BEACH EROSION AND COASTAL PROTECTION PLAN ALONG THE SOUTHERN ROMANIAN BLACK SEA SHORE

BEACH EROSION AND COASTAL PROTECTION PLAN ALONG THE SOUTHERN ROMANIAN BLACK SEA SHORE Keiji Kuroki, 1 Yoshimi Goda, 2 Nicolae Panin, 3 Adrian Stanica, 4 Danut I. Diaconeasa, 5 and Gheroghe Babu 6 The state of beach erosion of the southern unit of the Romanian Black Sea shore is described with explanation of its mechanism. Beaches in the northern part of the study area are composed of terrigenous sand from the Danube, while the sediment in the southern beaches is made of shell fragments. Shoreline changes over years have been well reproduced with the one-line model and the net sediment transport rate is evaluated at several coastal units INTRODUCTION Romania has the coastline of 245 km along the northwestern shore of the Black Sea together with Bulgaria, Turkey, Georgia, Russia, and Ukraine in the counterclockwise direction. Similar as any coast in the world, the Romanian coast has been suffering from the problem of beach erosion. The Government of Romania requested the Government of Japan to extend technical assistance for a study to formulate the coastal protection plan for the Southern Romanian Black Sea shore and to make a feasibility study on priority project(s). On behalf of the Japanese Government, the Japan International Cooperation Agency (JICA) undertook the study, which began in March 2005 and will be completed in early 2007. The present paper is based on the interim report of the study submitted to the Government of Romania in February 2006. The Romanian coastline in divided into the northern and southern units. The northern unit is the deltaic coast of the Danube, which is the second largest river in Europe. The sediment transport load of the Danube became less than one half in recent several decades owing to construction of many dams in her mainstream and tributaries (Bondar and Panin, 2000). Extension of two entrance jetties by 8 km at the Sulina Channel near the border with Ukraine, which is opened to the international fairway through the Central Europe, has aggravated the beach erosion severely. Some sections of the northern unit indicate the shoreline recession with the rate of 10 m/year with the maximum of 19 m/year. The southern unit of the Romanian Black Sea Coast, which is the area under study, is also experiencing severe beach erosion. Figure 1 shows the location map 1 ECOH CORPORATION, Tokyo, Japan: k-kuroki@ecoh.co.jp 2 ECOH CORPORATION, Tokyo, Japan: goda@ecoh.co.jp 3 National Institute of Marine Geology and Geo-ecology (GeoEcoMar), Bucharest, Romania: panin@geoecomar.ro 4 GeoEcoMar: Adrian_stanica@geoecomar.ro 5 National Institute for Marine Research and Development Grigore Antipa, Constanta, Romania, ddan@datanet.ro 6 Apeloe Romane Doborogea Litoral, Constanta, Romania: babu@dadl.rowater.ro 3788

COASTAL ENGINEERING 2006 3789 of the study area, which extends from Midia to Vama Veche next to the border with Bulgaria over some 80 km. Midia, Constan a, and Mangalia are the port areas, while Mamaia, Eforie, Constine ti and Vama Veche are the famous resort beaches. ROMANIA Danube River BULGARIA Midia Mamaia Constanta Eforie Costinesti Mangalia Vama Veche Figure 1. Location map of the Southern Romanian Black Sea shore. PRESENT STATE OF BEACH EROSION AND ITS MECHANISM The shoreline position change has been investigated by two methods: one with comparison of topographic maps and satellite images and another with regression analysis of the beach profile data, which have been surveyed by the National Institute for Marine Research and Development over years. Figure 2 is an example of the regression analysis of the shoreline distance from the bench mark M-14 at Mamaia Beach. Mamaia Beach was around 100 m wide in the 1960s, but severe erosion began in the late 1970s, when the north breakwater of Midia Port, which is located at the northern end of the 11 km long beach, began to be extended to the water of 10 m deep. The terrigenous sand from the Danube has been transported southwestward along the northern unit of the Romanian Black Sea Coast over tens of millenniums. The transported sand contributed to elongation of sand spits at a number of embayments and finally to their closure with barrier beaches. The barrier beach of Mamaia seems to have landlocked the embayment of Siutghiol around the second century. Being alarmed by disappearance of beaches of Mamaia, the Romanian Government built six detached breakwaters and took a beach fill operation in 1988 to 1990. The decrease of the shoreline distance before 1986 in Figure 2 indicates the beach erosion caused by the extension of the Midia breakwater, a jump of the distance between 1986 and 1991 is the result of beach fill operation, and a straight line fitted to the data between 1989 and 2005 indicates the shoreline regression rate of 2.25 m/year.

