of different watercourses and stretches. In the considered analysis, the computational network consists of a sequence of lines that represent the flow of the river Morača, the direction of lake Skadar and the Bojana river flow respectively. The figure shows that there are branching stretches of the network as well as stretches which are merging parts of the flows. A good feature of the program MIKE11 is that it enables automatic calculation of both branched and cyclic parts of the network.

2.2. GEOMETRY OF RIVER SECTIONS AND THE RIVERBED ROUGHNESS PARAMETERS. The positions of computational points in the computation of linear flow correspond to the position of cross-sections which are given in a separate program editor called Cross-section editor. In this editor, the cross-sections of the river bed are given in the local coordinate system, while the topological sorting of the section is performed on the basis of giving the chainage which represents the spatial coordinates of the section in the calculation.

52

The representation of the Morača flowing into the Lake

The representation of the Bojana flowing out of the lake Figure 2.1.2. The representations of the Morača flowing into the Lake and the Bojana flowing out of the Lake

Under natural conditions in watercourses a non-uniform and unstable hydraulic regime is established. Non-uniformity of flow, and volatility of the characteristics of the hydraulic regime on the flow length is determined by morphological changes of the

53 river bed (longitudinal decline and cross-section), while non-uniformity of flow, especially of high-flow levels, is caused by changes in the hydrological regime over time. The existing riverbed of the Morača, both in terms of primary and high-flow channels, has variable dimensions, so that the flow of water within the entire range of hydrologic conditions is characterized by spatial variability of morphological parameters (longitudinal decline and cross-section of the riverbed), which means that the flow takes place under conditions of non-uniform hydraulic regime. Due to the small slope of the natural riverbed (average slope value on the entire section is approximately 0:11%O), water flow occurs in conditions of a steady hydraulic regime. On the other hand, the variability of the hydrological regime (water level and discharge), caused an unstable hydraulic regime. By regulating a watercourse the natural morphological characteristics of a river bed (declines and dimensions of cross sections) are balanced, whereby equalization of the parameters of the hydraulic regime (the range of their variation along the flow is reduced) is achieved, so that the flow can be treated as quasi-uniform, while the unstability is approximated by a number of stable states. This was one of the reasons for the mentioned proposed regulated channel to be adopted for the analysis of the river Morača since it meets the conditions that the model requires. Cross sections of the river Morača, as already mentioned, have been defined on the basis of data from the regulation project, while the sections of the river Bojana are given based on the earlier recording of the riverbed. The flow in Lake Skadar is described in the same way as in the rivers Morača and Bojana using a linear model. A heuristic approach was applied, where the bed of Lake Skadar was approximated using an optical square whose volume curve corresponds to the real lake volume curve obtained by recording the lake bottom. In addition, the length of the optical square corresponds to the distance between the downstream boundary of the Morača and the upstream boundary of the Bojana. The cross section of the optical square which in the computational model is the section of the Skadar Lake is shown in Figure 2.2.5. The data that were used for the previous definition of the Skadar Lake optical square were obtained by processing the bathymetric measurements carried out on the lake in the first phase of the project and helped to define the volume curve. Such an approximation can be considered fully satisfactory given the features of the models used and the accuracy of the measured data that served as the basis for all these calculations . The aim of the present treatment of the lake in the computational model was not as much getting an accurate representations of flow (it can be obtained only by using spatial or plane model) as it was to a apply a simple model on using retention capacity and length of the lake in assessing the mutual impacts of the Morača and Bojana. Also, since the volume is a function of depth this made it possible to, with sufficient accuracy, monitor also the lake level changes in different inflow and outflow variants. In addition to data on the geometry of the bed flow resistance coefficients are also given in the Cross Section Editor. The analysis employed the Manning coefficient whose value 0.03 m-1/3s is taken from the relevant resources [2, 5]. This value was used in the entire river network. The adopted value integrates typical values f or an uncoated bottom and slopes coated with gabion mattresses.

