Simulation of Oil Spill Trajectory Using Gnome Model; Skikda port, Algeria

Transcription

1 Middle-East Journal of Scientific Research 24 (5): , 2016 ISSN IDOSI Publications, 2016 DOI: /idosi.mejsr Simulation of Oil Spill Trajectory Using Gnome Model; Skikda port, Algeria Sawsen Saad and Rachida Hamzi Institute of Heal and Industrial Safety, university of Batna2, Algeria Abstract: The Algerian coast has reatened by many severe oil spills at occurred because of e activities of e transportation of e crude oil and oer hydrocarbon products. In our study, an oil spill trajectory model is used to measure how much e oil spill reats e environment, how far it will travel, to predict what area will be affected and see how spilled oil is predicted to change chemically and physically during e time [1, 2]. GNOME (General NOAA Operational Modeling Environment) is a model used to forecast e trajectory of e oil slick. The end result is a map of e prediction of movement of oil. In fact, GNOME creates and displays an oil spill "movie" showing e provided trajectory of e oil slick. In is work, we describe e meod to simulate e trajectory of one of e oil spill events happened in skikda port in Algeria. Key words: Oil spill Oil slick Modeling GNOME Marine accident PANTELIS accident Trajectory INTRODUCTION In e present paper, we try to follow e oil slick trajectory in e first 24 hour In e last decades, e Algerian coast has been affected by many severe oil spills: M/T KAPTAIN Meodology: In e oil spill model ere are ree main MARKOS, JUAN ALAVALLEJA in 1980; M/T modules: 1-input module provides information about SOUTHERN CROSS in 1986; M/T MAASSLOUIS in initial conditions of e model: initial data about 1989, ERATO in 1991, COUGAR in 2003 and most environment (wind and current field, temperature, ice, etc.) recently, PANTELIS in 2008 [3]. and initial data about oil spill (type, volume, properties, A big number of incidents have occurred since location and time of occurrence, etc.), 2- transport 1980 and after 2008 but we haven t mentioned (trajectory) module and 3- fate (weaering) module [8]. em because of many reasons: e lack of data; e difficulty in obtaining e report containing GNOME Model: The General NOAA Oil Modeling e information or because e quantities of e oil Environment (GNOME) is a two-dimensional model [9] [1] spilled were less important; So e quantities spilled at designed for e rapid modeling of pollutant in e accidents quoted above were more an 250 tones trajectories in e marine environment; it s used as a [4-7]. forecasting tool for bo actual spill events and a In July, 29, 2008, e oil tanker PANTELIS caused planning tool to examine what-if scenarios [10]. an oil spill during e loading of e crude oil in Skikda GNOME uses an Eulerian/Lagrangian spill-trajectory port, e accident was because of a cargo transfer failure model in which e regional physics are simulated as and e quantity released was 280 tons. According to e Eulerian (continuous) fields wiin which e slick s reporting, e Cleanup done e same day by e Lagrangian Elements move [1]. Wiin GNOME e oil is intervention of private company, environmental and divided into a large number of small particles which move tourism services, civil protection, coast guards and under e influence of ocean currents, wind drift and municipal services and e beach was accessible e day turbulent motions. The fate and transport of oil spilled after [3]. into e marine environment also depends on a suite of Corresponding Auor: Sawsen SAAD, Institute of Heal and Industrial Safety, university of Batna2 (05000), Algeria. 1823

2 physical, chemical and biological processes weaering ; GNOME model includes simple algorims for oil weaering [11]. GNOME has ree model modes: Standard mode, GIS Output mode and Diagnostic mode [2]. In Our Study We Have Used e Diagnostic Mode: The diagnostic mode provides additional capabilities like full access to all model parameters and scaling options, tailor e model to represent real-time data, develop location files wi advanced training and hydrodynamic modeling experience [12]. Data and Location Research: Data required for modeling was obtained from multiple sources (table.1) and loaded into e GNOME software. The data included e ree main components necessary for running e model: maps, movers and input. Table 1: Model inputs and sources Inputs Source Location file [2][13] Current data [14] [15] Wind data [16][17] Diffusion GNOME default value [18] Spill info [3] Time of release User selection Map Data: GNOME uses only e map file format known as BNA, is format is available rough GNOME Online Oceanographic Data Server (GOODS) Fig. 1. Fig. 1: Location data available rough GOODS 1824

