Problems in Data Synthesis: Towards a dynamic understanding of marine habitat use

Allison Nancy Cramer

Integrating data from a variety of sources, or data synthesis, can increase the scale and dimension of ecological data sets, providing a dynamic window into animal habitat use. When combining data from disparate sources researchers must be cognizant of how the underlying data structure constrains their ecological inference. For the environmental and animal data commonly used to investigate habitat preference, characteristics of both the animals and the environment restrict how data represent ecological events. The marine environment represents a unique case where biology and the environment dramatically impact the structure and scale of available data that in turn can affect resulting inferences. In this dissertation I use examples from the marine environment to explore how data representation impacts data synthesis and ecological inference. In Chapter 1 I examine how the data representation of animal movement obtained from passive acoustic telemetry impacts calculations of habitat preference. I find that to distinguish habitat preference using passive acoustic telemetry researchers should incorporate a “burn in” window into their study design, and develop an analytical solution to estimate the burn in window for any target species. In Chapter 2 I address scaling issues between passive acoustic telemetry data and a high-resolution biogeochemical model to estimate habitat suitability for a rarely studied species of deep-water shark, Hexanchus griseus. To account for sampling mismatches, I alter the standard spatial sampling approach within Species Distribution Modeling (SDM) to estimate shark probability of presence and develop habitat suitability maps. In Chapter 3 I use data synthesis to re-examine marine habitat classification schemes in relation to mechanistic habitat variables. This analysis revealed a classification scheme which sorts biomes and their constituent communities based on energy flow and habitat stability and supports a new model of ecological community organization. Though the examples explored here stem from the marine environment, the constraints of sampling, scale, and classification occur across ecosystems. The methods discussed here demonstrate a variety of approaches research can take to understand, account for, and re-examine the impact of the underlying data structure on ecological inference.

Problems in Data Synthesis: Towards a dynamic understanding of marine habitat use

Files and links (1)

Abstract

Metrics

Details