54

Fig. 2.2.1. A bathymetric map of Lake Skadar resulting from the recording of the Montenegrin and Albanian side during the first phase of the project

Fig. 2.2.3. A 3-D model of Lake Skadar, obtained on the basis of bathymetric measurements, which served as the basis for obtaining the volume and area curves of the lake

55

Sl. 2.2.4 . The volume and area curves of Skadar Lake obtained on the basis of processing the lake bathymetric measurements (blue –area, red – volume)

Fig. 2.2.5. The adopted cross section of Lake Skadar for the simulation purposes in the model

56

2.3. INITIAL AND BOUNDARY CONDITIONS OF THE MODEL In the computational experiments, whose results are presented in this paper, the initial conditions are defined by a constant initial discharge in the entire section and level elevations corresponding to a steady flow regime, for a given flow rate, assuming that in the downstream end of the flow field the elevation level equals 0 m.a.s.l (the mouth of the Bojana in the Adriatic Sea). The values of 136 m3/s and 800 m3/s were used as the initial flow rate values in the Morača. The first value represents the average perennial flow of the Morača at the “Botun” hydrometeorological station while the other value resulted from the requirement that the initial level for the Skadar Lake simulation is approximately 7.5 m.a.s.l. The boundary conditions are given by defining time series of discharge and levels. Thus, at the upstream boundary hypothetical hydrographs were given while at the downstream boundary the hypothetical levelgraphs were given. At the upstream boundary, which is the location of the confluence of the river Zeta in the Morača either constant flow or hypothetical flood waves were given, set so that the flow increases linearly up to the value of the largest recorded flow rate on the Morača (HS “Botun”) for 24 hours. At the downstream end, either a constant elevation of the Adriatic Sea or a hypothetical marigraph were given according to which sea level elevation reaches a maximum value of 1m.a.s.l In addition to giving the time series of discharge and levels at the borders, the MIKE11 software, among other things, allows assigning the inflow values at any other part of the calculation section (the so-called point source inflow). Thus, the influence of the river Drim has been analyzed precisely by defining the additional inflow at the point where it flows into the Bojana, just after it starts to flow out of Lake Skadar. The variants have been analyzed in which the flow rates amount to 2000, 3000 and 4000 up to the above-mentioned maximum of 7165 m3/s. In all calculation variations, the values of the inflow by the Drim increase linearly from the value 0 to the values specified for a period of 3 days.

3. NUMERIC SIMULATIONS RESULTS 3.1. VARIATION 1 This variation has analyzed the impact of flood wave appearing on the river Morača on the flow in other parts of the river-lake-river system, if the initial discharge in the system equals the mean annual flow on the Morača River. Under this initial discharge the calculated (initial) lake elevation is approximately 5.40 m.a.s.l. Initial condition: Q=136 m3/s and Zmore=0. Upstream boundary condition: Flood wave defined by hydrograph presented in Figure 3 (Qmax=2040 m3/s) Downstream boundary condition: Zmore=0.

57 The results of calculation indicate: 1. Partial wave attenuation in the surveyed section of the river Morača. The calculated peak in the downstream end of the section is 1980 m3/s. 2. Elevation of lake level to 5.75 m.a.s.l. 3. An increase in discharge on the river Bojana to about 215 m3/s.

Figure 3.1.1. Hypothetical hydrograph at the upstream end

Figure 3.1.2. Peak discharge on the Morača River in the most downstream part

58

Figure 3.1.3. Changes in the Lake water levels

Figure 3.1.4. Changes in discharge on the Bojana River

Figure 3.1.5. Typical cross-section of the Bojana river bed, water level remains within the river bed limits

59

3.2. VARIATION 2 The other examined variation differs from Variation 1 only in initial discharge of 800 m3/s, under which the lake elevation is 7.48 m.a.s.l. The time of wave peak appearance and the peak value in the upstream end are the same as in the first variation. The results of calculation: 1. Partial wave attenuation in the examined section of the river Morača. The calculated peak in the downstream end of the section is 2010 m3/s. 2. Elevation of lake level to 7.68 m.a.s.l. 3. An increase in discharge on the river Bojana to approximately 885 m3/s.

3.3. VARIATION 3 The third variation explores the impact of inflow through the river Drim defined by flood wave (Figure 3.1.6) whose peak value is 4000 m3/s under the initial lake elevation of 7.48 m.a.s.l. Initial condition:

Figure 3.1.6. Inflow through the river Drim at a value of 4000 m3/s

60 Q=800m3/s and Zmore=0. Upstream boundary condition: Q=800 m3/s (constant discharge) Downstream boundary condition: Zmore=0. The results of calculation: 1. In the section between the confluence of the Drim into the Bojana and Skadar Lake water flows towards the lake. The highest discharge in “negative” direction is 1700 m3/s. 2. Lake elevation level reaches approximately 8.90 m.a.s.l as compared to the initial 7.48 m.a.s.l.

Figure 3.1.7. Discharge through lake, the highest “negative” discharge of 1700 m3/s

Figure 3.1.8. Changes in the Lake water levels, approximately 8.9 m.a.s.l.