3790 COASTAL ENGINEERING 2006 Figure 2. Regression analysis of the distance of shoreline position at Mamaia Beach. The shoreline position change rate analyzed for 34 bench mark data is shown in Figure 3. A large recession rate at MM-7 is due to the local erosion at the northern boundary zone of a series of detached breakwaters. At the southern area of Mamaia Beach, an average erosion rate of 2 m/year is registered. Along Mamaia Beach, the northward sediment transport is predominant, but no clear shoreline advance is observed at its north end; there is a possibility of offshore loss of sediment there. In the south of Constan a Port, low sea cliffs occupy the majority of the coastline. Constantinescu (2005) revealed the mean cliff erosion rate of 0.5 to 0.7 m/year on average through comparison of topographic maps in 1924 and the satellite Ikonos images taken in 2002. A shoreline advance at EF-1 at Eforie Nord is the result of tombolo formation by construction of a marina after 1986. A conspicuous shoreline recession of 2.5 m/year at EF-6 seems to be caused by reflected waves from a collapsed seawall located south of EF-7. The shoreline recession of about 1.5 m/year at a beach of SN-1 and -2 would have been compensation of suppressed beach erosion at both the north and south of the beach by shore protection facilities built there in the 1970s. At the bench mark VV-1 ofvama Veche, the shoreline recession rate of 0.7 m/year corresponds to cliff erosion that has been taken over millenniums. SEDIMENT SOURCES REVEALED BY MINERAL CONTENT ANALYSIS Several dozens of sediment samples were taken from the foreshore of various locations in the study area. The median diameter of sediment in Mamaia Beach was 0.18 to 0.27 mm, while the beaches from Eforie to Vama Veche varied from 0.37 to 0.70 mm.

COASTAL ENGINEERING 2006 3791 Figure 3. Shoreline position change rate in the study area. Sediment samples were exposed to the fluorescence X-ray analysis for measurement of their mineral contents. Figure 4 exhibits the alongshore variations of silicon dioxide (SiO 2 ), calcium oxide (CaO), titanium dioxide (TiO 2 ), and manganese oxide (MnO). Silicon dioxide constitutes quartz, while calcium oxide represents calcium carbonate, which is the major component of limestone and mollusk shells. Titanium dioxide (TiO 2 ) and manganese oxide (MnO) originate from the mountain area, and thus they are characteristics of terrigenous sand. The data are arranged from north to south, but the top sample marked R2 is the Danube river sand at the distance of 370 km from the Sulina entrance. The point Z2 is at a beach north of Midia, the points A1 to B3 are at Mamaia and Tomis Beaches, while the points C1 to G2 are located at Eforie and its south. Between B3 and C1 there is the harbor area of Constan a Port. As clearly

3792 COASTAL ENGINEERING 2006 shown in the left diagram of Figure 4, the sediment in the north has quartz as the major component, while the sediment in the south of Constan a Port is dominated by calcium carbonate. The right diagram demonstrates the absence of titanium Ti and manganese Mn in the south of Constan a Port. Figure 4. Alongshore Distributions of mineral contents Thus it is proved that the terrigeneous sand from the Danube has been brought to the beaches down to Cape Constan a, but it was not transported beyond it. The beach sand in the coast south of Constan a Port is made of shell fragments and some fragments of limestone that form the base of loess deposited coastal cliffs. WAVE CLIMATE AND ENERGY AVERAGED WAVES Offshore wave conditions have been analyzed with the hindcast data of the European Centre for Medium Range Forecasting (ECMWF) for the period of November 1991 to June 2002. The analysis was supplemented with the visual wave observation data having been conducted by the National Institute for Marine Research and Development, which provided the data supporting the validity of ECMWF hindcast wave data.

