The study of habitat is essential for understanding the biological and ecological requirements of animals as well as the strategies they employ to fulfill their needs [1] . In addition, habitat information is a fundamental prerequisite for the implementation of both management and conservation strategies [2] . However, the definition of habitat is still a contentious one, and its use is far from being consistent [3] . In particular, there is a general lack of unified definition in both terrestrial and marine ecological studies (Table 1 and Table 2). As a further example, a review of the use of the term in terrestrial studies found that 82% of articles reviewed, used habitat terminology imprecisely [4] .

Definition and research into habitat have frequently been identified as crucial for cetacean management and conservation (e.g. [2] [5] - [8] ). However, a consistent definition and understanding of what habitat actually means for cetaceans is still lacking. As a consequence, there is limited information and understanding of habitat characteristics for most species [9] . In order to advance our understanding of the underlying drivers and proces- ses that influence cetacean habitat, studies need to develop more a standardized and objective approach in which to examine them. This broadened approach will ultimately assist in the development and implementation of effective management, conservation and threat mitigation strategies.

In this context, the aims of the present work are: 1) to provide what we believe are representative examples between terrestrial and marine environments; 2) to highlight the specific features of marine environments that may contribute to the current lack of consensus in defining cetacean habitat, 3) provide a non-exhaustive review of how cetacean habitat has previously been studied, including modeling approaches and 4) to provide objective recommendations on how to develop an approach to studying habitat in order to advance cetacean ecology, and ultimately conservation and management efforts.

Typically, habitat in its simplest terms is defined as the physical environment, where an organism actually or potentially lives [10] . In addition, it has also been expanded to include the resources and environmental features present in an area which influences occupancy [4] . Habitat can also be thought of as a concept, used to link potential relationships between an organism and its physical and chemical environment [11] . However, a mechanistic understanding of this concept and how particular features influence organisms is still critically lacking [10] .

Accurately describing and understanding the processes that determine the distribution of organisms is often constrained by the environment itself. Terrestrial and marine ecosystems are both spatially heterogeneous, comprised of ecological entities such as forests, hills, deserts, sea grass beds, seamounts and coral reefs, but also vary in time from diel to annual cycles (e.g. [12] ). In terrestrial ecosystems, habitat is often defined by the presence of relatively persistent vegetation and animal life [13] . For example, the boundaries between the biotic and abiotic properties characterizing structurally diverse terrestrial environments (Figure 1) are easily observed and identifiable (e.g. vegetation patches, sedimentary rocky areas, gorges and slopes of cobbles and boulders). In many cases, the relatively immediate accessibility and visibility of the terrestrial environment, enhances our capacity to identify and observe environmental differences.

In contrast, most marine environments are characterized by a limited number of landmarks both above and beneath the surface (Figure 2). Marine organisms typically live in a fluctuating and heterogeneous three-di- mensional water mass. In addition, the inaccessibility of most of the World’s ocean, and the logistical considerations inherent in effectively studying marine organisms once underwater, places additional limitations on how to define habitat for an organism, a species or a community. The characteristic wide-ranging and migratory nature of many marine animals, including cetaceans, often means that habitat boundaries are difficult to define ( [7] ; Figure 2), and may change on a temporal basis. Furthermore, regions within the World’s Oceans are often defined by broad, general definitions such as open ocean or coastal waters, although specific sea surface temperature signatures such as warm and cold core eddies (Figure 3(A)) and thermal frontal zones (Figure 3(B)) can be specified. These broad classifications are frequently applied to species such as cetaceans, particularly those rarely sighted or cryptic species [14] . While these areas may be relatively distinct (Figure 3(C)), general classifications still lack a definitive understanding of what habitat actually means. As a consequence, the definition of marine habitat often seems arbitrary and in most cases non-existent. The application of the term habitat is often inconsistent even between marine animals of the same species or taxa (Table 2). These definitions highlight those potential factors (e.g. environmental factors) considered to be essential for the animals but again lack a thorough consideration of how the animal actually interacts with and relies on its environment.

In addition, habitat for many organisms (e.g. migratory birds, cryptic species), is often characterized using a limited number of observations recorded at specific encounter locations. Cetaceans are no exception, with habitat often described using only sightings or environmental measurements recorded at the surface, when the animals are exposed [15] - [19] . This strategy, however, disregards the properties and characteristics of the habitat concealed underneath, vertical structure of the water column. In contrast to terrestrial systems where environmental features are readily accessible and visible (Figure 1), in the marine environment it is considered much more difficult to gather relevant habitat information at depth. This again, potentially highlights the inaccessibility and the logistics of conducting research within these areas. Whilst some techniques (e.g. remote sensing imagery, animal borne sensors) can offer new perspectives and insight into detailed understanding of the vertical structure of the water column, they do not directly address the issue of habitat in order to provide sufficient information. Hence, in the marine environment, little information about the relations between species and their specific environments exists, despite their significance [20] .