61

Figure 3.1.9. Elevation of water level on the river Bojana at more than 3 m

Figure 3.1.10. The Bojana river bed section, water level exceeds the limits of embankment, occurrence of floods on both sides of the bed

3. An increase in discharge on the river Bojana downstream from the Drim confluence to approximately 2450 m3/s. 4. Water level elevation on the river Bojana at more than 3 meters. 5. Water level elevation on the river Morača at approximately 1 meter.

3.4. VARIATION 4 The next two calculation variations differ from the previous one in values of peak inflow through the river Drim. In Variation 4 the peak value is 3000 m3/s, while in Variation 5 the peak value is 2000 m3/s. The results of calculation:

62 1. In the section between the confluence of the Drim into the Bojana and Skadar Lake water flows towards the lake. The highest discharge in “negative” direction is 1150 m3/s. 2. Lake elevation level reaches approximately 8.60 m.a.s.l 3. An increase in discharge on the river Bojana downstream from the Drim confluence to about 2300 m3/s. 4. Water level elevation on the river Bojana at approximately 2.8 m. 5. Water level elevation on the river Morača at approximately 0.7 m.

Figure 3.1.11. The highest discharge in the Lake in “negative” direction is 1150 m3/s

Figure 3.1.12. Water level elevation reaches approximately 8.60 m.a.s.l.

63

Figure 3.1.13. Increase in discharge on the river Bojana downstream from the Drim confluence to approximately 2030m3/s

3.5. VARIATION 5 The results of calculation: 1. In the section between the confluence of the Drim into the Bojana and Skadar Lake water flows towards the lake. The highest discharge in “negative” direction is 550 m3/s. 2. Lake elevation level reaches approximately 8.20 m.a.s.l 3. An increase in discharge on the river Bojana downstream from Drim confluence to approximately 1650 m3/s. 4. Water level elevation on the river Bojana at approximately 1.9 m. 5. Water level elevation on the river Morača at approximately 0.5 m.

3.6. VARIATION 6 Variation 6 looks into the coincidence of flood waves on river Morača (from variation 2) and the river Drim (from variation 3) with occurrence of floodtide which results in elevation of the Adriatic Sea level to level 1 (Figure 3.1.14). The initial discharge through the river Morača is 800 m3/s. The results of calculation: 1. In the section between the confluence of the Drim into the Bojana and Skadar Lake water flows towards the lake. The highest discharge in “negative” direction is 1700 m3/s. 2. Lake elevation level reaches approximately 9.05 m.a.s.l 3. An increase in discharge on the river Bojana downstream from the Drim confluence to approximately 2500 m3/s. 4. Water level elevation on the river Bojana at approximately 3.4 m.

64

Figure 3.1.14. Fluctuations of the Adriatic Sea level as a boundary condition in model variation

Figure 3.1.15. The highest discharge in the “negative” direction is 1700 m3/s

65

Figure 3.1.16. Water level elevation on the Lake reaches approximately 9.05 m.a.s.l.

Figure 3.1.17. Increase in discharge on the river Bojana downstream from the Drim confluence to approximately 2500 m3/s

Figure 3.1.18. Elevation of water level on the river Bojana at approximately 3.4 m

66

3.7. VARIATION 7 This variation explores the “catastrophic” scenario of occurrence of maximum wave on the river Drim of 7165 m3/s in the moment when a maximum wave appears on the river Morača (from variation 2) with the occurrence of floodtide which results in elevation of the Adriatic Sea to level 1 (Figure 3.1.14). The initial discharge in the system is 800 m3/s. This variation has been aimed at determining system state closest to the catastrophic flood which hit the whole area in 2010-2011. The results of calculation: 1. In the section between the confluence of the Drim into the Bojana and Skadar Lake water flows towards the lake. The highest discharge in “negative” direction is 3500 m3/s.

Figure 3.1.19. The highest discharge in the “negative” direction is 3500 m3/s

Figure 3.1.20. Water level elevation on the Lake reaches approximately 10.20 m.a.s.l.

67

Figure 3.1.21. Increase in discharge on river Bojana downstream from the Drim confluence to approximately 3500 m3/s

Figure 3.1.22. Elevation of water level on the river Bojana at approximately 5.0 m

Figure 3.1.23. Typical cross-section of the Bojana river, water level elevates to 5.0 m above the initial level, flood wave covers both banks

68 2. Lake elevation level reaches approximately 10.20 m.a.s.l 3. An increase in discharge on the river Bojana downstream from the Drim confluence to approximately 3500 m3/s. 4. Water level elevation on the river Bojana at approximately 5.0 m.