COASTAL ENGINEERING 2006 3793 Table 1. Monthly significant wave height (period) off Constan a Month Mean 10% exceedance 1% exceedance Jan. 1.11 m (5.1 s) 2.0 m (6.7 s) 4.4 m (9.1 s) Feb. 1.06 m (5.1 s) 2.0 m (6.7 s) 3.6 m (8.4 s) Mar. 1.05 m (5.2 s) 1.8 m (6.8 s) 3.7 m (8.2 s) Apr. 0.88 m (5.1 s) 1.6 m (6.6 s) 2.9 m (8.0 s) May 0.71 m (4.8 s) 1.4 m (6.4 s) 2.6 m (7.8 s) Jun. 0.65 m (4.5 s) 1.3 m (5.8 s) 2.2 m (7.3 s) Jul. 0.68 m (4.5 s) 1.4 m (6.1 s) 2.3 m (7.5 s) Aug. 0.73 m (4.8 s) 1.4 m (6.2 s) 2.3 m (7.8 s) Sep. 0.89 m (4.9 s) 1.7 m (6.6 s) 3.1 m (8.0 s) Oct. 1.03 m (5.2 s) 2.0 m (6.9 s) 3.4 m (8.6 s) Nov. 1.18 m (5.3 s) 2.4 m (7.2 s) 4.4 m (9.4 s) Dec. 1.31 m (5.5 s) 2.7 m (7.4 s) 4.9 m (9.5 s) Annual 0.95 m (5.1 s) 1.8 m (6.6 s) 3.6 m (8.4 s) Table 2. Estimate of return wave height and period Return period Wave height (m) Wave period (s) 5 years 6.08 9.9 10 years 6.52 10.2 50 years 6.52 10.2 100 years 7.83 11.0 Table 3. Waves selected for shoreline simulation Waves Northerly waves Southerly waves Wave direction N64.0ºE N115.2ºE Wave height, H1/3 (m) 1.65 1.11 Wave period, T1/3 (s) 6.2 6.2 Directional spreading, smax 25 25 Table 1 lists the monthly average of significant wave height and period for the mean, 10% exceedance, and 1% exceedance waves. Waves are high during winter and low during summer. The months of December and January have the roughest waves and the months of June and July have mildest waves. If the winter season is defined to include November to March, the mean significant wave height is 1.16 m, while the summer season (April to October) has the mean significant wave height of 0.79 m. Extreme wave analysis has yielded the estimates of significant wave heights and periods corresponding to the return periods of 5, 10, 50, and 100 years as listed in Table 2. The 100-year significant wave height is 7.8 m. The wave direction is northeasterly during the winter season, while it is southeasterly during the summer season. However, the northerly and southerly waves appear throughout the year with their ratio varies depending on the season. Based on the direction-wise wave energy calculation, two representative waves for shoreline simulation have been selected as listed in Table 3.

3794 COASTAL ENGINEERING 2006 NUMERICAL SIMULATION OF SHORELINE CHANGE WITH ONE-LINE MODEL Methodology of Numerical Simulation Shoreline changes over years have been calculated with one-line model that employs the following formula by Ozasa and Brampton (1979): 1 H b Q ( EC g ) b K1 sin b cos b K 2 cot cos b ( (1) s ) g(1 ) y where K 1 and K 2 are empirical constants. Their values have been set at K 1 = 0.154 and K 2 = 0.125 for Mamaia Beach and K 1 = 0.109 and K 2 = 0.0.088 for the beaches south of Eforie through calibration with the records of shoreline changes. The smaller values of the latter beaches reflect the larger grain size than the former beach. Directional wave transformation from the deepwater to offshore the surf zone was analyzed with the wave energy balance equation due to Karlsson (1969). Variation of wave height due to diffraction by breakwaters and other barriers is estimated by means of the angular spreading method for the sake of simplicity. Wave breaking is judged with the criterion of H b = 0.8h, and a search is made to find out the location to satisfy this criterion. Inshore the line of wave breaking, monochromatic wave approach was taken for the analysis of refraction and other transformations. The size of computation domain varied from a sector to another. The sector from Midia to Constan a (Mamaia and Tomis Beaches) was the largest with the distance of 20,000 m divided into 1001 grid points with the spacing of 20 m. Computation was executed continuously over 20 years or longer with the time step of 50 to 200 minuets. Monthly variations of the northerly and southerly waves were introduced by means of the monthly ratio of wave energy to the annual average waves listed in Table 3. The depth of closure h c was calculated by the following formula by Hallermeier (1978): 2 H e hc 2.28H e 68.5 2 (2) gte where H e denotes the significant wave height to be exceeded for 12 hours per year. The wave climate off Constan a yields H e = 5.0 m and T e = 9.1 s, and the closure depth is calculated as h c = 9.3 m. For the coast south of Eforie, the closure depth has been set at h c = 7.1 m in consideration of larger grain size than the sand in Mamaia Beach. Calibration of One-line Model with Field Data Along Mamaia Beach, the topographic maps of shoreline were available for the years 1976, 1980, 1990, 1995, and 1997. Mutual comparison of the shoreline position has quantitatively revealed the advance and retreat of the shoreline.

