MAPRES Marine Pollution Monitoring and Mitigation by Remote Sensing FINAL TECHNICAL IMPLEMENTATION REPORT

Transcription

1 MAPRES Marine Pollution Monitoring and Mitigation by Remote Sensing MAPRES is co-financed by the European Commission under the Community framework for cooperation in the field of accidental or deliberate marine pollution. FINAL TECHNICAL IMPLEMENTATION REPORT

2 TABLE OF CONTENTS Chapter 1: MAPRES presentation... 4 Chapter 2: General Summary of project implementation process... 5 General overview of the process... 5 Comparative analysis of initial and actual time schedule... 5 Comparative analysis of planned and used resources... 5 Comparative analysis of expected and actual results... 6 Chapter 3: Evaluation of project management and implementation process... 7 Positive aspects and opportunities... 7 Partnership and core group cooperation... 8 Cooperation with the Commission... 8 Comments on European value added... 8 Lessons learnt and possible improvements... 9 Chapter 4: Activities Chapter 4: Activities Comparison between initially planned and actually implemented activities, including monitoring, evaluation and dissemination Qualitative evaluation of the activities Chapter 5: Presentation of the technical results and deliverables TASK A: Sensors, Platforms and Related Methodologies to detect oil spill by Remote Sensing: field of active sensors Action A.1: Operational aspects and methods used in the maritime member States to monitor and prevent oil spill Action A.2: Technical report on the bibliographic research on the suitable platforms Action A.3: Technical report on the bibliographic research on the existing methodologies Action A.4: Definition of the best operational chain for image processing and the methodologies to oil spill detect with limitate errors TASK B: Sensors, Platforms and relative Methodologies for detect the oil spill by Remote Sensing: field of Passive sensors (UV, NIR, TIR). 16 Action B.1: A review of the state-of-art of the detection of oil spill in terms of sensors, wavelengths and image processing methods. 16 Action B.2: Compiled tables of the most optimal and suitable wavelengths, sensors and processing methods for the detection of oil spills Action B.3: The identification of key wavelengths of reflected and emitted radiation which are correlated with oil concentration and oil composition. The identification of the generality of thermal and optical techniques for their separation from the signatures of the surrounding water

3 Action B.4: An image library of suitable and key remotely sensed datasets from a variety of sensors concentration on European coastal waters and covering oil spills in a range of different situations Action B.5: A range of image processing techniques which show efficacy to estimate oil extent and thickness in a range of conditions and situations presented in the images in the image archive Action B.6: Tested, draft robust and general processing chain for the general detection of oil slicks Action B.7: A full recommended of oil spill identification protocol for incorporation into the operating manual TASK C: Hydrodynamic modelling to forecast the fate of oil spills Action C.1: Coastal model implementation around the Maltese Islands Action C.2: Testing of β version of oil spill model Action C.3: Full implementation of oil spill model TASK D: Impact of coastal environment and best practices for mitigation and damage recovery Action D.1: Technical report on the bibliographic research on the impacts of oil pollution on coastal ecosystems Action D.2: Best Practices release on the intervention in case of accident TASK E: Exercise Action E.1: Exercise on image processing of an oil spill event Action E.2: Exercise on hydrodynamic modelling of oil fate Action E.3: Exercise on mitigation measures and best practices TASK F: Project Management and Reporting to the EC Action F.1: Action Plan Action F.3: Periodic Progress Reports Action F.4: Final Report TASK G: Dissemination Action G.1: Realization and on line publishing of the web site Action G.2: Editing and printing of MAPRES flyer Action G.3: Report on the restricted workshop in Edinburgh Action G.4: Report on the dissemination activities in Malta Action G.5: Guidelines on oil pollution monitoring and detection procedures used at European level Action G.6: Report on the open and final conference in Italy Chapter 7: Follow up Comparison between initial and current follow up measures Additional follow up approaches

4 Chapter 1: MAPRES presentation The aim of MAPRES concerns the cooperation in the field of accidental or deliberate marine pollution, in particular the main activity relates to Marine Pollution Monitoring and detection by aerial surveillance and satellite images. Large spills of oil and related petroleum products in the marine environment can have serious biological and economic impacts. Around 20% of oil transported by sea traverses the Mediterranean Sea, amounting to an annual flux of 350 million tons of crude oil and refined products. Most of this maritime traffic travels across the Malta Channel, the coastal zones near Sicily Island and Genoa City. North Sea is also a region threatened by this type of pollution. Public and media scrutiny is usually intense following a spill, with demands that the location and extent of the oil spill be identified. Remote sensing is playing an increasingly important role in oil spill response efforts. Through the use of modern remote sensing instrumentation, oil can be monitored on the open ocean around the world. With a knowledge of slick locations and movement, response personnel can more effectively plan countermeasures in an effort to lessen the effects of the pollution. An operating manual on methods, techniques, sensors and procedures for remote sensing and Oil Spill detection, on Hydrodynamic Numerical Simulations for spill propagation forecasting and mitigations procedures, will be the deliverable of the project. An exercise on these procedures will be devoted to the crew members of Coast Guard and aerial surveillance aircrafts and personnel working in the field of Satellite Imageries valuation. 4

