Annotated Bibliography

Bickers, A.N. (2003), Cost Effective Marine Habitat Mapping from Small Vessels using GIS, SSS and Video. Coastal GIS, Wollongong, Australia.

This paper looks at cost effective techniques for seafloor visualization and mapping. Mapping continental shelf habitats, even in shallow water, is particularly challenging, and this paper outlines methods of using the newest technologies of the time in the most cost effective way for the large, unmapped regions. Combining sidescan sonar, video towing equipment and GPS were found to be efficient, and the data was collected and processed in ArcGIS. The paper covers the equipment used in detail, and the processing procedure to produce maps of the territories. Methods for the processing and analysis in ArcGIS are also presented here.

This paper is particularly useful for looking at the setup and implementation of a commonly used seafloor mapping technique. The paper goes into detail about both the equipment and procedures used for this project, which can serve as a useful introduction to the science of seafloor mapping techniques, and as a comparison to currently used methods. This paper is particularly useful for information regarding sidescan sonar equipment, both the science behind the sonar, and the analysis and processing of the information given by the sidescan sonar.

Populus, J. , More, F. A U. , Coquelet, D. and Xavier, J.-P.(1995) An assessment of environmental sensitivity to marine pollutions: solutions with remote sensing and Geographical Information Systems (GIS), International Journal of Remote Sensing, 16: 1, 3 — 15

This paper looks at remote sensing images and image analysis to analyse marine habitats, in this case the effect of pollutants from oil to the Loire estuary habitats. Here, the data is in part collected from satellite imagery, which has a very different set of challenges when compared to seafloor mapping methods for marine habitat analysis. Data collection for this paper was from a variety of sources, SPOT data, ESR-1 data, aerial photography and land and sea surveys. The classification of the images for analysis was also outlined. These multiple data sources were compbined into one product for the region, and the paper demonstrated how a number of data sets could be integrated with each other to give a comprehensive description of an environment.

The paper looks at combining a variety of remote system data sets into a single map or data packet. This proves useful, as it shows a variety of methods for marine habitat mapping that can be used, aside in situ measurements taken by sonar. The paper covers satellite imagery and the analysis used for this, which can be very useful in mapping the shallower water regions of a marine habitat, and can provide a much larger area of coverage in a much shorter time period. However, the compromise for the large spatial coverage and speed of data availability is lost in data resolution and use of this technique is very much dependent on the resolution that is required for marine habitat analysis.

Urbanski, J. A., Szymelfenig, M., (2003) GIS-based mapping of benthic habitats. Estuarine, Coastal and Shelf Science 56: 99-109

This paper looked at methods to improve analysis of benthic habitats using HELCOM classifications. Thematic maps were generated, using two fuzzy set maps, one which used the seafloor position relative to the euphotic depth, and the second describing the sediment type on the seafloor. This paper looks at the methods behind combining data sets to gain more information about marine habitats. Research cruises to collect data with Secchi disc units, and AVHRR satellite images were some of the data sets included in this database. The paper also goes onto describe the methodology behind the GIS used to solve the physical and biological oceanographic problems of the lack of boundaries between the complex marine systems. The final map produced brings the descriptions of habitats closer to the reality, however, more data sets can be added.

This paper also looks at the GIS potential to combine multiple data sets into one large scale habitat map. This paper looks particularly at variety throughout both the boundaries in the water column and sediment present on the seafloor. The combination of these data sets, and the methods used are a useful reference for combining multiple data sets to further analyze marine variability in habitats.

Sotheran, I. S., Foster-Smith, R. L., and Davies, J. (1997). Mapping of marine benthic habitats using image processing techniques within a Raster-based Geographic Information System. Estuarine, Coastal and Shelf Science, 44 (Suppl. A): 25S–31S.

This paper looks at mapping marine benthic communities in the littoral and sublittoral zones. This paper focusing on processing the data, looking at techniques used with a raster based GIS. The raster image for the sublittoral system acoustic data was generated by a RoxAnn system, which was converted into a raster image. For the littoral system, data was collected as aerial photographs from which were derived digital raster images. Images were validated with in situ collected data to enable a supervised classification of the data. The method described here look at using data from the sublittoral and littoral systems and processed to allow similar analyses. However, spatial and scale problems first had to be resolved.

