Cetacean Data Hierarchy

Overview 

Temporally and spatially explicit estimates of species or population abundance, density, or distribution are typically the most valuable types of information to resource managers. The type and quantity of information available on cetacean distribution, abundance, and density varies in space and time. Typical sources of primary information include sighting, photo identification, acoustic, tagging, and genetic data. Estimates of absolute abundance or density can often be derived from sighting and photo identification data, but these data cannot be collected at night or during poor environmental conditions that limit visibility. Acoustic data can provide year-round information on the presence or location of calling animals, but species identification can be difficult, and currently group size can only be estimated reliably for a few species; therefore, acoustic data alone are often not sufficient for estimating absolute abundance or density. Tagging data provide detailed information on the movements of individuals, sometimes for periods greater than one year, but the sample sizes are usually relatively small due, among other things, to cost and logistics, and it is not possible to estimate absolute abundance or density from tag data alone. Genetic data provide information on population structure, sex composition, relatedness, site fidelity, and distribution, but estimates of absolute abundance or density cannot be derived from genetics alone. Furthermore, primary information can be analyzed to derive model-based estimates of abundance, density, or probability of occurrence.

The CetMap Working Group undertook an effort to compile the available information on cetacean abundance, density, and distribution throughout the US Exclusive Economic Zone (EEZ). Furthermore, members of the CetMap Working Group developed, or are in the process of developing, new spatially explicit models of cetacean density for 20 species, populations, or guilds along the US East Coast, 16 in the Gulf of Mexico, and 3 in the Alaskan Arctic, in addition to a spatiotemporally explicit model of the eastern Pacific gray whale migration along the US West Coast. The CetMap Working Group used the hierarchical classification system detailed below to rank the individual data or model products according to their ability to address spatially and temporally explicit questions on cetacean abundance, density, or distribution. In a few cases, more than one data set or model was available for a given species, population, or guild/area/time period. The criteria used by modelers to select data for creating the models were similar to the criteria used by the CetMap Working Group to select a single model or dataset to represent a species, population, or guild/area/time period, including: 1) age of the data; 2) consistency of survey protocols and platforms; 3) ability to correct for sighting biases; 4) taxonomic resolution of the data; 5) scientific rigor of survey protocols; and, specific to model selection, 6) whether and how the model was validated.

 

Information Hierarchy

Tier 1: Habitat-based Density Models - Density predicted heterogeneously across a surface of cells based on environmental covariates. Probability of detection was accounted for as a function of distance from the transect line to the sighting, and possibly other covariates such as group size or Beaufort sea state. Density was derived from the estimates of encounter rate, effective strip width, average group size, and an assumed or estimated value for the trackline detection probability, any of which may have been assumed to vary with the environment.

Tier 2: Stratified Density Models - Density predicted homogeneously across an area. Probability of detection was accounted for as a function of distance from the transect line to the sighting, and possibly other covariates such as group size or Beaufort sea state. Density was derived from the estimates of encounter rate, effective strip width, average group size, and an assumed or estimated value for the trackline detection probability.

Tier 3: Probability of Occurrence - Spatially heterogeneous predictions of the probability of encountering the species, population, or guild across a surface of cells, based on environmental covariates.

Tier 4: Records Exist - Presence-only observations in the CetMap data system, or known to exist in published literature or with other investigators. Principally, these observations were collated at OBIS-SEAMAP based on visual observations aboard ships or aircraft on scientific surveys. Acoustic observations and opportunistic visual observations were also included. The observations were not corrected for heterogeneity in survey effort or detection probabilities across the study area.

Tier 5: Expert Knowledge - Species status (Present, Likely Absent) for a given region. This tier was populated with information from regional species lists generated by the CetMap Working Group and regional experts. This tier was created at the temporal resolution of an entire year, such that if a species occurs at any time during the year, then all cells for that species/region that are not populated with Tier 1-4 data are designated as “expert-based presence”. It is important to note that this does not provide information on species presence during specific months.