PROCESS BASED DATA FUSION FOR MARINE MAMMAL SCIENCE
Abstract
The US Navy needs to know where marine mammals are distributed throughout timeand space in order to understand the impacts of anthropogenic sound. Knowledge ofthe spatio-temporal distribution of whales can be developed by combining observationsfrom several disparate data sources. However, most spatio-temporal analysesfocus on one dataset at a time, thereby limiting the scope of the inference and introducingbias. We propose methodology for combining, or #fusing,# disparate datasetsusing a hierarchical modeling framework that can be applied to other marine species,and can also be extended to incorporate observations from novel data sources in thefuture.We will develop formal data-fusion models using different space-time data sources,including line-transect data, passive acoustic monitoring and photo-identification.Data fusion, or data assimilation has become an increasingly studied issue as diversedatasets are collected often to learn about a common problem. For example, inthe field of environmental exposure, recent data fusion efforts have fused exposuresources such as monitoring stations, satellite data, and computer model output todevelop contaminant surfaces for ozone or particulate matter. To date, the datafusion work in ecology has lagged behind, and though recent efforts are being seen,all of this work is in the terrestrial realm. This represents an important gap betweenthis work in spatial statistics and marine mammal science.We will build hierarchical data-fusion models to a) better understand the distributionof North Atlantic right whales and two species of beaked whales; and b)introduce a new methodology for fusing the types of data that are increasingly common,e.g. fixed and towed passive acoustic data. A central goal of this work is to fullyunderstand the true spatial intensity of whale abundance, given that we imperfectlyobserve it with a variety of sampling techniques. We will build a data simulationthat allows us to better understand how the observation process can be linked to thistrue, but hidden, intensity surface. While we focus on NARW and beaked whales,these methods can be extended to other marine mammal systems where knowledgeof the spatial distribution is critical. With a more complete representation of thedistribution of marine mammal species, we can better characterize the consequencesof sound exposure on individual marine mammals as well as on populations.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Nov 08, 2024
- Source ID
- N000142412501
Entities
People
- Robert S Schick
Organizations
- Office of Naval Research
- Southall Environmental Associates (United States)
- United States Navy