5 Chapter 2: General Summary of project implementation process General overview of the process The project activities began after the administrative procedures were over. The kick off meeting in Rome and the two-day meeting in Brussels saw the participation of Mapres representatives. All the suggestions from the Commission were taken in consideration to make the interaction and the implementation of the process easier. The implementation of the activities followed 7 main sections: a) Sensors, platforms and related methodologies to detect oil spill by remote sensing: field of active sensors; b) Sensors, platforms and related methodologies to detect oil spill by remote sensing: field of passive sensors; c) Hydrodynamic modelling to forecast the fate of oil spills; d) Impact on coastal environment and best practices for mitigation and damage recovery; e) Exercise; f) Management and reporting to the European Commission; g) Dissemination. All these themes are strictly connected one to another; the implementation of the project was realized in connection with the deadline of the different deliverables. For instance the data coming from remote sensing observation of oil slick will be used to feed, together with meteoclimatic information, hydrodynamic models applied to forecast threatened areas. Moreover the activities related to the Dissemination were focused to spread the knowledge of the project results in the scientific arena. The activities were planned on a year basis, from January to December In order to monitor the development of the project, the mid term meeting was organised on July, in Edinburgh. The implementation process was checked also thanks to the periodic reports (which will be discussed in the relevant paragraph below) Comparative analysis of initial and actual time schedule The schedule foreseen during the elaboration of the proposal were generally respected. The work on the tasks was continuous and fruitful like the elaboration of the reports foreseen for each task. The sending to the Commission of these documents happened with some delay sometimes, due to force majeure. As far as the exercise in Sicily is concerned, it was postponed because of bad weather conditions in the open sea. In the first months of 2008 the activities were concluded. The work was mainly focused on collecting and preparing the documents to be sent to the Commission. Comparative analysis of planned and used resources The funds allocated to the various activities were spent without any particular delay or modification. No substantial changes were made in the budget. 5

6 Comparative analysis of expected and actual results The expected results were reached without any significant problem. All the expected deliverables were produced and the procedure chain for the monitoring and impact reduction of oil spills has been completely implemented. The actual results generally match the expected ones. 6

7 Chapter 3: Evaluation of project management and implementation process Positive aspects and opportunities The analysis of project management activities is made by taking into consideration the five processes that compose it. Initiating: the beginning of the construction of Mapres partnership and scheme was made easier by the fact that all the partners involved shared a common view on the core issue of the project. The positive aspect of these phase was that European excellence centres, i.e. University of Palermo, University of Genoa, University of Malta and University of Edinburgh, gather together to pursuit a common goal, that is analysing marine pollution monitoring and mitigation procedure. Planning: the planning phase mainly focused on the selection of activities to carry out. Each work package developed a specific topic which is correlated with the others but which is at the same time self standing. The planning of the activities was organised along 12 months during which different output were foreseen. Executing: the real implementation of the project began in January 2007 after two kick off meetings in Rome and Brussels. The execution of the activities followed the foreseen timetable, even if some short delay was due to the complex work of the experts. Controlling: the control of the progresses of the project took into consideration a variety of aspects. First of all the punctuality of the deliverables was welcomed as well as their accuracy. The work of the personnel was monitored by means of personal monthly timesheets; the financial aspects were noted in the Excel file provided by the Commission. By filling it day by day at the end of the year the release of the financial statement was made easier. Closing: the closing procedure followed 2 streams of intervention: the technical and the financial. The closing procedure of the technical intervention focused on the writing of the last reports, the operating manual and this final technical implementation report. On the other hand. the closing procedure for the financial aspects included the completion of the final financial statement. 7

