Marine species distribution modeling has seen explosive growth in recent years, and the Endangered Species Research Theme Section entitled 'Beyond marine mammal habitat modeling: applications for ecology and conservation' demonstrates that the field of marine mammalogy has been no exception. For the past decade, marine mammal ecologists have been developing habitat models with increasing proficiency and sophistication. However, these efforts have largely focused on correlative analyses of observed species−environment associations, which often have low explanatory power due to the absence of critical, but unaccounted for processes that are important drivers of animal distributions. Here we provide an overview of these processes, advocate for directed studies (e.g. tagging, prey sampling, focal follows, physiological assessment) to address how the processes influence species' distributions, and challenge the modeling community to incorporate these results into their efforts. We also identify a progression of modeling stages from correlative to confirmatory to mechanistic that should lead us to formulate increasingly robust and accurate predictions of species distributions rooted in greater ecological understanding. Given the on-going risks to marine mammals from human activities and climate change, such models are needed for conservation and management now more than ever.
FREE REE ACCESS CCESS
Contribution to the Theme Section 'Beyond marine mammal habitat modeling'Endang Species Res 22: 191-203, 2013 ecology and conservation' clearly demonstrate that the marine mammal research community has embraced these tools, and is applying them with increasing proficiency and sophistication (Gregr et al. 2013, this Theme Section).Marine mammal habitat models typically predict the spatial distribution or abundance of a species based on correlations between animal observations and attributes of the physical marine environment (e.g. seafloor terrain, sea surface temperature, proximity to the ice edge) or proxies of dynamic oceanographic processes that may directly or indirectly aggregate prey (Redfern et al. 2006). This approach often makes 2 implicit assumptions with regard to the environmental variables: (1) they adequately capture the underlying environmental gradients to which a species responds, and (2) they reasonably represent conditions that lead to enhanced feeding opportunities. These assumptions, and the focus on foraging success, reflect the limits of our understanding of how marine mammals interact with their environment as well as our ability to directly measure the prey field at scales relevant to a foraging predator.The correlative approach has generated considerable insight into the distribution of marine mammal species and their potential foraging areas. However, our models continue to capture only a small portion of the observed variability in species' occurrence or abundance. For example, the deviance explained (a measure of model fit) by models published in this Theme Section was generally be...