Figure 3.1.24. Longitudinal section of water discharge, the moment when the capacity of the Bojana becomes insufficient and a so-called negative discharge through the lake i.e. creating backwater on river Morača commences

Figure 3.1.25. Longitudinal section of water discharge, the moment when the lake elevation reaches maximum level, backwater on the river Morača reaches maximum length in the upstream, the capacity of the Bojana River insufficient

69

3.8. MODEL RESULTS – SUMMARY A few important conclusions can be drawn from the aforementioned model variations of different conditions of discharge and water levels in the Morača-Skadar LakeBojana system: The proposed regulation cross section of the Morača river bed meets all the requirements for protection of the riparian zone of this river in the area subject to modelling. It passes maximum river waters without them flooding the banks. When a “negative” direction of water discharge in Skadar Lake occurs, caused by the appearance of high waves from the river Drim, a certain “bung “ is formed at the mouth of the river Morača into the lake which slows down the river runoff into the lake and which is the main reason why flooding or effusion of the Morača river water in its most downstream part, in length of about 5 km, takes place. Moreover, in this part of the Morača river bed, even the proposed regulated channel would be inundated if a maximum wave of 7160 m3/s on the river Drim occurred. If discharge on the river Drim is below 5000 m3/s, “negative” discharge does not result in flooding i.e. the regulated channel on the Morača river takes in that “negative” wave with no impact on the environment. Fluctuations of Skadar Lake levels are primarily caused by the influence of the outflow velocity of the river Bojana from it and, to a lesser extent, by the influence of the inflow through the river Morača. A key factor in this situation is, as it turns out, once again, the river Drim. If the influence of the river Drim is excluded from the system, virtually none of the problems afflicting the riparian area of these watersheds would exist, including the effect of the aforementioned “negative” or upstream discharge through Skadar Lake. This effect leads to an increase in water level in the lake which occurs approximately three days upon the occurrence of high waves through the river Drim, in case of its maximum discharge. Through reconstruction of the known maximum water levels on the lake, the model has determined discharge and outflow values through the Morača and Bojana rivers. Thus, based on the up-to-now measured maximum value of 9.84 m.a.s.l., the model has shown that discharge through the Drim should be over 5500 m3/s. A variation of maximum discharge and levels has shown that the lake level of about 10.40 m.a.s.l is reached when the Drim discharge maximum is 7160 m3/s. Engaging in all future activities on the lake should be founded on knowledge that controlling the water level on the lake is possible, if it is possible to bring a key factor of its fluctuation – the river Drim, under control. Since it can be often heard in public that the development of an integral system of dams on the river Morača would put Skadar Lake into jeopardy, it is of utmost importance to emphasize that this is an arbitrary and absolutely incorrect qualification. If jeopardizing environmental conditions of the lake is the main argument against dam construction, it should be pointed out that in order to stabilize environmental conditions of the lake, preserve them and maintain their “good” status, the first and the best measure to be undertaken is stabilizing its level. Relying on natural inflow regimes has indicated that level stabilization cannot be carried out in a controlled manner since it is

70 a stochastic process. That is why reservoirs on the river Morača, which would provide reduction of flood flows and increase of low waters, and envisaged regulation in the section to the mouth of the river Sitnica, would be highly efficient especially when it comes to Skadar Lake ecosystem. It is of crucial importance to improve quality indicators in potentially critical environmental conditions in compliance with the results of monitoring of physical indicators of water quality in the lake and through increased purposeful discharge of waters from a reservoir. This is the exact reason why, throughout the world, reservoirs are purposefully built on rivers which flow into lakes. The River Bojana has a limited channel capacity due to its bed and longitudinal slope. The model has shown that water level on the river Bojana remains within the river bed in all discharge conditions in the Morača-Skadar Lake system, if the river Drim is excluded. Since, essentially, the river Drim is a tributary of the Bojana River, sudden discharge increase through the river Drim automatically poses a problem in the river bed of the Bojana, as a situation arises in which a river with higher discharge flows into a river with lower channel capacity. All variations which include higher discharge through the river Drim have indisputably shown that a rapid and significant increase in discharge and water level in the Bojana river bed occurs, which results in its effusion and flooding of its riparian zone. The value of 2000 m3/s can be set as a bottom inflow limit of the river Drim, assumed not to cause problems on the river Bojana. Any inflow value above this causes river bed overflow and flooding. It is necessary to carry out certain measurements in parts of the system, under different conditions, in order to calibrate the model results. There is a reliable set of data regarding the so-called mean discharge values on the river Moraca and water systems on Moraca and Skadar Lake and these data were the basis for previously explained specific calculation variations. Unfortunately, neither was it possible to carry out discharge measurements on the rivers Morača and Bojana during high flood wave nor do such measurements exist. It has already been pointed out that no discharge measurements have been carried out on the Bojana at all. There are certain assumptions about discharge values during a flood or maximum channel capacity of the Bojana river bed, presented in the outline of hydrological features of this river. Data based on which certain conclusions can be drawn refers to water levels on Skadar Lake. Since the commencement of undertaken surveys, maximum water level recorded at the hydrological station Plavnica has been 9.86 m.a.s.l. and water level value during the last 2010 floods reached 10.44 m.a.s.l. At the same time, the maximum water level on the Morača, as recorded by the measurement station in Podgorica, did not reach maximum level. The previously recorded maximum level was 1226 cm and during the last 2010 floods, the recorded level was 1177 cm, meaning it was lower by 49 cm. Based on this data, it has been concluded that the river Morača had no crucial impact on the recent floods around Skadar Lake but that floods were caused by discharge form the river Drim, which reached Skadar Lake through the Bojana. The calculation results in variation 7, aimed at presenting a scenario closest to the aforementioned disastrous flood, indicate that the measured lake level value is very close to the actual registered value (10.32 – 10.44) and that “ negative “ discharge through the lake i.e. upstream discharge towards the Morača occurs and the whole riparian zone of river Bojana and Morača is

