Flood hazards as a result of the Chańcza dam disaster floodplain zones designation

Flood hazards as a result of the Chańcza dam disaster floodplain zones designation K. Kępski Introduction Geoinformation systems associated with the hydrodynamic modeling software allow to obtain information needed for analysis to determine floodplain zones, determine the depth of the water, to identify endangered objects and simulation of flood wave and its effects. Knowledge of the areas that as a result of the floods could be flooded is the basis of management of flood protection. These are typically areas where specific management is conducted according to established principles of land use and construction. In April 2011 the team of specialists from the The Regional Water Management Board in Krakow in cooperation with OTKZ in Warsaw finished the two-year lasted project: Flood hazards as a result of dams disasters. The project of floodplain zones designation consists of the following steps: Acquisition of cross sections and hydrological data, Execution model calculations, Construction of a digital area model and digital model of water surface, The combination of water surface model with the model of the area and the extent and depth of flooding Visualization of the obtained floodplain zones. The current paper presents three final stages of work and the final effect, ie the area in the valley of Czarna Staszowska River (left tributary of the Wisła River), which could be flooded by the disaster of Chańcza dam, located 50 37'50, 08 "N 21 03'26, 78 "E, 36+000 km of Czarna Staszowska River. The Chańcza dam data Chańcza dam, typical soil ground, is heaped with fine sand with admixtures of medium and silty sands. In the middle of the dam a four-span concrete overflow block is situated of a length of 36 m. The block is dividing the dam into two parts: the left and the right one. The total length of the dam is 412 m and the maximum height is 15 m. Fig. 1. Cross sections through the overflow block of the dam. Model calculations Two basic models of the disaster formation have been adopted according to the previous assumptions the first one as a result of water overflow over the dam crest (breach failure), and the 1

second one as a result of breakdown of the dam body (piping failure). For the simulation calculations it has been decided to use the implemented in the program Mike11 algorithm of the breach formation according to the Engelund-Hansen formula of piping. This algorithm allows to calculate the time of the breach formation on the basis of the actual parameters of the dam ground without taking into account the real course of the breach formation including for instance its geometry. Regardless of the disaster scenario, the following geotechnical data based on the dam project have been assumed: GRAIN DIAMETER = 0.002 (2 mm) SPECIFIC GRAVITY = 2.6 (2.6 t/m3) POROSITY = 0.35 (35 %) CRITICAL SHEAR STRESS = 0.03 (wg Kryterium Shieldsa) SIDE EROSION INDEX = 1 The parameters of the dam geometry: UPSTREAM SLOPE = 3 DOWNSTREAM SLOPE = 3 TOP WIDTH = 10 (10 m) The restrictions on the breach final size: FINAL BOTTOM LEVEL = 207.00 m above sea-level FINAL BOTTOM WIDTH = od 59 do 111 m BREACH SLOPE = 0.5 (nachylenie ścian 2:1) For the purpose of the breach failure simulation it has been defined as follows: INITIAL LEVEL = 222.25 m above sea-level INITIAL WIDTH = 0.5 m The figure below shows the above mentioned parameters: Fig. 2. The view of the Mike11 window with the geotechnical parameters inserted of the Chańcza dam in the case of its breach failure. For the purpose of the piping failure simulation it has been defined as follows: STARTING LEVEL = 208.00 lub 213.00 m above sea-level INITIAL DIAMETER = 0.05 m ROUGHNESS = 0.001 m COLLAPSE RATIO = 0.5 VOLUME LOSS RATIO = 0.25. The figure below shows the above mentioned parameters: 2

