How Do Continuous High-Resolution Models of Patchy Seabed Habitats Enhance Classification Schemes?

Predefined classification schemes and fixed geographic scales are often used to simplify and cost-effectively map the spatial complexity of nature. These simplifications can however limit the usefulness of the mapping effort for users who need information across a different range of thematic and spatial resolutions. We demonstrate how substrate and biological information from point samples and photos, combined with continuous multibeam data, can be modeled to predictively map percentage cover conforming with multiple existing classification schemes (i.e., HELCOM HUB; Natura 2000), while also providing high-resolution (5 m) maps of individual substrate and biological components across a 1344 km2 offshore bank in the Baltic Sea. Data for substrate and epibenthic organisms were obtained from high-resolution photo mosaics, sediment grab samples, legacy data and expert annotations. Environmental variables included pixel and object based metrics at multiple scales (0.5 m–2 km), which improved the accuracy of models. We found that using Boosted Regression Trees (BRTs) to predict continuous models of substrate and biological components provided additional detail for each component without losing accuracy in the classified maps, compared with a thematic model. Results demonstrate the sensitivity of habitat maps to the effects of spatial and thematic resolution and the importance of high-resolution maps to management applications.

[1]  Luciano E. Fonseca,et al.  Geocoder: An Efficient Backscatter Map Constructor , 2005 .

[2]  B. Costa,et al.  Shallow-water benthic habitats of Northeast Puerto Rico and Culebra Island , 2015 .

[3]  Trevor Hastie,et al.  Novel methods for the design and evaluation of marine protected areas in offshore waters , 2008 .

[4]  Brett R. Scheffers,et al.  Biodiversity redistribution under climate change: Impacts on ecosystems and human well-being , 2017, Science.

[5]  D. Ruppert.  The Elements of Statistical Learning: Data Mining, Inference, and Prediction , 2004 .

[6]  Pål Buhl-Mortensen,et al.  Using EUNIS habitat classification for benthic mapping in European seas: present concerns and future needs. , 2012, Marine pollution bulletin.

[7]  M. Zettler,et al.  The potential exceptional role of a small Baltic boulder reef as a solitary habitat in a sea of mud , 2019, Aquatic Conservation: Marine and Freshwater Ecosystems.

[8]  M. Kendall,et al.  The Influence of Thematic and Spatial Resolution on Maps of a Coral Reef Ecosystem , 2008 .

[9]  M. Zettler,et al.  Biotope map of the German Baltic Sea. , 2015, Marine pollution bulletin.

[10]  Gregory S. Okin,et al.  A Toolkit for Ecosystem Ecologists in the Time of Big Science , 2017, Ecosystems.

[11]  Glenn De ' ath.  BOOSTED TREES FOR ECOLOGICAL MODELING AND PREDICTION , 2007 .

[12]  C. Ricotta,et al.  Accounting for uncertainty when mapping species distributions: The need for maps of ignorance , 2011 .

[13]  Thomas Brey,et al.  On the Art of Classification in Spatial Ecology: Fuzziness as an Alternative for Mapping Uncertainty , 2018, Front. Ecol. Evol..

[14]  Max Kuhn,et al.  Applied Predictive Modeling , 2013 .

[15]  Bodo Rosenhahn,et al.  Seabed sediment classification of side-scan sonar data using convolutional neural networks , 2017, 2017 IEEE Symposium Series on Computational Intelligence (SSCI).

[16]  Craig J. Brown,et al.  Spatial scale and geographic context in benthic habitat mapping: review and future directions , 2015 .

[17]  C. Stauss,et al.  Drawing lines at the sand: evidence for functional vs. visual reef boundaries in temperate Marine Protected Areas. , 2013, Marine pollution bulletin.

[18]  D. Wright,et al.  1.09 – Application of Estuarine and Coastal Classifications in Marine Spatial Management , 2011 .

[19]  J. Friedman.  Greedy function approximation: A gradient boosting machine. , 2001 .

