"Ground-validation for remotely sensed marine habitat mapping and species distribution modelling"

 

Background: The increased number of marine mapping programs around the world in recent years has resulted in the collection of a vast amount of spatial data that extend from intertidal habitats to the deep ocean. Despite the range of depths and habitats surveyed using a wide range of remote sensing tools from optical to acoustic sensors, there is a common need across mapping programs for in-situ ground-validation to characterise benthic habitats and allow predictions of species distributions. In part, this issue has also become more important in recent years due to an increasing focus on providing measures of error and uncertainty on predictions. Recent marine and terrestrial studies have identified the importance of understanding the underlying assumptions, characteristics and uncertainties associated with data availability and quality when choosing and applying habitat mapping approaches or species distribution models. The issue of accurate and representative in-situ ground-validation is especially pertinent in the marine environment where researchers are often working in cross-disciplinary projects with multiple objectives (containing geologists, hydro-surveyors, acousticians and ecologists). With often multiple objectives there is a need to examine ways of optimising survey design in a variety of environmental conditions based on depth and clarity limitations, survey extent and habitat complexity. Questions that are logically challenging to answer with certainty include: 

 

• Has a representative sample of habitats that occurs within a study locality been obtained? 

• Have any species/habitats being overlooked or under-sampled?

 

The extent to which these issues can be reduced will largely depend on the availability of time and gear, logistics, expertise and availability of pre-survey data. It is clear that how these issues affect resultant habitat maps and predictions of species distributions warrants further discussion. Hence, these issues surrounding collecting adequate in situ ground-validation samples are highly relevant to the GeoHab community and will form the basis of a one-day pre-conference workshop.

 

Workshop Objective: The objective of this workshop is to provide attendees and contributors with an opportunity to present recent advances in acquisition tools, techniques and approaches to improve in-situ ground-validation of marine remote sensing data, including gear developments, statistical approaches and developing research activities within this field. It will provide contributors with a forum to discuss varying approaches, and provide attendees with a basis for discussion on optimising ground-validation design to reduce uncertainty in habitat classification and application of techniques such as species distribution modelling.

 

Workshop presenters who would like to talk about tools or techniques can base their presentation on a common datasets that will be made available. 

 

Workshop Themes:

 

1. Issues with ground-validation and applications

• New and emerging tools for in situ seafloor ground-validation of remote sensing data 

• Sampling design methods for ground-validation 

• Use of acoustic and optical data for stratifying biological surveys

 

2. Ground-validation influence on species distribution modelling and habitat mapping approaches

• Influences of ground-validation design and technique on modelling outputs 

• Contrasting current approaches in habitat mapping and species distribution modelling 

• Novel techniques and applications

 

Workshop Format: This workshop will consist of talks by invited speakers followed by presentations from individuals/groups who have accepted an invitation to present work based on either their own data or a common dataset. The morning sessions will cover topics relating to the application of in-situ ground-validation for the purpose of mapping seafloor habitats, while afternoon sessions will focus on the effects of ground-validation on habitat mapping and species distribution modelling outputs. We will be seeking contributions from individuals/groups working on tools, techniques and related research, especially new or emerging approaches.

 

Common Dataset: Ningaloo Marine Park, Western Australia. The Ningaloo Reef is Australia’s largest fringing coral reef, extending across 300 kilometres of coastline between Exmouth and Carnarvon in northern Western Australia. In 2005 the WA Government initiated a major collaborative research effort to support management of Ningaloo Marine Park. The research was described as ‘intense, coordinated and collaborative research that has produced an unprecedented wealth of knowledge about the iconic Ningaloo coastal region’. This data has since been made publically available to encourage its ongoing use by planning and management agencies, industries and other stakeholders to assist in making wise decisions about and for the region. The comprehensive datasets (including multibeam, towed video, benthic sled and baited camera) provide many opportunities to showcase the implementation of ground-validation techniques and the development of different applications of habitat mapping and species distribution modelling, offering a diversity of research questions and problems to examine in detail.

More information on the research and data can be found here.

 

Venue & Abstracts: The workshop will take place in the Cumberland Lorne Resort in Lorne, Victoria, Australia.

 

For more information, contact Alan (alan.jordan@dpi.nsw.gov.au) or Cordelia (cordelia.moore@uwa.edu.au).

 

-> UPDATE 2 (Apr. 25th, 2014): Download the workshop agenda here <-