71 hit by flooding. This is actually quite in line with what really happened in the area i.e. in line with what was recorded by either measurements or consequences of the flood. Since the model is a “live” substance, constantly evolving and improving, its exploitation calls for additional calibration through measurements which should be carried out primarily in periods of flood flows occurrence. The undertaken project is and has been aimed at using a contemporary ultrasound Doppler device for discharge measurements in conducting proposed measurements in such situations, thus evaluating the model results. Very short project duration, as well as the hydrological situation during its undertaking, in all parts of the catchment, has not allowed for that to be done at this stage of the model development but it is realistic to expect that with first flood waters in the catchments it can be conducted and evaluated.

4. SHORT MANUAL FOR USING MIKE 11 COMPUTER PROGRAM ON THE MORAČA – SKADARSKO JEZERO– BOJANA MODEL A computational model of unsteady flow is created by defining river network, channel geometry, specifying flow resistance parameters, defining initial and boundary conditions and entering computational scheme parameters. Data needed for unsteady flow calculation in MIKE11 program are input in multiple different windows which form independent program interfaces (so-called editors). All windows used for creating a computational model through MIKE 11 program are opened from the environment named MIKE Zero.

4.1. NETWORK EDITOR (FILE->NEW->FILE->MIKE11-> RIVER NETWORK) Defining a computational grid is performed in a window called “Network Editor”. First, in this part of the program, a coordinate system is defined by choosing option “Workspace area and network projection” in the drop-down menu “Network.” River network is defined by a pecked line, which is set by using a group of tools “River Network Tools” (Figure 1). These tools are used to interactively set network line grid points, to define sections and flow branches points (or points in the network marking a confluence). First, by using an “Add New Points” tool flow grid points are defined and after that, by using a “Define Branch” tool, sections are defined through marking the created points. The point marked first in the process of section defining is assigned chainage 0 (in MIKE11 program due to initial settings, chainages are measured from the upstream end of a section). A tabular view of grid points coordinates as well as sections view is generated by choosing option “Tabular view” from the drop down menu “View”. It is important to note that grid points are not computational points as they affect topological sorting of sections as well as values in a particular chainage section. The position of computational points is determined by assigning cross sections chainages in the “Geometric Editor” window. Important computational data input in a tabular view is the maximum permissible distance between computational points (cross sections) which is entered in the column “Maximum dx” in the “Network/ =Branches” table. If at a certain section, the

72

Figure 4.1. Network Editor for system of the rivers Morača and Bojana and Skadar Lake

distance between cross sections is larger than the Maximum dx value, program inserts interpolated cross sections, prior to computational simulation. In order to provide an easier input of a river network, introducing georeference maps is made possible by choosing the “Add/Remove” option from the drop down menu “Layers”. Alternatively, it is possible to specify a river network via ascii text files (see the official manual). The file in which data created in the “Network Editor” window are stored has nwk11 extension (File type is M11Net.Document).