[20]  Simon J. Pittman,et al.  Multi-Scale Approach for Predicting Fish Species Distributions across Coral Reef Seascapes , 2011, PloS one.

[21]  C. Yesson,et al.  Towards High-Resolution Habitat Suitability Modeling of Vulnerable Marine Ecosystems in the Deep-Sea: Resolving Terrain Attribute Dependencies , 2012 .

[22]  B. Calder,et al.  Automatic processing of high‐rate, high‐density multibeam echosounder data , 2003 .

[23]  Zhenkui Ma,et al.  Tau coefficients for accuracy assessment of classification of remote sensing data , 1995 .

[24]  M. Kendall,et al.  Seascape models reveal places to focus coastal fisheries management. , 2018, Ecological applications : a publication of the Ecological Society of America.

[25]  Tore Söderqvist,et al.  Ecosystem Goods and Services from Swedish Coastal Habitats: Identification, Valuation, and Implications of Ecosystem Shifts , 2007, Ambio.

[26]  David J. Kriegman,et al.  Automated annotation of coral reef survey images , 2012, 2012 IEEE Conference on Computer Vision and Pattern Recognition.

[27]  Simon J. Pittman,et al.  Using Lidar Bathymetry and Boosted Regression Trees to Predict the Diversity and Abundance of Fish and Corals , 2009 .

[28]  G. Fader,et al.  Benthic habitat mapping on the Scotian Shelf based on multibeam bathymetry, surficial geology and sea floor photographs , 2001 .

[29]  J. Paruelo,et al.  How to evaluate models : Observed vs. predicted or predicted vs. observed? , 2008 .

[30]  Stuart R. Phinn,et al.  Scaling up Ecological Measurements of Coral Reefs Using Semi-Automated Field Image Collection and Analysis , 2016, Remote. Sens..

[31]  J Elith,et al.  A working guide to boosted regression trees. , 2008, The Journal of animal ecology.

[32]  Marvin N. Wright,et al.  Random forest as a generic framework for predictive modeling of spatial and spatio-temporal variables , 2018, PeerJ.

[33]  J. Strong,et al.  A review of the influence of marine habitat classification schemes on mapping studies: inherent assumptions, influence on end products, and suggestions for future developments , 2018, ICES Journal of Marine Science.

[34]  J. Teilmann,et al.  Basin-scale distribution of harbour porpoises in the Baltic Sea provides basis for effective conservation actions , 2018, Biological Conservation.

[35]  Max Kuhn,et al.  Building Predictive Models in R Using the caret Package , 2008 .

[36]  Craig J. Brown,et al.  Developments in the application of multibeam sonar backscatter for seafloor habitat mapping , 2009 .

[37]  H. Kautsky,et al.  Scale‐dependent influence of environmental variables on species distribution: a case study on five coastal benthic species in the Baltic Sea , 2013 .

[38]  C. Menza,et al.  Sampling design tool for ArcGIS: Instruction manual.[for ESRI ArcGIS 10.0 Service Pack 3 or higher] , 2013 .

[39]  Markus Diesing,et al.  Towards Quantitative Spatial Models of Seabed Sediment Composition , 2015, PloS one.

[40]  Simon J. Pittman,et al.  Quantifying seascape structure: extending terrestrial spatial pattern metrics to the marine realm , 2011 .

[41]  Mohammed J. Zaki.  Data Mining and Analysis: Fundamental Concepts and Algorithms , 2014 .

[42]  Martin Ludvigsen,et al.  Underwater hyperspectral imaging as an in situ taxonomic tool for deep-sea megafauna , 2018, Scientific Reports.

[43]  Kristjan Herkül,et al.  Applying multibeam sonar and mathematical modeling for mapping seabed substrate and biota of offshore shallows , 2017 .

[44]  B. M. Costa,et al.  The semi-automated classification of acoustic imagery for characterizing coral reef ecosystems , 2013 .

[45]  Senén Barro,et al.  Do we need hundreds of classifiers to solve real world classification problems? , 2014, J. Mach. Learn. Res..