Cetacean species exhibit a wide range of distribution patterns across all parts of the World’s Oceans [5] [21] - [26] As individuals and populations range widely and are not easy to observe directly, the concept of habitat is therefore difficult to grasp and define [27] . The wide diversity of cetacean species in general, makes our ability to understand their habitat more problematic. For example, many delphinids are widely distributed, with smaller populations inhabiting various locations and climatic regions, whilst in contrast larger mysticetes follow migrational routes each year to familiar calving areas [28] . The intrinsic difference in distributions across temperate and tropical, and coastal and offshore waters between the cetacean species, not only demonstrates their great ecological flexibility [29] , but also ultimately links them to their habitat. In addition, the distinction between life history strategies and the biological requirements of the different species potentially influences their choice, and utilization of specific habitats in the marine environment; something which is also commonly dismissed in many habitat studies.

Critical cetacean habitat in a broad sense has previously been defined as “those parts of a cetacean’s range either a species or population that are essential for the day-to-day survival and maintenance of a steady population growth rate, including those areas essential for specific behaviors such as mating, feeding and migrational routes” [7] . Cetacean habitat, and more specifically delphinid habitat, has typically been defined by investigating a number of abiotic and biotic factors ranging over various spatial and temporal scales thought to drive their distribution [30] - [36] . Factors previously investigated range from the physical and chemical features of the environment, such as water temperature, depth, salinity, topography and distance from shore, benthic habitat characteristics, and the presence of vessels, and preys and predators (Table 3). Measurement of these habitat variables was typically obtained using a wide variety of methodologies (e.g. in situ measurements, remotely sensed, obtained by boat or land based techniques), levels of precision (in situ vs. remotely sensed) and scales (temporal, spatial). Furthermore, variables used to assess habitat (even for a given species) were typically author and study dependent. For instance, a non-exhaustive review of pertinent studies of bottlenose dolphins (Tursiops sp.) clearly demonstrates the variety of both habitat measurements and spatio-temporal scales used [17] [37] - [40] .

Among the variables used to assess cetacean habitat, sea surface temperature represents a common measurement that is often measured with a variety of scales and resolutions (Figure 3). For instance, [17] compared the distribution of white-sided (Lagenorhynchus acutus) and common (Delphinus delphis) dolphins off the coast of New Zealand with sea surface temperature and salinity measured from the survey vessel at the time of a dolphin sighting. In comparison, water temperature measurements accessed from remote sensing data was used to investigate seasonal distribution changes in striped dolphins (Stenella coeruleoalba) in the Ligurian Sea [41] . Of further note is the typically small number of environmental variables measured in many studies despite the high plausibility of other factors being instructive in describing habitat [42] - [44] . This consideration and narrowed selection of assessed variables should therefore caution our application of some habitat studies for progressing management objectives and conservation strategies. More specifically, a thorough understanding of constraints, methodology and objectives needs to be made to ensure that the results of habitat studies are not being confounded by the concentration of the researcher’s effort and measurement tools or access [45] .

Statistical habitat modeling, although still a relatively recent topic of research, is increasingly being applied to help answer questions regarding the ecology of many cetacean species [46] . Since the first paper there has been a significant growth and increase in this topic (Figure 4), which suggests a field in rapid development.

Typically, the aim of statistical habitat modeling is to help predict and explain variation in the distribution and density of cetaceans, as well as to predict key locations by correlating observations of animals with various environmental variables [47] [48] . However, these efforts may describe correlations between variables, but they generally lack the ability to elucidate our ecological understanding of the relationships between cetaceans and their marine environment.

^aSST: Sea Surface Temperature; ^bChl. a: chlorophyll a concentration.

From a non-exhaustive review, we show as aforementioned for field-based habitat studies that modeling habitat studies do not converge in their approaches, methodologies, spatial and temporal scales and analyses even when they target the same species (Table 4). Some studies are vague in their definition of a potential focus species as well as an ecological question, and often the focus species is then defined afterwards depending upon what species were observed during surveys. The overall objective of many studies is then often limited to predict where and when cetaceans are present (e.g. [49] ). However, some studies do attempt to explain this presence further by linking them to features of the physical and biological oceanographic properties of their environment [48] ; these properties have either been assessed using remote sensing data (e.g. sea surface temperature, sea surface height), variables measured in situ (e.g. depth, mixing layer thickness) or even modeled environmental data such as prey densities [50] , hence allow to cover a very wide range of spatial and temporal scales; see Table 4 for further examples. There is, however, a strong study-to-study variability in the abiotic properties considered even in modeling studies dealing with similar environments and species (Table 4). Biotic variables are also dramatically under-represented (especially when compared to physical variables) in most of the studies reported here (Table 4). Similarly, biotic factors that may be critical to understand cetacean habitat use such as behavioral and life history strategies, have still been seldom used in habitat modeling studies ( [48] ; Table 4). Besides, studies that incorporate field-based visual and acoustic surveys [51] - [53] often lack information about the physical and vertical properties of the environments (Table 4). Also note that most synoptic studies that used remote sensing data, critically lack information about the vertical structure of the water column [54] . The aforementioned limitations of habitat modeling studies―which are by no means a criticism of their results and do not detract from the central point of their work―hence suggest that although habitat modeling studies provide valuable information on where and when cetaceans may be over space and time, they still ultimately lack the power in which to truly understand the mechanistic links between the presence and behavior of cetaceans and the nature of their environment.

