Project Title: Small Project: Modeling support for Southeast Deep Coral Initiative
Project Summary

Lead: Matt Poti and Peter Etnoyer

Models are expected to serve as a valuable integrative tool for the SEDCI. This effort will include:

1 the application of existing predictive habitat models developed at NCCOS from 2011-2013 to site selection and cruise planning,

2 the development of new models to bring all three sub-regions of the Southeast to a similar level,

3 the testing and validation of model predictions,

4 the use of models to help synthesize and generalize results of field surveys. 

Anticipated Management Application(s)

Because of the patchy nature of deep-sea coral distribution and the time and expense of deep-sea coral habitat exploration, models are essential tools to generalize from field survey results to the comprehensive spatial information needed for effective deep-sea coral habitat management and conservation. Predictive habitat modeling combines large suites of potential habitat variables with data on deep-sea coral presence locations in order to create predictive maps of the relative suitability of different areas for the different deep-sea coral groups of interest and management regions.

  •  None Defined.
Point of Contact: Office of the Point of Contact:
Project Title: Small Project: Modeling support for Southeast Deep Coral Initiative

The MaxEnt presence-only modeling approach for deep-sea coral predictive habitat modeling is described in detail in Bauer et al. 2016 and Kinlan et al., in review. This approach is identical to that used to develop previous models in the U.S. South Atlantic and Gulf of Mexico. This same approach will be used to develop a Leiopathes model for the Gulf of Mexico and models of deep-sea coral taxonomic groups of interest in the U.S. Caribbean.

Poti has previously conducted model accuracy assessment and integration of model predictions with regional field survey efforts in the U.S. Northeast/Mid-Atlantic (2012-2015), in the U.S. South Atlantic (2014-2016; for hard-bottom model predictions developed using the same MaxEnt method), and in the Gulf of Mexico (2014 NOAA Ship Okeanos Explorer cruises). Model performance will be assessed by receiver operating characteristic (ROC) curve analysis, assessment of positive and negative predictive value, and other model accuracy statistics useful in the application of model results. Testing these models will enhance their value to fishery management councils and other planning bodies tasked with deep-sea coral management/conservation.

Methods for overlaying fishing effort patterns with deep-sea coral model predictions are described in a forthcoming report (expected February 2017) on the FY15 DSCRTP-funded small project on fishing effort and predicted deep-sea coral habitat overlap in the Gulf. The same methods will be broadened to areas where similar data exist in the U.S. Caribbean and South Atlantic.

Project Results and Management Outcomes  None Defined
Project Title: Small Project: Modeling support for Southeast Deep Coral Initiative
Internal References:
  •  None Defined

There are currently no items to list.