This is a useful example of the problems of combining two different systems to gain a complete map. This paper looks specifically as the processing systems to allow two different data sets to be compared to allow general analysis to take place. While these maps are not high resolution, they do give a good idea of sublittoral and littoral variation, and show methods for combining vastly different data sets.

Cervenka, P., De Moustier, C. Sidescan Sonar Image Processing Techniques (1993), IEEE Journal of Oceanic Engineering, Vol. 18, No. 2, 108-122

This paper looks at the four step method required to process and analyse data collected from a sidescan sonar unit. The method of this paper looks at first the data collected by each ping of the acoustic backscatter, and the removal of outliers which can skew the data. Spectral analysis takes place to remove artifacts, and the contrast is then enhanced. Pixels generated are then mapped on a geographically referenced grid taking account the geometry and spacing between the pings to minimize smearing of the data. The results obtained through a single survey are then given to illustrate this methodology.

This is useful to greater understand one of the most common forms of data collection regarding marine habitats. A better understanding of the processing of sidescan sonar data allows for a better understanding of the data that is being used to create maps, and this paper extensively covers the processing of sidescan data into a ‘clean’ data set.

Fawcett, J., Myers, V., Hopkin, D., Crawford, A., Couillard, M., Zerr, B., (2010) Multiaspect Classification of Sidescan Sonar Images: Four Different Approaches to Fusing Single-Aspect Information, IEEE Journal of Oceanic Engineering, Vol. 35, No. 4, 863-876

This paper looks as an object on the seafloor from two separate sidescan sonar images. This allows comparison of an object from different viewpoints. The challenge lies in fusing these two views from individual looks. Four different approaches to this problem are described in this paper. The first method uses kernel regression using combined feature vectors. The other three approaches are based on the Dempster-Shafer fusion from a single look kernel based classifier. These methods are covered in the paper, are found to give similar results, however the use of a second sonar is found to significantly improve the classification performance.

Allowing understanding of the use of sidescan sonar, and ways to improve data is vital. The findings on the use of a second sonar – the combination of the produced data, and the improved classification it allows are important aspects of data collected this way. This paper allows development of the sidescan sonar system.

Capus, C.G., Banks, A.C., Coiras, E., Tena Ruiz, I., Smith, C.J., Petillot, Y.R. (2008) Data correction for visualisation and classification of sidescan SONAR imagery. IET Radar Sonar Navig., Vol. 2, No. 3, 155–169

This paper covers acoustic methods for marine data collection. The paper looks at the necessary corrections that need to be made to the data. Sidescan sonar is typically affected by factors of either range dependency, or of angular dependency. Varying altitude of the sensor over the seafloor causes variations in both of these factors. Corrections made to the data for these changes can be applied over the whole survey area, and they allow more stable means and variances across the sonar swath width. Results of using this method of data correction shows an improvement in the classification performance of a sample area. This indicates that using these corrections allow for improvements in accuracy, robustness, usability and execution time of data.

These corrections are vital to allow for better and clearer data collection using a side scan sonar. This paper can be used as a reference for better data processing, to allow for clearer visualization of marine habitats.

Bakran-Petricioli, T., Antoni, O., Bukovec, D., Petricioli, D., Janekovi´c, I., Kriˇzan, J., Kuˇsan, V., Dujmovi´c, S., Modelling spatial distribution of the Croatian marine benthic habitats (2006) Ecological Modelling 19, 196–105

The supralittoral and mediolittoral regions of the Croatia region were mapped using a variety of data sets. Spatial modeling, fieldwork, bathymetric models, and satellite imagery and in situ measurements of seawater variables were all combined in the hopes of giving a comprehensive picture of the Croatian marine environment. This model of the environment was built within the framework of a raster GIS. The developed model by this method explains a significant part of the infralittoral variability of the region. This method and data set can be used as a basis to expand the coverage to the whole Croatian marine environment. This project also shows that even with limited field data, there is the possibility to explain complex spatial phenomena using this data.