As a conclusion, statistical habitat modeling is undeniably a useful and promising tool to predict cetacean distributions as a function of range of descriptors (Table 4), in particular for those large whale and offshore cryptic species which lack baseline data and are often difficult to access. However, this approach still does not converge in the approach followed (Table 4), hence may prevent future progress in our ability to provide further insight into animal ecology. As stressed in the present work and in the recommendation below, there is a genuine need to refine modeling methods to move beyond correlations towards a mechanistic understanding of the processes that interact to create cetacean habitat and try to provide a more ecological explanation for their presence. Ultimately, this may also help to bridge the gap between fundamental research and conservation and management initiatives.

To increase our understanding of cetacean habitat, we suggest to develop a more systematic and objective approach to cetacean habitat research. In particular, we stress the need to identify the underlying influences driving habitat, for example physical and chemical environmental features, social and behavioral factors, predation and anthropogenic pressures in order to determine how cetaceans interact with and use their environment. With this

^aSea surface temperature referred to as SST_sat hereafter from satellite data, and as SST_{in situ} when measured in situ. ^bChlorophyll a concentration referred to as chla_sat from satellite data, as chla_{in situ} when measured in situ. ^cIVF in vivo fluorescence measured in situ as a proxy of chlorophyll a concentration. The letters M and O respectively stands for Mysticetes and Odontocetes; M₁: Eubalaena glacialis, M₂: Eubalaena australis, M₃: Balaenoptera musculus, M₄: Balaenoptera physalus, M₅: Balaenoptera borealis, M₆: Balaenoptera acutorostrata, M₇: Balaenoptera bonaerensis and M₈: Megaptera novaeangliae. O₁: Physeter macrocephalus, O₂: Delphinapterus leucas, O₃: Berardius bairdii, O₄: Ziphius cavirostris, O₅: Hyperoodon ampullatus, O₆: Orcaella brevirostris, O₇: Orcinus orca, O₈: Globicephala melas, O₉: Globicephala macrorhynchus, O₁₀: Pseudorca crassidens, O₁₁: Feresa atenuata, O₁₂: Peponocephala electra, O₁₃: Steno bredanensis, O₁₄: Lagenorhynchus obliquidens, O₁₅: Lagenorhynchus obscurus, O₁₆: Lagenorhynchus acutus, O₁₇: Lagenorhynchus australis, O₁₈: Grampus griseus, O₁₉: Tursiopus truncatus, O₂₀: Stenella attenuate, O₂₁: Stenella frontalis, O₂₂: Stenella longirostris, O₂₃: Stenella longirostris orientalis, O₂₄: Stenella coeruleoalba, O₂₅: Delphinus delphis, O₂₆: Delphinus capensis, O₂₇: Lagenodelphis hosei, O₂₈: Lissodelphis borealis, O₂₉: Celphalorhyncus comersonnii, O₃₀: Celphalorhyncus eutropia, O₃₁: Phocoenoides dalli, O₃₂: Phocoena phocoena and O₃₃: Platanista gangetica.

in mind, a number of factors should be considered on how to approach and develop methodologies in which to investigate cetacean habitat. The following six recommendations―synthesized as a logical flow chart in Figure 5―demonstrates how future studies could become more targeted and effective in defining and describing habitat. The purpose of this outline is to progress towards a more standardized and objective approach to habitat studies. Specifically, the six recommendations developed hereafter are illustrated using selected case studies from the primary literature on both well-documented and more cryptic species to demonstrate the generality of the proposed approach.

Given the difficulty and complexity of adequately understanding the meaning of habitat for cetaceans, the development of a sound approach incorporating suitable techniques and methodologies is critical to enable the quantification of appropriate variables. Understanding the influences and the inter-relationships between cetaceans and their surrounding environment will not only greatly improve our understanding, but also ultimately allow us the ability to develop targeted and more effective mitigation and conservation measures.

The authors wish to thank C. Chapperon for providing comment on an earlier draft of this paper. This research was supported under Australian Research Council’s Discovery Projects funding scheme (projects number DP0664681 and DP0988554). Professor Seuront is the recipient of an Australian Professional Fellowship (pro- ject number DP0988554).

NardiCribb,CaraMiller,LaurentSeuront, (2015) Towards a Standardized Approach of Cetacean Habitat: Past Achievements and Future Directions. Open Journal of Marine Science,05,335-357. doi: 10.4236/ojms.2015.53028