COASTAL ENGINEERING 2006 3795 Figure 5 shows the comparison of the shoreline changes evaluated from the topographic maps and the computation results during the four periods between the five survey years (N vodari is located at the north of Mamaia). The top diagram is the plan of the computed area which has the dimensions of 20 km and 5 km. The left edge is Midia Port bounded by its south breakwater. The right edge is the wharf area of Constan a Port. Six detached breakwaters are indicated at the center. The second diagram from the top compares the change shoreline positions in 1976 and 1980. The full line is the computed shoreline in 1980 relative to the shoreline of 1976, while the dashed line is the measured shoreline. The accreted portion according to numerical simulation is represented with small dots, while the eroded portion is shown with hatches. In the period between 1976 and 1980, beach erosion at the southern part of Mamaia Beach appeared and the change is reproduced in the simulation. Construction of six detached breakwaters and operation of beach fill began in 1988 at the southern part of Mamaia Beach. The shoreline advance by beach fill and start of salients behind detached breakwaters are clearly recognized in the third diagram (1976 to 1990) from the top of Figure 5; the pattern of changes in the simulation well fit to those by the measurement. The advance of shoreline by beach fill at the southern end of Mamaia Beach disappeared in the data of 1995 as seen in the second diagram from the bottom of Figure 5. The diagram at the bottom of Figure 5 indicates the changes of the shoreline position between 1976 and 1997. The measured and simulated shoreline changes agree well on the whole, even though there remain some differences. The shoreline advance predicted by simulation in the northern part but not materialized in reality is one of the differences. As described later, the net alongshore sediment transport is northward for the whole area of Mamaia Beach; sediment is deprived from the southern part (right) and transported to the northern part (left). Because no cross-shore loss of sediment was introduced in the numerical simulation, the sediment transported alongshore did accumulate in the northern area. The difference between the measurements and the simulation suggests an appreciable cross-shore transport toward the offshore in the northern area. Alongshore Sediment Transport Rate The numerical simulation also yields estimates of the sediment transport rate along the coast. For the coastal sector of N vodari to Tomis, the sediment transport rate is calculated as shown in Figure 6 for the shoreline shape in 2006. The top diagram is the plan shape of the coastal sector, which is same as in Figure 5. The middle diagram shows the estimated transport rate in the southward (upper) and northward (lower) directions.

3796 COASTAL ENGINEERING 2006 Figure 5. Comparison of the shoreline changes from measurements and simulation in the coast from N vodari to Tomis from 1976 to 1997

COASTAL ENGINEERING 2006 3797 Figure 6. Alongshore sediment transport rate in the coast from N vodari to Tomis