8 Internal and external difficulties encountered As a general consideration, it can be said that no major difficulties were encountered during the evolution of the project. The difficulties encountered at the internal level could be linked to the location of the partners; sometimes it was complicated to keep all the participants linked. However the will to co-operate was strong and by means of some compromises all the difficulties were solved. At the external level the difficulties were linked to some suppliers conditions and to the fact that organising events, like the final conference, could be made difficult because of casualties. Partnership and core group cooperation The partnership of Mapres project is well structured and responded to each need of the activities. The highly qualified scientific personnel worked is synergy in order to accomplish a final work which is remarkable. The link among European Universities was very strong and crossed the borders, reaching a community of international scientific interests. The will to co-operate was clear and declared from the beginning of the initiating phase. The project management activities thus were made easier and during the year in which the project developed there were many cases of mutual support and understanding. Cooperation with the Commission The role of the Commission officers was crucial. Mapres representatives participated together with the representatives of the other projects financed to the kick off meeting in Brussels. During the event the Commission officers gave useful hints and suggestion to the participants in order to help them in putting into practice the correct procedure requested by the system. The participants were given a handy book with all the information about deliverables, costs, administrative and financial rules which helped everyone. The Commission referent officers have been always present during the life of Mapres project. Their advice was sought in case of doubts and need. Their understanding was well appreciated when some minor delay in presenting the deliverables occurred. Comments on European value added Marine pollution is a transnational issue that must be addressed by an international approach. Like air, water, and in particular marine water, is a global resource. Considering that the Mediterranean sea is sailed every year by boats transporting 20% of transported oil, the activities of MAPRES, focusing on how to face accidental or deliberate marine pollution, seem to be particularly relevant. The partners developed the tasks of the project by highlighting the need of a European cooperation: the activities in the Malta channel for instance saw the cooperation of the University of Malta and the University of Palermo. The experts developed a software that predicts how the spills move on the water surface. This shows how technological improvements can be made by opening the borders to scientific cooperation and by making the European funds fruitful. 8

9 Lessons learnt and possible improvements The MAPRES project provided a complete depiction of the state of the art on the oil spill monitoring and environmental recovering systems. The lessons learnt is the need of an inter-disciplinary approach to the problem in order to setup an efficient system. The possible improvements are related to a near real time availability of remotely sensed images in order to provide a quick reaction to the pollution event. 9

10 Chapter 4: Activities Comparison between initially planned and actually implemented activities, including monitoring, evaluation and dissemination The activities of the project were implemented without delay and problems. From the kick off meeting in Rome (January 2007) to the drafting of the final technical implementation report, the partners worked in synergy to promote the project and to prepare documents of important scientific value. The monitoring, evaluation and dissemination activities lasted for all the 12 months of the project: they were functional to the project because they allowed the partners to analyse their work. Moreover the dissemination activities made it possible to widespread the awareness of the Mapres. Qualitative evaluation of the activities In our opinion, the quality of the activities carried out during the MAPRES project is good. The inter-disciplinary approach improved in a significant way the quality and the efficiency of the programmed actions. 10

11 Chapter 5: Presentation of the technical results and deliverables TASK A: Sensors, Platforms and Related Methodologies to detect oil spill by Remote Sensing: field of active sensors Action A.1: Operational aspects and methods used in the maritime member States to monitor and prevent oil spill Description This document shows what are the main aspects and operational methods used in Member States to prevent oil spill accident. This document identifies what are the main international laws and conventions concluded among Countries to control and to protect the marine environment and shows what are the main prevention methods adopted at national and international level. This paper shows as Member States over these years have faced several serious oil accidents at sea (Erika, near the Breton coasts (1999) and the Prestige tanker (2001),) by making conventions and laws more and more effective to prevent sea water oil spill. This document also identifies the main methods used by some Countries in the event of oil spill in order to obtain a complete cognitive framework on extra-european experiences, on European ones, and on the current national situation. Purpose This paper wants to clarify what are the law references that identify the liability for sea water oil spill and what are the responsibilities of individual actors who are involved in the event of national or international waters accidents, Indeed, without a proper legal and technical framework of oil spill phenomenon there is a risk of an overlap of competences that may be a cause of inefficiency. With MAPRES project we also intendes to provide simple rules for the involvement of the authorities in order to improve efficiency in areas of greatest risk. Evaluation The document has reached the aim required for the project, as were identified existing methodologies and operational aspects that can be used in the context of the project MAPRES. Were examined several different methods used to monitoring areas with high risk and to face the oil spill phenomenon. Through the bibliographic research Was demonstrated that the path regulatory push towards more prevention activities in the transport of hydrocarbons introducing important innovations such as double-hulled ships (safer against spills but with a bigger explosion risk). Value-added This document explains in a practical way regulatory references governing the transportation of oil over the sea and show methods and operational aspects adopted by the European and not European States in case of oil spill. The European Community, thanks to it, has the foundation for maximizing the efficiency of actions and to know how to interact with the actors involved in a sea accident. 11

12 Dissemination This document has been disseminated on the occasion of final MAPRES project conference occurred on 6 th of December 2007 at the MUVITA congress centre of Arenzano (Genoa Italy). Dissemination activities included the publication of the report on the project web-site 12