This paper can be used in reference to collating data sets – and this paper develops this idea into expanding a limited field data set into a much more expansive and complex system. Using the combined set of limited data, a complex system was able to be explained. This shows the potential to explain marine systems with the expansion of limited known variables.

Alexandridis, T. K., Topaloglou, C. A., Lazaridou, E., Zalidis, G. C., (2008) The performance of satellite images in mapping aquacultures. Ocean & Coastal Management Vol. 51, 638–644

This paper analyses the use of commercial satellite images such as the QuickBird bundle, SPOT-5 multispectral, Landsat 7 ETMþ, RADARSAT SAR, and ENVISAT ASAR. These images have various spatial and spectral characteristics, and have been assessed in their ability to effectively map mussel farms off a coast of Northern Greece. The ability to identify these farms was tested separately in each of the image’s study areas. The process looked at the validity of each of these image processes. The results showed that the sensor that best showed the variability of the farm was not necessarily the sensor with the best resolution.

This paper gives much more detailed information on the process and variety of satellite imagery options for marine habitat mapping. This paper looks at a number of different sources for imagery, and is useful for deciding the best satellite imagery to use when looking for marine habitat variation. Different sources of data are discussed, and different resolutions of data are also covered.

Cerdeira-Estrada, S., Lorenzo-Sánchez, S., Areces-Mallea, A., Martínez-Bayón, C.,  (2008), Mapping of the spatial distribution of benthic habitats in the Gulf of Batabanó using Landsat-7 images Ciencias Marinas 34 (2): 213–222

This paper looks at the distribution of a variety of different benthic habitats in the Gulf of Batabano. The five identified benthic habitats are medium- to high-density seagrass, low-density seagrass, sand with scarce vegetation, mud with scarce vegetation, and rock. These areas were identified from five images from the Landsat-7 enhanced Thematic Mapper Plus ETM+ satellite sensor. The area covered by thee sensor was a total underwater area of 21,305 km2 and is presented at the 1:250,000 scale. Results from the Landsat images were confirmed from in situ measurements from cruises taken in the region between 2003 and 2005.

This paper looks at classification from remote sensor images, and collates them with in situ measurements. This is a common practice for mapping a region and establishing the variability in the marine habitats. The system of classification used is comprehensive, and this allows the satellite data to be extensively analysed. The use of in situ measurements also provide validity to the data set, and are something that should be considered for clarification when working with any remote sensor data.

Brown, C. J., Blondel, P., (2009) Developments in the application of multibeam sonar backscatter for seafloor habitat mapping, Applied Acoustics 70, 1242–1247

This paper looks at multibeam echosounders, and their role in re-evaluating the way we survey the seafloor. This increasing human impact has increased the need to have a better understanding of our role upon the seafloor. Using multibeam echosounders, it is now possible to survey large areas of delicate marine habitats in a much more cost-effective process. These surveys now provide the baseline data from which thematic maps can be created of these large regions. With in situ ground truthing data, the data was verified. This paper covers the common data set and methodology used for this type of work and which studied used this new methodology.

The use of this paper can be seen as one of the first papers to use the new multibeam echosounder methodology to analyse and process data. This paper outlines the techniques agreed upon and used, and gives details of this refined method of sampling the seafloor and marine habitat locations. This paper can be used to compare the cost effective nature of this technique, and the ease of use in comparison to older techniques.

Mayer, L. A., Frontiers in seafloor mapping and visualization (2006) Marine Geophysical Researches 27: 7–17

This paper covers the history and the developments of the techniques in seafloor mapping, including -: sonars, ancillary sensors (motion sensors, positioning systems, and sound speed sensors), platforms upon which they are mounted, and the products that are produced. Development of new instrumentation and better development of corrections to variations in the data develop higher resolution from the equipment. Limiting factors on the development of sonars often comes from the sound speed structures of the water column that the sonar wave is travelling through. The products produced by these technological advances will also change. These allow massive data sets to be cleaned rapidly and efficiently, and the programming delopment allows for the interpretation an analysis of very large, complex and multicomponent data sets.