3798 COASTAL ENGINEERING 2006 Because of wave sheltering effects of the detached breakwaters which are shown with dots, the transport rate shows large variations locally. Outside the breakwater zone, the sediment transport varies smoothly. The variation of the net transport rate is shown in the bottom diagram with an enlarged scale of transport rate. The northward transport is up to about 160,000 m 3 per year while the southward transport is up to about 140,000 m 3 per year. The difference is about 20,000 m 3 per year in the northward direction. The maximum rate of the net northward transport at 24,000 m 3 per year is observed at the location from 9,000 to 10,000 m, which is behind the northernmost detached breakwater. Around this location, the alongshore gradient of wave height is largest. The significant erosive tendency at the benchmarks MM-7 and MM-8 shown in Figure 2 corresponds to the largest net transport rate there. Numerical simulation carried out at other coastal sectors has yielded the estimated values of alongshore sediment transport rate there. At Eforie Middle, the one-directional (northward or southward) alongshore sediment transport rate amounts to about 100,000 m 3 per year. The net sediment transport is southward with the maximum rate of 7,000 m 3 per year. At Costine ti, the northward and southward sediment transport rates are both about 100,000 m 3 per year, and they are almost equal so that the net transport is slightly southward with the rate of 60 m 3 per year only. The very small rate of the net sediment transport explains a healthy stability of Costine ti beach. At Vama Veche, The one-directional sediment transport rate amounts up to about 130,000 m 3 per year, but the northward and southward transport rate is almost equal and there is little net transport. If not limited by cliffs, the net transport rate may become 250 m 3 per year southward, but the presence of cliffs limits the net transport to 40 m 3 per year northward. COASTAL PROTECTION PLAN The study area has the beach and cliff coasts of about 59 km long out of the total length of about 80 km; the rest belongs to the port areas of Midia, Constan a and Mangalia. The beach and cliff coasts are divided into seven sectors and 20 subsectors. Five sectors are regarded as the independent coastal littoral cells by themselves and two sectors have two coastal littoral cells within them. In total, nine independent coastal littoral cells are recognized. Necessity for implementing coastal protection and rehabilitation projects has been evaluated on the basis of the beach erosion rates, the state of coastal utilization, and the viewpoint of environmental protection. Thus, project implementation has been recommended for nine sub-sectors among twenty. The proposed measures of coastal protection and rehabilitation include beach fill with the volume of 200,000 to 800,000 m 3 per project, construction of jetties of 200 to 400 m long that can function as artificial headlands, series of submerged, widecrested breakwaters (sometimes called the artificial reefs), and others. The total

COASTAL ENGINEERING 2006 3799 volume of beach fill sand will exceed 3 million m 3, but the plan is to be executed by one or two projects for one time over a long time, say 20 years or more. Among nine sub-sectors recommended for project implementation, two sites at Mamaia South and Eforie Nord have been selected for priority implementation and a feasibility study has been carried out by the JICA study team. The one-line model calibrated with the past records of shoreline changes is utilized for evaluation of the beach protection capacity of various alternative facility plans and selection of the optimum plan. It is hoped that the priority projects will be implemented at the earliest time possible. CONCLUDING REMARKS The study on beach erosion and coastal protection planning along the Southern Romanian Black Sea shore has successfully been carried out, thanks to the invaluable database of geophysical observation data accumulated over years at the National Institute for Marine Research and Development, the National Institute of Marine Geology and Geo-ecology, and other institutions. The study is an example of highlighting the necessity of long-range field measurements in the coastal area. The one-line theory has again proved to be an effective model to simulate the shoreline change over the coast of several tens of kilometer long and the period of scores of years, when field measurement records of beach topography over many years are available for providing the good data for model calibration. ACKNOWLEDGEMENT The authors wish to express their sincere gratitude to the Japan International Cooperation Agency for its permission to publish the present paper. REFERNCES Bondar, C. and Panin, N. 2000. The Danube delta hydrologic database and modelling, GEO-ECO-MARINA, 5-6/2000-2001, pp. 5-52 Constantinescu, St. 2005. Analiza geomorfologic a rmului cu falez intre Capul Midia i Vama Veche pe baza modelelor numerice altitudinale, Unpublished PhD thesis, Faculty of Geography, University of Bucharest (in Romanian). Hallermeier, R.J. 1978. Uses for a calculated limit depth to beach erosion, Proc. 16th Int. Conf. Coastal Eng., ASCE, Hamburg, pp. 1493-1512. Karlsson, T. 1969. Refraction of continuous ocean wave spectra, Proc. ASCE, 95 (WW4), pp. 437-448. Ozasa, H. and Brampton, A.H. 1979. Numerical modeling for the shore-line changes in front of the seawall, Rep. Port and Harbour Res. Inst., 18(4), pp. 77-103 (in Japanese).