13 Action A.2: Technical report on the bibliographic research on the suitable platforms Description This document describes the mainly satellite platforms used for the oil spill identification. Through an accurate literature research phase Synthetic Aperture Radar (SAR) has been individuated as the best active sensor. At the present days, only eight platform mounted sensors based on this kind of technologies and only few platforms are launched in the recent times. Moreover, the new COSMO-SkyMed constellation will be operational only in the next times. After a brief introduction the document describes these eight sensors referring the equipments, the functionality and the service offered. Purpose The purpose of deliverable is the literature research about the suitable platforms for oil spill detection. Evaluation The document has reached the prefixed aims. Many suitable platforms for oil spill detection have been individuated and moreover has been noted that the improvement of technologies will give in the near future available systems characterized by high temporal and spatial resolution. Value-added This document represents for EC an accurate literature research on suitable platforms for oil spill monitoring using active Radar sensors images. Dissemination This document has been disseminated on the occasion of final MAPRES project conference occurred on 6 th of December 2007 at the MUVITA congress centre of Arenzano (Genoa Italy). Dissemination activities included the publication of the report on the project web-site 13

14 Action A.3: Technical report on the bibliographic research on the existing methodologies Description After a brief introduction, this document shows the exist methodologies for the oil spill detection. This deliverable is organized in sections: primarily, the capability of SAR and Laserfluorsensor to detect an oil spill is explored and finally an operational chain of processes for automatic oil spill detection is described. Inside the SAR section the look-alike problem is discussed with particular mark on the optimal wind condition and optimal wind speed range are individuated. Also, the problem of detection in time using low-temporal resolution images are explained considering the oil spill evolution caused by weathering processes. The presence of look-alike is the main limitation for operational oil spill detection using SAR system and for this reason the last section is dedicated on operational method for oil spill detection and automatic look-alike discrimination. Purpose The aim of the deliverable is the literature research for the best operation method for oil spill monitoring and detection using remotely sensed data. Evaluation The document has reached the prefixed aims. The best technologies for oil spill monitoring and detection are pointed out. Also, the main problems of operational oil spill monitoring have been individuated. In addition, the literature research has pointed out that the present days algorithms need some improvements referring the look-alike discrimination phases. Value-added This document shows the operational methods for oil spill detection and represents an important document for EC. The methodologies, the advantages and limitations are pointed out. Dissemination This document has been disseminated on the occasion of final MAPRES project conference occurred on 6 th of December 2007 at the MUVITA congress centre of Arenzano (Genoa Italy). Dissemination activities included the publication of the report on the project web-site 14

15 Action A.4: Definition of the best operational chain for image processing and the methodologies to oil spill detect with limitate errors Description In this document three different methodologies useful for the aim of MAPRES project are discussed. These methodologies are improved for minimize the errors both on the oil spill detection and also on the look-alike discrimination. The first algorithm, Evaluation of features for automatic detection of oil spills in ERS SAR images, has been developed by some researchers of Norwegian Computing Center. The second one, SAR polarimetry to observe oil spills, has been developed by a research group of University Phartenope of Naples. The last one, Neural networks for oil spill detection, is the most complex. All above mentioned models belong on the semi-automatic processes methods that are considered the most efficient. Purpose The purpose of the document is the individuation of the best methodology for operational oil spill detection. Evaluation The document has reached the prefixed aims. Three operational algorithms for oil spill monitoring have been explained. These methodologies resulted the more efficient and it is really difficult to choice the best one method. Moreover the literature research showed that further development is the setup of automatic processes for real-time oil spill monitoring. Value-added The present document has pointed out the most efficient operational methodologies for oil spill monitoring and look-alike discrimination. Dissemination This document has been disseminated on the occasion of final MAPRES project conference occurred on 6 th of December 2007 at the MUVITA congress centre of Arenzano (Genoa Italy). Dissemination activities included the publication of the report on the project web-site 15

16 TASK B: Sensors, Platforms and relative Methodologies for detect the oil spill by Remote Sensing: field of Passive sensors (UV, NIR, TIR) Action B.1: A review of the state-of-art of the detection of oil spill in terms of sensors, wavelengths and image processing methods Description Satellite and aerial remote sensing can, in principle, provide a convenient means to detect and precisely map marine oil spills, and provide timely information for guiding recovery operations. Passive remote sensing techniques were evaluated. UV remote detection generally operates in the 250 to 380 nm range. A number of satellite sensors exist which are capable of passive measurement in the UV region but which are not suitable for oil slick detection. The optical wavelength range covers the range 400 to 2500 nm and encompasses the visible, near and middle infrared wavelength regions. There are numerous spaceborne sensors for in operational use for detection within this wavelength region, although none have been specifically designed for oil slick detection. These sensors range from the very high resolution to coarse global monitoring systems. Visual and camera based systems also exist. Limitations were also considered which had to do with timing and frequency of overpasses, restriction to daylight hours and the need for cloud free skies. Remote sensing of oil slicks in thermal infrared wavelengths utilises differences in the thermal characteristics between oil and water. In contrast to optical remotes sensing, thermal remote sensing benefits from imaging capabilities during both daytime and nighttime. Satellite thermal sensor studies for oil spill detection were reviewed. The review concluded that whilst with regular passes over the Earth s land and oceans, the theoretical ability for such satellite based sensors to obtain frequent information for oil slick identification and location is an attractive one. Numerous studies reviewed have shown evidence that passive remote sensing methods can be successfully applied, but that coverage and techniques for automated detection are far from satisfactory and far from operational use. The review has shown that limitations in the use of passive detectability of oil slicks is common due to the spectral and thermal similarities between slicks and surrounding waters and confusion with other slick-like phenomena. The successful determination of an oil slick on an image appears largely to depend on oil type, differences in between the slick and background water colour, differences in illumination (sun angle and sunny vs. cloudy conditions), lack of field measurements or other in-situ data that would allow site-specific algorithm calibration. There currently exists no reliable method to accurately measure oil spill thickness using passive remote sensing techniques. Coordinated networks may be the only reliable way of using passive remote sensing to detect oil spills. Purpose Action B1 reviewed the current state-of-the-art of passive remote sensing of marine oil slicks. This Action built on previous reviews and updated them on the basis of the more recently published literature incorporating developments in technology and image processing. Evaluation 16