4.2. CROSS SECTION EDITOR (FILE->NEW->FILE->MIKE11-> CROSS SECTIONS) Data on river bed geometry and flow resistance are set in the program section called “Cross Section Editor” (Figure 4.2). A new cross section is specified by choosing the “Insert Cross Section” option. When creating a new cross section, it is necessary to enter a cross section chainage and the name of the cross section location. Based on these data, topological cross section sorting is carried out. Once again, it should be noted that according to “default” settings, a chainage is a distance along the cross section flow from the upstream end of the section.

73

Figure 4.2. Cross Section Editor of the river Bojana

Geometry is defined in a tabular view by means of x and z coordinates. Coordinate x is a horizontal distance between a cross section point and the first point (end point) on a left bank, while coordinate z is an absolute elevation of a cross section point. In addition to coordinates, the value of corresponding flow resistance is to be input in the table edited in this section of the program. The method in which flow resistance is modelled is defined by choosing the “Radius Type” option and the “Resistance Type” option which can be accessed via the “Cross-Sections” menu by choosing the“Apply to All Sections/Apply to Selected Sections” options (depending on whether the changes apply to all cross sections or only the marked ones). It is important to note that in this section of the program, prior to computation, a command “Recompute All” is to be executed. The command is located in the same window where the type of resistance radius is entered. Details on methods of flow resistance modeling may be found in the official model manual (Reference Manual). It is recommended that in case of sudden changes in width of water tables, with a change of depth, flow resistance rate values be changed in the points where the change occurs or alternatively, that the so-called Resistance Radius be used instead of hydraulic radius. Database file which stores data created in the “Cross Section Editor” window has xns11 extension ( File type is XSec.Document ).

74

Figure 4.3. Boundary Conditions Editor for Morača-Lake-Bojana system

4.3. BOUNDARY EDITOR (FILE->NEW->FILE->MIKE11-> BOUNDARY CONDITION) Boundary conditions are entered in the “Boundary Editor” window. Figure 3 shows the window with inserted boundary conditions on an example of unsteady flow modeling on rivers Bojana and Morača. As the figure shows, it is a description of the boundary condition based on which the type is chosen that is entered in the window first. The presented example illustrates that there are three boundaries which are “open” and variations in time of water level elevation and discharge are inserted, while one boundary is a point in cross section at which flow load is added (hence the name “Point Source”). “Open type” boundaries must represent the ends of cross sections i.e. chainages which are entered in column Chainage must be the same as end point chainages of a respective cross section, inserted when the river network has been defined (in the “Network Editor” window). Needless to say, it is necessary to specify, in the “BranchName” column, which cross section the boundary conditions refer to (name of the cross section must match the name determined when a river network has been defined). By editing the type in the upper table in the “Boundary Editor” window, the table below changes for a respective boundary condition.The table below is the table in which the file, that represents a time series of physical item whose change is a boundary condition, is selected. Time series are entered in a separate MIKE11 window (”Time Series Editor”) which is accessed via the “File” menu by choosing the “New” option (File>New->File->MIKE Zero->Time Series). If a boundary condition is an item whose value

75 does not change in the computation, instead of a time series file, option “Constant” in the column “TS Type” is chosen, after which an appropriate value can be entered. The database file which stores data created in the “Boundary Editor” window has bnd11 extension (File type is Boundary.Document).

4.4. PARAMETER EDITOR (FILE->NEW->FILE->MIKE11-> HD PARAMETERS) Initial conditions and parameters of a computational scheme, among other items, are input in the “Parameter Editor” window. The window in which initial conditions are input appears upon choosing the“Initial” tab. Initial conditions are set by defining discharge values and water levels (or depth) at the river network points defined by the name of the cross section and chainage (see Figure 4.4). Rows in the table are created by pressing the Insert button on the keyboard. The program assigns linear interpolated respective initial values for network sections between points at which initial conditions are defined. It is important to note that another method of initial conditions input may be chosen in the window in which simulation is run, so the program does not take into account data unseen in the “Initial “ tab (described as follows). Computational scheme parameter values are input by choosing the “Default Values” tab in the “Parameter Editor” window. It should be noted that the choice of com-

Figure 4.4. Parameter Editor for Morača-Lake-Bojana system, initial values

76

Figure 4.5. Scheme Parameters Values used in the analysis of discharge of the rivers Morača and Bojana

putational scheme parameter values depends on the case in question. General recommendations can be found in the official manual for using the MIKE11program. Figure 4.5 shows the scheme parameters values used in the analysis of discharge of the rivers Morača and Bojana. The database file which stores data created in the Parameter Editor window has hd11 extension (File type is HDPar.Document).