Fig. 3. The view of the Mike11 window with the geotechnical parameters inserted of the Chańcza dam in the case of its piping failure. For the simulation calculations the hypothetical waves hydrographs of the likelihood p = 1%, 0.1% and 0.01% have been prepared. These hydrographs were the upper boundary conditions in the model of Chańcza dam disaster and have been established by Reitz- Kreps method based on historical waves collections. As mentioned above the two basic types of disasters of the Chańcza dam have been adopted: disaster as a result of the breach failure disaster as a result of the piping failure For both types of disaster the detailed scenarios have been developed. In the case of a disaster resulting from the breach failure the scenarios additionally have been differentiated according to: parameters of a hypothetical flood wave which flows into the reservoir at the time point of the disaster - Q 0.1%, Q 0.01% initial water level in the reservoir at the start of simulation calculations - NPP or MaxPP the time point at which the disaster occurs (piping failure) - the moment the water reaches the level of the dam top or the moment at which the inflow into the reservoir reaches its culminating level. the place of the breach formation the middle of the right side of the dam body (as the most probable variant) or both parts of the dam body at the same time - i.e. the middle of the right side of the body and the contact of the overflow block and the left side of the dam body (extreme variant). On the base of these assumptions 16 scenarios in the case of breach failure have been developed as shown below: Fig. 4. The file of scenarios for the case of breach failure of the Chańcza dam. 3

In the case of a disaster resulting from the piping failure the scenarios have been differentiated depending on the following factors: parameters of a hypothetical flood wave which flows into the reservoir at the time of the disaster - Q 0.1%, Q 0.01% initial water level in the reservoir at the start of simulation calculations - NPP or MaxPP the time point at which the disaster occurs- the level of NPP or MaxPP the height on a downstream slope, where breakdown occurs - two possible levels where assumed: H1 = 208.00 m - the level of the downstream slope base; H2 = 213.00 m - the level of the berm on the downstream slope. In this type of disaster, the additional variant of the scenario has been included, i.e. the piping failure in a situation of maximum possible reservoir filling, when the inflow into the reservoir is equal to the biological outflow - this option reflects the scenario in terms of so-called "solar day" disaster. On the base of these assumptions 14 scenarios in the case of piping failure have been developed as shown below: Fig. 5. The file of scenarios for the case of piping failure of the Chańcza dam. In the next stage selected flood scenarios have been implemented for two-dimensional model. The first step were the preliminary simulation calculations for the chosen scenarios for the breach failure and for the piping failure as well. Fig. 6. The view of the floodplain zone for Chańcza river as a result of the chosen scenario of breach failure. The calculations for the breach and piping failures were made for a hypothetical water flood wave Q0,01% and the initial filling of the reservoir equal to the maximum water level in the reservoir 4

(MaxPP). These calculations were made to examine the stability of the hydraulic model 2D, to correct the hydraulic parameters and to choose the optimal time interval. The stability was tested for the time interval of 5, 2 and 1 second, but finally the chosen time interval was 1 second. Additional simulations have confirmed that the model is stable in the case of quickly changing water flow, with which we deal in the case of a dam disaster. Fig. 7. Determination of the time interval to model a dam disaster. Fig. 8. Determination of the hydraulic parameters to model a dam disaster. 5

Fig. 9. The result of calculations 2D modeling for assumed time interval 5 s. Based on the results of the calculation of 14 scenarios for the piping failure and 16 scenarios for the breach failure, 3 representative (varying significantly from each other) scenarios were selected and for each of them two-dimensional modeling in two variants hove been performed: for the model based on digital terrain model (DTM) according to the photogrammetric method (Land Parcel Identification System - LPIS) and for the model based on digital terrain model developed by Light Detection and Ranging (LIDAR). Determination of floodplain zones. For this purpose the Digital Model of Water Surface (DMWS) based on water surface profiles in cross sections has been built. In the next step determining the range of floodplain zones was based on combined results of the simulation of water levels in DMWS and a digital terrain models. As a result a set of lines has been obtained as shown on the figure 10 and 11. The intersections of DMWS and DTM were connected and line designated the potential flood plain zones was obtained. Then the lines were introduced to the topographic map to illustrate clearly floodplain zones. NMT NMPW H 1 H 2 H i Fig. 10. Diagram showing the principle of setting floodplain zones. 6