17 The document has reached and exceeded its original aims. The physics and chemistry of oil spills and their evolution have also been reviewed in detail, with respect to UV, optical and thermal wavelength domains. Potentially useful approaches are highlighted (e.g. coordinated sensor networks), limitations have been discussed and potentially useful image processing techniques for testing in later workpackages have been highlighted. Value-added The document has reviewed the state of the art in passive remote sensing techniques and represents the most up-to-date and complete review presently available. Dissemination The document was disseminated in draft form at the July 2007 MAPRES progress meeting held in Edinburgh. The final version is available in pdf format on the MAPRES website ( 17

18 Action B.2: Compiled tables of the most optimal and suitable wavelengths, sensors and processing methods for the detection of oil spills Description Lists of the common sensors covering the passive UV, optical and thermal wavelength ranges available were compiled, along with their principal characteristics (e.g. wavelength coverage, spatial resolution, temporal frequency etc). These were presented in tabular format. There are few passive satellite sensors which are capable of detection in the UV wavelength range. Those in existence are mostly for ozone monitoring. There is wide availability of image data in the optical wavelength range although of the sensors available, none were expressly designed for oil slick detection. There are a number of suitable spaceborne thermal imaging sensors, several with multispectral capabilities. Purpose To compile and present tables of the most relevant earth observation sensors for the detection and monitoring of oil spills, for sensors covering the different wavelength regions (UV, Optical, Thermal). Evaluation The tables are extensive and list not only the satellite sensors themselves, but their principal characteristics, including: sensor name, satellite, launch date, country/organisation, swath width, spatial resolution, number of wavebands, spectral coverage, temporal resolution, radiometric resolution and pricing. Value-added These tables represent the state of the art for multispectral information commonly available from spaceborne data at the present time. Dissemination Tables of satellites and suitable sensors and their characteristics were included as part of the report compiled under Action B-1. The final version of this report is available in pdf format on the MAPRES website ( 18

19 Action B.3: The identification of key wavelengths of reflected and emitted radiation which are correlated with oil concentration and oil composition. The identification of the generality of thermal and optical techniques for their separation from the signatures of the surrounding water Description Relevant literature was reviewed and reported of the relevant wavelengths useful for passive detection of oil spills. Measurements of the reflectance properties of oils in the optical domain were performed to confirm results previously obtained. UV detection of oil spills relies on sunlight reflected from the sea surface. Oil slicks display a high reflectivity of ultraviolet (UV) solar radiation from the sea surface even at very thin layers (<0.01 µm). There is little evidence to suggest that quantitative estimates of the film thickness can be estimated using UV detection of spills. False positives and their separation from real oil spills remain a problem. Oil pollutants modify optical light fields both above and below the water surface, manifest by the attenuation of the light passing through an oiled water surface, by changes in light absorption in the seawater column due to the formation of an emulsion, and by the scattering of light by particles of such an emulsion. Oil has a higher surface reflectance than water in the visible region, but also shows limited nonspecific absorption tendencies. Thus oil can be generally seen in the visible spectrum and can be separated from the water itself and from oil dispersants. Sheens show up as silvery and reflect light over a wide spectral region down to the blue. However, the literature and our measurements confirm that oil has no specific spectral characteristics that distinguish it from the background. In the thermal domain, theory suggests that an oil slick should show a lower temperature than the surrounding clear water. Oil also shows differences in heat capacity, thermal conductivity, and thermal inertia, compared with sea water, which should aid its discrimination. Salisbury et al. (1993) comprehensively measured thermal infrared spectra of a range of oil slicks, all of which were similar and little affected by thickness, or other factors. Thus, oil slicks provide a flat and remarkably unvarying spectral signature in the thermal infrared. The apparent lack of unique spectral structures for different oils in the 8 to 14 µm region would appear to limit the effectiveness of monitoring oil slicks in the thermal infrared. Purpose To identify key wavelengths of reflected and emitted radiation which are correlated with oil concentration and oil composition from both the literature and from laboratory measurements of the optical reflectances of different oils. To draw conclusions on the generality of thermal and optical techniques for their separation of oil from water on the basis of spectral characteristics. Evaluation This action has concluded that there is little scope for the discrimination of oil spills in the optical and thermal domains on the basis of spectral signature data. Although 19

