The First PhD Student Symposium, 7-8 December 2011, Dunarea de Jos University of Galati.

Robert Toderascu and Eugen Rusu Dunarea de Jos University, Faculty of Mechanical Engineering, Galati, Romania

Abstract
On November 12, 2007 a fierce accident occurred in the Black Sea when Volga-Neft tanker, carrying 4,000 tons of crude oil, broke in two during a storm, when it was performing a stopover in the Russian port Kavkaz. The officials reported on this incident as a very serious environmental disaster. It spilled at least 1,300 tons of oil. The severe weather prevented emergency workers from collecting the oil, which authorities said was sinking to the seabed. Hours after Russian oil tanker incident, two other cargo ships carrying sulfur sank near the port Kavkaz. The increasing traffic on the Black Seas surface rises considerably the risks of accidents like the above mentioned one. In order to avoid or minimize potential accidents its impossible not to observe the necessity of implementing an operational numerical model in order to prevent and minimize the damage to the Black Sea. This work refers to the development of a software for processing data related with MOHID (Water Modeling Software) model, in order to asses currents speed and directions for the Black Sea basin.

1. Introduction
MOHID is a three-dimensional water modeling system developed by MARETEC (Marine and Environmental Technology Research Center) at the Technical University of Lisbon. This system allows for the adoption of an integrated modeling philosophy, not only of processes, but also of different scales (allowing the use of nested models) and systems (estuaries and watersheds), due to the adoption of an object oriented programming philosophy. The integration of MOHID different tools, (MOHID Water, MOHID Land and MOHID Soil) can be used to study the water cycle in an integrated approach. Since these tools are based on the same framework, their coupling can be easily achieved. Currently, MOHID is under implementation for the Black Sea basin. The purpose of this work is to focus on the development of an operational software able to process temperature and salinity data in the format needed for MOHID input.

MyOcean website

Etopo website MOHID GIS MATLAB MOHID Digital Terrain

HDF5 repack MOHID GUI

2. Formulation of the problem

Black Sea being and enclosed sea, tides are extremely small and dont interfere with current circulation. In order to force current movement for the Black Sea basin there is need for temperature and salinity data. MOHID can accept input only from HDF5(Hierarchical data file) data type, and the temperature and salinity data can be found in various datasets over the internet on specialized websites. In this case a request has been performed on the MyOcean website and access to BLACKSEA_ANALYSIS_FORECAST_PHYS_007_001 data has been granted. This kind of data contains temperature, salinity, depth and time archived in a NETCDF CF 1.4 format. The temperature and salinity are produced on a regular grid of 238x132x35 grid points (~5 km horizontal resolution), depth being on 35 levels that vary from 5 to 5 meters close to surface and 100 meters close to the bottom layer. Inside one file there are data for 6 time intervals, each corresponding to one day. These data needs to be transformed and coupled with bathymetry, latitude and longitude files for MOHID required input and written in an HDF5 file. For data manipulation and transformation, Matlab 2009b has been used.