3 Current Data: GNOME accepts various formats for Oer Inputs: For all oer information related to e current data, a selection of recent measured and product spilled, ey were provided by The Regional forecast currents compatible wi GNOME are Marine Pollution Emergency Response Centre for e available for download via GOODS. For e historical Mediterranean Sea [3]. currents data, it was obtained from Ocean Surface Current Analyses-Real time (OSCAR) as NetCDF file PANTELIS Scenario: In our work, a scenario for 280 format. metric tons of crude oil spilled is run on e model. According to CEDRE (Centre of Documentation, Research Wind Data: GNOME supports bo ASCII and NetCDF, and Experimentation on Accidental Water Pollution) [19], in our case; we have created e data file using e e accident occurred in e evening so e release start variable wind dialog boxes. For e historical data, it was time selected by us is 7:00pm. The release point is: latitude obtained from WEATHER UNDERGROUND. 36, 53.4 N, longitude 6, 54 E. The current data file Diffusion: Random diffusion is done by a simple provided by OSCAR website was filtered for a mon (from 15, july to 15, august) and loaded into e model random walk wi a square unit probability. The random inputs, The Wind was selected as variable so half hourly walk is based on e diffusion value set in e model wi and hourly values (from 7pm, 29, july to 7pm, 30, july) presents e horizontal eddy diffusivity in e water, a low are gaered from e weaer history database which has 2-1 value would be 1000 cm s and a high value would be provided by e airport RABEH BITAT, ANNABA. 2-1, between and cm s e model default is 2 (Figure2), and a value of cm s 1 is used for e cm s [18]. diffusion'' Fig. 2: Wind speed and direction data available rough UW Simulation: GNOME runs for 24 hours to find e The minimum regret splot locations indicate a roughly direction of e oil slick trajectory (Fig. 3 and 4). The 90% probability of e maximum extent of e spill [21]. spilled oil is represented by dots on e map which are called splots; Black dots indicate GNOME s best guess RESULTS AND DISCUSION trajectory estimate of e oil spilled from e tanker while red dots indicate GNOME s larger minimum regret GNOME footprint (Fig. 4) shows e spill movement trajectory estimate [2], we use e minimum regret and direction depending to e movers. solution to show where else e spill could go if e currents, winds, or oer model inputs were a little bit Depending to e Wind Direction: Depending to e wind different [20]. direction, we can see at: 1825

4 Fig. 3: (280) metric ton of crude oil, 24hrs after release A screen capture is taken every time e oil slick changes its direction. Fig. 4: Oil slick movement during e time depending to e movers 1826

5 In e first hour (8:00pm), e wind was 5.1 m/s from 2.97 km along latitude and 3.4 km along longitude, noted Nor, e major portion of e spill has moved in e at an additional portion of 0.1% was off map, e sou direction so 91.6% of e quantity of e oil spilled evaporated portion was 15.6% and e beached portion has beached (Tabel4) and almost 1.4 km along e coast was 62.2% was touched. In e next 7 hour e dominant wind direction was 8 hours after release (3:00am), e wind changed e from Nor to Nor-noreast wi a speed which varies direction to Souwest wi a velocity of 1.5m/s; a part of between 3.6 to 6.7 m/s, at e end of ese 7 hours, e e spill (11.7%) has propagated in e noreast most important portion of oil was reached e 4.48 km direction. We should note at a quantity of 28 metric along e coast for e best guess. For e minimum tons (10%) was Evaporated and dispersed. regret, more an 4.8 km along e coast was touched At e end of e 8 hours (4:00am), e wind direction (from N E to N E). for e changed to Nor-norwest, floating quantity was 40 evaporated quantity was 57 metric tons, e floating metric tons (14.4%), beached quantity was 209 (74.7%) quantity was 10 metric tons and 213 metric tons quantity metric tons and evaporated quantity was 30 metric tons. beached. During e next 3 hours, e spill stretched over a By e end of e next 3hours (24 hours after release), region of 2.8 km along latitude and 3.1 km along e major portion of spill has beached wi a quantity longitude, for e minimum regret e spill reached 4.8 km of 209 metric tons, Almost 62 metric tons got evaporated along e coast from N E to 36 54,71N and dispersed. Remaining 9 metric tons of spill is floating 6 53, 17 E, The portion of evaporated oil was 13.8% and and a portion of 0.2 was off map. e floating portion was 23.3% From e end of e 12 to e 13 hours (after release) e wind speed was 0 m/s, so Depending to e Surface Current: The current vectors ere is no movement, except movement due to e aren t seen on our map, at s because of e weakness of diffusion. e currents in is area (near e port), to prove at we At 9.00am (14 hours after release), e wind speed downloaded anoer current data file from HYCOM was 1.5 m/s and its direction was from Nor to Nor- (Hybrid Coordinate Ocean Model) [22] for recent days noreast, for e minimum regret e spill extended to (Fig. 5). Fig. 5: Current in e map of our study's area 1827