20 somewhat disappointing, these results begin to indicate which image processing techniques will be useful for subsequent Task B actions (e.g. thresholding). Value-added Measurements have been performed which have confirmed that oil slicks are relatively featureless phenomena in the optical and thermal domains. This has important implications, notably in determining appropriate approaches using image processing techniques. But equally the results suggest that the image processing techniques used can rely on simple thresholding and contrast enhancement methods and that a wide range of panchromatic, as well as multispectral sensors, can be used. Dissemination Elements of this workpackage (UV, optical and thermal properties of oils and oil emulsions on water) were included as part of the report compiled under Action B.1. A second summary together with measurement results is provided in the second report on specific tasks to Task B (Sensors, Platforms and relative Methodologies for detect the oil spill by Remote Sensing: field of Passive sensors (UV, NIR, TIR)). The final versions of these reports are available in pdf format on the MAPRES website ( 20

21 Action B.4: An image library of suitable and key remotely sensed datasets from a variety of sensors concentration on European coastal waters and covering oil spills in a range of different situations Description A library of images was developed where images were sourced mostly from publicly available image archives. The library was developed not to be a definitive set of images covering all known oil spill events in Europe for which images are available but to show images typical from a range of different satellite and airborne sensors and to test image processing algorithms over a range of resolutions and environmental situations. The library consists of ASTER and Landsat satellite images, along with airborne datasets. The archive search also highlighted the lack of a significant coordinated effort of sensor acquisitions known to coincide with oil spill events there is apparently no single organisation routinely acquiring images as soon after known oil spill events are known. Many false positive images were found. Purpose To compile a library of suitable and key remotely sensed datasets from a variety of sensors, with concentration on European coastal waters and covering oil spills in a range of different situations. Evaluation The archive of images was not intended to be exhaustive, but selective and focussed on European waters to show images typical from a range of different satellite and airborne sensors and to test image processing algorithms over a range of resolutions and environmental situations. No high resolution images (e.g. Quickbird and IKONOS) covering oil spills were found for European waters in available image archives. No images were found from sensors capable of sensing in UV. The library thus consists of optical and thermal image datasets. The archive search also highlighted the lack of a significant coordinated effort of sensor acquisitions known to coincide with oil spill events there is apparently no single organisation routinely acquiring images as soon after known oil spill events are known. Many false positive images were found. Nevertheless, the archive is a comprehensive and useful one. Value-added A library of relevant and recent images of oil slick pollution and related events has been compiled for a number of satellite and airborne sensors and covering the optical and thermal domains. This archive is suitable for further experimentation by EU Member Sates, for the further development and testing of image processing techniques and processing chains. 21

22 Dissemination The development of the archive was reported at the MAPRES project conference on the 6 th of December 2007 at the MUVITA congress centre of Arenzano (Genoa Italy). Tables of archived images sourced for the image library for European coastal waters and reporting sensor, acquisition date, image name, imaged location, resolution, known oil spill event, a quicklook image and associated comments are provided in the second report on specific tasks to Task B (Sensors, Platforms and relative Methodologies for detect the oil spill by Remote Sensing: field of Passive sensors (UV, NIR, TIR)). The final version of this report is available in pdf format on the MAPRES website ( 22