This paper outlines the future of the science, and the variety of ways that the science has evolved through the development of technology. New technological advances are covered, as are processing techniques which increase the efficiency of data cleaning and processing whilst dealing with an increase in data volume. This paper is useful for looking at the development of the science and the development of processing techniques.

Costa, B. M., Battista T.A., Pittman, S.J., Comparative evaluation of airborne LiDAR and ship-based multibeam SoNAR bathymetry and intensity for mapping coral reef ecosystems (2009) Remote Sensing of Environment, 113, 1082–1100

Bathymetric data was collected from two of the conventional methods of remote sensing data collection techniques – LiDAR (Airborne Light Detection and Ranging) and ship based multibeam Sound and Navigation and Ranging (SoNAR). Differences between these two data sets over the same mapped region were found, with LiDAR was found to be consistently shallower than SoNAR readings. The data sets were not significantly correlated, indicating that improvement needed to be made to the LiDAR technique used. However, a cost comparison indicates that the LiDAR system is the more cost effective, although it appears that the data is not as sound, with the SoNAR system providing a clearer picture of the seafloor.

This paper is key in understanding the difference between these two systems, and is especially useful in developing a cost comparison and data suitability function regarding future marine habitat studies. These techniques, LiDAR and SoNAR are both used extensively to map marine regions, and further analysis of their processing and comparisons of their accuracy is extremely useful for future decisions of their suitability.

Schaeffer, B. A., Morrison, J. M., Kamykowski, D., Feldman, G. C., Xie, L., Liu, Y., Sweet, W., McCulloch, A., Banks, S., Phytoplankton biomass distribution and identification of productive habitats within the Galapagos Marine Reserve by MODIS, a surface acquisition system, and in-situ measurements, (2008) Remote Sensing of Environment 112, 3044-3054

This paper looks at the identification of phytoplankton biomass distribution in the Galapagos islands, with the hope of better understanding the ecosystem of this region. Data collected was both Moderate Resolution Imaging Spectroradiometer (MODIS) and hyperspectral surface acquisition system deriving the chlorophyll, in-situ chlorophyll fluorescence, nitrate, salinity, and temperature of the water column. These helped define the productive regions during an El Nino event in the region. The models created allowed the assessment of the potential productivity and ability to support an ecosystem of this marine system. These techniques both combined provided a picture of the region, with the in situ measurements providing clarification of the larger scale remote sensor readings of the MODIS technology.

This paper is another example of the combination of remote sensing and in situ measurements to gain a better understanding of an ecosystem. In this paper, the MODIS system is covered, which provides a different set of data, and also requires a variety on in situ measurements for comparison in this case. This helps build a robust model of the area which can be used to assess the potential biological viability.

Keefer, M. L.. Peery, C. A., Wright, N., Daigle, W. R., Caudill, C. C., Clabough, T. S., Griffith, D. W., Zacharias, M. A., Evaluating the NOAA Coastal and Marine Ecological Classification Standard in estuarine systems: A Columbia River Estuary case study, (2008) Estuarine, Coastal and Shelf Science 78, 89-106

This paper outlines the importance of classification systems – which are key to many aspects of marine habitat study. It is important the classification scale should have broad applicability across the spatial and temporal scale of the system, and that the system is also found to be both flexible and adaptable. This paper outlines the CMECS classification standard, being developed by NOAA. The most difficult region seen to be classified are estuarine systems, and so this classification system was tested in several scales within an estuarine system, across a wide range of conditions. Strengths of the system were acknowledged, and improvements were also suggested.

This paper convers an important aspect of marine habitat mapping – and that is the classification of the different marine environments. It is important that a standardized system is used, so there is reduced ambiguity in the science community. However with so much variation not only locally but on a global scale, it can be difficult to develop a clear standardized system. This paper is important in that it defines a standard system for marine habitat mapping – and these classifications will be incorporated into any GIS work to further develop maps and continue analysis on these areas and habitats.