6 Depending to e Diffusion: To know e role which e and almost 2.7 km along latitude for e minimum regret. diffusion has played in our simulation, we run Pantelis And just a 0.6km along e coast was touched for e scenario and we make e random diffusion inactive minimum regret. (Fig. 6). By e 24 hours, 0.4 km along e coast was affected It can be seen from Fig. 6 at e spill by oil for e best guess and 2.66 km along e coast for took a different trajectory, by e 12 hour after e minimum regret. release e spill has moved away from e coast at a So e diffusion plays an important role along wi distance of 0.45 km along latitude for e best guess e wind Fig. 6: The oil spill trajectory wiout using e diffusion mover Oil Weaering: The oil fate during 24 hrs is presented in Table 2 and Fig. 7 Table 2: GNOME output quantities per hour Evaporated After *hour & Dispersed Beached Floating Off map * * * * * * * * * * * * * * * * * * * * * * * * Fig. 7: Oil crude fate during 24hrs, 280 metric tons As we have said before, GNOME includes just a simple algorim for e weaering processes; in fact GNOME uses a simplistic 3-phase evaporation algorim, which is appropriate for simple drills and educative comparisons. ADIOS (Automated Data Inquiry for Oil Spills) is a software simulation developed by NOAA and used to predict e oil weaering [23, 24]. So we have seen to use ADIOS for comparing our GNOME results wi ADIOS estimate. 1828

7 ADIOS model asks for information on e spill itself, e product spilled and its proprieties; and e environmental condition such as wind speed, wave heights, water temperature and water salinity. Wi e same scenario using in GNOME we run e ADIOS but it s required to add more information like e water proprieties and selecting e oil type from e ADIOS list (Fig. 8). Fig. 8: ADIOS inputs (spill scenario) Fig. 9: ADIOS outputs 1829

8 Fig. 10: Oil budget table Figures 9 and 10 show e ADIOS results 2. GNOME user s manual, January 2002 According to our GNOME and ADIOS results, ere 3. REMPEC is a big difference in e evaporation values, so ADIOS AccidentsDatabase have e capability to predict e oil weaering better 4. CLEAN UP THE MED Oil pollution in e an GNOME. Mediterranean Sea (in French) 5. Mohamad ALBAKJAJI 2010., Pollution of e CONCLUSION Mediterranean Sea by oil related to maritime traffic, doctoral esis, PARIS-EST University (in French) Based on e Results Obtained in Our Study: The 6. REMPAC, List of alerts and accidents in e trajectory of e oil slick depends on e wind and e Mediterranean, Regional information system / part c diffusion coefficient. / section 2, The role of e surface current stays not important in 7. REMPAC, Statistical analysis Alerts and accidents e area of study. database, Regional Information System RIS C2, The simulation using GNOME is just predictions and 2011 not reality, many oer factor can effect. 8. Alexander Kileso, Boris Chubarenko, Petras Zemlys ADIOS model can estimate e expected and Igor Kuzmenko, Oil spill modelling characteristics and behavior of spilled oil basing on e meods: application to e sou eastern part of e physical properties of e oil spilled and e sea water, Baltic Sea, BALTICA Volume 27 Special Issue: so it can help us to correct e oil weaering results 9. Poojia, D. Yapa, Hao Xie, C.J. Beegle-Krause and given by GNOME Caitlin O Connor, Improvements to e Work REFERENCES on Integration of NOAA s GNOME Model wi CDOG Model, a final report submitted to e coastal response research center. 1. Beegle Krause, C.J., General NOAA Oil 10. William, B. Samuels, David E. Amstutz, Rakesh modeling Environment (GNOME): a new spill Bahadur & Christopher Ziemniak, Development of a trajectory model, International oil spill conference Global Oil Spill Modeling System, Ear Science Research; Vol. 2, No. 2;

9 11. Amy MACFADYEN, Eugene WEI, Christopher 17. ERSAN BAÞAR M/V strontsiy gemi kazasi WARREN, Charlie HENRY and Glen sonrasindaki dökülen yakit daðiliminin benzetim WATABAYASHI (2014) Utilization of e Norern modeli, Dokuz Eylül University. (In turkish) Gulf Operational Forecast System to Predict 18. GNOME technical documentation, October 2012 Trajectories of Surface Oil from a Persistent Source 19. CEDRE, Technical Letter sea coast n 23, (in Offshore of e Mississippi River Delta, International French) Oil Spill Conference 20. Gnome_Boston_UGex, Remyalekshmi, R. and Arkal Vittal Hegde, restoration.noaa.gov/oil-and-chemical-spills/oil- Numerical Modeling of Oil Spill Movement along spills/response-tools/gnome-location-files-atlantic- Nor-West Coast of India Using GNOME, region.html International Journal of Ocean and Climate Systems, 21. Farzingohar, M., Zelina Z. Ibrahim and M. Yasemi, volume 4, number1, Oil spill modeling of diesel and gasoline wi 13. NOAA, GNOME around Rajaee Port of Bandar Abbas, Iran 14. NASA Langley Research Center (2013), Utilizing 22. HYCOM, S.H. Tan, N. GNOME and ArcGIS for Pumice Raft Trajectories: A Lerdphornsuttirat, A. Tisavipaksakul e role of User s Guide. pre-impact site assessment in emergency oil spill response planning, Oil Spill Response and East Asia 16. WEATHER UNDERGROUND, Response Limited (OSRL/EARL) 2 Jalan Samulun, Singapore NOAA,