23 Action B.5: A range of image processing techniques which show efficacy to estimate oil extent and thickness in a range of conditions and situations presented in the images in the image archive Description A range of image processing techniques were tested for oil slick and slick-like feature detection using the images compiled under Action B4. These techniques consisted of many of the methods highlighted under Action B1. Many of these were tested for their efficacy to identify oil slicks and their ability to discriminate oil from look-alikes (false positives). The ability of these techniques to estimate extent, (perimeter and area) under a range of conditions was also investigated. The key steps in the image processing process can be identified as: image pre-processing, identification of the feature, discrimination from look-alikes, and estimation of area, perimeter and intensity. Taking these into account, we developed a rationale for image processing, which emphasised speed, few processing steps and simplicity. Thus, we conclude that simple contrast enhancement and thresholding methods should be enough to identify the slicks if they are going to be detected. The key pre-processing steps were highlighted and ranked in importance. Examples of the application of image preprocessing and enhancement methods were reported for both optical and thermal datasets. Best detection on optical imagery was found in the presence of sunglint on the seasurface, an approach similar to the detection of slicks in Synthetic Aperture Radar data. Removal of look-alikes is still a problem and methods here are in significant need of further development. Purpose To test a range of image processing techniques for their effectiveness to estimate oil extent and thickness in a range of conditions and situations presented in the images in the image archive. Evaluation A rationale for image processing of images for oil slick detection was developed, which emphasised speed, few processing steps and simplicity. Removal of look-alikes is still a problem and methods here are in significant need of further development. Development of methods for detection on optical data under sunglint conditions warrants further investigation. Value-added A wide range of image processing steps has been tested and evaluated on real image data showing real oil slicks. For the first time, a rationale for the processing of passive remotely sensed data for oil slick detection is presented. Image pre-processing steps are also ranked in relative importance. The utility of sunglint conditions in passive optical remotely sensed data represents a new discovery for detection of oil slicks using these techniques. 23

24 Dissemination Results of our tests on image processing techniques on images from the archive dataset, were reported at the MAPRES project conference on the 6 th of December 2007 at the MUVITA congress centre of Arenzano (Genoa Italy). A report on the image processing techniques tested, with examples of their application is provided in the second report on specific tasks to Task B (Sensors, Platforms and relative Methodologies for detect the oil spill by Remote Sensing: field of Passive sensors (UV, NIR, TIR)). The final version of this report is available in pdf format on the MAPRES website ( 24

25 Action B.6: Tested, draft robust and general processing chain for the general detection of oil slicks Description A range of recommended processing chains was presented for imagery obtained in each wavelength range explored in this action on passive remote sensing. The principal image processing technique to detect oil on water in the UV is most often visual on-screen detection followed by a simple thresholding technique where the high reflectivity oil spill may be highlighted to determine its approximate areal coverage. The lack of standard algorithms for data processing and interpretation remains a problem in the processing of optical data for oil spills monitoring. A number of methods were tested. Evidence suggests that extensive effort on data processing improves the chances of oil detection but this can be very time consuming and runs counter to the philosophical approach adopted in this project that the routines used should be simple and rapid (Task B-5). Despite theory for the detection of oil slicks in the thermal infrared region being relatively well established our experience suggests a lot of difficulty in detecting known spill events in this wavelength domain. Purpose Based on the rationale developed under Task B-5, processing chains for the identification of oil slicks in UV, optical and thermal imagery were presented. Evaluation A range of recommended processing chains for imagery obtained in each wavelength range explored in this action on passive remote sensing was reported. Ultimately the goal of an oil slick detection approach using passive techniques should be the development of an automated feature recognition system. Such a system could automatically recognize oil spill features from a range of different remotely sensed data types, but was beyond the scope of this current project. The need to eliminate false oil spills still remains and would need inclusion. Automated feature recognition systems would also need to take into account information about locality; knowledge about the region, environment, and water current events to interpret the results seen after processing the satellite imagery in context. Value-added Efficient operational processing chains have been recommended taking into account the results of previous Task B actions. Our recommendations are made on the basis of speed, simplicity and a minimum number of image processing steps. Dissemination 25

26 Recommended image processing chains techniques are included in the second report on specific tasks to Task B (Sensors, Platforms and relative Methodologies for detect the oil spill by Remote Sensing: field of Passive sensors (UV, NIR, TIR)). The processing chains were also included in the Operational Manual (Action G-5). The final versions of the Task B second report and Operational Manual is available in pdf format on the MAPRES website ( The recommended processing chains were also presented and discussed at the MAPRES project conference on the 6 th of December 2007 at the MUVITA congress centre of Arenzano (Genoa Italy). 26

27 Action B.7: A full recommended of oil spill identification protocol for incorporation into the operating manual Description This task effectively concludes the work and results found of Tasks B-1 to B-6. Despite the number of processing steps that must be completed to process remotely sensed data, it is recommended that the choice of methods applied to the imagery is both rapid and simple. This recommendation was based upon 1) the need for early detection and monitoring to aid management and mitigation of the problem, and 2) the fact that oils lack specific spectral features in either reflectance and emittance which means that image processing tools that rely on spectral identification mostly won t be very useful. Contrast enhancement and thresholding methods should be enough to identify the slicks if they are going to be detected. The removal of lookalikes is still a problem and methods to separate real oil spills from other phenomena are in significant need of further development. Key limitations of passive remote sensing were also reported (these include: low temporal resolution, poor spatial resolution of some systems, cloud contamination, the need for good atmospheric conditions, daytime acquisition for UV and optical data, costs of acquisition, lack of standard image processing algorithms and false target identification). Purpose Recommended image processing chains and important factors in the use of passive remote sensing techniques for oil slick detection were reported in the Operational Manual (Action G-5). Evaluation A balanced view of the utility of passive remote sensing techniques for oil slick detection is presented. This considers both advantages and limitations of the different methods. Recommended processing chains are reported. To overcome some of the limitations, coordinated networks of sensors, would seem the only the solution to meeting the need for high temporality to monitor the rapidly changing conditions associated with oil spills and to overcome the problems of cloud cover. This will ultimately involve a multi-wavelength approach and a combination of passive and active techniques. Sensor combinations, as employed from aircraft, show probably the greatest promise and utility for the operational detection of oil spills. Value-added Effective tools for the rapid and simple detection of oil slicks using passive remote sensing techniques are recommended. The intention of these recommendations is to improve the capabilities of EU Member States to detect slicks using such techniques. Dissemination 27