Fig. 11. DMWS in the form of isolines. Determination of the river valley development Determination of the Czarna Staszowska river valley development, on the section between the Chańcza dam and Połaniec has been determined on the following data: topographic map on a scale of 1:10 000 orthophotomap scale 1:5 000 digital terrain model made by photogrammetric method (LPIS) digital terrain model made by Light Detection and Ranging (LIDAR) These materials are the raster maps, thus not allowing a direct analysis of spatial objects depicted on them. The mere presentation of the floodplain zones on this type of material does not allow for unambiguous definition of the number of objects exposed to floodwater inundation and to estimate the number of objects protected by the proper reservoir functioning. In order to carry one spatial analysis it has been decided to vectoriz selected classes of objects from raster maps. As the primary source of information topographic maps were selected for quick and unambiguous identification of classes of objects. Then, using the orthophotomaps produced on a smaller scale and with a greater degree of timeliness, but not allowing a clear identification of all objects, a more careful classification of objects was done. Vectorization was performed in the zone corresponding to approximate range of the various floodplain zones. They were five classes of objects determined Outbuildings Residential buildings Industrial buildings Public buildings Cemeteries Additionally, having a topographical map of the vector VMap Level2 of a level of details corresponding to a topographic map at scale of 1:50 000 following 3 classes of objects were determined: Roads Railways Embankments 7

From many attempts of flood wave hydraulic modeling in the program MIKE11after the initial selection for further analysis were selected as follows: the 40 scenarios of the wave after piping failure, 4 scenarios of the wave resulting from the breach failure. For these scenarios, floodplain zones were generated and the area of flooding was determined for each case. At the same time, a comparison of two methods of digital terrain model creating were made: photogrammetric LPIS and the Light Detection and Ranging LIDAR. The result are shown in the table below. SCENARIO Flooded area [km 2 ] L.p. Nazwa LIDAR LPIS różnica F % 1 PIP max PP max PP 39.56 38.56 1.00 97.5 2 PIP max PP max PP 39.56 38.56 1.00 97.5 3 PIP max PP max PP 39.41 38.43 0.98 97.5 4 PIP max PP max PP 39.42 38.44 0.98 97.5 5 PIP max PP max PP 37.38 36.36 1.02 97.3 6 PIP max PP max PP 37.39 36.37 1.02 97.3 7 PIP max PP max PP 37.38 36.36 1.02 97.3 8 PIP max PP max PP 37.36 36.35 1.01 97.3 9 PIP nor PP max PP 40.75 40.40 0.35 99.1 10 PIP nor PP max PP 41.12 40.78 0.34 99.2 11 PIP nor PP max PP 40.47 40.12 0.35 99.1 12 PIP nor PP max PP 40.77 40.42 0.35 99.1 13 PIP nor PP max PP 38.78 37.13 1.65 95.7 14 PIP nor PP max PP 39.07 37.34 1.73 95.6 15 PIP nor PP max PP 38.84 37.16 1.68 95.7 16 PIP nor PP max PP 39.13 37.41 1.72 95.6 17 PIP nor PP nor PP 36.20 35.40 0.80 97.8 18 PIP nor PP nor PP 36.41 35.56 0.85 97.7 19 PIP nor PP nor PP 37.42 36.41 1.01 97.3 20 PIP nor PP nor PP 37.66 36.60 1.06 97.2 21 PIP nor PP nor PP 33.40 33.68-0.28 100.8 22 PIP nor PP nor PP 33.64 33.07 0.57 98.3 23 PIP nor PP nor PP 34.93 34.65 0.28 99.2 24 PIP nor PP nor PP 35.15 34.85 0.30 99.1 25 PIP cul max PP 45.14 45.62-0.48 101.1 26 PIP cul max PP 48.40 48.90-0.50 101.0 27 PIP cul max PP 44.79 44.82-0.03 100.1 28 PIP cul max PP 46.37 46.80-0.43 100.9 29 PIP cul max PP 41.67 42.07-0.40 101.0 8