Figure 1. Logical scheme of the proposed system

2. Solution
HDF5 is a data model, library, and file format for storing and managing data. It supports an unlimited variety of datatypes, and is designed for flexible and efficient I/O and for high volume and complex data. HDF5 is portable and is extensible, allowing applications to evolve in their use of HDF5. The required HDF5 file format for MOHID will have to contain 3 groups: Grid, Results and Time. The grid group will include the following data sets: Bathymetry, Connection X, Connection Y, Latitude, Longitude, Vertical Z and Water Points. Bathymetry dataset stores the depths of the sea. The file is downloaded from Etopo website with a resolution of ~ 1.8 km, then transform with Geodas program to display the depths with positive values and the heights with negative. Next the data is loaded into MOHID GIS (a geographical information system designed to support MOHID). The coastlines will have to be either drawn using Google earth or directly in MOHID GIS if possible. The horizontal grid will be set according to the desired resolution (in this case it has been chosen the same resolution as the temperature and salinity data 238x132, in order to avoid errors). The MOHID Digital Terrain creator will be next run in order to produce the Bathymetry data for the HDF5 file, as well as for the actual simulation with MOHID (figure 2). Latitude and Longitude data will be interpolated in Matlab to correspond to exact latitude and longitude of the temperature and salinity data. Connection X and Connection Y will be transformed from latitude and longitude data, since the refer to UTM coordinates (however they are not mandatory). Vertical Z dataset is formed from the assumed depth plus one more for the surface which will have to be zero. Water Points will tell MOHID if a given point is to be represented or no, so for every point inside the data that has a value will be equal to 1 and for every land points will be 0. Water Points will be a matrix of 238x132x35. Results group will contain the initial temperature or salinity values as well as the ones computed by MOHID after the simulation, for every time step of the simulation. And Time group will contain all the time steps for which the simulation will run, in our case 31 days, for the month of January 2010. In order to maintain a small data size we chose to make separate scripts for temperature and salinity. Aside from the different results groups there is no other difference among them. After the HDF5 files has been created it is needed to pass the file trough an HDF5 executable which will check for possible errors and perform a level 6 GZIP archiving, since MOHID will not read any other format. Once the GZIP archiving is finished, the HDF5 file is ready to be introduced in MOHIG GIU and used for analysis. A scheme of the described processes is presented above in figure 1, figure 3 showing the inspection of certain depth levels of the resulted temperature file: a) level 1, b) level 3, c) level 7, d) level 11, e) level 15, f) level 19, g) level 23, h) level 27 and i) level 35.

Figure 2. The Black Seas resulted bathymetry after the run of MOHID Digital Terrain creator.

3. Future development directions and problems met

This example has been made using the original grid level from temperature and salinity data (238x132x35) but the matrixes can be interpolated or extrapolated to fit any desired area. This method will be used for level 2, 3 and 4 generation towards the Romanian coasts, once the first simulation is finished. The work at the processing software is still in progress since is enough for one single point or quote to be misplaced and the MOHID simulation will crash. Another problem met was the different formats used for HDF5 files, since MOHID uses HDF5 library version 1.6.4 and Matlab 2009b uses 1.8 library versions. Downgrading the HDF5 library for Matlab is impossible, so the only valid solution was to recompile a MOHID executable file in order to include newer HDF5 libraries. For future improvements we are considering including wind and wave data, as well as minimizing the scripting code size by using iteration methods, since its currently around 100 pages for one month. At the end this software will be part of an integrated system which will include MOHID for water circulation and SWAN for wave generation in order to be able to predict and asses oils spill movement for the Black Sea basin.

a)

b)

c)

4. Acknowledgments
The PhD grant of Toderascu Robert is supported from EU funding by Project SOP HRD EFICIENT 61445

5. References
1. Booij, N. Ris, R. C and Holthuijsen, L. H., 1999. A third generation wave model for coastal regions. Part 1: Model description and validation, J. Geophys. Res. 104, C4, pp. 7649-7666. 2. INSTITUTO SUPERIOR TCNICO MARETEC, MOHID Hydrodynamic Module User Guide 3. A. Svendsen, Kevin Haas, and Qun Zhao, Center for Applied Coastal Research University of Delaware, Quasi-3D Nearshore Circulation Model SHORECIRC 4. E. Rusu, F. Onea, R. Toderascu, Galati University Dunarea de Jos, Romania, Dynamics of the Environmental Matrix in the Black Sea as Reflected by Recent Measurements and Simulations with Numerical Models, Nova Publishers, 2010 5. Rusu, E, 2010: Modeling of wave-current interactions at the Danube s mouths, Journal of Marine Science and Technology article in press. http://www.springerlink.com/content/pqq215x867882424/

d)

e)

f)

g) h) i) Figure 3. Resulted temperature file on different depth levels: a) level 1, b) level 3, c) level 7, d) level 11, e) level 15, f) level 19, g) level 23, h) level 27 and i) level 35.