28 Recommended image processing chains and important factors in the use of passive remote sensing techniques for oil slick detection were reported in the Operational Manual (Action G-5). A more extended summary was also presented in the second report on specific tasks to Task B (Sensors, Platforms and relative Methodologies for detect the oil spill by Remote Sensing: field of Passive sensors (UV, NIR, TIR)). The final versions of the Task B second report and Operational Manual is available in pdf format on the MAPRES website ( The recommended processing chains and general issues associated with the use of passive remote sensing techniques were also presented and discussed at the MAPRES project conference on the 6 th of December 2007 at the MUVITA congress centre of Arenzano (Genoa Italy). 28

29 TASK C: Hydrodynamic modelling to forecast the fate of oil spills Action C.1: Coastal model implementation around the Maltese Islands Description This paper describes the results of the bibliographic study on the physical, chemical, and biological processes, that occur in the transport and fate of spilled oil in seas, and the equations that govern the hydrodynamic processes of coastal areas, Particularly, was performed the implementation of a finite element coastal model in the area around the Maltese Islands, nested to the shelf scale forecasting model already running at IOI-MOC, University of Malta. In the model the time-marching of the solution is achieved using a fractional-step method in order to overcome the incompressible flows problem of pressure-velocity decoupling. Also, the model across nested models (shelf scale and coastal scale) was implemented with successively smaller grid embedded sub-domains in order to be able to predict the track of a spill from an open sea source down to a coastal area with precision. Purpose The aim of this deliverable was the implementation of the model developed at the DIIAA of the University of Palermo, that resolves the free-surface Reynolds and continuity equations in 3D which are discretised using the Finite Volume method. Evaluation The paper has successfully respected the project aim, recogninazing physical, chemical, and biological processes, that occur in the transport and fate of spilled oil in seas, and the equations that govern the hydrodynamic processes simulated with the coastal model implemented in the Mapres project. The system has shown a good capacity in transfer of numerical outputs between submodels and it has demonstrated the ability of sub-models to reside and run on different computers Value-added This paper describes the main physical, chemical, and biological processes, that occur in the transport and fate of spilled oil in seas, in order to give EC members the tools for the analisys of oil spill hydrodynamic. Dissemination This document has been disseminated on the occasion of final MAPRES project conference occurred on 6 th of December 2007 at the MUVITA congress centre of Arenzano (Genoa Italy). Dissemination activities included the publication of the report on the project web-site 29

30 Action C.2: Testing of β version of oil spill model Description This paper mainly describes the test-runs to gain an accurate insight on the functionality of the oil spill model. To have realistic predictions within the Malta Shelf region, forecasted data arranged in the required format was imported. This was generated from the ROSARIO II marine forecast as well as hourly wind fields from the SKIRON forecasts, which were averaged and interpolated. Was,also, effected an oil spill simulation offshore the coast of Pozzallo in Sicily, using a by-product of rice. The main aim was to compare the actual path followed by an artificial slick (spilled) to the forecasted trajectory by the oil spill model. Since this does not evaporate or emulsify, methods to bypass the effect of such mathematical models had to be found. Given that the spill source was chosen to be very close to the coast, a new high resolution Pozzallo Coast region was defined and implemented within MEDSLIK. Purpose The aim of this deliverable was testing of the model, through an artificial oil spill simulation on the field, offshore the coast of Pozzallo in Sicily, using a by-product of rice. Evaluation Considering the constraints that the slick was not oil, that atmospheric conditions were severe, and that exact geo-referencing of observations was dubious, the model performed well. The simulation served to verify the skill of the oil spill model in predicting the slick movement and hence its ability to use in actual emergency situations. Value-added Notwithstanding the quick decay of the artificial slick, the MEDSLIK model proved to be a very effective tool for EC members to predict the slick movement. Dissemination This document has been disseminated on the occasion of final MAPRES project conference occurred on 6 th of December 2007 at the MUVITA congress centre of Arenzano (Genoa Italy). Dissemination activities included the publication of the report on the project web-site 30