30 PIP cul max PP 43.58 43.33 0.25 99.4 31 PIP cul max PP 41.65 42.05-0.40 101.0 32 PIP cul max PP 43.56 43.30 0.26 99.4 33 PIP biol max PP 40.10 38.89 1.21 97.0 34 PIP biol max PP 39.10 38.14 0.96 97.5 35 PIP biol max PP 37.10 36.03 1.07 97.1 36 PIP biol max PP 37.10 35.95 1.15 96.9 37 PIP biol nor PP 2.64 2.56 0.08 97.0 38 PIP biol nor PP 3.01 2.56 0.45 85.0 39 PIP biol nor PP 3.27 2.88 0.39 88.1 40 PIP biol nor PP 2.97 2.56 0.41 86.2 7 OVT max PP 46.51 46.93-0.42 100.9 8 OVT max PP 47.71 47.64 0.07 99.9 15 OVT max PP 47.41 47.65-0.24 100.5 16 OVT max PP 49.15 47.64 1.51 96.9 For spatial analysis the most negative option has been selected, i.e. at which the flood wave could flood the largest area. That scenario was called number 26 piping failure at the initial max PP in the reservoir and assuming the synchronization the culminate flood wave into the reservoir with the moment of launch disasters which in turn cause the dam body destruction, and the scenario number 16 - the breach failure at initial max PP in the reservoir, the wave of the likelihood of 0.01%, both sides of the dam body destruction and modified control rules, which ultimately also leads to a destruction of the dam body. Floodplain zone resulting from the scenario No. 16 gave the largest flooded area. For that area the spatial analysis was performed to determine potential flood losses that may arise in the infrastructure. In truth, for the method LPIS the flooded area was higher in scenario No. 26, but it was smaller than the area obtained in the scenario No. 16 using the method of LIDAR, wherein the difference was only 0.25 km2 and occurred in a completely undeveloped area - the forest. Therefore for further analysis the zone generated by the scenario number 16 was used - the breach failure. In the next stage the analysis of area Czarna Staszowska River valley was performed taking into account the administrative boundaries and the type of buildings and land use. Results of the analysis are summarized in tables, of which examples are presented in this analysis. The area in the Czarna Staszowska River valley downstream the dam contains 7 municipalities belonging to the Staszowski County. They are: Szydłów, Staszów town, Staszów rural area, Rytwiany, Połaniec town, Połaniec rural area, and Osiek. The area, which could be flooded as a result of the Chańcza dam disaster, detailed into areas of individual municipalities is presented in the table below. POWIERZCHNIA [km 2 ] GMINA Skaning Całej gminy Zalana F % Szydłów 107.0 Staszów miasto 20.0 1.7 1.59 LIDAR 1.7 1.59 LPIS 4.0 20.00 LIDAR 3.9 19.50 LPIS Staszów obszar 206.0 5.7 2.77 LIDAR 9

wiejski 5.7 2.77 LPIS Rytwiany 125.5 20.0 15.94 LIDAR 21.2 16.89 LPIS Połaniec miasto 17.0 7.1 41.76 LIDAR 7.1 41.76 LPIS Połaniec obszar 9.3 16.12 LIDAR 57.7 wiejski 9.2 15.94 LPIS Osiek 119.8 0.1 0.04 LIDAR 0.1 0.04 LPIS From the above table the following conclusions were made: 1. Differences in the flooded area for photogrammetric methods and laser scanning are negligible, therefore no detailed analysis was performed in this field. 2. The largest area of flooding occurs in the municipality Połaniec, and covers almost 50% of the town. It should be noted that this is a town located about 40 km (along the river) below the dam. This localization allows to carry one at least a partial evacuation of threatened areas. 3. The town Staszów is also seriously threatened about 20% of the town area could be flooded. 4. The area in the municipality Osiek is safe and would not be flooded by the floodwaters. An area of about 0.1 km2 is undeveloped and covered with forest. The spatial analysis regarding cubature objects such as outbuildings, residential and public buildings (schools, kindergartens, hospitals, offices, etc.) and cemeteries is presented in the following table. GMINA Szydłów Staszów miasto Staszów obszar BUDYNKI, OBIEKTY [szt.] Gospodarcze Mieszkalne Przemysłowe Publiczne Cmentarze Skaning 30 22 1 0 1 LIDAR 30 20 1 0 1 LPIS 354 347 0 26 0 LIDAR 356 352 0 27 0 LPIS 101 102 4 3 0 LIDAR wiejski 102 105 4 3 0 LPIS Rytwiany Połaniec miasto Połaniec obszar 850 611 25 4 0 LIDAR 868 628 23 4 0 LPIS 255 317 9 8 3 LIDAR 253 316 9 8 3 LPIS 285 164 8 2 0 LIDAR wiejski 288 162 10 2 0 LPIS Osiek Powiat staszowski 0 0 0 0 0 LIDAR 0 0 0 0 0 LPIS 1875 1563 47 43 4 LIDAR 1897 1583 47 44 4 LPIS 10