4.5. SIMULATION EDITOR (FILE->NEW->FILE->MIKE11-> SIMULATION) The window in which simulation is run is called “Simulation Editor”. It comprises five tabs. The type of computation model is chosen in the first tab on the left side, called “Models”. If “only” complete Saint Venan equations are solved, it is necessary to choose “Hydrodynamic” option in “Models” group of options and the option “Unsteady” in the “Simulation Mode “group. The input tab is used to “link” files created in the previously described windows. Thus, paths to files created in the “Network”, “Cross-Sections” and “Boundary Data” fields are specified in the corresponding windows of the same name, while a path to the

77 file created in “Parameters Editor” field is specified in the “HD Parameters” window. These four files are required to solve Saint Venan equations, i.e. to carry out unsteady flow simulation. By using the “Simulation” tab, the time step value, simulation period value and method of specifying an initial condition are chosen. Recommendations regarding time step selection can be found in the official manual. Initial conditions type is chosen via a drop-down menu “Type Conditions” (a group of options “Initials Conditions”). A program user can choose a level line based on steady flow computation (“Steady State” option) as initial condition, pre-defined data in the “Initial” tab in the “Parameter Editor” window (option “Parameter”), a combination of previous two options or elevation levels and discharge rates in a particular moment of time measured through some of previously conducted computational simulations (option “Hotstart”). It is recommended that in cases where an initial condition is defined through steady flow, “Hotstart” option be used (the use of this option is explained in the appendix to the text). Once all the necessary computation data are determined, it is necessary to define a database file in the tab “Results” in which computation results and the time step at which results are displayed will be stored. Finally, a simulation is run in the “Start” tab, using “MIKE 11 Classic” or “MIKE 1D” commands. The database file in which the results are stored has res11 extension and it is opened, just like all the files created in the MIKE software package, from the main window of MIKE Zero environment.

4.6. APPENDIX: USING HOTSTART OPTION FOR OBTAINING INITIAL CONDITION WHICH CORRESPONDS TO STEADY DISCHARGE

Figure 4.6. An example of calculated discharge in one cross section

78 In case the “Hotstart” option is chosen when initial conditions are defined, it is necessary to carry out a flow simulation in which boundary conditions are set in the form of a constant value corresponding to initial conditions for ”real” simulation prior to conducting “real” unsteady flow computation. Thus, in the example of unsteady flow simulation of the rivers Bojana and Morača, a computation was performed in which constant flow was specified as a boundary condition in the downstream end and a constant elevation level (sea elevation level) was specified as the boundary condition in the downstream end of the section. The duration of this “anterior” simulation was chosen iteratively, by reviewing the results, based on condition that the computed values of the physical items are steadied (an example of calculated discharge in one cross section from “anterior” simulation is presented in Figure 4.6). The database file that stores computational results of the “anterior” simulation is loaded during the process of “real” simulation parameters setting by defining file paths in the “Hotstart Filename” field (the “Simulation” tab in the “Simulation Editor” window). It is important to note that the start-time of the “real” simulation must correspond with the time of the “anterior” simulations in which a steady flow has been reached (as for the example in Figure 4.6, that would mean that the start-time of a new simulation should not be before 18. 02. 2005.)

REFERENCES [1] „Glavni projekat regulacije rijeke Morače na potezu od ušća Sitnice u Botunu do Ponara”, 2010. god [2] „Glavni projekat regulacije rijeke Morače na potezu od Vrbiša do Mišurice“, 1987. god. [3] „Glavni projekat regulacije rijeke Morače na potezu od Vrbiša do ušća Cijevne“, 1988. god, „Idejno riješenje regulacije reke Morače na potezu od Mišurice do Ponara“, 2007. godine [4] „Glavni projekat regulacije rijeke Morače na potezu od Vrbiša do ušća Cijevne“, 1988. god, „Idejno riješenje regulacije reke Morače na potezu od Mišurice do Ponara“, 2007. godine [5] „Projekt održavanja protočnosti desnog rukavca reke Bojane u zoni njenog ušća“, Republika Crne Gore UPRAVA ZA VODE, Podgorica, 2011. godine [6] „Regulacija Skadarskog jezera, Drima i Bojane“, Idejni projekat - izvod, Vodoprivredna organizacija „Zeta“, Titograd, 1981. godine [7] „Tenderska dokumentacija na korekciji i čišćenju kanala Port Milena i pročišćavanju desnog rukavca korita Bojane od ušća u Jadransko more do razdjelne račve sa međudržavnom granicom“, Republika Crne Gore UPRAVA ZA VODE, Podgorica, 2006. godine [8] Branislav Đorđević, Goran Sekulić, Mićko Radulović, Milinko Šaranović; Vodni potencijal Crne Gore, Monografija CANU, ISBN 978-86-7215-243-6, Podgorica 2010. god. [9] British Standard Code of practice for Maritime structures, Part 1. General criteria, 1984. [10] Chow, T. C., Maidment, D. R. and Mays, L.W., 1988. Applied Hydrology, McGraw-Hill Series in Water Resources and Environmental Engineering, pp.201-210, 257-263ž [11] CORINE, 1998. Coastal Erosion, European Communities Publication [12] Danish Hydraulic Institute, 2011, MIKE 11 Reference manual, Appendix A. Scientific background, Danish Hydraulic Institute