From the above table the following conclusions were made: 1. Differences in the number of buildings in all classes of objects are low - 1 to 5 objects. In the extreme case the difference is 18 outbuildings. However, it occurs when the total number of flooded buildings is 850 - LIDAR and 868 - LPIS. In 25 cases out of 35 comparisons the difference was 0. 2. The highest number of residential, business and industrial buildings that can be flooded in the event of a disaster is in the municipality Rytwiany. The second to the number of flooded buildings is the city Staszów town. Here most public objects - 26 buildings are localized in the floodplain zone. 3. Three cemeteries of the four that are in potential floodplain zone are in the area of the city Połaniec, and one in the municipality Szydłów. Line objects that were analyzed in potential floodplain zones are presented in the following table: GMINA Szydłów Staszów miasto Staszów obszar OBIEKTY LINIOWE [km] Drogi Linie kolejowe Wały Skaning 4.500 0.000 0.592 LIDAR 4.500 0.000 0.592 LPIS 13.800 0.205 0.000 LIDAR 13.900 0.205 0.000 LPIS 7.850 0.000 0.867 LIDAR wiejski 7.750 0.000 0.835 LPIS Rytwiany Połaniec miasto Połaniec obszar 32.310 0.184 0.000 LIDAR 33.690 0.186 0.000 LPIS 18.860 3.300 3.830 LIDAR 18.830 3.200 1.450 LPIS 17.060 13.130 3.010 LIDAR wiejski 16.940 13.070 4.040 LPIS Osiek 0.000 0.000 0.000 LIDAR 0.000 0.000 0.000 LPIS From the above table the following conclusions were made: 1. Differences in the length of the flooded line objects using the method of photogrammetric and laser scanning are negligible, therefore no detailed analysis was performed in this field. 2. The total road length that can be flooded in the case of dam disaster is about 95 km. 3. The highest total road length that can be flooded in the case of dam disaster is in the Rytwiany tmunicipality - about 33 km. Second on the length of flooded roads is the Połaniec town - about 19 km. 4. The total length of railway lines that can be flooded in the case of dam disaster is about 17 km. At the same time it should be noted that the majority of this lines is located in the municipality of Połaniec - 13 km of railways. 5. The track runs along the border of floodplain zone. Because the railway tracks are usually located on a high embankment, they might not be flooded. However, it should be remembered that water is close to the embankment, which may be damaged. 6. The highest discrepancy between the methods of photogrammetric and laser scanning affected the length of embankments. This resulted from the fact that the embankments are located on the borders of the zones and have not exactly been included in the digital model of the terrain. For this analysis the 11

most unfavorable variants were adopted. 7. The total length of the embankments that can be flooded in the case of dam disaster is about 8 km (LIDAR) and 6 km (LPIS). 8. In the area of the Połaniec municipality about 4 km of the embankment may be flooded (LPIS), in the Połaniec town as well - embankment of the length about 4 km (LIDAR). In addition for the scenario number 21 and 33 piping failure at various boundary conditions, a spatial analysis was performed in the same way as for Scenario No. 16 The results are shown in the following tables. GMINA Dla scenariusza przelanie 33 BUDYNKI, OBIEKTY [szt.] Gospodarcze Mieszkalne Przemysłowe Publiczne Cmentarze Skaning Szydłów 30 22 1 0 1 LIDAR 30 20 1 0 1 LPIS Staszów 354 347 0 26 0 LIDAR miasto 356 352 0 27 0 LPIS Staszów 101 102 4 3 0 LIDAR obszar wiejski 102 105 4 3 0 LPIS Rytwiany 850 611 25 4 0 LIDAR 868 628 23 4 0 LPIS Połaniec 255 317 9 8 3 LIDAR miasto 253 316 9 8 3 LPIS Połaniec 285 164 8 2 0 LIDAR obszar wiejski 288 162 10 2 0 LPIS Osiek 0 0 0 0 0 LIDAR 0 0 0 0 0 LPIS Dla scenariusza przelanie 33 12