79 [13] DHI. 2012. Mike-SHE. Short description. Hørsholm, Denmark, 2012 [14] Drin River hydropowerplants, Controls and Monitoring Systems, 2006. www.news. admin.ch [15] Eurosion, 2003. Living with coastal erosion in Europe. www.eurosion.org [16] Komar, P.D., 1976. Beach Processes and Sedimentation, Prentice- Hall, Inc. New Jersey [17] MIKE 11 (2011) ‘A Modelling System for River and Channels’, Reference Manual, DHI Water & Environment. [18] Nacrt temeljne studije za uspostavljanjem zaštićenog područja Delte Bojane, Berlin, 2008. [19] Niko PANO, Bardhyl AVDYLI, A Method to Estimate Buna River Discharge, Albania, Hydrology Days 2009 [20] Rapid Assessment of proposed Hydropower Plants on Drin River near Ashta (south of Shkodra). Euronatur, 2009. [21] Regional Environmental Action Plan, Drini river delta, Shkodra – Lezhe, Regional Environmental Center (REC), Albania, 2006. [22] Regulacija Skadarskog jezera, Drima i Bojane“, Idejni projekat, R.O. VODOPRIVREDA BIH, Zavod za vodoprivredu, Sarajevo, 1981. godine, [23] Schneider-Jacoby, M., U. Schwarz, P. Sackl, D. Dhora, D. Saveljic and B. Stumberger (2006): Rapid assessment of the Ecological Value of the Bojana-Buna Delta (Albania / Montenegro). Euronatur, Radolfzell. [24] Slavoljub Jovanovic and Nenad Djordjevic, Problems of water balance and trend in the water level of Lake Skadar, IAHS Publication No. 109, 1974 [25] Wind and Wave Atlas of the Mediterranean Sea, WESTERN EUROPEAN UNION, WESTERN EUROPEAN ARMAMENTS ORGANISATION, 2004. [26] Zavod za hidrometeorologiju seizmologiju Crne Gore: Arhivska dokumentacija sektora za hidrologiju [27] http://www.mikebydhi.com/Products/WaterResources/MIKESHE.aspx

MONTENEGRIN ACADEMY OF SCIENCES AND ARTS

IPA PROJECT

„DEVELOPMENT OF HYDROLOGICAL AND HYDRAULIC STUDY OF REGULATION OF SKADAR LAKE AND BOJANA RIVER WATER REGIME” VOLUME II

Prepress Bojаn R. Popović Medeon – Podgoricа Copies ____________________ Printed by _________________________ 2014.

CIP – Каталогизација у публикацији Национална библиотека Црне Горе, Цетиње ISBN COBISS.CG-ID

ipa project

ipa project

Suggest Documents

IPA

Hakikat IPA dan Pembelajaran IPA

Hakikat IPA dan Pendidikan IPA

PERKEMBANGAN PENDIDIKAN IPA DI INDONESIA Kurikulum IPA ...

FIRST CALL FOR PROJECT PROPOSALS UNDER IPA CROSS ...

Template of Project Fiche for IPA programmes / component I

Template of Project Fiche for IPA programmes / component I

Template of Project Fiche for IPA programmes / component I

Template of Project Fiche for IPA programmes / component I

Template of Project Fiche for IPA programmes / component I

CHINOOK IPA

ADE IPA

IPA SMP

IPA 5b

Dunia IPA

B IPA

IPA - WordPress.com

IPA-AR_2016

Dunia IPA

Matematika IPA

Pronunciation Notes Jason A. Zentz IPA Garner Examples IPA ...

Ulangan IPA Kelas X “Hakikat IPA & Metode Ilmiah” - WordPress.com

RPP IPA Kelas 8 semester 2 - BELAJAR IPA ONLINE

Kurikulum Prodi Pendidikan IPA S1 (2010) - Pendidikan IPA | Unnes