GMINA OBIEKTY LINIOWE [km] Skaning Drogi Linie kolejowe Wały Szydłów 4.500 0.000 0.592 LIDAR 4.500 0.000 0.592 LPIS Staszów 13.800 0.205 0.000 LIDAR miasto 13.900 0.205 0.000 LPIS Staszów obszar wiejski 7.850 7.750 0.000 0.000 0.867 0.835 LIDAR LPIS Rytwiany 32.310 0.184 0.000 LIDAR 33.690 0.186 0.000 LPIS Połaniec 18.860 3.300 3.830 LIDAR miasto 18.830 3.200 1.450 LPIS Połaniec obszar wiejski 17.060 16.940 13.130 13.070 3.010 4.040 LIDAR LPIS Osiek 0.000 0.000 0.000 LIDAR 0.000 0.000 0.000 LPIS GMINA Dla scenariusza przelanie 21 BUDYNKI, OBIEKTY [szt.] Skaning Gospodarcze Mieszkalne Przemysłowe Publiczne Cmentarze Szydłów 30 22 1 0 1 LIDAR 30 20 1 0 1 LPIS Staszów 354 347 0 26 0 LIDAR miasto 356 352 0 27 0 LPIS Staszów obszar wiejski 101 102 102 105 4 4 3 3 0 0 LIDAR LPIS Rytwiany 850 611 25 4 0 LIDAR 868 628 23 4 0 LPIS Połaniec 255 317 9 8 3 LIDAR miasto 253 316 9 8 3 LPIS Połaniec obszar wiejski 285 288 164 162 8 10 2 2 0 0 LIDAR LPIS Osiek 0 0 0 0 0 LIDAR 13

0 0 0 0 0 LPIS GMINA Dla scenariusza przelanie 21 OBIEKTY LINIOWE [km] Skaning Drogi Linie kolejowe Wały Szydłów 4.500 0.000 0.592 LIDAR 4.500 0.000 0.592 LPIS Staszów 13.800 0.205 0.000 LIDAR miasto 13.900 0.205 0.000 LPIS Staszów obszar wiejski 7.850 7.750 0.000 0.000 0.867 0.835 LIDAR LPIS Rytwiany 32.310 0.184 0.000 LIDAR 33.690 0.186 0.000 LPIS Połaniec 18.860 3.300 3.830 LIDAR miasto 18.830 3.200 1.450 LPIS Połaniec obszar wiejski 17.060 16.940 13.130 13.070 3.010 4.040 LIDAR LPIS Osiek 0.000 0.000 0.000 LIDAR 0.000 0.000 0.000 LPIS The results of the analysis are also shown in a graphical way, on the background of the raster topographic map on a scale of 1:10 000 and orthophotomap on a scale 1:5 000, on which were inserted the layers of objects and the area of water depth as a shade of blue color. The examples of the results are shown on map fragments below. 14

Fig. 12 Fragment of the floodplain aarea, highlighting the various layers of objects. Mapping The maps have been developed showing the spatial extent of floodplain zone for one selected scenario of the dam disaster on the basis of the Digital Terrain Model of flooded areas and layers containing the characteristics of the river valley development. As explained above, this scenario shows maximum flooded area downstream the dam. These maps were made on the background of raster ortophotomap, on which was inserted the area of water depth as a shade of blue color, and cross sections of the valley wit the entered values of water ordinate. To develop and visualize floodplain zones ArcGIS software has been used. An example map is shown below. 15

Fig. 13 Map of designated flood zone - the final project product. Summary and conclusions The basic information used in the analysis of the flood risks is the spatial extent of potential flood wave, water depth in areas potentially flooded and the flow velocity of flood waters. This information, combined with data on the current development of potentially threatened areas allow to estimate the losses, the designation of evacuation routes, the analyze of the best available methods of flood protection and the selection of preventing methods. All these information on flood risk collected in one system with data on flood wave simulation create a decision support system for authorities responsible for civil defense. The first step in creating such a system is to determine the potential floodplain zones. The combination of the possibilities offered by the hydraulic models with the tools available in GIS software is an excellent tool to determie and analyze the flood risk. Konrad Kępski MSc Eng. THE REGIONAL WATER MANAGEMENT BOARD IN KRAKOW Coordination and Information Centre for Flood Protection e-mail: kkepski@krakow.rzgw.gov.pl 16